Sarcoma

Chondrosarcoma

Chondrosarcoma

Sarcoma

Chondrosarcoma

Copyright © 2011 Hindawi Publishing Corporation. All rights reserved.

This is a focus issue published in volume 2011 of “Sarcoma.” All articles are open access articles distributed under the Creative Com-mons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original workis properly cited.

Editorial Board

A. Abudu, UKIrene Andrulis, CanadaJose Casanova, PortugalQuincy Chu, CanadaA. Craft, UKCyril Fisher, UKGeorge Gosheger, GermanyRobert Grimer, UK

Alessandro Gronchi, ItalyNora Janjan, USAMichael Leahy, UKOle Steen Nielsen, DenmarkBrian O’Sullivan, CanadaAlberto Pappo, USAR. E. Pollock, USAAjay Puri, India

Chandrajit Premanand Raut, USAMartin H. Robinson, UKLuca Sangiorgi, ItalyCharles R. Scoggins, USABeatrice Seddon, UKDavid Spooner, UKClement Trovik, NorwayC. Verhoef, The Netherlands

Contents

Surmounting Chemotherapy and Radioresistance in Chondrosarcoma: Molecular Mechanisms andTherapeutic Targets, Anne C. Onishi, Alexander M. Hincker, and Francis Y. LeeVolume 2011, Article ID 381564, 8 pages

Involvement of the Soluble Urokinase Receptor in Chondrosarcoma Cell Mobilization, Katia Bifulco,Immacolata Longanesi-Cattani, Maria Teresa Masucci, Annarosaria De Chiara, Flavio Fazioli,Gioconda Di Carluccio, Giuseppe Pirozzi, Michele Gallo, Antonello La Rocca, Gaetano Apice,Gaetano Rocco, and Maria Vincenza CarrieroVolume 2011, Article ID 842842, 10 pages

A “Proteoglycan Targeting Strategy” for the Scintigraphic Imaging and Monitoring of the Swarm RatChondrosarcoma Orthotopic Model, Caroline Peyrode, Francois Gouin, Aurelien Vidal,Philippe Auzeloux, Sophie Besse, Marie-Melanie Dauplat, Serge Askienazy, Dominique Heymann,Jean-Michel Chezal, Francoise Redini, and Elisabeth Miot-NoiraultVolume 2011, Article ID 691608, 8 pages

Chondrosarcoma: With Updates on Molecular Genetics, Mi-Jung Kim, Kyung-Ja Cho, Alberto G. Ayala,and Jae Y. RoVolume 2011, Article ID 405437, 15 pages

Technical Notes on Endoscopic Transnasal Transsphenoidal Approach for Clival Chondrosarcoma,Atsushi Kuge, Shinya Sato, Kaori Sakurada, Sunao Takemura, Zensho Kikuchi, Yuki Saito,and Takamasa KayamaVolume 2011, Article ID 953047, 6 pages

Chondrosarcoma of the Mobile Spine and Sacrum, Ryan M. Stuckey and Rex A. W. MarcoVolume 2011, Article ID 274281, 4 pages

Chondrosarcoma of the Thorax, Philip A. Rascoe, Scott I. Reznik, and W. Roy SmytheVolume 2011, Article ID 342879, 7 pages

The Bone Niche of Chondrosarcoma: A Sanctuary for Drug Resistance, Tumour Growth and also aSource of New Therapeutic Targets, E. David, F. Blanchard, M. F. Heymann, G. De Pinieux, F. Gouin,F. Redini, and D. HeymannVolume 2011, Article ID 932451, 8 pages

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 381564, 8 pagesdoi:10.1155/2011/381564

Review Article

Surmounting Chemotherapy and Radioresistance inChondrosarcoma: Molecular Mechanisms andTherapeutic Targets

Anne C. Onishi, Alexander M. Hincker, and Francis Y. Lee

Department of Orthopaedics, College of Physicians and Surgeons, Columbia University, Black Building, 14-1412, 650 West 168th Street,New York, NY 10032, USA

Correspondence should be addressed to Francis Young-In Lee, fl127@columbia.edu

Received 29 August 2010; Accepted 15 November 2010

Academic Editor: Alberto Pappo

Copyright © 2011 Anne C. Onishi et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Chondrosarcoma, a primary malignancy of bone, has eluded successful treatment with modern chemotherapeutic and radiationregimens. To date, surgical resection of these tumors remains the only curative treatment offered to patients with this diagnosis.Understanding and exploring the nature of chemotherapy and radiation resistance in chondrosarcoma could lead to newmolecular targets and more directed therapy for these notoriously difficult-to-treat tumors. Here we review the most currenthypotheses regarding the molecular mechanisms mediating chemotherapy and radiation resistance and the future direction ofchondrosarcoma therapy research.

1. “Know thy Enemy”: Treatment Obstacles inPatients with Chondrosarcoma

Tzu, the 6th century BC Chinese war theorist, is most famousfor his military treatise The Art of War and for havingwritten, “Know thyself, know thy enemy; one thousandbattles, one thousand victories” [1]. Chondrosarcoma, aheterogeneous group of tumors with extraordinarily diversepresentations and morphologies that have an enormousrange of clinical behaviors, is a complex and difficult enemyto know. To date, patients who receive a diagnosis ofchondrosarcoma are treated primarily with wide resectivesurgery since these tumors are notoriously resistant to bothchemotherapy and radiation treatment. Dramatic excisionsand even amputations are required, since local control oftumor is of paramount importance in order to prevent futuremetastasis.

Chondrosarcoma is the second most common typeof primary bone malignancy in the United States afterosteosarcoma. It represents 26% of all primary bone cancers,with approximately 2000 new cases every year [2]. These

tumors are characterized by cartilage-forming cells with noevidence of direct osteoid formation [3] and are gradedbased on local invasiveness and metastatic potential. Theycan be further classified by histologic characteristics suchas cellularity, matrix changes, cell character, and replicativeactivity [4]. Clinically, the most common sites of tumorformation include the pelvis and appendicular long bones,though reports in the literature have documented other sitesincluding the distal appendicular bones, temporomandibularregion [5], and thoracic spine [6].

While wide resection with limb salvage surgery or ampu-tation are definitive therapies for patients presenting withappendicular tumors [7], such surgeries introduce disabilityand morbidity into the lives of patients. Furthermore, suffi-ciently wide resection is not always possible in large tumors,tumors growing in the pelvis or axial skeleton, or tumorsthat have already metastasized. In such cases, tumor location,chemoresistance, and radioresistance lead to insurmountabletreatment obstacles and very poor outcomes. Understandingthe molecular mechanisms of resistance to chemotherapyand radiation in these tumors could lead to new targets

2 Sarcoma

for treatment or adjuvant therapies to surgery, therebyimproving the length and quality of life in patients sufferingfrom chondrosarcoma.

In this paper, we will provide a summary of the mostrecent research regarding the molecular pathways and genesresponsible for chemotherapy and radiation resistance inchondrosarcoma tumors (see Table 1). We will also explorethe implications of these molecular mechanisms with respectto their role in the development of new therapies for thisdifficult-to-treat cancer.

2. The Expression of P-GlycoproteinAllows Chondrosarcoma Cells toWithstand the Cytotoxic Effects ofMost Chemotherapy Agents

P-glycoprotein is an ATP-dependent membrane-boundpump that excretes small hydrophobic molecules and isnormally expressed in cells with secretory functions such asthe proximal tubular cells of the kidney and the epitheliallining of the bile duct. It is also expressed in the hypertrophicregion of the epiphyseal growth plate, providing additionalevidence for its important physiologic excretory roleinprotecting the cell from extracellular insults [8].

The expression of P-glycoprotein in chondrosarcomatumors has been well-established, and it has been proposedthat this expression is an extremely important mechanismin the development of chemoresistance (see Figure 1).P-glycoprotein is encoded by the gene Multiple DrugResistance-1, or MDR-1. Expression of this protein is com-mon in both benign and malignant cartilaginous lesions, andin one study, 90% of tumors stained with specific antibodieswere positive for P-glycoprotein expression [9]. The authorspostulated that MDR-1 expression was a marker for thepresence of drug resistance to multiple chemotherapy agentsincluding cyclophosphamide, doxorubicin, vincristine, cis-platin, methotrexate, and dacarbazine.

It has been shown using fluorescence microscopy andother in vitro techniques that chondrosarcomacells express-ing P-glycoprotein accumulate lower levels of intracellulardoxorubicin thanchondrosarcoma cells that do not expressP-glycoprotein, and that the P-glycoprotein-positive cellsare insensitive to the cytotoxic effects of doxorubicin atphysiologic concentrations in vitro [10]. Rosier et al. havepostulated that the development of chemoresistantchon-drosarcoma tumors is a result of selective pressure against P-glycoprotein-negative cells in the presence of cytotoxic agents[8]. Alternatively, some authors suggest that toxin exposurecauses direct upregulation of P-glycoprotein expression [13].

More recently, the importance of P-glycoprotein inmediating chemoresistance has been highlighted throughexperiments in which P-glycoprotein was inhibited bothpharmacologically and with gene silencing techniques. UsingsiRNA, Kim et al. showed that knocking down the expressionof MDR-1 increased chemosensitivity by up to 4.1 fold[12]. Of note, a combination of the inhibition of P-glycoprotein with the inhibition of the antiapoptotic proteins

Bcl-2, Bcl-xL, and XIAP showed an even higher increasein chemosensitivity, up to 5.5 fold [12]. These findingssuggest that though P-glycoprotein is clearly an importantmediator of chemoresistance in chondrosarcoma cells; othermechanisms are also involved.

3. Telomerase Activity MediatesChondrosarcoma Cell Immortality

DNA polymerase cannot faithfully complete replication tothe end of the superstructure of the macromolecule, whichleads to the loss of nucleotides with each replication cycle—this is known as the “end replication problem.” Telomeres areregions of noncoding DNA at the end of the double strandthat protect coding regions from this degradation. Withevery newgeneration, a piece of the telomere is lost, even-tually causing the cell to exit the replication cycle [14]. Germcells express telomerase, an enzyme capable of synthesizingnew telomeres via its own RNA template, thereby allowingtelomeres to be faithfully replicated and preventing the cellfrom exiting the replication cycle. Many cancer cell typesincluding testicular, ovarian, breast, endometrial, and basalcell carcinoma and lymphoma express telomerase, and this isone suggested mechanism of attaining cellular immortality[15].

Many chondrosarcoma cells also express telomerase,and it has been established that telomerase expression inchondrosarcoma correlates significantly with tumor gradeand rates of recurrence. Martin et al. have proposed thatusing immunohistochemical techniques to identify thosetumors that express telomerase could be a useful adjuvantto traditional prognostic grading systems [16]. Dr. Martin’sgroup has also explored the role of telomerase combinedwith the loss of tumor suppressor protein p16 as a means ofmalignant transformation of cartilage neoplasms and foundthat these two mutations in combination lead to a moreaggressive and invasive phenotype in vitro [17].

Because of the important relationship between telom-erase expression and cancer cell immortality, inhibition oftelomerase has been an expanding area of cancer therapyresearch. It is hypothesized that by preventing the continuedelongation of telomeres, malignant cells can reenter thenormal cell cycleand and undergo the normal genetic andmolecular checks on replication. Additionally, these cellswould become susceptible to the apoptotic mechanisms thatnormally execute the process of programmed cell death whengenetic damage is induced by chemotherapeutic agents.

The specific pharmacologic inhibitor of telomerase,BIBR1532, has been tested as a means to resensitize chon-drosarcoma cells to traditional modes of chemotherapy.Parsch et al. have established telomerase expression in thechondrosarcoma cell line SW1353 and used the TRAP assayto show that BIBR1532 inhibits telomerase in SW1353 cells.Furthermore, it was demonstrated that long-term dosageof the telomerase inhibitor slowed the rate of growthof SW1353 cells, but did not arrest growth completely;the authors contend that this incomplete growth arrest iscaused by chemotherapy-induced natural selection of clones

Sarcoma 3

Table 1: A summary of the mechanisms involved in chemotherapy and radiation resistance in chondrosarcoma cells. Summarized hereare the various molecular mechanisms responsible chemotherapy and radiation resistance in chondrosarcoma, strategies to overcome thesemechanisms, and the results of studies testing these strategies as possible adjuvants to traditional surgical treatments. See text for morein-depth discussions of each topic.

Resistancemechanism

Effect within the cell Therapeutic strategies Has treatment been tested?

P-glycoproteinexpression

Exports chemotherapydrugs from within thecell

Gene silencing using siRNA pharmacologicinhibition using C-4

In vitro: dramatic increases inchemosensitivity

Telomeraseactivity “Immortal” cell

phenotype

BIBR1532- pharmacological telomeraseinhibitor

BIBR152- in vitro: slowed growthand increased sensitivity tocisplatin.

GRN 163L- pharmacological telomeraseinhibitor

GRN 163L- in vivo: but not inchondrosarcoma—helpfuladdition in leukemia treatment

AngiogenesisSupports larger tumors,allows metastasis

SU6668-inhibits receptors for VEGF, FGF,PDGF ET-743 and plasminogen-relatedprotein B-chemotherapeutic andendothelial-cell-metabolism downregulator

In vivo murine model: decreasedtumor size, vascularization.

In vivo murine model:inductionof profound necrosis, inhibitionof neovascularization

COX-2expression

Unclear, but associatedwith poor prognosis

COX-2 inhibitors-mechanism of actionremains unclear

In vivo: slows cancer growth,although growth relapses aftersix weeks of treatment

Melovonatesynthesis

Shift bone remodelingbalance towardresorption

Bisphosphonates-inhibit melovonatesynthesis in the bone microenvironment

In vivo: induction of apoptosis,inhibition of metalloproteinaseactivity, and reduction of VEGFlevels

TumorSuppressor p16Mutation

Decreased tendencytoward apoptosis

p16-restoring virusOncolyticviruses-selectively target immune-inducingmolecules to cells with pathway defects

In vitro: increasedradiosensitivity in p16-deficientcells.

In vivo, but not inchondrosarcoma-results showstrong selectivity, desiredefficacy, no serious side effects.

Increasedexpression ofBcl-2, Bcl-xL, andXIAP

Decreased tendencytoward apoptosis

Downregulation viapharmacotherapy/siRNA-shift cellularbalance toward apoptosis

In vitro: increasedradiosensitivity

In vivo: increasedchemosensitivity

HypoxiaDecreased ROS creationby radiation

Acridine orange-enhances ROS creationIn vivo: significantly increasedradiosensitivity

with strong telomerase activity. Moreover, data from thisstudy show that cell lines with low-telomerase activity weremore sensitive to cisplatin-induced apoptosis while cell lineswith high levels of telomerase activity were nearly resistant[18].

In clinical trials, telomerase inhibition by the specificinhibitor GRN 163L has been shown to have additive oreven synergistic effects when used in combination with otherchemotherapies and radiation regimens against leukemia[19]. These promising clinical results taken together withthe aforementioned in vitro studies suggest that telomeraseinhibition may be a beneficial strategy in the treatment ofchemoresistant chondrosarcoma.

4. Targeting Angiogenesis inChondrosarcoma Slows Tumor Growth

One important area of cancer research has been the roleof angiogenesis in tumor growth and metastasis. As tumorsgrow beyond a few millimeters in size, oxygen can nolonger diffuse to every cell, and the tumor becomes hypoxic.Angiogenesis, or the growth of new blood vessels, mustoccur for the tumor to grow beyond a certain size. Hypoxiainduces angiogenesis via increased levels of the transcriptionfactor hypoxia inducible factor-1α (HIF-1α). This smallmoleculeactivates the transcription of several proangiogenicfactors, the most important of which is the cytokine vascular

4 Sarcoma

Geneticdamage

If repair mechanisms fail. . .

Apoptosis

Mediated by

Tumor suppressors

ChemotherapyRadiation

P-glycoprotein

Repair mechanisms fail, but. . .

SurvivalMediated by

p16 inactivating mutationsChondrosarcoma cell

Normal cell

Proapoptotic proteins

Antiapoptotic proteins

O2•−

Figure 1: An overview of how chondrosarcoma cells evade the cytotoxic effects of chemotherapy and radiation. In normal cells, radiationand chemotherapy cause cell death by inducing genetic damage either directly, or in the case of radiation, through a reactive oxygen species(ROS) intermediate. For example, the drug doxorubicin intercalates with DNA, preventing replication, while ROS cause strand breaks. Thisdamage is sensed by the cell, and then through the actions of tumor suppressor proteins such as p16 or p53 and the proapoptotic proteinsincluding Bax, Bak, and Bim, the cell undergoes apoptosis, becomes senescent, or necroses. The chondrosarcoma cell’s main defense againstchemotherapeutic agents is P-glycoprotein, a membrane-bound pump that extrudes small, hydrophobic molecules from within the cell[10]. The action of P-glycoprotein can lower intracellular concentrations of chemotherapeutic agents beyond a point at which they exacttheir cytotoxic effects. Though radiation treatment still induces genetic damage in chondrosarcoma cells, several mutations allow them tosurvive. These mutations include inactivation of the gene encoding the important tumor suppressor p16 via methylation or deletion [11],and upregulation of the antiapoptotic proteins Bcl-2 and XIAP [12]. Figure adapted from Motifolio Cell and Nucleic Acid Toolkit.

endothelial growth factor (VEGF). Because tumors dependon angiogenesis for survival and further growth, this pathwayhas become a popular drug target in many cancers, andchondrosarcoma is no exception.

Immunostaining and siRNA experiments have been usedto demonstrate that chondrosarcoma tumors of all gradesexpress HIF-1α at higher levels than do normal chondrocytesand benign cartilaginous tumors [20], and that VEGFexpression is dependent on HIF-1α [21]. Another studyconcluded that in chondrosarcoma, HIF-1α expression wasonly present in 40% of tumors sampled; however, this studywas unique in staining for nuclear HIF-1α, the active form.Tumors positive for nuclear HIF-1α were associated withdecreased disease-free survival, suggesting that nuclear HIF-1α could serve as a prognostic factor in addition to histologicgrade [20].

In terms of therapeutic potential, inhibiting angiogenesismay be a promising strategy for inducing growth arrest asan adjuvant to surgical removal of chondrosarcoma tumors.Klenke et al. have demonstrated that inhibition of the angio-genic tyrosine kinase receptors for VEGF, fibroblast growthfactor (FGF), and platelet-derived growth factor (PDGF)with the small-molecule inhibitor SU6668 has a beneficial

effect on tumor size and neovascular density in vivo. In thisstudy, mice with chondrosarcoma xenografts implanted inthe calvarium and treated with SU6668 had tumors thatwere 53% smaller than those tumors in mice treated withcontrol injections at the end of the 28 day growth period.Additionally, tumors in mice treated with the inhibitor hadvessel densities reduced by 37% compared to untreated mice[22]. In another in vivo xenograft model, ecteinascidin-743(ET-743), a sea squirt-derived chemotherapeutic, elicitedprofound tumor necrosis and prevented neovascularizationwhen used in combination with plasminogen-related proteinB, an endogenous molecule that downregulates endothelialcell metabolism [23].

5. Beyond Traditional Chemotherapy:Alternative Drug Targets inChemoresistant Chondrosarcoma

Though they present unique treatment obstacles, chon-drosarcoma tumors also present unique treatment oppor-tunities because they maintain many of the phenotypiccharacteristics of the chondrocytes from which they are

Sarcoma 5

derived. One interesting area of research involves treatingthese phenotypic similarities as drug targets. Some excitingexamples of these targets include cyclooxygenase-2 (COX-2)expression, melovonate synthesis, and estrogen signaling.

Chondrocytes are known to express the inflammatorycytokine COX-2 when exposed to other inflammatorycytokines and free radicals [24]. COX-2 expression has beendemonstrated in peripheral chondrosarcoma tumors inseveral studies, and higher levels of expression are associatedwith poor prognosis [25, 26]. Schrage et al. hypothesized thatthere may be a role for COX-2 inhibition in chondrosarcomatreatment. Dr. Schrage demonstrated that the COX-2inhibitor Celecoxib decreased cell viability in vitro. Celecoxibslowed tumor growth in vivo initially. Interestingly, thesemechanisms were independent of COX-2 activity sinceseveral cell lines showed these responses but did not expressCOX-2 according to ELISA assays [25]. Further study isneeded in this area because the in vivo growth arrest relapsedafter six weeks of treatment.

Like Celecoxib, bisphosphonates are typically used in thetreatment of other musculoskeletal conditions, but may haveindications in the treatment of chondrosarcoma. Bispho-sphonates inhibit melovonate synthesis and are extremelyspecific to the bone microenvironment. The mechanism ofaction involves inhibiting osteoclast-mediated bone resorp-tion and the promotion of osteoblast differentiation. Inter-estingly, bisphosphonates may have an anticancer effectby inducing apoptosis, inhibiting metalloproteinase activity,and reducing VEGF levels [27]. It has been shown thatthe bisphosphonate zoledronic acid acts synergistically withpaclitaxel in osteosarcomas [28]. In the context of chon-drosarcoma, one case report demonstrates that zoledronicacid significantly reduces bone pain and improves the qualityof life of patients with chondrosarcoma and chordoma [29].Furthermore, another study showed that the third generationbisphosphonate minodronate decreases chondrosarcoma cellgrowth in a dose-dependent manner in vitro [30].

Finally, the role of estrogen signaling in skeletal devel-opment has implications for chondrosarcoma treatment.Estrogen signaling is involved in longitudinal bone growth,chondrocyte differentiation, and epiphyseal growth platematuration. Therefore, Cleton-Jansen et al. hypothesizedthat targeting estrogen signaling could inhibit chondrosar-comacell proliferation. Dr. Cleton-Jansen demonstrated thatinhibiting estrogen signaling with the aromatase inhibitorexemestane inhibits chondrosarcoma cell growth in vitro[31]. Estrogen signaling has been an important target inother cancers such as breast carcinoma, and this studysuggests that it may also be an important target in chon-drosarcoma.

6. Radiation Resistance in Chondrosarcoma

Radiation treatment is used widely and commonly in cancertherapy and incites its cytotoxic effects via the productionof reactive oxygen species (ROS). These ROS then causestrand breaks in DNA. Under normal physiologic condi-tions, healthy cells have DNA repair machinery that senses

ROS-induced genetic damage and initiates attempts to repairthis damage (see Figure 1). If the damage is irreparable,proteins called tumor suppressors will signal to the cell tostop dividing and the cell will either undergo apoptosis,necrose, or become senescent. These mechanisms preventgenetic damage from being passed to future generations.

The efficacy of radiation treatment is contingent uponseveral things: first, formation of ROS is necessary tomediate genetic damage; second, the genetic damage-sensingmechanisms must be intact; third, the tumor suppressionactivity of the cell must be functional. Radiation resistance intumor cells could feasibly develop at any one of these threesteps due to mutations in the myriad proteins involved inthese complex cascades. The exact mechanisms of radiationresistance in chondrosarcoma have remained somewhatunclear, though several culprits, which will be reviewed here,have been identified.

7. Loss of Tumor Suppressorp16 in Chondrosarcoma Leads toRadioresistance

The role of the loss of tumor suppressor genes is one well-known cause of radioresistance in numerous tumors. Inchondrosarcoma, it has been shown that genetic changesto the p16 coding gene CDKN2 are quite common inhigh-grade tumors. Asp et al. have shown using PCRand gene sequencing that 41% of chondrosarcoma tumorssampled had some sort of change to CDKN2, while notumors sampled had any change to the well-known p53tumor suppressor gene [11]. In addition, almost all ofthe highly malignant tumors sampled had alterations tothis gene, suggesting that changes to the CDKN2 gene areone important mechanism contributing to high malignancy.Such mutations to p16 are not seen in benign cartilaginoustumors such as enchondroma, suggesting a role for thisgene in the transition to an aggressive, malignant phenotype[32]. These findings support evidence linking altered p16expression to malignant phenotypes in other types of cancerincluding breast carcinoma and oropharyngeal squamouscell carcinoma [33, 34].

The role of p16 in mediating radioresistance was demon-strated by experiments that showed that reintroducing p16expression with a viral vector can increase radiosensitivityin vitro by inciting mitotic catastrophe [35]. While nomethod of restoring the functionality of the p16 pathwayhas been developed in vivo, research is underway to use thedefective p16/retinoblastoma (RB) pathway as the markerfor targeting viral vectors to cancer cells. A virus couldtheoretically be derived that would only infect cells with adefect in the pathway; alternatively, a virus could be designedthat would infect indiscriminately but would only replicateand/or produce its immune-modulatory product in tumorcells [36].

Of course, treatments using viral vectors present theirown issues such as ensuring viral genomic stability and selec-tivity for cancer cells versus healthy cells. Early trials of theuse of such an oncolytic virus in the treatment of many types

6 Sarcoma

of cancer are currently underway. This particular adenovirusselectively targets cells with p16/RB pathway dysfunctionand stimulates the production of granulocyte-macrophagecolony-stimulating factor (GM-CSF), an immune modulatorthat locally induces CD8+ T cell and natural killer (NK)cell activity. Though the initial results suggest that the virusis capable of increasing tumor chemosensitivity withoutcausing severe side effects, these trials have not includedchondrosarcoma patients [37].

8. The Antiapoptotic ProteinsBcl-2, Bcl-xL, and XIAP Can MediateRadioresistance

Within any cell, survival and proliferation are mediated bya delicate balance between pro- and antiapoptotic signals.When the balance is tipped in favor of antiapoptotic signals,mediated by proteins such as Bcl-2, Bcl-xL, and XIAP, thecell will survive, proliferate, differentiate, and migrate. Con-versely, when the balance is tipped in favor of proapoptoticsignals, such as Bax, Bak, and Bim, the cell will exit the cellcycle, disrupt its nucleus, and begin to break apart into smallparts commonly referred to as “blebs.”

Given that ionizing radiation causes cytotoxicity byinducing apoptosis, one theory regarding radiation resis-tance is that some cancer cells may tip their own internalbalance to favor the anti-apoptosis signals. That way, nomatter how much genetic damage is induced by radiation,the cell will continue to differentiate, divide, and survive.

Kim et al. have shown that chondrosarcoma cells expresshigher basal levels of the antiapoptotic proteins Bcl-2 andXIAP than do normal chondrocytes in vitro, though thetwo cell types express similar amounts of the antiapoptoticprotein Bcl-xL [38]. In addition, radiation induces increasedexpression of these proteins in a dose-dependent manner,suggesting an important role of antiapoptotic signalingin radioresistance [38]. Finally, Dr. Kim demonstrated anincrease in radiosensitivity by using siRNA to silence theexpression of these proteins—silencing them individuallycould lead to up to 9.2 fold increases in radiation sensitivity.Silencing two of these genes simultaneously enhanced sensi-tivity to 5 and 10 Gy doses of radiation by 11.3 and 11.2 fold,respectively [38].

Clinically, gene silencing of Bcl-2 has great poten-tial as an adjuvant therapy. In Phase III trials, it hasalready been shown that the antisense oligonucleotideG3139 combined with the chemotherapy agent dacarbazineincreased progression-free and overall survival in patientswith advanced melanoma when compared to patients treatedwith just dacarbazine alone [39]. Such clinical studies haveshown that treating humans with antisense oligonucleotidesagainst Bcl-2 is feasible and potentially beneficial.

Little is known about the possible additive or synergisticeffects of Bcl-2 antisense oligonucleotides in the contextof radiation treatment, as no clinical trials have beenundertaken. Several studies have shown promising resultsin xenograft models: Bcl-2 antisense oligonucleotides haveinherent antitumor activity and can enhance the effects of

local tumor irradiation in a colon carcinoma model [40] andin a nasopharyngeal carcinoma model [41].

Given these findings, it would appear that targeting theexpression of these antiapoptotic proteins is a promising newavenue of inducing radiation sensitivity in chondrosarcomatumors. Previous clinical studies have already established thefeasibility and safety of antisense oligonucleotide treatmentand pharmacologic inhibition in humans, though in vivostudies in a chondrosarcoma model are lacking.

9. Technological Advancesin Radiation Therapy: ImprovingOutcomes in Chondrosarcoma

Many improvements in the field of radiation oncology ingeneral can be applied to treatment of chondrosarcoma. Inshort, these advances can be classified into those that allowthe delivery of greater amounts of radiation to the tumor andthose that enhance the killing power of a given amount ofradiation. With regard to the former category, improvementsin proton therapy have allowed the safe delivery of moreradiation to tumors. Unlike traditional radiotherapy, protontherapy does not deliver radiation to tissue beyond the tumor[42]. This is especially important in tumors occurring inareas such as the spine and the skull (tumors which aretraditionally very difficult to resect with a wide marginand, as such, most often in need of radiation). Additionaladvances in this technology have allowed increased focusingof the radiation (“spot-scanning” technology), allowing theuse of more intense radiation without increasing the toxicityto the patients. This technology was recently tested inchondrosarcomaof the skull base, and the results suggestthat the technique is relatively safe and can be an effectiveadjuvant to surgery when used at high doses [43].

In addition to delivering more radiation directly totumors, improving the efficiency of a given dose of radiationis another strategy for dealing with radioresistant tumors.Because of the hypoxic conditions inherent to the microen-vironment of a chondrosarcoma tumor, the generationof ROS is severely curtailed. Acridine Orange (AO) iscapable of producing damaging oxidants in the presence ofgamma radiation even in oxygen-poor environments. Thiscapability makes AO an attractive addition to radiotherapyfor chondrosarcoma. Moussavi-Harami et al. have shownthat AO in and of itself is only slightly cytotoxic but inducessignificant increases in sensitivity to low-dose radiation invitro [35].

10. Conclusion

We have reviewed here the numerous mechanisms thatmake chondrosarcoma a challenging cancer to treat. AsTzu suggested, knowing the enemy is a vital part of anysuccessful confrontation. Clinical experience has shown usthat chondrosarcoma is strongly resistant to traditionalmeans of cancer treatment, save for dramatic surgeries. YetTzu also notes that “the way to avoid what is strong is to strikeat what is weak” [1]. Cutting edge research in the field of

Sarcoma 7

chondrosarcoma has exposed some of these key weaknesses.Exploiting these vulnerabilities will allow us to underminethe resistance that has made these tumors such a frustratingclinical entity and to offer much needed treatment options topatients who would otherwise face a very poor prognosis.

References

[1] S. Tzu, The Art of War, El Paso Norte Press, El Paso, Tex, USA,2005.

[2] http://www.mayoclinic.org/chondrosarcoma/.[3] S. W. Sung, R. Fabrizio, and F. Y. Lee, “Chondrosarcoma,”

in Orthopedic Knowledge Update Musculoskeletal Tumors 2, S.Herbert and M. D. Schwartz, Eds., pp. 185–195, AmericanAcademy of Orthopedic Surgeons, 2007.

[4] F. Y. Lee, H. J. Mankin, G. Fondren et al., “Chondrosarcomaof bone: an assessment of outcome,” Journal of Bone and JointSurgery A, vol. 81, no. 3, pp. 326–338, 1999.

[5] P. Garzino-Demo, G. Tanteri, P. Boffano et al., “Chondrosar-coma of the temporomandibular joint: a case report andreview of the literature,” Journal of Oral and MaxillofacialSurgery, vol. 68, no. 8, pp. 2005–2011, 2010.

[6] N. Kawahara, K. Tomita, H. Murakami, S. Demura, K. Yosh-ioka, and T. Miyazaki, “Total excision of a recurrent chon-drosarcoma of the thoracic spine: a case report of a seven-year-old boy with fifteen years follow-up,” Spine, vol. 35, no. 11, pp.E481–E487, 2010.

[7] R. Wesolowski and G. T. Budd, “Use of chemotherapy forpatients with bone and soft-tissue sarcomas,” Cleveland ClinicJournal of Medicine, vol. 77, supplement 1, pp. S23–S26, 2010.

[8] R. Rosier, R. J. O’Keefe, L. A. Teot et al., “P-glycoproteinexpression in cartilaginous tumors,” Journal of Surgical Oncol-ogy, vol. 65, no. 2, pp. 95–105, 1997.

[9] R. M. Terek, G. K. Schwartz, K. Devaney et al., “Chemotherapyand P-glycoprotein expression in chondrosarcoma,” Journal ofOrthopaedic Research, vol. 16, no. 5, pp. 585–590, 1998.

[10] M. Kawakami, T. Matsumoto, K. Kuribayashi, and T. Tamaki,“Multidrug resistance-1 and P-glycoprotein in human chon-drosarcoma cell lines: expression correlates with decreasedintracellular doxorubicin and in vitro chemoresistance,” Jour-nal of Orthopaedic Research, vol. 17, no. 6, pp. 935–940, 1999.

[11] J. Asp, L. Sangiorgi, S. E. Inerot et al., “Changes of the p16gene but not the p53 gene in human chondrosarcoma tissues,”International Journal of Cancer, vol. 85, no. 6, pp. 782–786,2000.

[12] D. W. Kim, K. O. Kim, M. J. Shin et al., “siRNA-basedtargeting of antiapoptotic genes can reverse chemoresistancein P-glycoprotein expressing chondrosarcoma cells,” MolecularCancer, vol. 8, article 28, 2009.

[13] P. M. Chaudhary and I. B. Roninson, “Induction of multidrugresistance in human cells by transient exposure to differentchemotherapeutic drugs,” Journal of the National CancerInstitute, vol. 85, no. 8, pp. 632–639, 1993.

[14] M. Z. Levy, R. C. Allsopp, A. B. Futcher, C. W. Greider, and C.B. Harley, “Telomere end-replication problem and cell aging,”Journal of Molecular Biology, vol. 225, no. 4, pp. 951–960, 1992.

[15] J. W. Shay and S. Bacchetti, “A survey of telomerase activity inhuman cancer,” European Journal of Cancer Part A, vol. 33, no.5, pp. 787–791, 1997.

[16] J. A. Martin, B. R. DeYoung, S. Gitelis et al., “Telom-erase reverse transcriptase subunit expression is associatedwith chondrosarcoma malignancy,” Clinical Orthopaedics andRelated Research, no. 426, pp. 117–124, 2004.

[17] J. A. Martin, E. Forest, J. A. Block et al., “Malignanttransformation in human chondrosarcoma cells supported bytelomerase activation and tumor suppressor inactivation,” CellGrowth and Differentiation, vol. 13, no. 9, pp. 397–407, 2002.

[18] D. Parsch, U. Brassat, T. H. Brummendorf, and J. Fellenberg,“Consequences of telomerase inhibition by BIBR1532 onproliferation and chemosensitivity of chondrosarcoma celllines,” Cancer Investigation, vol. 26, no. 6, pp. 590–596, 2008.

[19] C. Harley, “Telomerase and cancer therapeutics,” NatureReviews—Cancer, vol. 8, pp. 167–179, 2008.

[20] T. Kubo, T. Sugita, S. Shimose, T. Matsuo, K. Arihiro, and M.Ochi, “Expression of hypoxia-inducible factor-1α and itsrelationship to tumour angiogenesis and cell proliferation incartilage tumours,” Journal of Bone and Joint Surgery - Series B,vol. 90, no. 3, pp. 364–370, 2008.

[21] C. Lin, R. McGough, B. Aswad, J. A. Block, and R. Terek,“Hypoxia induces HIF-1α and VEGF expression in chon-drosarcoma cells and chondrocytes,” Journal of OrthopaedicResearch, vol. 22, no. 6, pp. 1175–1181, 2004.

[22] F. M. Klenke, A. Abdollahi, E. Bertl et al., “Tyrosine kinaseinhibitor SU6668 represses chondrosarcoma growth viaantiangiogenesis in vivo,” BMC Cancer, vol. 7, article 49, 2007.

[23] H. Morioka, L. Weissbach, T. Vogel et al., “Antiangiogenesistreatment combined with chemotherapy produces chon-drosarcoma necrosis,” Clinical Cancer Research, vol. 9, no. 3,pp. 1211–1217, 2003.

[24] S. Pallu, P. J. Francin, C. Guillaume et al., “Obesity affectsthe chondrocyte responsiveness to leptinin patients withosteoarthritis,” Arthritis Research & Therapy, vol. 12, p. R112,2010.

[25] Y. M. Schrage, I. Machado, D. Meijer et al., “COX-2 expressionin chondrosarcoma: a role for celecoxib treatment?” EuropeanJournal of Cancer, vol. 46, no. 3, pp. 616–624, 2010.

[26] M. Endo, T. Matsumura, T. Yamaguchi et al., “Cyclooxy-genase-2 overexpression associated with a poor prognosis inchondrosarcomas,” Human Pathology, vol. 37, no. 4, pp. 471–476, 2006.

[27] D. Santini, B. Vincenzi, G. Avvisati et al., “Pamidronateinduces modifications of circulating angiogenetic factors incancer patients,” Clinical Cancer Research, vol. 8, no. 5, pp.1080–1084, 2002.

[28] B. Kubista, K. Trieb, F. Sevelda et al., “Anticancer effects ofzoledronic acid against human osteosarcoma cells,” Journal ofOrthopaedic Research, vol. 24, no. 6, pp. 1145–1152, 2006.

[29] L. Montella, R. Addeo, V. Faiola et al., “Zoledronic acid inmetastatic chondrosarcoma and advanced sacrum chordoma:two case reports,” Journal of Experimental and Clinical CancerResearch, vol. 28, no. 1, article 7, 2009.

[30] T. Kubo, S. Shimose, T. Matsuo et al., “Inhibitory effects ofa new bisphosphonate, minodronate, on proliferation andinvasion of a variety of malignant bone tumor cells,” Journalof Orthopaedic Research, vol. 24, no. 6, pp. 1138–1144, 2006.

[31] A. M. Cleton-Jansen, H. M. Van Beerendonk, H. J. Baelde, J. V.G. M. Bovee, M. Karperien, and P. C. W. Hogendoorn, “Estro-gen signaling is active in cartilaginous tumors: implicationsfor antiestrogen therapy as treatment option of metastasizedor irresectable chondrosarcoma,” Clinical Cancer Research, vol.11, no. 22, pp. 8028–8035, 2005.

[32] H. M. van Beerendonk, L. B. Rozeman, A. H. M. Taminiau etal., “Molecular analysis of the INK4A/INK4A-ARF gene locusin conventional (central) chondrosarcomas and enchondro-mas: indication of an important gene for tumour progression,”Journal of Pathology, vol. 202, no. 3, pp. 359–366, 2004.

8 Sarcoma

[33] O. L. Bohn, M. Fuertes-Camilo, L. Navarro, J. Saldivar, andS. Sanchez-Sosa, “p16 INK4a expression in basal-like breastcarcinoma,” International Journal of Clinical and ExperimentalPathology, vol. 3, no. 6, pp. 600–607, 2010.

[34] P. M. Weinberger, Z. Yu, B. G. Haffty et al., “Prognosticsignificance of p16 protein levels in oropharyngeal squamouscell cancer,” Clinical Cancer Research, vol. 10, no. 17, pp. 5684–5691, 2004.

[35] F. Moussavi-Harami, A. Mollano, J. A. Martin et al., “Intrinsicradiation resistance in human chondrosarcoma cells,” Bio-chemical and Biophysical Research Communications, vol. 346,no. 2, pp. 379–385, 2006.

[36] E. A. Chiocca, “Oncolytic viruses,” Nature Reviews Cancer, vol.2, no. 12, pp. 938–950, 2002.

[37] A. Koski, L. Kangasniemi, S. Escutenaire et al., “Treatmentof cancer patients with a serotype 5/3 chimeric oncolyticadenovirus expressing GMCSF,” Molecular Therapy, vol. 18,no. 10, pp. 1874–1884, 2010.

[38] D. W. Kim, S. W. Seo, S. K. Cho et al., “Targeting of cell sur-vival genes using small interfering RNAs (siRNAs) enhancesradiosensitivity of grade II chrondrosarcoma cells,” Journal ofOrthopaedic Research, vol. 25, no. 6, pp. 820–828, 2007.

[39] A. Y. Bedikian, M. Millward, H. Pehamberger et al., “Bcl-2antisense (oblimersen sodium) plus dacarbazine in patientswith advanced melanoma: the oblimersen melanoma studygroup,” Journal of Clinical Oncology, vol. 24, no. 29, pp. 4738–4745, 2006.

[40] N. Wiedenmann, M. Koto, U. Raju, L. Milas, and K. A. Mason,“Modulation of tumor radiation response with G3139, a bcl-2antisense oligonucleotide,” Investigational New Drugs, vol. 25,no. 5, pp. 411–416, 2007.

[41] K. W. Yip, J. D. Mocanu, P. Y. B. Au et al., “Combination Bcl-2 antisense and radiation therapy for nasopharyngeal cancer,”Clinical Cancer Research, vol. 11, no. 22, pp. 8131–8144, 2005.

[42] S. J. Frank and U. Selek, “Proton beam radiation therapy forhead and neck malignancies,” Current Oncology Reports, vol.12, no. 3, pp. 202–207, 2010.

[43] C. Ares, E. B. Hug, A. J. Lomax et al., “Effectiveness and safetyof spot scanning proton radiation therapy for chordomas andchondrosarcomas of the skull base: first long-term report,”International Journal of Radiation Oncology Biology Physics,vol. 75, no. 4, pp. 1111–1118, 2009.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 842842, 10 pagesdoi:10.1155/2011/842842

Research Article

Involvement of the Soluble Urokinase Receptor inChondrosarcoma Cell Mobilization

Katia Bifulco,1 Immacolata Longanesi-Cattani,1 Maria Teresa Masucci,1

Annarosaria De Chiara,2 Flavio Fazioli,3 Gioconda Di Carluccio,1

Giuseppe Pirozzi,1 Michele Gallo,3 Antonello La Rocca,3 Gaetano Apice,4

Gaetano Rocco,3 and Maria Vincenza Carriero1

1 Department of Experimental Oncology, National Cancer Institute of Naples, Via Mariano Semmola 1, 80131 Naples, Italy2 Department of Human Pathology, National Cancer Institute of Naples, Via Mariano Semmola 1, 80131 Naples, Italy3 Department of Surgical Oncology, National Cancer Institute of Naples, Via Mariano Semmola 1, 80131 Naples, Italy4 Department of Medical Oncology, National Cancer Institute of Naples, Via Mariano Semmola 1, 80131 Naples, Italy

Correspondence should be addressed to Maria Vincenza Carriero, mariolinacarriero@yahoo.it

Received 26 August 2010; Revised 3 November 2010; Accepted 1 December 2010

Academic Editor: Charles Scoggins

Copyright © 2011 Katia Bifulco et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

High levels of urokinase receptor (uPAR) in tissue and serum of patients with chondrosarcoma correlate with poor prognosis.First, we analyzed the uPAR levels in tissues and plasma of five patients affected by chondrosarcoma. Interestingly, very highlevels of uPAR and its soluble forms (SuPAR) were found on tumor cell surfaces and plasma, respectively, of two patients withlung metastases. Therefore, to investigate the role of SuPAR in chondrosaromas, we generated a primary cell culture from achondrosarcoma tissue overexpressing uPAR on cell surfaces. We found that chondrosarcoma-like primary culture cells releasea large amount of SuPAR in the medium. In vitro, SuPAR elicits chondrosarcoma cell migration likely through its uPAR88−92

sequence, since the DII88−183 or DIIDIIR88−284 uPAR domains retain motogen effect whereas DI1−87 or DIII184−284 domains, bothlacking the uPAR88−92 sequence, are ineffective. Chondrosarcoma cells cross matrigel in response to SuPAR, and their invasioncapability is abrogated by RERF peptide which inhibits uPAR88−92 signalling. These findings assign a role to uPAR in mobilizingchondrosarcoma cells and suggest that RERF peptide may be regarded as a prototype to generate new therapeutics for thechondrosarcoma treatment.

1. Introduction

Chondrosarcomas are a heterogeneous group of neoplasmshaving in common the production of cartilage matrixby the tumor cells [1]. Chondrosarcoma accounts forapproximately 20% of bone sarcomas with a peak incidencein the fifth to seventh decade of life. Because there areno effective treatments for patients with unresectable ormetastatic disease, there is a pressing need to develop newtargeted approaches [2].

Chondrosarcomas can progress from low grade to highgrade, which is reflected by increased cellularity, nuclearatypia, mucomyxoid matrix changes, and increased vascu-larization [3, 4]. Low-grade chondrosarcomas are locallyaggressive but rarely metastasize [5]. By contrast, high-grade

chondrosarcomas often metastasize and are lethal in mostpatients [3]. The molecular mechanisms involved in theprogression to high-grade chondrosarcoma are beginning tobe unravelled [1]. Furthermore, processes such as tumor cellattachment, migration, and invasion, which are known to befundamental in carcinoma, have not been similarly exploredin chondrosarcoma.

Proteolytic degradation of the extracellular matrix con-stituents and modification of cell adhesion properties arerequired for tumor invasion and metastasis. The urokinaseplasminogen activator (uPA) system have an importantrole in tumorigenesis, extracellular matrix degradation,and migration and invasion of tumor cells [6–10]. Uponbinding to uPA, the cell-surface urokinase receptor (uPAR)elicits a variety of cell responses, including cell migration

2 Sarcoma

and invasion [11]. Many malignant cultured cell lines andhuman neoplasms have been characterized by their increaseduPAR expression [12, 13], thus suggesting that the inhibitionof uPAR activity could be a promising strategy to preventcancer invasion and metastasis.

uPAR is a glycosylated glycosylphosphatidylinositol-anchored protein [14] formed by three domains DI, DII, andDIII connected by short linker regions [15]. The Ser88-Arg-Ser-Arg-Tyr92 (uPAR88–92) linker region between DI and DIIdomains is a protease sensitive region which retains chemo-tactic activity, even in the form of an isolated SRSRY peptide[16, 17]. The flexibility of this region enables its interactionwith a wide variety of ligands [18]. uPAR engagementwith uPA favours the exposure of the uPAR88–92 sequencewhich, in turn, promotes cytoskeletal rearrangements anddirectional cell migration by binding to the G-protein-coupled formyl-peptide receptors (FPRs) [16, 17, 19]. Bya drug-design approach based on the conformational anal-ysis of the uPAR88–92 sequence, we have recently developeda family of peptides which revealed to be uPAR antagonistsin virtue of their ability to prevent uPAR/FPR interaction.Among these, we found that RERF peptide potently inhibitsin vitro and in vivo cell migration and invasion of highlyinvasive human fibrosarcoma HT1080 cells [20].

In tumor tissues, shedding and/or enzymatic cleavage ofuPAR generate truncated forms of uPAR (SuPAR), which aresecreted in the extracellular milieu [21–24]. Soluble formsof uPAR have been identified, in vitro, in conditioned cellculture medium and, in vivo, in serum and urine of patientsaffected by several type of solid tumors, including sarcomasand chondrosarcamas, and have been significantly associatedto a bad prognosis [21–25]. In particular, codetection of ahigh expression level of uPA, uPAR, and PAI-1 in tumourtissue and of SuPAR in serum of patients affected by soft-tissue sarcoma has been reported to significantly correlatewith a shortened overall survival [25]. To gain some insighton the role of soluble forms of uPAR in determining anaggressive phenotype of chondrosarcoma, we have analysedthe effects of SuPAR on a primary cell culture derived froman uPAR expressing chondrosarcoma case.

2. Materials and Methods

2.1. Patients, Tissue and Plasma Collection. Six patients withchondrosarcoma were studied. Surgical removed tumorswere routinely processed for the histopathological diagnosisperformed according to the WHO classification [26]. Arepresentative sample from each tumor excision was imme-diately frozen in liquid nitrogen and stored at −80◦C untilused for immunocytochemistry. A sample from the tumorexcision of patient no. 6 was immediately processed forpreparation of a primary cell culture. Plasma samples wereobtained just before surgery and were stored at −80◦C untilassayed.

2.2. Immunohistochemistry. Frozen sections, correspondingto the largest cross-sectional area of the tumor, were cut,placed on glass slides and subjected to immunostaining

with the streptavidin-biotin-peroxidase method, as pre-viously described [27]. Briefly, sections were fixed with2.5% formaldehyde in phosphate buffered saline (PBS)and incubated overnight at 4◦C with diluents (negativecontrol), or 2 μg/mL R4 anti-uPAR monoclonal antibody(mAb), kindly provided by G. Hoyer-Hansen (Finsen Insti-tute, Copenhagen, Denmark). After several washes in PBS,1 : 200 diluted biotinylated goat anti-mouse immunoglob-ulins were applied to sections at 23◦C for 60 min. There-after, sections were incubated with streptavidin-biotinylatedhorseradish peroxidase complex for additional 30 minand the peroxidase-dependent staining was developed bydiaminobenzidine. Slides were counterstained with Mayer’shaematoxylin.

2.3. Primary Cell Culture. A representative sample fromthe tumor excision (∼1 cm × 1 cm) from patient no. 6was immediately minced by scalpel under sterile conditionsand incubated with 1.0 mg/mL collagenase XI (Sigma) for3 h at 37◦C under gentle agitation, as previously described[28]. Cells, recovered by centrifugation at 1500 rpm, werecultured in 6-well multidish plates in Dulbecco ModifiedEssential Medium (DMEM) with the addition of 10% foetalbovine serum (FBS), 100 IU/mL penicillin and 50 μg/mLstreptomycin. Isolated cell clusters were further amplifiedin growth medium until an adherent, homogeneous cellpopulation was obtained.

2.4. Cell Lines and Conditioned Media. Mouse fibroblastLB6, LB6 cells stably transfected with cDNA encodinghuman SuPAR (LB6/hSuPAR) [12], and human fibrosar-coma HT1080 cell line were grown in DMEM supplementedwith 10%FBS, 100 IU/mL penicillin and 50 μg/mL strepto-mycin. To prepare conditioned media, LB6, LB6/hSuPAR,or chondrosarcoma cells were grown to 80% confluence on10 cm Ø plates. Growth medium was removed and cells, afterextensive washing with PBS, were incubated in serum-freemedium. After 18 h, the medium was recovered, cleared bycentrifugation, and analysed for the SuPAR content applyinga commercially available enzyme-linked immunosorbentassay kit (ELISA) purchased by R&D System, as previouslydescribed [12]. Antigen concentrations were expressed as nganalyte per μg proteins.

2.5. Immunofluorescence. Chondrosarcoma cells, plated onglass slides (30%–40% confluence), were fixed and perme-abilized with 2.5% formaldeyde-0.2% Triton X-100 in PBSfor 10 min at 4◦C, then incubated overnight at 4◦C with5 μg/mL anti-vimentin (Dako), anti-cytokeratin (ZymedLaboratories Inc.) or 2 μg/mL R4 anti-uPAR mAbs. Asubset of experiments was performed on fixed with 2.5%formaldeyde, nonpermeabilized cells. Immunofluorescencewas carried out by incubating slides with 1 : 100 diluted Alexa488-conjugated F(ab’)2 fragment of rabbit anti-mouse IgG(Molecular Probes) for 1h at 22◦C. After nuclear stainingwith 4′6-diamidino-2-phenylindole dye (DAPI), cells wereanalysed by a fluorescence inverted microscope connected toa videocamera (Carl Zeiss), as described [20].

Sarcoma 3

2.6. Flow Cytometry. Cells were detached using 200 mg/LEDTA, 500 mg/L trypsin (Cambrex). Nonspecific bindingsites, possibly due to any Fc receptor, were blocked by normalrat serum. Cells (0.5× 106 cells/sample) were incubated with1 : 40 normal rat serum added to PBS (CTL) or 2 μg/mLR4 anti-uPAR mAb for 30 min at 4◦C. After extensivewashing with PBS, cells were incubated with Alexa 488-conjugated F(ab’)2 fragment of rabbit anti-mouse IgG andfinally resuspended in 0.6 mL PBS. Samples were analysedby flow cytometry using a FACS Vantage cell sorter (Becton& Dickinson). All data were analysed using CellQuestsoftware.

2.7. Immunoprecipitation. Conditioned medium of LB6,LB6/hSuPAR or chondrosarcoma cells (500 μL/sample) wereprecleared with 10 μL Protein G-Sepharose (Ge-Healthcare)for 1 hr at 4◦C, immunoprecipitated with 100 μL rabbit399 anti-uPAR conjugated to sepharose beads (0.5 mg IgGper mL of beads) diluted 1 : 1 for 18 h at 4◦C. Beads werewashed and then boiled in SDS-PAGE sample buffer. Proteinswere separated by a 12.5% SDS-PAGE followed by Westernblotting with 2 μg/mL R4 anti-uPAR mAb.

2.8. Cell Migration and Invasion Assays. Cell migration andinvasion assays were performed using Boyden chambersand 8 μm pore size polyvinyl-pyrrolidone-free polycarbonatefilters (Nucleopore) as previously described [20, 27]. Theability of primary cell culture to migrate or to cross matrigelwas assessed between the VI and the IX passage. For cellinvasion assays, filters were coated with 50 μg/mL matrigel,a reconstituted basement membrane (BD Biosciences). Cellswere preincubated with DMEM, 2 μg/mL normal rabbitserum (NRS), blocking 399 anti-uPAR Ab [29, 30], oranti-uPAR84–95 Ab which specifically recognizes the uPARchemotactic Ser88-Arg-Ser-Arg-Tyr92 sequence [31] for 1 hat 37◦C, prior to seeding in the upper chamber at 3 × 104

cells/well. In a subset of experiments, cells were exposedto 10 nM RERF or ERFR peptides (Primm) which wehave previously reported to inhibit uPAR88–92-dependentsignalling without affecting cell proliferation [20]. Theindicated chemoattractants were placed in the bottom well.Recombinant uPAR domains (Calbiochem) were employedat 10 nM concentration. Cells were allowed to migrate orinvade matrigel at 37◦C in humidified air with 5% CO2

for 4 h or 18 h, respectively. At the end of the assay, cellsin the upper chamber and on the upper filter surface wereremoved whereas cells on the lower filter surface were fixedwith ethanol and stained with haematoxylin. The numberof migrating or invading cells was determined by count-ing cells in 10 random fields/filter at 200x magnification.HT1080 cells were employed as an internal control. Datawere calculated as a percentage of migrated or invadingcells in the absence of chemoattractant, considered as100%.

2.9. Statistical Analysis. The data were analysed for signif-icance using Student’s t-test. Differences were consideredstatistically significant at a level of P < .05.

3. Results and Discussion

3.1. uPAR Expression and SuPAR Release in Chondrosarcomas.The age of the patients at diagnosis ranged from 34 to 72years. Surgical removed tumors were routinely processed forthe histopathological diagnosis performed according to theWHO classification [26]. Table 1 reports the pathologicalfindings of 6 primary chondrosarcomas: 5 were primarybone lesions including femur (3) and sternum (2) and 1was extraskeletal lesion involving pelvis (Table 1). The mainclinical features at diagnosis are summarized in Table 2.To investigate the molecular mechanisms underlying theactivity of uPAR in chondrosarcoma, we first analysed theuPAR expression on chondrosarcoma tissues by immuno-histochemistry, using R4 anti-uPAR mAb. The intensity ofuPAR staining of tumor cells was graded as faint (grading1), moderate (grading 2), or intense (grading 3) (Table 2).Except for the benign/low grade lesion which did not showany reactivity to R4 anti-uPAR mAb, all tumors, althoughat a different extent, exhibit a heterogeneous pattern ofstaining, mainly localized on tumor plasma cell membranes(Figure 1). Several tumor cells have been reported to shedsoluble forms of uPAR [21–25]. Therefore, we performed aquantitative analysis of the of SuPAR content in the plasma ofpatients using a commercially Elisa Kit. As shown in Table 2,an appreciable amount of SuPAR was detected in all theplasma tested. Interestingly, patients with lung metastases(#3 and #4) exhibited higher levels of SuPAR (Table 2). Thesedata encouraged us to further analyze the role of SuPAR inchondrosaroma invasiveness.

The patient no. 6 underwent a surgery for an extensivesternal mass (Table 1). Preoperative workup showed bilateralnodules suspicious for pulmonary metastases. Followingmultidisciplinary consultation, it was decided to submitthe patient to complete sternal resection and pulmonarymetastasectomy. For the reconstruction of the anterior chestwall defect, three cadaveric cryopreserved ribs were used. Theresected tumor measured 18 × 15 × 8 cm. Histology of theprimary tumor yielded a diagnosis of grade 2, focally grade3 chondrosarcoma, characterized by frank hypercellularity,with elongated hyperchromatic and sometimes binucleatednuclei (Figure 2(a)). The pulmonary lesions were confirmedto be pulmonary metastases from chondrosarcoma. Thepatient was followed up to 8 months after primary surgery,when multiple extraskeletal metastases were detected. At thispoint, he received chemotherapy, but died soon after due tochemotherapy-related complications.

3.2. Isolation and Characterization of Chondrosarcoma Cells.As described in the methods, a representative sample fromthe tumor excision was minced and subjected to enzymaticdigestion; cell suspension was recovered and cultured inmultidish plates until to the third passage (Figure 2(b)). Sub-cloning of the isolated cell clusters (Figure 2(c)) and sevenfurther passages resulted in an adherent, homogeneous cellpopulation mainly characterized by small chondrosarcoma-like cells (Figure 2(d)), resembling, in shape and size, thoseobserved on haematoxilin/eosin stained section (Figures 3(a)and 3(b)). Immunocytochemical analysis of cells grown on

4 Sarcoma

No. 1

(a)

No. 2

(b)

No. 3

(c)

No. 4

(d)

No. 5

(e)

No. 6

(f)

Figure 1: uPAR protein is expressed in chondrosarcoma tissues. Immunohistochemistry was performed on frozen sections with thestreptavidin-biotin-peroxidase method using 2 μg/mL R4 anti-uPAR mAb. Sections were counterstained with haematoxylin. Originalmagnifications: x400.

Table 1: Histopathological findings of enrolled chondrosarcoma patients.

Patients Age (yr) Gender Site Size (cm) Histologya Grade

1 42 FRightfemur

5.5 × 2 × 2.5 Low-grade chondrosarcoma

2 34 F Pelvis 10 × 8 × 8 Mesenchymal chondrosarcoma G2

3 72 MLeft

femur21 × 12 × 12 Dedifferentiated chondrosarcoma G3

4 63 FRightfemur

13 × 9 × 28 Dedifferentiated chondrosarcoma G3

5 69 M Sternum 10 × 7 × 5 Chondrosarcoma G2, focallyG3

6 58 M Sternum 18 × 15 × 8 Chondrosarcoma G2, focallyG3aHistopathological diagnosis was performed according to the WHO classification. F, female; M, male.

Table 2: Clinicopathological parameters and uPAR levels in tumour tissues and in plasma of patients with chondrosarcoma.

PatientsFirst clinicalevaluation

TherapySurvival from

diagnosis (months)uPAR

gradingaSuPAR

(pg/mL)b

1 P Surgery 6 0 1094

2 P Surgery 87 1 1.508

3 P + Lung M Surgery 3 3 3.645

4 P + Lung MNeoadiuvant chemotherapyand surgery

14 3 5.152

5 P Surgery 9 2 1.556

6 P + Lung M Surgery and chemotherapy 8 3 ND

P, primary tumor; M, metastasis; aImmunohistochemical staining of tumor frozen sections with R4 anti-uPAR mAb was graded as absent (grading 0), faint(grading 1), moderate (grading 2), or intense (grading 3). bDetermination of plasmatic SuPAR content by Elisa, expressed as pg SuPAR/mL plasma.

Sarcoma 5

(a) (b)

(c)(d)

Figure 2: Generation of a primary cell culture from a chondrosarcoma tissue. (a) Haematoxylin/eosin stained section of a human primarygrade 2, focally grade 3 chondrosarcoma of sternum (patient no. 6). (b–d) A representative sample from the tumor excision was subjected toenzymatic digestion. The isolated cells, recovered by centrifugation at 1500 rpm, were cultured in DMEM 10% FBS until to the third passage.At this passage, cell population showed an evident cellular heterogeneity (b). Subcloning of the isolated cell clusters (c) and amplificationfor six further passages resulted in an adherent, homogeneous cell population characterized by small, chondrosarcoma-like cells (d). Imageswere captured by an inverted microscope connected to a video camera. Original magnification: x200 (a and c), x400 (b and d).

glass slides revealed the total absence of epithelial (cytoker-atin) cell marker whereas a strong staining was observed inthe 95% of cells exposed to anti-vimentin mAb (Figures 3(c)and 3(d)).

3.3. Identification of Membrane-Anchored uPAR on Chon-drosarcoma Cells. First, uPAR expression was analysed inchondrosarcoma cells by immunofluorescence and cytoflu-orimetry. Accordingly to immunohistochemical findings(Figure 1), chondrosarcoma cells express high levels of uPARmainly localized on plasma cell membranes (Figure 4(a)). Inpermeabilized cells, a discrete, intracytoplasmic amount ofuPAR was also found, thus indicating that chondrosarcomacells effectively synthesize uPAR (Figure 4(b)). Interestingly,flow cytometry revealed that chondrosarcoma cells expresshigher levels of uPAR as compared to HT1080 cells (25 and23,9 mean fluorescence intensity, resp., in the 98% of totalcells) (Figure 4(c)).

3.4. Chondrosarcoma Cells Shed Soluble Forms of uPAR intothe Medium. We investigated whether soluble forms of uPARare produced and released by chondrosarcoma cells in theculture medium. To this purpose, we took advantage byemploying, as a negative and a positive controls, conditionedmedium of uPAR lacking wild-type LB6 and transfected

LB6/hSuPAR cells, respectively. We measured by Elisa theamount of SuPAR antigen released in the conditionedmedium of LB6, LB6/hSuPAR and chondrosarcoma cells.As expected, LB6/hSuPAR cells released a large amount ofSuPAR as compared to the wild-type LB6 cells [12]. Inkeeping with the expression of high levels of uPAR on plasmacell membranes, chondrosarcoma cells released a very largeamount of SuPAR in the medium as compared to thatproduced by LB6/hSuPAR cells (10.3 ng/μg and 5.2 ng/μg ofproteins, resp.) (Figure 4(d)). To ascertain the occurrenceof cleaved forms of SuPAR in the conditioned medium ofchondrosarcoma cells, serum-free medium was subjectedto immunoprecipitation with rabbit 399 anti-uPAR Abwhich recognizes all soluble forms of SuPAR, followed byWestern blotting with R4 anti-uPAR mAb which recognizesfull-length DIIDIII and DIII cleaved forms of uPAR [24].LB6 and LB6/hSuPAR conditioned media were employedas negative and positive control, respectively. According toSidenius et al. [24], we found in the conditioned mediumof both chondrosarcoma and LB6/hSuPAR, but not LB6cells, a fragment with an approximate 45 kDa molecularweight, comigrating with the purified full-length SuPAR(Figure 4(d)). R4 anti-uPAR mAb specifically recognizedin the conditioned medium of LB6/hSuPAR but not LB6cells an additional fragment having molecular weight ofabout 35 kDa, compatible with the cleaved DIIDIII fragment

6 Sarcoma

H & E

(a)

H & E

(b)

CK

(c)

Vimentin

(d)

Figure 3: Immunophenotyping of chondrosarcoma cells. (a, b) haematoxilin/eosin stained chondrosarcoma cells resemble in shape andsize those of chondrosarcoma tissue section. (c, d) chondrosarcoma cells grown on glass slides to semiconfluence were stained with anti-cytokeratin (CK) or anti-vimentin mAbs and with Alexa 488-conjugated F(ab’)2 fragment of rabbit anti-mouse IgG (green). Nuclei werestained blue with DAPI. Original magnification: x400.

of SuPAR. As shown in Figure 4(e), a fragment with anapproximate 35 kDa molecular weight, comigrating with theDIIDIII fragment of SuPAR was found in the conditionedmedium of chondrosarcoma cells. Although we can notassess whether other fragments of SuPAR do exist in thechondrosarcoma cell medium, this findings indicate thatat least two fragments, both containing the chemotacticsequence of uPAR, may influence cell behaviour.

3.5. SuPAR Promotes In Vitro Migration of Human Chon-drosarcoma Cells. Soluble forms of uPAR, including theDIIDIII fragment, have been shown to strongly chemoattracta variety of cell types [16, 17, 20, 32]. We investigatedwhether chondrosarcoma cells specifically respond to solubleforms of uPAR. In these experiments, we took advantageby employing the uPAR expressing human fibrosarcomaHT1080 cells as an internal control. Cell migration assayswere carried in Boyden chambers using conditioned media ofwild-type LB6 or LB6/hSuPAR cells. We found that similarlyto HT1080, chondrosarcoma cells exhibited a strong abilityto migrate toward the LB6/hSuPAR conditioned medium,reaching 363 ± 26% of the random cell migration (Table 3).On the contrary, wild-type LB6 conditioned mediumdid not exert any effect, indicating that chondrosarcomacells specifically respond to SuPAR. Interestingly, 10 nMrecombinant DII88–183 or DIIDIII88–284 uPAR domains aswell as full length DIDIIDIII1–284 triggered an appreciablecell migration (215% ± 5, 221% ± 13, and 240% ± 10,

resp.) whereas recombinant DI1–87 or DIII184–284 uPARdomains, both lacking the uPAR88–92 sequence, wereineffective at 10nM (Table 3). These findings suggest thatSuPAR is able to mobilize chondrosarcoma cells through itsuPAR88–92 sequence. To further investigate the role of theuPAR88–92 sequence in mobilizing chondrosarcoma cells, asubset of experiments was performed in the presence of 399anti-uPAR Ab or a polyclonal antibody which specificallyrecognized the uPAR84–95 sequence [30, 31]. As shown inTable 3, in the presence of 399 anti-uPAR or anti- uPAR84–95

Abs, both HT1080 and chondrosarcoma cells failed torespond to the conditioned medium of LB6/hSuPAR cells orto recombinant uPAR domains containing the chemotacticsequence. The inhibition was specific, as the presenceof nonimmune serum did not abrogate cell motility. Alltogether, these results clearly indicate that soluble formsof uPAR containing the uPAR88–92 sequence are able tomobilize chondrosarcoma cells.

3.6. uPAR Promotes In Vitro Invasion of Human Chon-drosarcoma Cells. Cell migration is a prerequisite for cancerinvasion. Therefore, we performed in vitro matrigel inva-sion assays [33] to quantify, the relative invasive potentialof chondrosarcoma cells. In these assays, 10% FBS wasemployed as a source of chemoattractants and the basal cellinvasion, assessed in the absence of any chemoattractant, wastaken as 100%. We employed the uPAR expressing humanfibrosarcoma HT1080 cells as an internal control. According

Sarcoma 7

(a)

(b)

Cou

nts

Log FL1 intensity

(c) (e)

IPPT : 399 WB : R4Mr ×103

SuPA

R

Non

e

LB

6

LB6/

hSu

PAR

Ch

ondr

osar

com

a

0 5 10

(d)

Con

diti

oned

med

ia

SuPAR content(ng/μg proteins)

LB6

∗

∗

LB6/hSuPAR

Chondrosarcoma

46−30−

66−

100 101 102 103 104

0

30

60

90

120

150

Figure 4: Chondrosarcoma cells express uPAR and release its soluble forms in the medium. Chondrosarcoma cells were fixed with 2.5%formaldeyde (a) or fixed with 2.5% formaldeyde and permeabilized with 0.2% Triton X-100 (b) for 10 min at 4◦C, incubated overnight at4◦C with 2 μg/mL R4 anti-uPAR mAb and then exposed to Alexa 488-conjugated F(ab’)2 fragment of rabbit anti-mouse IgG (green). Nucleiwere stained blue with DAPI. Original magnification: x1000. (c) Flow cytometry analysis of uPAR on chondrosarcoma and HT1080 cellsurfaces. HT1080 and chondrosarcoma cells were harvested, incubated with normal rat serum added to PBS (green and blue curves, resp.)or 2 μg/mL R4 anti-uPAR mAb (black and red curves, resp.), stained with Alexa 488-conjugated F(ab’)2 rabbit anti-mouse IgG and analyedby FACS. (d) Chondrosarcoma cells release soluble forms of uPAR in the medium. LB6, LB6/hSuPAR, or chondrosarcoma cells were grown to80% confluence. Growth medium was removed, cells were extensively washed with PBS and then incubated in serum-free medium for 18 h.The medium was recovered, cleared by centrifugation, and analysed for SuPAR content by ELISA. Antigen concentrations were expressed asng SuPAR per μg proteins. Columns, mean of two independent experiments; bars,±SD, ∗P < .001 against the control (conditioned mediumof LB6 wild-type cells). (e) 500 μL conditioned media of LB6, LB6/hSuPAR or chondrosarcoma cells were immunoprecipitated with 100 μLrabbit 399 anti-uPAR Ab conjugated to sepharose beads for 18 h at 4◦C. The eluted proteins were separated by a 12.5% SDS-PAGE andanalysed by Western blot using R4 anti-uPAR mAb. 1 μg purified SuPAR or 10 μL 399 anti-uPAR conjugated to sepharose beads (none) wereloaded as a control. Arrows indicate the full-length SuPAR and the DIIDIII fragment of SuPAR.

to their previously reported highly invasive capability [20],HT1080 cells exhibited a strong ability to cross matrigel (597± 30% of the basal level). Chondrosarcoma cells exhibited avery high ability to invade matrigel as compared to HT1080cells, reaching 721 ± 91% of the basal level (Figure 5(a)).Interestingly, cell exposure to blocking 399 anti-uPAR oranti-uPAR84–95 Abs strongly reduced cell invasion ability ofboth HT1080 and chondrosarcoma cells. To further elucidatethe role of uPAR in promoting cell invasiveness, a subsetof experiments were performed in the presence of RERFpeptide which we have previously reported to specificallyinhibit uPAR88–92-dependent signalling without affecting cellproliferation [20]. We found that cell exposure to 10 nM

RERF strongly reduced the ability of both HT1080 andchondrosarcoma cells to cross matrigel whereas the ERFRcontrol peptide failed to exert any inhibitory effect. Thesefindings strongly support the role of uPAR in promotingcell invasion. To assess whether also soluble forms of uPARmay be involved in cell invasion capability, experimentswere performed using conditioned medium of LB6/hSuPARcells as a source of intact and DIIDIII uPAR fragment, orconditioned medium of LB6 cells as a negative control. Wefound that both HT1080 and chondrosarcoma cells respondto LB6/hSuPAR conditioned medium, the effect being pre-vented by 399 anti-uPAR as well as by anti-uPAR84–95 Abs(Figure 5(b)).

8 Sarcoma

Table 3: SuPAR-dependent migration of chondrosarcoma cells.

Supplements EffectorCell migration (%)

HT1080 cells Chondrosarcoma cells

DMEM

DMEM 100 100

CM LB6 110 ± 5 112 ± 8

CM LB6/SuPAR 401 ± 12∗∗ 363 ± 16∗∗

10nM DI1–87 115 ± 6 105 ± 8

10nM DII88–183 265 ± 9∗∗ 215 ± 5∗∗

10nM DIII184–284 98 ± 9∗ 100 ± 6

10nM DIIDIII88–284 259 ± 11∗∗ 221 ± 13∗

10nM DIDIIDIII1–284 270 ± 15∗ 240 ± 10∗∗

NRS

DMEM 100 ± 3 100 ± 2

CM LB6/SuPAR 398 ± 9∗∗ 351 ± 6∗∗

10nM DII88–183 271 ± 5∗∗ 225 ± 11∗∗

10nM DIIDIII88–284 262 ± 9∗∗ 200 ± 8∗∗

10nM DIDIIDIII1–284 250 ± 9∗ 236 ± 13∗

399 Anti-uPAR Ab

DMEM 100 ± 7 100 ± 5

CM LB6/SuPAR 145 ± 12 133 ± 10

10nM DII88–183 115 ± 6 124 ± 9

10nM DIIDIII88–284 106 ± 11 99 ± 13

10nM DIDIIDIII1–284 110 ± 17 102 ± 14

Anti-uPAR84–95 Ab

DMEM 100 ± 6 100 ± 5

CM LB6/SuPAR 124 ± 9 130 ± 6

10nM DII88–183 104 ± 13 101 ± 7

10nM DIIDIII88–284 102 ± 9 119 ± 15

10nM DIDIIDIII1–284 107 ± 11 110 ± 13

Cells incubated with diluents (DMEM) or 5 μg/mL the indicated antibody for 1 h at 37◦C were seeded in Boyden chambers for cell migration assays asdescribed in the Material and Method section, in the presence or absence of the indicated effectors. Conditioned medium (CM) of LB6/hSuPAR cells was usedas a source of SuPAR. For quantitative analysis of cell migration, the basal value (DMEM) was taken as 100% and all values were reported relative to that. Dataare the means ± SD of two independent experiments, performed in duplicate. Statistical significance with P values was calculated against the control DMEM.∗Statistical significance with P < .05. ∗∗Statistical significance with P < .001.

4. Conclusions

uPAR plays a key role in pathological processes sustainedby an altered cell migration [11]. High levels of uPARand SuPAR in tissue and serum of patients with sarcoma,including chondrosarcoma, correlate with a poor prognosis[25]. We generated a primary cell culture derived froma uPAR overexpressing chondrosarcoma tissue. We foundthat chondrosarcoma-like primary culture cells express highlevel of uPAR on plasma cell membranes and release alarge amount of intact SuPAR as well as DIIDIII uPARfragment in the medium. Our findings revealed that,in vitro, SuPAR: (i) elicits chondrosarcoma cell migra-tion through its uPAR88–92 sequence and (ii) promoteschondrosarcoma cell invasion, the effect being reducedto the basal level by anti-uPAR blocking antibody. Takentogether, our findings raises the possibility that solubleuPAR released by chondrosarcoma cells in the extracel-lular matrix may generate a chemotactic gradient which,in turn, stimulates tumor cells to migrate and invade

the surrounding tissues. There are few studies investigatingthe clinical impact of uPAR expression and its correlationto prognosis in chondrosarcoma [25]. Our findings sug-gest that the determination of plasmatic SuPAR contentin patients with chondrosarcoma could be helpful for aprognostic evaluation. Furthermore, there are currently nouniversally effective therapies for unresectable or metasta-sized chondrosarcomas [1]. It can be envisaged that theinhibition of uPAR activity could be a promising strategyto prevent chondrosarcoma invasion and metastasis. Wefound that, in vitro, the chondrosarcoma cell invasionability may be abrogated by RERF peptide which specificallyinhibits the uPAR88–92-dependent signalling by preventingits interaction with the G-protein coupled formyl-peptidereceptor [20]. In conclusion, our data indicate that uPAR isrequired for the cell migration and invasion machinery ofchondrosarcoma cells and suggest that RERF peptide maybe regarded as a prototype to generate new therapeuticagents for the treatment of unresectable or metastasizedchondrosarcoma.

Sarcoma 9

∗∗

∗∗

0100200300400500600700800

DM

EM

DM

EM

10%

FBS

10%

FBS

DM

EM

10%

FBS

DM

EM

10%

FBS

DM

EM

10%

FBS

399 anti-uPAR Ab

RERF ERFR

Mat

rige

lcel

linv

asio

n(%

)

None anti-uPAR84−95

(a)

∗∗

0

100

200

300

400

500

600

700

800

Mat

rige

lcel

linv

asio

n(%

)

HT1080 cellsChondrosarcoma cells

DM

EM

CM

LB6

CM

LB6/

hSu

PAR

CM

LB6/

hSu

PAR

CM

LB6/

hSu

PAR

DM

EM

CM

LB6

DM

EM

CM

LB6

None 399 anti-uPAR Ab anti-uPAR84−95

(b)

Figure 5: uPAR-dependent matrigel invasion of chondrosarcoma.Cells human fibrosarcoma HT1080 and chondrosarcoma cellswere subjected to cell invasion assays. Cells were preincubatedwith DMEM (none), 2 μg/mL indicated anti-uPAR antibodies,10 nM RERF peptide or 10 nM ERFR peptide, plated in Boy-den chambers and allowed to cross matrigel-coated filters for18 h at 37◦C in humidified air with 5% CO2. 10% FBS (a),LB6, or LB6/hSuPAR conditioned media (b) were employed aschemoattractants. In all cases, data are reported as a percentageof invading cells in the absence of chemoattractant, consideredas 100% (DMEM), and represent the average of two experi-ments, performed in duplicate. Columns, mean of two indepen-dent experiments; bars, ±SD, ∗P < .001 against the control(DMEM).

Acknowledgments

The authors are grateful to Dr. Pietro Mugione for histechnical assistance. This work was supported in part byItalian Ministry of Health-FSN2007 and AIRC (AssociazioneItaliana per la Ricerca sul Cancro). K. Bifulco and I.Longanesi-Cattani equally contributed to this work.

References

[1] J. V. M. G. Bovee, P. C. W. Hogendoorn, J. S. Wunder, and B. A.Alman, “Cartilage tumours and bone development: molecularpathology and possible therapeutic targets,” Nature ReviewsCancer, vol. 10, no. 7, pp. 481–488, 2010.

[2] A. Y. Giuffrida, J. E. Burgueno, L. G. Koniaris, J. C. Gutierrez,R. Duncan, and S. P. Scully, “Chondrosarcoma in the UnitedStates (1973 to 2003): an analysis of 2890 cases from the SEERdatabase,” Journal of Bone and Joint Surgery - Series A, vol. 91,no. 5, pp. 1063–1072, 2009.

[3] F. Fiorenza, A. Abudu, R. J. Grimer et al., “Risk factors forsurvival and local control in chondrosarcoma of bone,” Journalof Bone and Joint Surgery - Series B, vol. 84, no. 1, pp. 93–99,2002.

[4] G. Ayala, C. Liu, R. Nicosia, S. Horowitz, and R. Lackman,“Microvasculature and VEGF expression in cartilaginoustumors,” Human Pathology, vol. 31, no. 3, pp. 341–346, 2000.

[5] H. Gelderblom, P. C. W. Hogendoorn, S. D. Dijkstra et al.,“The clinical approach towards chondrosarcoma,” Oncologist,vol. 13, no. 3, pp. 320–329, 2008.

[6] M. J. Duffy and C. Duggan, “The urokinase plasminogenactivator system: a rich source of tumour markers for theindividualised management of patients with cancer,” ClinicalBiochemistry, vol. 37, no. 7, pp. 541–548, 2004.

[7] A. Mondino and F. Blasi, “uPA and uPAR in fibrinolysis,immunity and pathology,” Trends in Immunology, vol. 25, no.8, pp. 450–455, 2004.

[8] V. Pillay, C. R. Dass, and P. F. M. Choong, “The urokinaseplasminogen activator receptor as a gene therapy target forcancer,” Trends in Biotechnology, vol. 25, no. 1, pp. 33–39, 2007.

[9] B. R. Binder, J. Mihaly, and G. W. Prager, “uPAR-uPA-PAI-1 interactions and signaling: a vascular biologist’s view,”Thrombosis and Haemostasis, vol. 97, no. 3, pp. 336–342, 2007.

[10] M. S. Pepper, “Role of the matrix metalloproteinase andplasminogen activator-plasmin systems in angiogenesis,” Arte-riosclerosis, Thrombosis, and Vascular Biology, vol. 21, no. 7, pp.1104–1117, 2001.

[11] H. W. Smith and C. J. Marshall, “Regulation of cell signallingby uPAR,” Nature Reviews Molecular Cell Biology, vol. 11, no.1, pp. 23–36, 2010.

[12] M. V. Carriero, P. Franco, S. Del Vecchio et al., “Tissuedistribution of soluble and receptor-bound urokinase inhuman breast cancer using a panel of monoclonal antibodies,”Cancer Research, vol. 54, no. 20, pp. 5445–5454, 1994.

[13] K. Danø, N. Behrendt, G. Høyer-Hansen et al., “Plasminogenactivation and cancer,” Thrombosis and Haemostasis, vol. 93,no. 4, pp. 676–681, 2005.

[14] M. G. Rasch, I. K. Lund, C. E. Almasi, and G. Hoyer-Hansen,“Intact and cleaved uPAR forms: diagnostic and prognosticvalue in cancer,” Frontiers in Bioscience, vol. 13, pp. 6752–6762,2008.

[15] M. Ploug and V. Ellis, “Structure-function relationshipsin the receptor for urokinase-type plasminogen activator:comparison to other members of the Ly-6 family and snakevenom α-neurotoxins,” FEBS Letters, vol. 349, no. 2, pp. 163–168, 1994.

[16] M. Resnati, I. Pallavicini, J. M. Wang et al., “The fibrinolyticreceptor for urokinase activates the G protein-coupled chemo-tactic receptor FPRL1/LXA4R,” Proceedings of the NationalAcademy of Sciences of the United States of America, vol. 99, no.3, pp. 1359–1364, 2002.

10 Sarcoma

[17] L. Gargiulo, I. Longanesi-Cattani, K. Bifulco et al., “Cross-talkbetween fMLP and vitronectin receptors triggered by uroki-nase receptor-derived SRSRY peptide,” Journal of BiologicalChemistry, vol. 280, no. 26, pp. 25225–25232, 2005.

[18] C. Barinka, G. Parry, J. Callahan et al., “Structural basisof interaction between urokinase-type plasminogen activatorand its receptor,” Journal of Molecular Biology, vol. 363, no. 2,pp. 482–495, 2006.

[19] C. Selleri, N. Montuori, P. Ricci et al., “In vivo activity of thecleaved form of soluble urokinase receptor: a new hematopoi-etic stem/progenitor cell mobilizer,” Cancer Research, vol. 66,no. 22, pp. 10885–10890, 2006.

[20] M. V. Carriero, I. Longanesi-Cattani, K. Bifulco et al.,“Structure-based design of an urokinase-type plasminogenactivator receptor-derived peptide inhibiting cell migrationand lung metastasis,” Molecular Cancer Therapeutics, vol. 8, no.9, pp. 2708–2717, 2009.

[21] N. Sidenius and F. Blasi, “The urokinase plasminogen activatorsystem in cancer: recent advances and implication for progno-sis and therapy,” Cancer and Metastasis Reviews, vol. 22, no.2-3, pp. 205–222, 2003.

[22] E. Henic, C. Borgfeldt, IB. J. Christensen, B. Casslen, and G.Høyer-Hansen, “Cleaved forms of the urokinase plasminogenactivator receptor in plasma have diagnostic potential andpredict postoperative survival in patients with ovarian cancer,”Clinical Cancer Research, vol. 14, no. 18, p. 5785, 2008.

[23] S. F. Shariat, C. G. Roehrborn, J. D. McConnell et al.,“Association of the circulating levels of the urokinase systemof plasminogen activation with the presence of prostate cancerand invasion, progression, and metastasis,” Journal of ClinicalOncology, vol. 25, no. 4, pp. 349–355, 2007.

[24] N. Sidenius, C. F. M. Sier, and F. Blasi, “Shedding andcleavage of the urokinase receptor (uPAR): identification andcharacterisation of uPAR fragments in vitro and in vivo,” FEBSLetters, vol. 475, no. 1, pp. 52–56, 2000.

[25] H. Taubert, P. Wurl, T. Greither et al., “Co-detection ofmembers of the urokinase plasminogen activator system intumour tissue and serum correlates with a poor prognosis forsoft-tissue sarcoma patients,” British Journal of Cancer, vol.102, no. 4, pp. 731–737, 2010.

[26] C. R. Antonescu and M. Ladanyi, Pathology and Geneticsof Tumors of Soft and Bone, World Health OrganizationClassification of Tumors, IARC Press, Lyon, France, 2002.

[27] I. Silvestri, I. L. Cattani, P. Franco et al., “Engaged urokinasereceptors enhance tumor breast cell migration and invasionby upregulating αvβ5 vitronectin receptor cell surface expres-sion,” International Journal of Cancer, vol. 102, no. 6, pp. 562–571, 2002.

[28] K. Bifulco, A. De Chiara, F. Fazioli et al., “Cell invasiveness insarcomas: a possibly useful clinical correlation,” Tumori, vol.94, no. 4, pp. 505–510, 2008.

[29] M. V. Carriero, S. Del Vecchio, M. Capozzoli et al., “Urokinasereceptor interacts with α(v)β vitronectin receptor, promotingurokinase-dependent cell migration in breast cancer,” CancerResearch, vol. 59, no. 20, pp. 5307–5314, 1999.

[30] K. Bifulco, I. Longanesi-Cattani, M. Gala et al., “The solubleform of urokinase receptor promotes angiogenesis throughits SER(88) -ARG-SER-ARG-TYR(92) chemotactic sequence,”Journal of Thrombosis and Haemostasis, vol. 8, no. 12, pp.2789–2799, 2010.

[31] N. Montuori, K. Bifulco, M. V. Carriero et al., “The cross-talkbetween the urokinase receptor and fMLP receptors regulatesthe activity of the CXCR4 chemokine receptor,” Cellular andMolecular Life Sciences. In press.

[32] C. Selleri, N. Montuori, P. Ricci et al., “Involvement of theurokinase-type plasminogen activator receptor in hematopoi-etic stem cell mobilization,” Blood, vol. 105, no. 5, pp. 2198–2205, 2005.

[33] A. Albini, Y. Iwamoto, and H. K. Kleinman, “A rapid in vitroassay for quantitating the invasive potential of tumor cells,”Cancer Research, vol. 47, no. 12, pp. 3239–3245, 1987.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 691608, 8 pagesdoi:10.1155/2011/691608

Research Article

A “Proteoglycan Targeting Strategy” for theScintigraphic Imaging and Monitoring of the Swarm RatChondrosarcoma Orthotopic Model

Caroline Peyrode,1 Francois Gouin,2 Aurelien Vidal,1 Philippe Auzeloux,1 Sophie Besse,1

Marie-Melanie Dauplat,3 Serge Askienazy,4 Dominique Heymann,2 Jean-Michel Chezal,1

Francoise Redini,2 and Elisabeth Miot-Noirault1

1 INSERM (UMR 990) Universite d’Auvergne, rue Montalembert, BP 184, 63005 Clermont-Ferrand Cedex, France2 INSERM (UMR S957), Universite de Nantes, 44035 Nantes, France3 Service d’Anatomo-Pathologie, CLCC Jean Perrin, 63001 Clermont-Ferrand, France4 Cyclopharma Laboratoires, 63360 Saint-Beauzire, France

Correspondence should be addressed to Elisabeth Miot-Noirault, elisabeth.noirault@inserm.fr

Received 1 October 2010; Revised 24 December 2010; Accepted 4 January 2011

Academic Editor: Irene Andrulis

Copyright © 2011 Caroline Peyrode et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Our lab developed 99mTc-NTP 15-5 radiotracer as targeting proteoglycans (PGs) for the scintigraphic imaging of joint. This paperreports preclinical results of 99mTc-NTP 15-5 imaging of an orthotopic model of Swarm rat chondrosarcoma (SRC). 99mTc-NTP15-5 imaging of SRC-bearing and sham-operated animals was performed and quantified at regular intervals after surgery andcompared to bone scintigraphy and tumoural volume. Tumours were characterized by histology and PG assay. SRC exhibiteda significant 99mTc-NTP 15-5 uptake at very early stage after implant (with tumour/muscle ratio of 1.61 ± 0.14), whereas nomeasurable tumour was evidenced. As tumour grew, mean tumour/muscle ratio was increased by 2.4, between the early and latestage of pathology. Bone scintigraphy failed to image chondrosarcoma, even at the later stage of study. 99mTc-NTP 15-5 imagingprovided a suitable set of quantitative criteria for the in vivo characterization of chondrosarcoma behaviour in bone environment,useful for achieving a greater understanding of the pathology.

1. Introduction

Our laboratory has developed the 99mTc-N-[triethylammo-nium]-3-propyl-[15]ane-N5 (99mTc-NTP 15-5) radiotracerable to link proteoglycans (PGs) of cartilage, which wasdemonstrated to allow scintigraphic imaging of joint [1–6]. PG appear also as key partners in bone cell biology,and might represent interesting targets for the assessmentof malignant pathological processes [7]. Chondrosarcomarepresent 25% of all bone sarcoma and are characterized bythe formation of a cartilage like extracellular matrix [8–10].To the orthopedic oncologists, it is essential to identify thelesions that are thought to constitute the “cartilage tumourgroup”, through the presence of identifiable elements thatresemble the cells and matrix of cartilage. For such analysis,

the three-grade histopathologic classification of Evans isbased on cell type/differentiation, matrix formation andarchitecture, and is at present considered useful for theprediction of clinical behaviour [11–13]. To date, manyquestions still remain unanswered and there is an urgentneed for markers of biologic phenotypic features, that couldbe determined in a noninvasive manner, and could beused as objective criteria of grading, followup, response totherapy and detection of recurrence. CT and MRI definemorphology, but are limited in distinguished postoperativeresidual and postchemotherapeutic lesions due to alteredtissue planes, edema and fibrosis. Contrast-enhanced MRI,while providing data for local staging and extraosseousinvolvment, appears also limiting in determining biologicbehaviour of lesions [14–16]. SPECT by the use of highly

2 Sarcoma

specific radiolabelled probes, allows a noninvasive andquantitative assessment of many biochemical pathways invivo [17]. In such context, scintigraphic imaging with 99mTc-NTP 15-5 would be a powerful tool for a direct in vivoquantitative evaluation of chondrosarcoma at the PG level,allowing therefore the detection of changes in relation topathological processes and/or response to therapies. Wetherefore initiated preclinical studies aiming at determiningwhether 99mTc-NTP 15-5 imaging could be useful for theevaluation of the tumoural pathology of cartilage in vivo.In the present study, the 99mTc-NTP 15-5 imaging wascharacterized in the Swarm rat chondrosarcoma implantedin paratibial location.

2. Materials and Methods

2.1. Animals. Male Sprague Dawley (Charles River, France)were included in this study. They were handled and caredfor in accordance with the guidelines for the Care and Useof Laboratory Animals (National Research Council, 1996)and European Directive 86/809/EEC. They were maintainedat 21◦C with a 12 h/12 h light/dark cycle. Protocols wereperformed under the authorisation of the French Directionof Veterinary Services (Authorisation N◦ C63-113-10) andconducted by authorized investigators in accordance withthe institution’s recommendations for the use of laboratoryanimals.

2.2. Chondrosarcoma Model. The Swarm rat chondrosar-coma (SRC) line was a generous gift from Dr. P. A. Guerne(Geneva, Switzerland) as tissue fragments that were frozenuntil use.

For tumoural implant, the rats were anaesthetized byinhalation of a combination isoflurane (Abbott, France)/air(1.5%, 1 L/min) associated with an intramuscular injectionof 100 mg/kg ketamine (Imalgene, Rhone Merieux, France).

Allograft transplantation of a tumour fragment (10animals) was performed on the right paw, the other pawbeing used as the controlateral reference, as previouslypublished [18].

Using a lateral approach, the cortical surface of the dia-physis was scarified laterally on 10 mm, a 10 mm3 fragmentof SRC was placed contiguous to the scarified surface, andthe muscular and cutaneous wounds were sutured. The sameprocedure was performed for the controlateral paw, but notumour fragment was implanted. Tumours appeared at thegraft site 7–11 days later.

2.3. Establishment of the Sham Model. Five animals under-went sham surgery, in which the right paw was submitted tothe same surgery procedure as SRC animals, but no tumourfragment was implanted.

2.4. Acquisitions for 99mTc-NTP 15-5 Imaging. Scintigraphicin vivo imaging was performed using a small-animal γ cam-era (CsI(Na) crystal) equipped with a 1.3/0.2/35 parallel-holecollimator (hole diameter/septum thickness/height in mm)(Gammaimager, Biospace, France). The energy resolution

H

H

H H

N

N

N

N

N

O OTc

Br−

N+

C2H5

C2H5

C2H5(CH2)3

Figure 1: Chemical structure of 99mTc-N-[triethylammonium]-3-propyl-[15]ane-N5 (99mTc-NTP 15-5).

and intrinsic planar resolution of the system are given as 11%at 140 keV and<2 mm full width at half maximum (FWHM),respectively.

The radiotracer 99mTc-NTP 15-5 (Figure 1) radiotracerwas prepared and radiolabelled as previously published[1–6], and administered by IV route to vigil animals(25 Mbq/animal) using dedicated contention box.

First of all, a 10-min planar acquisition (with a 15%window centered on the 140 keV photopeak of 99mTc) wasperformed on 3 animals with a well-established tumour(volume = 949.35 ± 223 mm3) at several intervals (30,60, 90 and 120 mins) to determine kinetics of tracer accu-mulation within chondrosarcoma, respectively, to articularcartilage and muscle. For acquisition, each posterior paw ofanimals was positioned over the collimator of the cameraas illustrated in Figure 2. Quantitative analysis of scinti-graphic scans was performed using the GAMMAVISION+software (Biospace, France). Regions of interest (ROIs)were delineated over tumour, femorotibial cartilage andmuscle patterns. For each animal and each ROI, averagecount in cpm per pixel was obtained. Two semiquantitativeparameters were determined:

TM= average count in tumour

average count in muscle,

TC= average count in tumour

average count in femorotibial cartilage.

(1)

Data were expressed as mean ± standard deviation.A serial imaging was then performed on primary

chondrosarcoma-bearing rats (n = 10) and sham-operatedanimals (n = 5) at regular intervals after surgery (fromday 4 to day 35), with 10-min planar acquisition beingstarted 30 mins after iv administration of 99mTc-NTP 15-5. Scintigrams were considered positive when tracer uptakeareas corresponded to sites of implant. All the positivescans were evaluated using the ROI method as describedabove, with T/C and T/ M parameters being determinedfor each animal at each time point, and data expressed asmean ± standard deviation. Semi-quantitative parametersdetermined at day 4 were used as the “threshold referencevalue” and were compared to the mean values determined ateach time point of study (paired 2-sided Student t-test witha level of significance set at P < .05).

Sarcoma 3

Figure 2: In vivo scintigraphic acquisition of chondrosarcoma bearing rats using a small animal dedicated gamma camera.

2.5. Acquisitions for 99mTc-HMDP Bone Imaging. Primarychondrosarcoma-bearing rats were also submitted to bonescintigraphies at day 12, day 38, and day 48 after implant.For each paw, “delayed images” (10-minute duration) wereacquired 2 hours post injection of 99mTc-HMDP radiotracer(Osteocys, IBA, France; 30 MBq/animal).

Bone scintigrams were considered positive when traceruptake areas corresponded to sites of implant.

2.6. Tumour Growth Assessment. The tumour volume wasdetermined from the measurement of 2 perpendicular diam-eters using a calliper. Tumour volumes (V) were calculatedaccording to the following formula:

V(mm3) = 0.5× L× (S)2, (2)

where L and S are, respectively, the largest and smallestperpendicular tumour diameters in mm [19].

2.7. Histology of SRC Tissue. At the end of the study, threechondrosarcoma-bearing rats were killed for histologicalcharacterization of the tumours. The femora and tumourof each rat were removed and fixed at 4◦C for 48 hours informol buffer (pH 7.4). The femora where cut longitudinallyor transversally. Decalcified femoral fragments were embed-ded in paraffin and 5 μm sections were mounted on glassslides for staining with Hematoxylin-Eosin-Safran (HES)and Alcian blue.

2.8. PG Content of SRC Tissue. At the end of the study,SRC were removed (n = 3 animals), dissected and digestedfor 24 h at 60◦C with EDTA-phosphate buffer solutioncontaining Papaine 0.60 mg/mL (Sigma-Aldrich, France)and DL-dithiothreitol (DTT) 0.25 mg/mL (Sigma-Aldrich,France). Digests were then assessed for PG content using thedimethylene blue protocol [20].

PG content of SRC tissue was compared to PG contentof mouse B16F0 melanoma as negative control: B16F0melanoma were removed from bearing mice (n = 3) (stageday 15 after subcutaneous inoculation of 300000 B16F0cells), and submitted to the same procedure of extraction anddigestion as SRC tumours.

3. Results

3.1. SRC Characterisation: Tumoural Growth, Histology andPG Assay. Tumour volumes were followed for 35 days

4 10 20 35

Days post tumor implant

0

1000

2000

3000

4000

5000

6000

Tum

oral

volu

me

(mm

3)

Figure 3: Tumour volume growth for Swarm rat chondrosarcoma(SRC) implanted in paratibial location. Mean values + standarddeviation are presented at each time point.

after primary implant (Figure 3). All the animals developeda tumour that became palpable from day 10. Histolog-ical examination (Figure 4(a)) at study end evidenced achondroid tumoural tissue, poorly vascularized, lobular inorganization with lobules containing chondroid stroma anddelimited by fine fibrous septa. Hypercellularity was alsoobserved. At this later stage of pathology, an extensiveinvasion of bone and surrounding tissues was present, as wellas osteolysis.

The overall presence of PG was first investigated byAlcian blue staining: as shown in Figure 4(b), a high densityof stained areas were observed in SRC tissue, revealing thepresence of PG in this tumour.

PG content of SRC tumour tissue was also assessed bybiochemical dosage. PG content of SRC tissue was comparedwith PG of murine B16F0 melanoma tumour as negativecontrol: As shown in Figure 5, a high PG content wasobserved in SRC tissue (7.46 ± 2.41 μg of PG/mg of tissue),respectively, to melanoma (0.45 ± 0.16 μg of PG/mg oftissue).

3.2. 99mTc-NTP 15-5 Distribution in SRC Tissue, Respectively,to Muscle and Cartilage. 99mTc-NTP 15-5 distribution inwell-established primary chondrosarcomas was character-ized at 30, 60, 90 and 120 mins after 99mTc-NTP 15-5administration, on the basis of tumour/muscle (T/M) andtumour/cartilage (T/C) ratios. As shown in Figure 6(a),99mTc-NTP 15-5 rapidly accumulated within SRC tissue,

4 Sarcoma

(a) (b)

Figure 4: Histomorphological features of the SRC model at study end showing: (a) a tumoural tissue with hypercellularity, and boneosteolysis (x10); (b) a high density of blue Alcian stained areas as a reflect of proteoglycan content (x5).

0

2

4

6

8

10

12

Pro

teog

lyca

nco

nte

nt

(μg/

mg

oftu

mor

)

Tumour type

SRCB16F0 melanoma

Figure 5: PG content of SRC tissue, respectively, to murinemelanoma tumour (as negative control); Mean values + standarddeviation are presented.

with T/M and T/C ratios being 1.44 ± 0.27 and 0.78 ±0.03 from 30 mins post injection. Radiotracer accumulationwithin tumoural tissue was observed to be as high as incartilage tissue from 60 mins pi, with T/C value of 0.90 ±0.23. No significant changes were observed in T/M and T/Cratios from 60 to 120 mins pi.

This high and selective accumulation of 99mTc-NTP 15-5 within tumour was at the origin of the highly contrastedchondrosarcoma imaging in vivo, as shown in (Figure 6(b))for a representative animal.

3.3. Diagnosis Capability of 99mTc-NTP 15-5 Imaging,Respectively, to 99mTc-HMDP Bone Scintigraphy and CalliperMeasurement. All the implanted animals were positive forchondrosarcoma at necropsy.

Figure 7 shows the disease incidence according to 99mTc-HMDP scintigraphy, 99mTc-NTP 15-5 imaging, and callipermeasurement during the whole study of monitoring.

A palpable and measurable tumour was observed fromday 10 after implant for 50% of the animals. From day 20,100% of the animals evidenced measurable tumours.

99mTc-NTP 15-5 scintigraphy was positive for 66.66% ofthe animals as early as day 4 post implant, and for 100% ofthe animals from day 10.

99mTc-HMDP bone scintigraphy was negative for SRCimaging during the whole duration of study, and even later(48 days post implant).

3.4. Monitoring of SRC Growth In Vivo Using 99mTc-NTP 15-5Imaging. As shown in Figure 8 for a representative animal,99mTc-NTP 15-5 radiotracer was observed to accumulatewithin tumoural tissue, respectively, to the contralateral paw(Figure 8(a)), as early as day 4 after implant: scintigramof the tumour bearing paw (Figure 8(b)) evidenced (i)uptake areas in femorotibial articular cartilage, with thetibial plateau uptake clearly distinguished from the femoralcondyle uptake, (ii) accumulation at the site of implant.99mTc-NTP 15-5 examination of the same animal at late stageof pathology (day 35, Figure 8(c)) evidenced that radiotraceraccumulation within the tumour-bearing paw was highlyincreased, respectively, to day 4.

99mTc-NTP 15-5 accumulation was quantitativelyassessed in vivo in the tumour, respectively, to muscle andcartilage (as T/M and T/C uptake ratios, resp.) as a functionof time after implant (Figure 9): As early as day 4 postimplant, the mean T/M and T/C values were 1.61 ± 0.14 and0.57 ± 0.06, respectively. From day 10, tumoural uptake of99mTc-NTP 15-5 was as high as cartilage uptake with T/C =1.08 ± 0.09; T/M value was 2.55 ± 1.15.

At study ending (day 35), mean T/C and T/ M values were1.61 ± 0.06 and 3.97 ± 0.74, respectively.

3.5. Monitoring of Sham-Operated Animals Using 99mTc-NTP15-5 Imaging. Serial 99mTc-NTP 15-5 imaging of sham-operated animals (n = 5) performed over 50 days after

Sarcoma 5

0

0.5

1

1.5

2

2.5

30 60 90 120

T/MT/C

Time after 99mTc-NTP 15-5 administration (mins)

99m

Tc-

NT

P15

-5u

ptak

era

tios

(a)

T

F

CHS

0.8 6

(b)

Figure 6: 99mTc-NTP 15-5 in vivo distribution in chondrosarcoma;(a) 99mTc-NTP 15-5 uptake ratios as a function of delay betweeniv administration of tracer and acquisition; (b) Scintigraphic imageobtained for a representative animal, for a delay of 30 min betweeninjection and acquisition. Abbreviations: F: femoral condyle; T:tibial plateau; CHS: chondrosarcoma.

surgery did not evidence any accumulation of the tracer atsite of surgery.

4. Discussion

This experimental study was conducted in the syngeneicmodel of SRC that has been the subject of extensivebiochemical studies on extracellular matrix and chondrocytemetabolism [21, 22]. Histological characterization of SRCimplanted in paratibial location, evidenced a chondroidtumoural tissue, poorly vascularized, lobular in organizationwith lobules containing chondroid stroma delimited by finefibrous septa. In the present study PG were confirmed (byboth alcian blue staining and biochemical assay) as being amajor component of SRC tissue. The high concentration ofPG in the tumour is at the origin of a high density of ECMstrong negative charges that may interact with the positivelycharged quaternary ammonium moiety of the 99mTc-NTP15-5 radiotracer. The ability of 99mTc-NTP 15-5 to imageSRC tissue was therefore demonstrated from 30 mins post

0

20

40

60

80

100

Posi

tive

diag

nos

is(%

)

4 7 10 20 25 35

Days post tumour implant

Calipper

99mTc-HMDP

99mTc-NTP 15-5

Figure 7: Disease incidence at each time point of study, according to99mTc-HMDP scintigraphy, 99mTc-NTP 15-5 imaging, and callipermeasurement.

injection with T/M and T/C ratios being 1.44± 0.27 and 0.78± 0.03.

Interestingly, tracer accumulation was detected at earlystage (day 4) and for about 70% of the animals investigated,whereas no palpable nor measurable tumour was evidenced.Such early accumulation raised the question of whether theimaging pattern represented an effective tumoural uptake ora surgery-induced inflammation. Negative 99mTc-NTP 15-5imaging observed for sham-operated animals definitely ruledout the hypothesis of inflammation contribution.

99mTc-NTP 15-5 accumulation within SRC was assessedat multiple time-points of tumoural growth using a semi-quantitative method. As tumour developed over time, meanT/M and T/C values were observed to be increased by a factorof 2.4 and 2.8, respectively, between the early stage day 4and the late stage day 35. Tumour to femorotibial cartilageuptake ratio (T/C parameter) is of particular importancefor clinical application (in patients, tumour is located in thejoint). From our results, the increase in T/C with pathologyargued in favour of a high tumour/cartilage contrast whichwould expect diagnosis of tumour occurrence within thejoint.

Another interesting result of this study is that 99mTc-HMDP bone scintigraphy commonly used in clinical practisefailed to image tumoural tissue throughout the study,whereas 99mTc-NTP 15-5 imaging was positive.

These preclinical results underlined the suitability andhigh sensitivity of 99mTc-NTP 15-5 scintigraphic imaging forassessing in vivo the PG component of chondrosarcomas,providing therefore criteria for a quantitative and functionalassessment of the tumoural pathology in vivo.

Our results raised the question of whether 99mTc-NTP15-5 imaging patterns in benign cartilage tumours, (whichare known to form a cartilaginous PG—rich matrix), wouldbe the same as for malignant tumours. According to manyhistological studies, the neoplastic cells of benign cartilagetumours such as chondroblastoma did not show, at any pointof their evolution, real cartilage matrix formation [13, 23]. As

6 Sarcoma

0.5 5.3

(a)

CHS

2.5 13

(b)

11.5

CHS

2.5

(c)

Figure 8: 99mTc-NTP 15-5 longitudinal in vivo examination of a representative SRC-bearing rat; (a) in vivo scintigraphic image of thecontralateral paw at day 4; (b) in vivo scintigraphic image of the tumour-bearing paw at day 4; (c) in vivo scintigraphic image of the tumourbearing paw of the same animal at day 35. A clear accumulation of radioactivity within tumoural tissue was observed as early as day 4 afterimplant (Image b, arrow); abbreviations: CHS = chondrosarcoma.

4 7 10 20 25 35

Days post implant

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Upt

akes

rati

os

T/CT/M

Figure 9: Quantitative analysis of 99mTc-NTP 15-5 accumulation inchondrosarcoma against time after implant. Results are expressedas Tumour/Muscle and Tumour/Cartilage uptake ratios (T/M andT/C, resp.). Mean values + standard deviation are presented at eachtime point.

a consequence, a differential accumulation of 99mTc-NTP 15-5 radiotracer could be expected. Since the biological natureof benign cartilage tumours is still debating, we will haveto elucidate this aspect by performing a preclinical 99mTc-NTP 15-5 imaging study of rodent models of enchondroma,(such as mutant mice for hedgehog signalling pathwaysor Ext1/Ext2 exostosin encoding gene) in parallel to thebiological characterization of PG of tissue [23–25].

From a clinical point of view, the distinction betweenbenign and low-grade malignant pathology, is known to beextremely difficult both at radiologically and histologically,and remains a challenge [9, 26]. All current imagingmodalities have shown their limitations. Radiographs, CT,MRI and contrast enhanced MRI provide morphologicaldata for local staging and extra-osseous involvement, but

appeared limited in determining functional markers ofpostoperative recurrence, response or relapse to therapies[9, 27–29]. In such context, 99mTc-NTP 15-5 radiotracermay serve as an adjunct to CT and MRI, by supplyingquantitative data which would allow imaging and follow-up to be functionally rather than morphologically based.For functional imaging of cartilage neoplasms, radiotracerscurrently available provide indirect evaluations of the pathol-ogy, such as bone remodeling and inflammation features,but do not allow biological assessment of the tumour. Manytumour seeking agents such as 201Tl, 99mTc-MIBI, 99mTc-Tetrofosmin, 99mTc-DMSA(V) and more recently 18F-FDGhave been found useful in an initial diagnosis and grading,but they have also demonstrated their limitations for imagingchondrosarcoma with low cellularity and low vascularity[30–33]. 99mTc-NTP 15-5 imaging would allow a regular invivo follow up, in order to evidence any “upregulation” or“downregulation” synthesis of PG, as the reflect of potentialmalignant transformation, recurrence, response or relapse totherapies. We strongly believe in such potential applicationsince 99mTc-NTP 15-5 imaging recently evidenced in the SRCmodel a “downregulation synthesis” of PG as a result ofanticancer treatment (unpublished data).

Another potential application of 99mTc-NTP 15-5 imag-ing in chondrosarcoma would be the early imaging of metas-tasis. As a consequence, the sensitivity of 99mTc-NTP 15-5imaging for the early evaluation of chondrosarcoma metasta-sis has to be determined in relevant preclinical animal modelsof chondrosarcoma with spontaneous metastasis [34, 35].

5. Conclusion

This preliminary work in the orthotopic SRC model under-lined the suitability and high sensitivity of 99mTc-NTP15-5 radiotracer for imaging chondrosarcoma at the PGlevel. 99mTc-NTP 15-5 imaging provided a suitable setof quantitative criteria for the in vivo characterization of

Sarcoma 7

chondrosarcoma behaviour in bone environment, whichcould be useful for achieving a greater understanding of thepathology.

In view of a potential clinical application, 99mTc-NTP 15-5 imaging appears of interest for (i) the establishment of thecartilaginous nature of “musculoskeletal” tumours (ii) the invivo assessment to PG changes associated to the evolutionof pathological process, local recurrence, response or relapseto therapies. Additional preclinical studies are needed toinvestigate the potential of 99mTc-NTP 15-5 imaging in thetumoral pathology of cartilage.

Acknowledgments

The authors received grants from Institut National duCancer, “projet libre intercanceropole CLARA/GrandOuest”, Ligue Contre le Cancer, Regional Innovation FundFRI2/OSEO, Auvergne Region for “Contrat de Projets EtatRegion”.

References

[1] J. C. Madelmont, I. Giraud, C. Nicolas et al., “inven-tors. Cyclopharma Laboratoires, assignee. Novel quaternaryammonium derivatives, method for preparing same andpharmaceutical use,” International patent PCT WO 01/00621A1. January 2001.

[2] C. Nicolas, M. Borel, J. C. Maurizis et al., “Synthesis of N-quaternary ammonium [3H] and [99mTc]Polyazamacrocycles,potential radiotracers for cartilage imaging,” Journal ofLabelled Compounds and Radiopharmaceuticals, vol. 43, no. 6,pp. 585–594, 2000.

[3] J. C. Maurizis, M. Rapp, C. Nicolas, M. Ollier, M. Verny,and J. C. Madelmont, “Disposition in rats of N-pyridinium-propyl-cyclam, N- triethylammoniumpropyl-cyclam, and N-[triethylammonium]-3-propyl-[15]ane-N5, potential carti-lage imaging agents,” Drug Metabolism and Disposition, vol.28, no. 4, pp. 418–422, 2000.

[4] M. Ollier, J. C. Maurizis, C. Nicolas et al., “Joint scintigraphyin rabbits with 99mTc-N-[3-(triethylammonio)propyl]-15ane-N5, a new radiodiagnostic agent for articular cartilage imag-ing,” Journal of Nuclear Medicine, vol. 42, no. 1, pp. 141–145,2001.

[5] E. Miot-Noirault, A. Vidal, P. Pastoureau et al., “Early detec-tion and monitoring of cartilage alteration in the experimentalmeniscectomised guinea pig model of osteoarthritis by 99mTc-NTP 15-5 scintigraphy,” European Journal of Nuclear Medicineand Molecular Imaging, vol. 34, no. 8, pp. 1280–1290, 2007.

[6] E. Miot-Noirault, A. Vidal, P. Auzeloux, J. C. Madelmont,J. Maublant, and N. Moins, “First in vivo SPECT imagingof mouse femorotibial cartilage using 99mTc-NTP 15-5,”Molecular Imaging, vol. 7, no. 6, pp. 263–271, 2008.

[7] F. Lamoureux, M. Baud’huin, L. Duplomb, D. Heymann, andF. Redini, “Proteoglycans: key partners in bone cell biology,”BioEssays, vol. 29, no. 8, pp. 758–771, 2007.

[8] World Health Organization, “Cartilage tumours,” in WorldHealth Organization Classification of Tumours: Pathology andGenetics, Tumours of Soft Tissue and Bone, C. D. M. Fletcher,K. K. Unni, and F. Mertens, Eds., pp. 234–257, IARC Press,Lyon, France, 2002.

[9] H. Gelderblom, P. C. W. Hogendoorn, S. D. Dijkstra et al.,“The clinical approach towards chondrosarcoma,” Oncologist,vol. 13, no. 3, pp. 320–329, 2008.

[10] W. A. Chow, “Update on chondrosarcomas,” Current Opinionin Oncology, vol. 19, no. 4, pp. 371–376, 2007.

[11] H. L. Evans, A. G. Ayala, and M. M. Romsdahl, “Prognosticfactors in chondrosarcoma of bone. A clinicopathologicanalysis with emphasis on histologic grading,” Cancer, vol. 40,no. 2, pp. 818–831, 1977.

[12] F. Schajowicz, L. H. Sobin, and World Health Organization,International Histological Classification of Tumours, Histologi-cal Typing of Bone Tumours, Springer, Berlin, Germany, 1993.

[13] T. Aigner, “Towards a new understanding and classification ofchondrogenic neoplasias of the skeleton—biochemistry andcell biology of chondrosarcoma and its variants,” VirchowsArchiv, vol. 441, no. 3, pp. 219–230, 2002.

[14] L. Ollivier, D. Vanel, and J. Leclere, “Imaging of chondrosar-comas,” Cancer Imaging, vol. 4, pp. 36–38, 2003.

[15] F. Feldman, R. van Heertum, and C. Manos, “18FDG PETscanning of benign and malignant musculoskeletal lesions,”Skeletal Radiology, vol. 32, no. 4, pp. 201–208, 2003.

[16] S. E. Anderson, L. S. Steinbach, S. Schlicht, G. Powell, M.Davies, and P. Choong, “Magnetic resonance imaging of bonetumors and joints,” Topics in Magnetic Resonance Imaging, vol.18, no. 6, pp. 457–465, 2007.

[17] R. J. Hicks, G. C. Toner, and P. F. Choong, “Clinicalapplications of molecular imaging in sarcoma evaluation,”Cancer Imaging, vol. 5, no. 1, pp. 66–72, 2005.

[18] E. Grimaud, C. Damiens, A. V. Rousselle, N. Passuti, D.Heymann, and F. Gouin, “Bone remodelling and tumourgrade modifications induced by interactions between boneand Swarm rat chondrosarcoma,” Histology and Histopathol-ogy, vol. 17, no. 4, pp. 1103–1111, 2002.

[19] F. Gouin, B. Ory, F. Redini, and D. Heymann, “Zoledronicacid slows down rat primary chondrosarcoma development,recurrent tumor progression after intralesional curretage andincreases overall survival,” International Journal of Cancer, vol.119, no. 5, pp. 980–984, 2006.

[20] R. W. Farndale, D. J. Buttle, and A. J. Barrett, “Improved quan-titation and discrimination of sulphate glycosaminoglycans byuse of dimethylmethylene blue,” Biochimica et Biophysica Acta,vol. 883, no. 2, pp. 173–177, 1986.

[21] G. K. Hascall, “Ultrastructure of the chondrocytes and extra-cellular matrix of the Swarm rat chondrosarcoma,” AnatomicalRecord, vol. 198, no. 2, pp. 135–146, 1980.

[22] D. Breitkreutz, L. Diaz de Leon, and L. Paglia, “Histologicaland biochemical studies of a transplantable rat chondrosar-coma,” Cancer Research, vol. 39, no. 12, pp. 5093–5100, 1979.

[23] T. Aigner, S. Loos, C. Inwards et al., “Chondroblastoma isan osteoid-forming, but not cartilage-forming neoplasm,”Journal of Pathology, vol. 189, no. 4, pp. 463–469, 1999.

[24] S. Hopyan, P. Nadesan, C. Yu, J. Wunder, and B. A. Alman,“Dysregulation of hedgehog sinalling predisposes to synovialchondromatosis,” Journal of Pathology, vol. 206, no. 2, pp. 143–150, 2005.

[25] J. V. M. G. Bovee, “EXTra hit for mouse osteochondroma,”Proceedings of the National Academy of Sciences of the UnitedStates of America, vol. 107, no. 5, pp. 1813–1814, 2010.

[26] X. L. Wang, L. H. De Beuckeleer, A. M. A. De Schepper, andE. Van Marck, “Low-grade chondrosarcoma vs enchondroma:challenges in diagnosis and management,” European Radiol-ogy, vol. 11, no. 6, pp. 1054–1057, 2001.

[27] J. V. M. G. Bovee and P. C. W. Hogendoorn, “Cartilage-forming tumours of bone and soft tissue and their differentialdiagnosis,” Current Diagnostic Pathology, vol. 7, no. 4, pp. 223–234, 2001.

8 Sarcoma

[28] J. V. M. G. Bovee, A. M. Cleton-Jansen, A. H. M. Taminiau,and P. C. W. Hogendoorn, “Emerging pathways in thedevelopment of chondrosarcoma of bone and implications fortargeted treatment,” Lancet Oncology, vol. 6, no. 8, pp. 599–607, 2005.

[29] J. V. M. G. Bovee, “Multiple osteochondromas,” OrphanetJournal of Rare Diseases, vol. 3, no. 1, pp. 1–7, 2008.

[30] S. G. Pneumaticos, S. N. Chatziioannou, W. H. Moore,and M. Johnson, “The role of radionuclides in primarymusculoskeletal tumors beyond the ’bone scan’,” CriticalReviews in Oncology/Hematology, vol. 37, no. 3, pp. 217–226,2001.

[31] G. Buyukdereli, O. Sargin, and S. Ozbarlas, “Tc-99m tetrofos-min imaging in chondrosarcoma,” Clinical Nuclear Medicine,vol. 25, no. 1, pp. 64–65, 2000.

[32] P. F. M. Choong, T. Kunisada, J. Slavin, S. Schlicht, and R.Hicks, “The role of thallium-201 and pentavalent dimercapto-succinic acid for staging cartilaginous tumours,” InternationalSeminars in Surgical Oncology, vol. 8, pp. 1–7, 2004.

[33] F. Y. I. Lee, J. Yu, S. S. Chang, R. Fawwaz, and M. V.Parisien, “Diagnostic value and limitations of fluorine-18fluorodeoxyglucose positron emission tomography for carti-laginous tumors of bone,” Journal of Bone and Joint Surgery A,vol. 86, no. 12, pp. 2677–2685, 2004.

[34] J. C. M. Clark, C. R. Dass, and P. F. M. Choong, “Developmentof chondrosarcoma animal models for assessment of adjuvanttherapy,” ANZ Journal of Surgery, vol. 79, no. 5, pp. 327–336,2009.

[35] J. C. M. Clark, T. Akiyama, C. R. Dass, and P. F. M. Choong,“New clinically relevant, orthotopic mouse models of humanchondrosarcoma with spontaneous metastasis,” Cancer CellInternational, vol. 20, pp. 1–10, 2010.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 405437, 15 pagesdoi:10.1155/2011/405437

Review Article

Chondrosarcoma: With Updates on Molecular Genetics

Mi-Jung Kim,1 Kyung-Ja Cho,1 Alberto G. Ayala,2 and Jae Y. Ro1, 2

1 Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea2 Department of Pathology, Weill Medical College of Cornell University, The Methodist Hospital, Houston, TX 77030, USA

Correspondence should be addressed to Jae Y. Ro, jaero@tmhs.org

Received 14 September 2010; Revised 23 November 2010; Accepted 17 December 2010

Academic Editor: C. Verhoef

Copyright © 2011 Mi-Jung Kim et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Chondrosarcoma (CHS) is a malignant cartilage-forming tumor and usually occurs within the medullary canal of long bonesand pelvic bones. Based on the morphologic feature alone, a correct diangosis of CHS may be difficult, Therefore, correlation ofradiological and clinicopathological features is mandatory in the diagnosis of CHS. The prognosis of CHS is closely related tohistologic grading, however, histologic grading may be subjective with high inter-observer variability. In this paper, we presenthistologic grading system and clinicopathological and radiological findings of conventional CHS. Subtypes of CHSs, such asdedifferentiated, mesenchymal, and clear cell CHSs are also presented. In addition, we introduce updated cytogenetic andmolecular genetic findings to expand our understanding of CHS biology. New markers of cell differentiation, proliferation, andcell signaling might offer important therapeutic and prognostic information in near future.

1. Introduction

Chondrosarcoma (CHS) is a rare malignant tumor thatproduces cartilage matrix. The estimated overall incidence ofCHSs is 1 in 200,000 per year [1], and it is the third mostfrequent malignant bone tumor after multiple myelomaand osteosarcoma. It is estimated that CHSs account forapproximately 3.6% of the annual incidence of all primarybone malignancies in the USA [2] and 20∼30% of primarymalignant bone tumors [3].

CHSs that arise de novo are called primary CHSs,whereas CHSs developing superimposed on preexistingbenign cartilage tumors such as an enchondroma or osteo-chondroma are referred to as secondary CHSs. CHSs area heterogeneous group of tumors that can be categorizedby anatomic location as central when they occur withinthe medullary canal or peripheral when they occur in thecartilage cap of an exostosis. In addition to conventionalCHSs that show hyaline cartilage differentiation, there areother types of CHSs such as dedifferentiated, mesenchymal,or clear cell subtypes which show distinct genetic andclinicopathologic characteristics [4] (Tables 1 and 2). MyxoidCHS is not included in this paper because its existence inbone is highly controversial.

Most (about 85%) of CHSs, however, are of conventionalCHSs, and the majority arises in the medullary cavity of longbone. The minority (up to 15%) of conventional CHSs issecondary peripheral CHSs which develop from the surfaceof bone as a result of malignant transformation within thecartilage cap of a preexisting benign cartilage tumor such asosteochondroma or develop de novo on the bone surface,and these are not uncommonly referred as juxtacortical orperiosteal CHSs.

2. Conventional Intramedullary CHSs

2.1. Epidemiology. Conventional CHS of bone is the mostcommon type of primary CHS. Primary CHS typically affectsan old population. The majority of patients are older than 50years. The peak incidence is in the fifth to seventh decades oflife. There is a male predilection of 1.5–2 to 1 [1].

2.2. Sites of Involvement. CHS can involve any bone; theincidence of axial and appendicular involvement is verysimilar. The bones of the pelvis, especially for ilium, are fre-quently involved. Long tubular bones are frequently affected.The proximal femur is the most preferred site, followedin frequency by proximal humerus, distal femur, and ribs.

2 Sarcoma

Other less frequently involved bones are the spine, scapula,and sternum. CHS rarely involves craniofacial bones, neck,forearm, clavicle, and sesamoids (including the patella). CHSin the small tubular bones is extremely rare (1∼4% of allcases) [3, 5, 6].

2.3. Clinical Features and Imaging. Clinical symptoms aremostly nonspecific. Localized pain is the most frequentpresenting symptom (about 80%) after local swelling. Thesymptoms are usually insidious, progressive, and worse atnight and have a long duration (several months or years).Pathologic fractures are also common at initial presentation(up to 27%).

Radiographs of conventional CHS typically reveal amixed lytic and sclerotic pattern with characteristic smallcalcifications, often referred as “popcorn” or “ringlets” calci-fications. In the long bones, primary CHS most commonlyinvolves the metaphysis (49%), followed by the diaphysis(36%) (Figure 1). The presence of typical calcifications isradiologically diagnostic of cartilage, but often does notdiscriminate between benign, borderline, or malignant typesof lesions. Size of the lesions (<5 cm), lack of break throughthe cortex, lack of infiltrative pattern, and lack of lyticcomponent favors a benign or borderline process, whilelocation in the axial skeleton and size greater than 5 cmare reliable predictor of low-grade CHS [7]. Radiographicfindings including cortical destruction, soft tissue extension,and permeative changes such as the “moth-eaten pattern” arecommonly associated to malignancy. A permeative patternis often seen with high-grade CHS. Endosteal scallopingis a sign of aggressiveness in intramedullary cartilaginouslesions, but it is not completely diagnostic of malignancy.According to Murphey et al. [3], endosteal scalloping greaterthan two-thirds of the normal thickness of the long bonecortex is strong evidence of CHS over enchondroma. Thus,enchondromas and intramedullary low-grade CHSs (border-line tumors) of long bones often share similar radiologicalfeatures. These lesions should be diagnosed by histologicexamination after a complete resection of the lesion is madewhether it is an excision or a complete curettage. Magneticresonance (MR) imaging is also a preferred modality indiagnosis of cartilaginous tumor as well as in evaluation ofthe extent of marrow involvement and presence of soft tissueextension. On T1-weighted images after gadolinium contrastinjection, marked “septal” or “ring-and-arc” enhancementis typical for enchondromas and low-grade CHSs whichcorresponds to fibrous bands between the confluent carti-lage lobules on the histologic analysis. Inhomogeneous orhomogeneous enhancement of high-grade CHSs correlateswith high cellular areas on the microscopic examination [8].In addition, fast contrast-enhanced MR imaging could assistin differentiation between enchondromas and CHSs. In theadult patient, both early and exponential enhancement ispredictors of CHSs [9].

2.4. Pathology. Conventional intramedullary CHSs are largelesions, usually greater than 4 cm in size [3]. CHSs growin a lobulated pattern and are usually firm but may be

Table 1: Classification of CHSs.

Anatomic classification

Intramedullary (central)

Peripheral

Juxtacortical (periosteal)

Primary versus secondary

Primary

Arise de novo

Secondary

Arise in a benign precursor, either an osteochondroma orenchondromaArise in a benign precursor, either anosteochondromatosis or enchondromatosis

Histologic classification

Conventional

Dedifferentiated

Mesenchymal

Clear cell

soft, mucoid, or even gelatinous (Figure 2). They are whiteor bluish gray and often focally gritty because of matrixcalcifications. The presence of gray hemorrhagic fish fleshytissue or myxoid change is not uncommonly associated to ahigh-grade lesion. Histologic examination reveals lobules ofhyaline cartilage with variable degrees of cellularity, myxoidchange, and calcification. The chondrocytes usually haveenlarged, hyperchromatic nuclei with binucleation (Figures3(a)–3(c)). Necrosis and mitoses are mostly seen in high-grade lesions.

2.4.1. Grading. The grading is primarily based on nuclearsize, hyperchromasia, cellularity, and mitoses [10]. Thenuclear size is evaluated by assessing the tumor cells whetherthose cells are small and dark staining, moderate sized withvisible intranuclear detail, or large and pleomorphic. Thebackground is considered chondroid if definite lacunae areobserved and myxoid if the cells are separated by basophilicintercellular substance without definite lacunae.

Grade 1 (low-grade) lesions are poorly cellular withhyperchromatic round nuclei the size of a mature lympho-cyte. There are no mitotic figures, or nuclear atypia and thecells retain the lacunar pattern. Myxoid background is notpresent, but there may be some degenerative myxoid change.Binucleated cells are rare if any. Dense cellularity, presenceof significant numbers of moderately sized or larger nuclei,and mitotic figures are not features of low-grade CHS and,if present, indicate a higher grade of CHS [10]. Grade 2(intermediate) tumors are more cellular lesions characterizedby cells with nuclear enlargement; the chromatin may befine with presence of nucleoli, and mitotic activity is rarelypresent. The cells also retain the lacunar pattern, and thereis no myxoid change, although degenerative myxoid maybe present. When myxoid stroma appears, this is clue thatthe tumor may become aggressive or is frankly malignant

Sarcoma 3

Table 2: Summary of frequency, age, sex, and prognosis of CHSs.

Subtype Frequency Gender (M : F) Peak age 5-year survival

Conventional 85% 1.5–2 : 1 40–60 G1:89%, G2, and G3:57%

Periosteal <2% Slight male predilection 20–30 83%

Dedifferentiated 10% 1 : 1 50–60 <10%

Mesenchymal <2–13% 1 : 1 10–20 54.6%

Clear cell 1-2% 2.6 : 1 20–30 >80%

Figure 1: A well-demarcated cartilaginous tumor, measuring6.5 cm in greatest dimension, in the metaphysis of distal femur. Themass extends through the periosteum to adjacent soft tissue.

especially if it is associated to mitotic activity. Grade 3 (high-grade) tumors characteristically display 2 or more mitosesper ten high-power fields in the most cellular areas. Thereis usually a myxoid background associated to spindle orpleomorphic cells and the lacunar pattern is predominantlylost. Foci of necrosis are usually seen. For the purpose ofclarification, myxoid change may associate to malignancyin cartilage tumors or may be degenerative; the latter ischaracterized by the presence of myxoid areas withoutcellularity, while the myxoid change associated to malignancyis characterized by a tumor without lacunar pattern withatypical spindle or stellate cells floating in a myxoid stroma(Figures 4(a)–4(d) and 5) [1, 3].

2.4.2. Differential Diagnosis. More than 90% of conventionalCHSs are low- to intermediate-grade tumors and shouldbe distinguished from enchondroma. Permeation of corticalbone and/or preexisting medullary bone is most importantto distinguish CHSs from enchondromas [11], for which itis crucial to take biopsy material consisting of cortical andmedullary bones, and one should observe the growth pattern(Figure 6). One should avoid misinterpreting enchondromaas CHS by regarding areas of cartilage crushed into sur-rounding marrow spaces as true permeation in curettage

Figure 2: An intramedullary cartilaginous mass, measuring 9.5 cmin greatest dimension, in the metadiaphysis of proximal femur. Notethe white to bluish gray and lobulated cut surface and corticaldestruction.

specimen. The frequency of cellularity, double nuclei, andmitoses is similar between enchondroma and low-gradeCHS. Recent study suggests that presence of myxoid matrix≥20% and/or host bone entrapment strongly suggests CHS[12]. Extensive myxoid change is an ominous sign in achondroid lesion, and in such cases one should try to searchfor other histologic features suggesting CHS (Figure 7). Ifone sees areas with undoubtful neoplastic osteoid, the lesionshould be considered osteosarcoma with chondroblasticdifferentiation.

2.5. Genetics. So far, several attempts have been made toidentify reliable molecular markers and therapeutic targetsfor CHS [13]. Collagen subtype has been proposed to reflectdifferentiation of CHS, specifically collagen types II and Xas well as the proteoglycan aggrecan as a marker for matureneoplastic phenotype and collagen type I as a marker rep-resenting proliferative, “dedifferentiated” phenotype [14].Cyclooxygenase-2 overexpression was also proposed as amarker associated with histologic grade and poor survival[15]. However, none of these biologic markers has beenproved to provide independent prognostic information. Theattempt to assess the effect of celecoxib (Celebrex) onCHS growth using xenograft model did fail to attain a

4 Sarcoma

(a)

(b)

(c)

Figure 3: (a) Atypical chondrocytes with binucleation andeosinophilic cytoplasm; (b) chondrosarcoma with myxoid change;(c) chondrosarcoma with necrotic tumor cells and calcification.

satisfactory result [16]. Hedgehog signaling pathway that isimportant for development of central CHS is a potentialtherapeutic target and has been under preclinical tests[17]. Of the Indian hedgehog (IHH)/parathyroid hormone-related protein (PTHLH) pathway, induction of the PTHLHpathway and reactivation of bcl2 have been implicated inpathogenesis and progression of conventional CHS, and bcl2was suggested to be a reliable marker for the distinctionbetween low-grade CHS and enchondromas [18].

There is increase of genetic aberrations as CHSs progressfrom low to high grade. Although the role of p53 in CHS

pathobiology remains obscure, the presence of overexpres-sion of the p53 protein, 17p1 alterations, and TP53 mutationsmainly in almost all high-grade CHSs suggest that thep53 mutation is a late event involved in CHS progression[17, 19, 20]. Amplification of 12q13 and loss of 9p21 aresome of the few consistent genetic aberrations found inconventional CHS. The 12q13 region harbors MDM2, anegative regulator of p53, and the 9p21 region harborstwo cell cycle regulators, CDKN21/p16/INK4A and INK4A-p14ARF. The loss of INK4A/p16 expression was shown to berestricted to high-grade CHS, suggesting the role for CHSprogression [17, 21].

2.6. Prognostic Factors. For the prognosis of CHSs, thesingle most important predictor of local recurrence and/ormetastasis is histological grade, although several histologicaland clinical parameters such as tumor necrosis, mitotic rate,type of surgery, and tumor location have been suggested tobe associated with prognosis. Grade 1 tumor has indolentclinical behavior and no metastatic potential. The five-yearsurvival by grade was 89% for patients with grade 1 and57% for the combined group of patients with grades 2 and 3tumors. Only high tumor grades (2 and 3) were significantlyassociated with the probability of metastasis [22, 23].

2.7. Treatment. CHS is considered relatively resistant tochemotherapy or radiotherapy, and the mainstay of treat-ment is surgical treatment. Wide, en bloc excision is thepreferred surgical treatment in grade 2 or grade 3 CHS. Ingrade 1 CHS confined to the bone, extensive intralesionalcurettage followed by local adjuvant treatment and fillingthe cavity with bone graft has promising long-term clinicalresults and satisfactory local control [24].

2.8. CHSs in Specific Anatomic Locations

2.8.1. CHSs of the Hands and Feet

(1) Epidemiology. The hands and feet are rare sites for centralCHSs, whereas enchondromas are extraordinarily commonin these sites. The median age of the patients at the timeof diagnosis is 67 years (range, 21–87 years), with a slightpreference for female in contrast with CHSs located in othersites of the skeleton [25, 26].

(2) Sites of Involvement. Occurrences in the hand are morecommon than in the foot, with the proximal phalanx affectedmost often. The fifth digit has the highest incidence of CHS,and the fourth digit is the least common site in the hands[27].

(3) Clinical Features and Imaging. Pain (usually withoutfracture) is usually present at presentation. The median sizeis approximately 3 cm (range, 1–8 cm). Radiologically, CHSsare predominantly lucent lesions, sometimes with areas ofpunctuate calcification, with irregular cortical destructionand extension into surrounding soft tissue.

Sarcoma 5

(a) (b)

(c) (d)

Figure 4: (a) Grade 1 CHS with chondroid matrix and low cellularity. Note the soft tissue extension. (b) Grade 2 CHS with increasedcellularity and soft tissue extension. (c) Grade 3 CHS with more increased cellularity and cellular atypia. (d) Pleomorphic tumor cells andfrequent multinucleated cells in grade 3 CHS.

Cellularity

Cellular spindling

Mitosis, necrosis

Chondroid matrix

Myxoid change

Nuclear atypia

Grade 1 Grade 2 Grade 3

Figure 5: Schematic representation illustrating grading system of conventional CHS (not all CHSs follow this scheme).

6 Sarcoma

Figure 6: CHS permeating between preexisting bony trabeculae.Based on the cellularity and nuclear atypia, this lesion correspondsto grade 2 CHS.

Figure 7: Extensive myxoid change in CHS.

(4) Pathology. The distinction between enchondroma andCHS is often difficult on histologic examination becauseenchondromas of the hands and feet can exhibit increasedcellularity, binucleated cells, and hyperchromasia. Therefore,the most important histologic features of CHS in these sitesare permeative growth pattern between preexisting bonesand extension to soft tissue or joint space. Other histologicfeatures suggesting malignancy are the presence of myxoidchange and peripheral spindling of neoplastic chondrocytes.

(5) Prognostic Factors. CHSs of the hands and feet aretypically low-grade CHSs with a propensity to recur, but alimited tendency to metastasize in contrast to CHSs locatedelsewhere and complete excision with margins of normaltissue is curative in almost all cases [25, 26]. CHSs of thecalcaneus and the talus were more likely to metastasize.

2.8.2. CHSs of the Craniofacial Region

(1) Epidemiology and Sites of Involvement. Craniofacial CHSsaccount for 2% of all CHSs and have a predilection forthe skull base. The mean age of the patients at the time ofdiagnosis is 39 years (range, 10–79 years) (mean, 39 years).

The temporo-occipital junction is the most preferred site infrequency followed by clivus and sphenoethmoid complex[28].

(2) Clinical Features and Imaging. Most patients present withsymptoms related to the central nervous system. CT and MRimaging reveal bone destruction and associated soft tissuemasses usually containing punctuate areas of chondroidmineralization [3].

(3) Pathology. The majority of the tumors are conventionalCHSs of low to intermediate grade.

(4) Prognostic Factors. When the skull base CHS involves theclivus, distinction from chordoma is important because CHShas a much better prognosis than chordoma. Chordomatends to occur in patients a decade older than do CHSs andgrow much more rapidly. Skull base CHS has an excellentprognosis, and the 5- and 10-year disease-specific survivalrates are reported to be both 99%. In contrast, the 5- and10-year survival rates of chordoma have been reported to beapproximately 51% and 35%, respectively [28].

3. Periosteal (Juxtacortical) CHS

Periosteal CHS is a rare malignant hyaline cartilage tumorarising from the external surface of bone and has also beenreferred to as parosteal CHS.

3.1. Epidemiology and Sites of Involvement. This tumoraccounts for less than 2% of all CHSs and 0.2% of all bonetumors [29]. The tumor tends to affect younger adults thanconventional CHS with peak incidences in the third to fourthdecade of life. There is a slight male predilection. The mostcommon site is the metaphyseal region of the long bones,especially the femur and the humerus.

3.2. Clinical Features and Imaging. The clinical signs andsymptoms are mostly nonspecific and present with painor slowly growing mass. The lesion appears to involvethe cortex with indistinct margins. On radiographs, thetumor often appears as a radiolucent juxtacortical soft tissuemass with sharply defined borders containing calcificationscharacteristic of cartilage tumors.

3.3. Pathology. The lesion is usually large (mean size, 8.1 cm)and covered by a fibrous pseudocapsule that is continuouswith the underlying pseudocapsule. The mass is usuallyround to oval, lobulated, and gritty white with areas ofenchondral ossification and scattered calcification. Whereasperiosteal osteosarcoma is commonly fusiform and showsless-constant chondroid features [30]. Histological featuresare similar to those of conventional CHS that is composedof solid nodules of hyaline cartilage with variable amount ofmyxoid stroma. Nodules of the tumor can invade surround-ing soft tissues. Almost all periosteal CHS corresponds to

Sarcoma 7

grade 1 or 2 CHS. By definition, tumor osteoid should notbe present within the tumor [30].

3.4. Differential Diagnosis. The differential diagnosisincludes periosteal osteosarcoma. The peak age of incidenceof periosteal osteosarcoma is 10 years younger than CHS.The most common anatomic site is the diaphysis ordiaphyseal-metaphyseal area of the proximal tibia, followedby the femur and humerus. Periosteal osteosarcomapresents as small radiolucent lesions on the surface, withformation of spicules of bone perpendicular to the boneshaft. Histologically, periosteal osteosarcoma is intermediategrade, predominantly chondroblastic osteogenic sarcoma[31, 32]. Periosteal chondroma also should be distinguishedfrom periosteal CHS. Periosteal chondroma is a slow-growing benign cartilaginous tumor arising within or underthe periosteum. The size of the tumor is usually 1–3 cmin diameter. The peak incidence is in the second and thirddecades of life. The most common anatomic site is themetaphyseal region of long tubular bones. The proximalhumerus is the most common site, followed by the femurand short tubular bones of the hands and feet. Histologically,the tumor consists of lobules of hyaline cartilage with fociof myxoid change. Although periosteal chondroma can becellular and show binucleated chondrocytes, penetration intocancellous bone and nuclear anaplasia is not identified [33].

3.5. Prognostic Factors. The prognosis for patients with peri-osteal CHS is favorable compared to that of intramedullaryCHS. The overall 5-year metastasis-free survival is approx-imately 83%. The 5-year metastasis-free survival is less forpatients with grade 2 tumors (50%) than for patients withgrade 1 tumors (94%). Invasion of the medullary cavity is notfrequent. Metastasis is exceptional and occurs very late [34].Dedifferentiation has been rarely reported and is associatedwith poor prognosis [35].

4. Secondary CHS

Secondary CHS is a CHS arising in a benign precursor, eitheran osteochondroma or enchondroma. Although secondaryCHS can be either central or peripheral, peripheral lesionsare more common.

4.1. Peripheral Secondary CHSs

4.1.1. Epidemiology. Malignant transformation to peripheralCHS can be seen in 1% of solitary osteochondroma and3∼5% of patients with osteochondromatosis (hereditarymultiple exostoses, HME) [36, 37]. HME is an autosomaldominant skeletal disease characterized by the formationof multiple cartilage-capped bone tumors growing outwardfrom the metaphyses of long tubular bones.

4.1.2. Genetics. HME is caused by mutations in either of twogenes: exostosin-1 (EXT1), which is located on chromosome8q24.11–q24.13, and exostosin-2 (EXT2), which is locatedon chromosome 11p11-12. Most of the mutations in these

two genes are inactivating mutations (nonsense, frame shift,or splice-site mutations), causing premature termination ofthe EXT proteins and the loss of protein function [38, 39].In accordance with Knudson’s two-hit model, both allelesof EXT seem to need to be inactivated for osteochondromaformation. Loss of heterozygosity (LOH) of EXT1 and/orEXT2 is shown in some solitary osteochondromas and HMEs(Figures 8(a) and 8(b)) [40, 41]. The absence of LOHin a proportion of osteochondromas seems to be becausecartilaginous cap of osteochondroma is mosaic. Therefore,detection of a second mutational event depends on thebalance between EXT mutated and wild-type cells [42, 43].

In the growth plate, IHH regulates chondrocyte prolif-eration and differentiation in a tightly regulated paracrinefeedback loop, together with PTHLH, and deregulated IHHsignaling has been implicated in the pathogenesis of osteo-chondromas. The EXT genes encode glycosyltransferasesinvolved in the biosynthesis of heparan sulfate (HS) chainsat HS proteoglycans (HSPGs). HSPGs have been shownto play a role in the diffusion of IHH, PTHLH, andfibroblast growth factor (FGF), all of which are involvedin chondrocyte proliferation and differentiation. Therefore,EXT inactivation affects hedgehog signaling by defective HS(Figures 8(a) and 8(b)). In addition, disturbed hedgehogsignaling can cause defect in the body collar becausehedgehog is important for the formation of the bonycollar. EXT−/− cells lose their ability to respond to polaritysignals, then grow out of the bone, and then recruit normalcells to form an osteochondroma. IHH/PTHLH and FGFsignaling molecules are mostly absent in osteochondromasand reexpressed with the progression of osteochondromatowards peripheral CHSs. Upregulation of PTHLH and Bcl-2characterizes malignant transformation of osteochondroma[42, 44–46].

4.1.3. Clinical Features and Imaging. In osteochondromas,lesions that continue to grow or cause pain after skeletalmaturity suggest malignant transformation since osteochon-dromas only rarely grow after skeletal maturation. A thickhyaline cartilage cap greater than 1.5∼2.0 cm thick (inosteochondroma: 6 to 8 mm thick) in a skeletally maturepatient has been cited as a sign of possible malignanttransformation (Figure 9). However, the key for the diagnosisis the histopathologic differentiation of the cartilaginousproliferation. Radiographic findings that suggest malignancyare growth of a previously unchanged osteochondroma ina skeletally mature patient, irregular or indistinct lesionalsurface, focal regions of radiolucency in the interior of thelesion, erosion or destruction of the adjacent bone, and asoft tissue mass with scattered or irregular calcifications.Malignant transformation develops earlier in patients withHME (average, 25∼30 years) than in those with solitaryosteochondroma (average, 50∼55 years). Malignant trans-formation before the age of 20 is very unusual [37, 47].

4.1.4. Pathology. CHS arising in osteochondroma is usu-ally solitary and low grade in type, but multifocalityand dedifferentiation have also been reported (Figure 10).

8 Sarcoma

ChondrocytedifferentiationEXT

HSPG FGF-FGFR+

+

Diffusion of IHH

Patched-IHH

5. Ossification

3. Transition zone

2. Proliferating zone

Growth plate zones

IHHBcl-2

1. Resting chondrocytes

PTHLH4. Hypertrophiczone

(a)

Peripheral secondary CHS, low grade

Growth plate chondrocytes

HME Sporadic OC

Germline mutation ofEXT1 or EXT2

LOH of EXT1and/or EXT2

Genetic instabilityReactivation of PTHLHsignaling

Homozygous EXT1deletions

(b)

Growth plate chondrocytes

PTHR1 mutation

EnchondromatosisEnchondroma

?

Active PTHLH signalingInactive hedgehog signaling

Central secondary CHS, low grade

(c)

Figure 8: (a) Schematic representations of IHH/PTHLH signaling in the growth plate. The growth plate is composed of chondrocytes atdifferent stages of differentiation, finally leading to longitudinal bone growth. This process is tightly regulated by IHH/PTHLH signaling.IHH, expressed by transition zone chondrocytes, diffuses and binds to Patched (Ptc) in the hypertrophic zone, stimulating PTHLHexpression. PTHLH then binds to its receptor in the transition zone and upregulates Bcl-2, which inhibits chondrocyte differentiation anddownregulates IHH secretion. EXT gene products play a role in the diffusion of hedgehog proteins and FGF-FGFR interaction. Therefore,defect or absence of EXT genes results in an abnormal IHH diffusion pattern, leading to an osteochondroma. (b) Proposed genetic modelfor peripheral secondary CHS. Analogous to Knudson’s two-hit model, both alleles of an EXT gene are inactivated for osteochondromaformation in both HME and solitary osteochondroma. Genetic instability and reactivation of PTHLH signaling characterizes the malignanttransformation of osteochondroma. (c) Proposed genetic model for central secondary CHS. Patients with enchondromatosis infrequentlyharbor PTHR1 mutation, which disrupts the normal IHH-PTHLH feedback loop, leading to constitutive hedgehog signaling. In mostenchondromas, causative genetic or epigenetic changes have not been identified.

Sarcoma 9

Figure 9: Secondary CHS arising in osteochondroma. Note thethickening of cartilage cap (2.5 cm, arrows).

On microscopic examination, loss of cartilaginous colum-nar architecture, fibrous bands between cartilage lobules,increased nuclear atypia, mitosis, or myxoid changes arefeatures suggestive of malignant transformation.

4.1.5. Prognostic Factors and Treatment. Malignant transfor-mation of osteochondroma is usually treated with surgery.Treatment of patients with HME is more complex than thatof patients with solitary osteochondroma. Because most ofthese lesions are low-grade CHS, the overall prognosis isgood, with long-term survival in 70∼90% of patients. Localrecurrence rate varies with adequacy of the tumor margins,from 0–15% in widely resected cases to 57∼78% in cases withmarginal or intralesional resection [37, 48]

4.2. Central Secondary CHSs

4.2.1. Epidemiology. Central secondary CHSs develop asmalignant transformation of enchondroma (extremely rare)or enchondromatosis such as Ollier disease or Maffucci syn-drome. Patients with Ollier disease and Maffucci syndromehave a 25∼30% risk of developing CHS.

4.2.2. Genetics. The exact cause of Ollier disease and Maf-fucci syndrome remains to be elucidated, although mutationin the PTHR1 gene, c.448C>T (p.R150C), has been suggestedto cause enchondromatosis [49, 50]. A mutation in PTHR1disrupts the normal IHH-PTHLH feedback loop, causingconstitutive hedgehog signaling (Figure 8(c)).

4.2.3. Clinical Features and Imaging. Ollier disease is anonhereditary developmental abnormality characterized bymultiple enchondromas throughout the epiphyses, meta-physes, and diaphyses of the skeleton. The size, number,location, and evolution of enchondromas are quite variable.Clinically, Ollier disease often shows asymmetric, unilateralinvolvement of the lower extremities, but it is often bilateralin the hands and feet. Any portion of the skeleton formedby endochondral ossification can be affected; however,Ollier disease rarely affects bones formed by membranous

Figure 10: Grade 1 CHS arising in osteochondroma with loss oforganized architecture and mild nuclear atypia.

ossification, such as the skull and facial bones. Maffuccisyndrome is a condition in which enchondromatosis isassociated with soft tissue hemangiomas [51, 52].

The development of pain as well as the appearance of softtissue mass, areas of bone destruction, endosteal scalloping,periosteal reaction, and fracture without significant traumaraises the suspicion of malignant transformation of enchon-dromatosis.

4.2.4. Pathology. CHS arising in enchondromatosis is usuallya low-grade tumor like that of osteochondroma or HME.Therefore, identification of invasion to surrounding tissuesor marked myxoid change is helpful to the diagnosis.

4.2.5. Prognostic Factors and Treatment. CHS in enchondro-matosis has the same prognosis as conventional CHS anddepends on the site and grade of the tumor. Malignanttransformation of enchondromatosis is greater in Maffuccisyndrome than Ollier disease, and the prognosis is worsethan that of Ollier disease.

5. Dedifferentiated CHS

Dedifferentiated CHS is a distinct variant of CHS containinga well-differentiated cartilage tumor, either an enchondromaor a low-grade CHS, with an abrupt transition to foci havinghigh-grade noncartilaginous sarcoma (Figures 11(a)–11(c)).

5.1. Epidemiology and Sites of Involvement. DedifferentiatedCHS makes up 10% of CHSs. The average age of presentationis between 50 and 60 years. Men and women are affectedequally. The most common affected sites are the femur andpelvis [53, 54]. The majority of lesions occur centrally inthe medullary cavity of bone, although there are reports ofdedifferentiation in juxtacortical CHS or from preexistingosteochondroma [35, 55].

5.2. Clinical Features and Imaging. The patients present mostfrequently with pain (90%), followed by pathologic fracture

10 Sarcoma

(a)

(b)

(c)

Figure 11: (a) An ill-demarcated lobulating firm mass (8.5 × 4.0× 4.0 cm) in the proximal femur. Most of the mass which is whiteto bluish gray is conventional chondrosarcoma. The yellowishgray area in the center (arrows) is dedifferentiated area. (b) Theabrupt transition between conventional CHS and dedifferentiatedcomponent. (c) The dedifferentiated component consisting ofmalignant spindle cells without matrix formation (malignantfibrous histiocytoma).

and soft tissue mass. The proportion of noncartilaginouscomponent varies greatly and may be frequently osteosar-coma, fibrosarcoma, or malignant fibrous histiocytoma.Rhabdomyosarcoma, leiomyosarcoma, and angiosarcomahave been reported as the dedifferentiated component.Dedifferentiated CHSs have a wide range of radiologicalappearance; however, the presence of “tumoral dimorphism”

with cartilaginous component and aggressive lytic compo-nent invading adjacent soft tissues suggests a diagnosis ofdedifferentiated CHS [56].

5.3. Histogenesis. There are at least three hypotheses explain-ing the origin of dedifferentiated CHS. One theory is thatthe high-grade noncartilaginous tumor component arises ina long-standing low-grade cartilaginous tumor, particularlywhen the tumor is recurrent. The second hypothesis isthat noncartilaginous component arises simultaneously withCHS with ability to differentiate. The third theory is thatnoncartilaginous sarcoma represents malignant transforma-tion of adjacent inflamed but otherwise normal tissue [53].

5.4. Genetics. So far, no specific aberrations seem to beassociated with dedifferentiated CHS, although dediffer-entiated component tends to show aneuploidy, loss ofheterozygosity, and amplification and deletion more fre-quently [57]. Recently an array-based comparative genomichybridization (Array-CGH) study demonstrated statisticallysignificant association between high-grade tumor (grade IIIand dedifferentiated) and the recurrent genetic deletions at5q14.2∼q21.3, 6q16∼q25.3, 9p24.2∼q12, and 9p21.3 [58].Regarding dedifferentiated peripheral CHSs, dedifferentiatedcomponent shows more frequent expression of cyclin D1,p53, plasminogen activator inhibitor 1 (PAI-1), and CD44.The PTHLH signaling seems to be downregulated in chon-drogenic component of dedifferentiated peripheral CHSswhereas FGF signaling pathway in active, compared withsecondary peripheral CHSs without dedifferentiation [59].

5.5. Prognostic Factors. Because the dedifferentiated compo-nent determines the prognosis, its identification is a key formanagement. In spite of aggressive treatment, the overallsurvival rate is less than 10% at five years, with a mediansurvival time of 7.5 months. While local control is achievedin the majority of cases, distant disease remains the greatestclinical challenge, developing in 90% of patients [60].

6. Mesenchymal CHS

Mesenchymal CHS is a rare highly malignant tumor thatarises in bone but can occur in extraskeletal sites and ischaracterized by highly cellular areas composed of undiffer-entiated small round or spindle cells admixed with lobules ofmature hyaline cartilage.

6.1. Epidemiology. Mesenchymal CHS makes up less than 2∼13% of all primary CHSs. Mesenchymal CHS occurs at anyage, with peak incidences in the second to third decades oflife. There is no significant sex predilection [61].

6.2. Sites of Involvement. The skeletal tumors show a wide-spread distribution. The craniofacial region is the mostfrequently affected site (15∼30%), specifically the mandibleand maxilla. Other common sites include femur, ribs, spine,pelvis, and humerus [1, 3]. About 7% of the osseous lesionsare reported to be multicentric. Up to one-third of the

Sarcoma 11

(a)

(b)

Figure 12: (a) Mesenchymal CHS showing bimorphic pattern,consisting of islands of hyaline cartilage and undifferentiatedcells. Note hemangiopericytomatous vascular pattern. (b) Theundifferentiated small cells with relatively uniform oval to roundnuclei and scanty amount of cytoplasm.

lesions primarily affect extraskeletal sites. The meninges(cranial > spinal) are the most common sites of extraskeletalinvolvement, followed by lower extremity [61].

6.3. Clinical Features and Imaging. Most patients presentedwith pain and/or swelling. Duration of symptoms prior tothe histologic diagnosis is quite variable, ranging from fewdays to several years. Oncogenic osteomalacia secondary tomesenchymal CHS has been reported.

Roentgenographically, mesenchymal CHSs in bone fre-quently resemble ordinary CHSs, showing osteolytic anddestructive appearances with stippled calcifications. Tumorsin extraskeletal sites are almost always identified as a masswith flocculent or stippled calcific densities. Sclerosis orperiosteal reaction is uncommon, while expansion of thebone, cortical destruction, or cortical breakthrough withextraosseous extension of soft tissue is common [61].

6.4. Pathology

6.4.1. Gross Findings. Grossly, the tumors are gray to tan,firm to soft, and usually well defined and well circumscribed.The size of the tumor ranges from 3 cm to 30 cm in diameter.

Figure 13: The undifferentiated small cell component in mesenchy-mal CHS strongly positive for CD99.

Lobulation is infrequent. Most lesions contain hard miner-alized deposits that vary in amount from scattered foci toprominent areas. Some tumors show a clearly cartilaginousappearance, even in a small area. Necrosis is uncommon butmay be prominent [62]. Bony expansion with cortical thin-ning or bone destruction and soft tissue invasion is frequent.

6.4.2. Microscopic Findings. Histologically, a bimorphic pat-tern with cellular zones of undifferentiated small or spindlecells and islands of hyaline cartilage is pathognomonic(Figure 12). The amount of cartilage is highly variable.Transition from cellular areas to zones with hyaline cartilageis usually abrupt but can be gradual. The undifferentiatedcells with oval nuclei frequently tend to be arranged in avague alveolar pattern or in solid sheets, resembling Ewingsarcoma. A hemangiopericytomatous vascular pattern is seenin most cases. Osteoclastic giant cells can be seen, usuallyadjacent to the cartilaginous islands [61].

6.4.3. Immunohistochemical Findings. Immunohistochemi-cally, the cartilaginous area is strongly positive for S-100protein, whereas only scattered single cells in the undiffer-entiated areas stain for this antigen. The undifferentiatedsmall cell component of mesenchymal CHSs is consistentlypositive for CD99 and may stain for vimentin and Leu7 whilenegative for osteocalcin, actin, cytokeratin, and epithelialmembrane antigen (EMA) (Figure 13). SOX9 is almostinvariably positive in both components [63, 64].

6.5. Genetics. At present, cytogenetic findings of mes-enchymal CHS are rarely reported. An identical Robert-sonian translocation involving chromosomes 13 and 21(der(13;21)(q10;q10)) has been detected in two cases ofmesenchymal CHSs, possibly representing a characteristicrearrangement for this histopathologic entity [65]. Thet(11;22) of the Ewing family of tumors is not seen inmesenchymal CHS. Although approximately 60% of thetumors demonstrate p53 overexpression, no mutation hasbeen found within exons 5∼9 regions [66].

12 Sarcoma

Figure 14: Clear cell CHS located in the epiphysis metaphysis ofthe proximal femur. The lesion is lytic, slightly expansile, and welldelineated from the adjacent normal bone.

6.6. Prognostic Factors. The prognosis of mesenchymal CHSis poor. However, the clinical course may be protracted.Because local recurrence or metastasis sometimes is encoun-tered even after more than 20 years, long-term follow-up isessential. The 5-year survival rate was 54.6%, and the 10-year survival rate was 27.3% in a group of 23 patients fromthe Mayo Clinic. The most frequent site of metastasis is thelung. Ablative surgical treatment seems to be the treatmentof choice [61].

7. Clear Cell CHS

Clear cell CHS is a rare, low-grade malignant tumorcharacterized by clear cytoplasm of the tumor cells.

7.1. Epidemiology and Sites of Involvement. The tumoraccounts for about 1-2% of all CHSs. The lesion affectsmales more commonly than females (2.6 : 1) and has apredilection for the end of long bones (epiphysis) in contrastto conventional CHS which tends to occur in the meta-diaphyseal regions of the bone. Although the proximal femurand humeral head are the sites most commonly affected inabout two-thirds of the cases by this lesion, most bonesincluding spine, rib, pelvis, and hands and feet can beinvolved. The age range is wide with peak incidences in thethird to fourth decades of life [67].

7.2. Clinical Features and Imaging. Clinically, clear cell CHSpresents one or two decades later than chondroblastoma. Theclinical symptoms are nonspecific; however, pain is the mostcommon presenting symptom. More than half of patientshave pain for longer than a year.

Figure 15: Gross finding of clear cell chondrosarcoma in theproximal femur.

Roentgenographically, the lesion is typically located inthe epiphysis metaphysis of long bone. The lesion is mostoften purely lytic and slightly expansile, with a sharpmargin between the tumor and the adjacent normal bone(Figure 14). Typically, there is no cortical destruction orperiosteal new bone formation. More than one-third of thelong bone lesions contained matrix mineralization with acharacteristic chondroid appearance. Pathologic fracture isoccasionally present. Flat bone lesions are typically lytic andexpansile with occasionally demonstrated areas of corticaldisruption. Typically, matrix mineralization, when present, isamorphous. Adjacent bone marrow edema is typically absentor only minimally observed [68].

7.3. Pathology. Grossly, the tumors are well circumscribedand may be either firm or soft. Grossly cartilage is not usuallypresent (Figure 15). The lesions consist of clear cells arrangedin an indistinct lobular pattern and having round, large,centrally located nuclei with clear cytoplasm and distinctcytoplasmic membranes (Figure 16). Clear cell componentsin clear cell CHS are accompanied by “conventional” fociof CHS in less than 50% of cases. Secondary findingsincluding areas of osteogenesis, osteoclast-like giant cells,and zones resembling aneurysmal bone cyst or giant celltumor of bone could be found. Mitotic figures are rare[69]. Dedifferentiation to high-grade sarcoma has beenrarely reported [70]. Clear cell CHS should be differentiatedfrom other osseous tumors which can show focal or diffuseclear cell changes such as osteosarcoma, chondroblastoma,chordoma, adamantinoma, and Ewing’s sarcoma/primitiveneuroectodermal tumor as well as metastatic renal clear cellcarcinoma. The clear cells are positive for type II collagen aswell as S-100 protein and aggrecan [71].

7.4. Genetics. Recent molecular genetic studies show thatgenetic alterations of p53 are infrequent in clear cell CHSin spite of substantial overexpression of p53 [72]. Tumor-specific cytogenetic change is currently unknown; althougha case report described clonal chromosomal abnormalities inthree of four cases of clear cell CHS [73].

Sarcoma 13

(a)

(b)

Figure 16: (a) Tumor cells with clear cytoplasm between irregulartrabeculae of woven bone. (b) Higher magnification of tumor cellswith abundant clear or faintly granular cytoplasm and central roundnuclei containing occasional prominent nucleoli.

7.5. Prognostic Factors. Clear cell CHS is a low-grade malig-nancy and usually curable by en bloc resection. About 25% ofpatients experience local recurrences or metastases. However,tumor-related death is uncommon, particularly when thelesion has been completely resected en bloc [69].

References

[1] C. D. M. Fletcher, K. K. Unni, and F. Mertens, World HealthOrganization Classification of Tumors: Pathology and Geneticsof Tumors of Soft Tissue and Bone, IARC Press, Lyon, France,2002.

[2] A. Jemal, R. Siegel, E. Ward et al., “Cancer statistics, 2006,” Ca:A Cancer Journal for Clinicians, vol. 56, no. 2, pp. 106–130,2006.

[3] M. D. Murphey, E. A. Walker, A. J. Wilson, M. J. Kransdorf, H.T. Temple, and F. H. Gannon, “From the archives of the AFIP:imaging of primary chondrosarcoma: radiologic-pathologiccorrelation,” Radiographics, vol. 23, no. 5, pp. 1245–1278,2003.

[4] A. Y. Giuffrida, J. E. Burgueno, L. G. Koniaris, J. C. Gutierrez,R. Duncan, and S. P. Scully, “Chondrosarcoma in the UnitedStates (1973 to 2003): an analysis of 2890 cases from the SEERdatabase,” Journal of Bone and Joint Surgery. American, vol. 91,no. 5, pp. 1063–1072, 2009.

[5] M. D. Murphey, D. J. Flemming, S. R. Boyea, J. A. Bojes-cul, D. E. Sweet, and H. T. Temple, “Enchondroma versuschondrosarcoma in the appendicular skeleton: differentiatingfeatures,” Radiographics, vol. 18, no. 5, pp. 1213–1237, 1998.

[6] D. J. Flemming and M. D. Murphey, “Enchondroma andchondrosarcoma,” Seminars in Musculoskeletal Radiology, vol.4, no. 1, pp. 59–71, 2000.

[7] M. U. A. Geirnaardt, J. Hermans, J. L. Bloem et al., “Usefulnessof radiography in differentiating enchondroma from centralgrade I chondrosarcoma,” American Journal of Roentgenology,vol. 169, no. 4, pp. 1097–1104, 1997.

[8] M. J. A. Geirnaerdt, J. L. Bloem, F. Eulderink, P. C.W. Hogendoorn, and A. H. M. Taminiau, “Cartilaginoustumors: correlation of gadolinium-enhanced MR imaging andhistopathologic findings,” Radiology, vol. 186, no. 3, pp. 813–817, 1993.

[9] M. J. A. Geirnaerdt, P. C. W. Hogendoorn, J. L. Bloem, A.H. M. Taminiau, and H. J. van der Woude, “Cartilaginoustumors: fast contrast-enhanced MR imaging,” Radiology, vol.214, no. 2, pp. 539–546, 2000.

[10] H. L. Evans, A. G. Ayala, and M. M. Romsdahl, “Prognosticfactors in chondrosarcoma of bone. A clinicopathologicanalysis with emphasis on histologic grading,” Cancer, vol. 40,no. 2, pp. 818–831, 1977.

[11] J. M. Mirra, R. Gold, J. Downs, and J. J. Eckhardt, “A new his-tologic appraoch to the differentiation of enchondroma andchondrosarcoma of the bones. A clinicopathologic analysis of51 cases,” Clinical Orthopaedics and Related Research, vol. 201,pp. 214–237, 1985.

[12] D. Eefting, Y. M. Schrage, M. J. A. Geirnaerdt et al., “Assess-ment of interobserver variability and histologic parametersto improve reliability in classification and grading of centralcartilaginous tumors,” American Journal of Surgical Pathology,vol. 33, no. 1, pp. 50–57, 2009.

[13] W. A. Chow, “Update on chondrosarcomas,” Current Opinionin Oncology, vol. 19, no. 4, pp. 371–376, 2007.

[14] T. Aigner, S. Muller, D. Neureiter, D. M. Illstrup, T. Kirchner,and J. Bjornsson, “Prognostic relevance of cell biologicand biochemical features in conventional chondrosarcomas,”Cancer, vol. 94, no. 8, pp. 2273–2281, 2002.

[15] M. Endo, T. Matsumura, T. Yamaguchi et al.,“Cyclooxygenase-2 overexpression associated with a poorprognosis in chondrosarcomas,” Human Pathology, vol. 37,no. 4, pp. 471–476, 2006.

14 Sarcoma

[16] Y. M. Schrage, I. Machado, D. Meijer et al., “COX-2 expressionin chondrosarcoma: a role for celecoxib treatment?” EuropeanJournal of Cancer, vol. 46, no. 3, pp. 616–624, 2010.

[17] J. V. M. G. Bovee, P. C. W. Hogendoorn, J. S. Wunder, and B. A.Alman, “Cartilage tumours and bone development: molecularpathology and possible therapeutic targets,” Nature ReviewsCancer, vol. 10, no. 7, pp. 481–488, 2010.

[18] L. B. Rozeman, L. Hameetman, A. M. Cleton-Jansen, A.H. M. Taminiau, P. C. W. Hogendoorn, and J. V. M. G.Bovee, “Absence of IHH and retention of PTHrP signallingin enchondromas and central chondrosarcomas,” Journal ofPathology, vol. 205, no. 4, pp. 476–482, 2005.

[19] R. M. Terek, J. H. Healey, P. Garin-Chesa, S. Mak, A.Huvos, and A. P. Albino, “p53 mutations in chondrosarcoma,”Diagnostic Molecular Pathology, vol. 7, no. 1, pp. 51–56, 1998.

[20] D. J. Papachristou, M. A. Goodman, K. Cieply, J. L. Hunt, andU. N. M. Rao, “Comparison of allelic losses in chondroblas-toma and primary chondrosarcoma of bone and correlationwith fluorescence in situ hybridization analysis,” HumanPathology, vol. 37, no. 7, pp. 890–898, 2006.

[21] H. M. van Beerendonk, L. B. Rozeman, A. H. M. Taminiau etal., “Molecular analysis of the INK4A/INK4A-ARF gene locusin conventional (central) chondrosarcomas and enchondro-mas: indication of an important gene for tumour progression,”Journal of Pathology, vol. 202, no. 3, pp. 359–366, 2004.

[22] J. Bjornsson, R. A. McLeod, K. K. Unni, D. M. Ilstrup, and D. J.Pritchard, “Primary chondrosarcoma of long bones and limbgirdles,” Cancer, vol. 83, no. 10, pp. 2105–2119, 1998.

[23] L. B. Rozeman, P. C. W. Hogendoorn, and J. V. M. G. Bovee,“Diagnosis and prognosis of chondrosarcoma of bone,” ExpertReview of Molecular Diagnostics, vol. 2, no. 5, pp. 461–472,2002.

[24] H. Gelderblom, P. C. W. Hogendoorn, S. D. Dijkstra et al.,“The clinical approach towards chondrosarcoma,” Oncologist,vol. 13, no. 3, pp. 320–329, 2008.

[25] J. V. M. G. Bovee, R. O. van der Heul, A. H. M. Taminiau,and P. C. W. Hogendoorn, “Chondrosarcoma of the phalanx:a locally aggressive lesion with minimal metastatic potential: areport of 35 cases and a review of the literature,” Cancer, vol.86, no. 9, pp. 1724–1732, 1999.

[26] M. L. Ostrowski and H. J. Spjut, “Lesions of the bones of thehands and feet,” American Journal of Surgical Pathology, vol.21, no. 6, pp. 676–690, 1997.

[27] A. Ogose, K. K. Unni, R. G. Swee, G. K. May, C. M. Rowland,and F. H. Sim, “Chondrosarcoma of small bones of the handsand feet,” Cancer, vol. 80, no. 1, pp. 50–59, 1997.

[28] A. E. Rosenberg, G. P. Nielsen, S. B. Keel et al., “Chondrosar-coma of the base of the skull: a clinicopathologic study of200 cases with emphasis on its distinction from chordoma,”American Journal of Surgical Pathology, vol. 23, no. 11, pp.1370–1378, 1999.

[29] S. Chaabane, M. C. Bouaziz, C. Drissi, L. Abid, and M.F. Ladeb, “Periosteal chondrosarcoma,” American Journal ofRoentgenology, vol. 192, no. 1, pp. W1–W6, 2009.

[30] S. Kenan, I. F. Abdelwahah, M. J. Klein, G. Hermann, and M.M. Lewis, “Lesions of juxtacortical origin (surface lesions ofbone),” Skeletal Radiology, vol. 22, no. 5, pp. 337–357, 1993.

[31] F. Bertoni, S. Boriani, M. Laus, and M. Campanacci,“Periosteal chondrosarcoma and periosteal osteosarcoma. Twodistinct entities,” Journal of Bone and Joint Surgery. British, vol.64, no. 3, pp. 370–376, 1982.

[32] K. K. Unni, D. C. Dahlin, and J. W. Beabout, “Periostealosteogenic sarcoma,” Cancer, vol. 37, no. 5, pp. 2476–2485,1976.

[33] T. Nojima, K. K. Unni, R. A. McLeod, and D. J. Pritchard,“Periosteal chondroma and periosteal chondrosarcoma,”American Journal of Surgical Pathology, vol. 9, no. 9, pp. 666–677, 1985.

[34] P. J. Papagelopoulos, E. C. Galanis, A. F. Mavrogenis et al.,“Survivorship analysis in patients with periosteal chondrosar-coma,” Clinical Orthopaedics and Related Research, no. 448, pp.199–207, 2006.

[35] A. Mitchell, J. R. Rudan, and P. V. Fenton, “Juxtacorticaldedifferentiated chondrosarcoma from a primary periostealchondrosarcoma,” Modern Pathology, vol. 9, no. 3, pp. 279–283, 1996.

[36] H. D. Dorfman and B. Czerniak, Bone Tumors, Mosby, St.Louis, Mo, USA, 1998.

[37] M. D. Murphey, J. J. Choi, M. J. Kransdorf, D. J. Flemming,and F. H. Gannon, “Imaging of osteochondroma: variantsand complications with radiologic-pathologic correlation,”Radiographics, vol. 20, no. 5, pp. 1407–1434, 2000.

[38] I. Jennes, E. Pedrini, M. Zuntini et al., “Multiple osteo-chondromas: mutation update and description of the mul-tiple osteochondromas mutation database (MOdb),” HumanMutation, vol. 30, no. 12, pp. 1620–1627, 2009.

[39] R. M. Terek, “Recent advances in the basic science ofchondrosarcoma,” Orthopedic Clinics of North America, vol.37, no. 1, pp. 9–14, 2006.

[40] J. V. M. G. Bovee, A. M. Cleton-Jansen, W. Wuyts et al.,“EXT-mutation analysis and loss of heterozygosity in sporadicand hereditary osteochondromas and secondary chondrosar-comas,” American Journal of Human Genetics, vol. 65, no. 3,pp. 689–698, 1999.

[41] L. Hameetman, K. Szuhai, A. Yavas et al., “The role of EXT1in nonhereditary osteochondroma: identification of homozy-gous deletions,” Journal of the National Cancer Institute, vol.99, no. 5, pp. 396–406, 2007.

[42] J. V. M. G. Bovee, “EXTra hit for mouse osteochondroma,”Proceedings of the National Academy of Sciences of the UnitedStates of America, vol. 107, no. 5, pp. 1813–1814, 2010.

[43] K. B. Jones, V. Piombo, C. Searby et al., “A mouse model ofosteochondromagenesis from clonal inactivation of Ext1 inchondrocytes,” Proceedings of the National Academy of Sciencesof the United States of America, vol. 107, no. 5, pp. 2054–2059,2010.

[44] C. E. de Andrea, M. Wiweger, F. Prins, J. V. M. G. Bovee, S.Romeo, and P. C. W. Hogendoorn, “Primary cilia organizationreflects polarity in the growth plate and implies loss of polarityand mosaicism in osteochondroma,” Laboratory Investigation,vol. 90, no. 7, pp. 1091–1101, 2010.

[45] C. M. A. Reijnders, C. J. F. Waaijer, A. Hamilton et al.,“No haploinsufficiency but loss of heterozygosity for EXT inmultiple osteochondromas,” American Journal of Pathology,vol. 177, no. 4, pp. 1946–1957, 2010.

[46] J. V. M. G. Bovee, L. J. C. M. van den Broek, A. M. Cleton-Jansen, and P. C. W. Hogendoorn, “Up-regulation of PTHrPand Bcl-2 expression characterizes the progression of osteo-chondroma towards peripheral chondrosarcoma and is a lateevent in central chondrosarcoma,” Laboratory Investigation,vol. 80, no. 12, pp. 1925–1933, 2000.

[47] E. W. Brien, J. M. Mirra, and J. V. Luck, “Benign andmalignant cartilage tumors of bone and joint: their anatomicand theoretical basis with an emphasis on radiology, pathologyand clinical biology. II. Juxtacortical cartilage tumors,” SkeletalRadiology, vol. 28, no. 1, pp. 1–20, 1999.

[48] P. I. J. M. Wuisman, P. C. Jutte, and T. Ozaki, “Secondarychondrosarcoma in osteochondromas. Medullary extension in

Sarcoma 15

15 of 45 cases,” Acta Orthopaedica Scandinavica, vol. 68, no. 4,pp. 396–400, 1997.

[49] A. Couvineau, V. Wouters, G. Bertrand et al., “PTHR1mutations associated with Ollier disease result in receptor lossof function,” Human Molecular Genetics, vol. 17, no. 18, pp.2766–2775, 2008.

[50] L. B. Rozeman, L. Sangiorgi, I. H. Briaire-de Bruijn et al.,“Enchondromatosis (Ollier disease, Maffucci syndrome) is notcaused by the PTHR1 mutation p.R150C,” Human Mutation,vol. 24, no. 6, pp. 466–473, 2004.

[51] C. Silve and H. Juppner, “Ollier disease,” Orphanet Journal ofRare Diseases, vol. 1, no. 1, article 37, 2006.

[52] S. Pannier and L. Legeai-Mallet, “Hereditary multiple exos-toses and enchondromatosis,” Best Practice and Research:Clinical Rheumatology, vol. 22, no. 1, pp. 45–54, 2008.

[53] A. D. Mitchell, K. Ayoub, D. C. Mangham, R. J. Grimer, S. R.Carter, and R. M. Tillman, “Experience in the treatment ofdedifferentiated chondrosarcoma,” Journal of Bone and JointSurgery. British, vol. 82, no. 1, pp. 55–61, 2000.

[54] R. J. Grimer, G. Gosheger, A. Taminiau et al., “Dedifferentiatedchondrosarcoma: prognostic factors and outcome from aEuropean group,” European Journal of Cancer, vol. 43, no. 14,pp. 2060–2065, 2007.

[55] F. Bertoni, D. Present, P. Bacchini et al., “Dedifferentiatedperipheral chondrosarcomas. A report of seven cases,” Cancer,vol. 63, no. 10, pp. 2054–2059, 1989.

[56] G. Bierry, A. Feydy, F. Larousserie et al., “Dedifferentiatedchondrosarcoma: radiologic-pathologic correlation,” Journalde Radiologie, vol. 91, no. 3, pp. 271–279, 2010.

[57] J. V. M. G. Bovee, A. M. Cleton-Jansen, C. Rosenberg, A.H. M. Taminiau, C. J. Cornelisse, and P. C. W. Hogendoorn,“Molecular genetic characterization of both components ofa dedifferentiated chondrosarcoma, with implications for itshistogenesis,” Journal of Pathology, vol. 189, no. 4, pp. 454–462,1999.

[58] M. Hameed, C. Ulger, D. Yasar et al., “Genome profiling ofchondrosarcoma using oligonucleotide array-based compara-tive genomic hybridization,” Cancer Genetics and Cytogenetics,vol. 192, no. 2, pp. 56–59, 2009.

[59] L. B. Rozeman, I. H. B. de Bruijn, P. Bacchini et al., “Dedif-ferentiated peripheral chondrosarcomas: regulation of EXT-downstream molecules and differentiation-related genes,”Modern Pathology, vol. 22, no. 11, pp. 1489–1498, 2009.

[60] I. D. Dickey, P. S. Rose, B. Fuchs et al., “Dedifferentiatedchondrosarcoma: the role of chemotherapy with updatedoutcomes,” Journal of Bone and Joint Surgery. American, vol.86, no. 11, pp. 2412–2418, 2004.

[61] Y. Nakashima, K. K. Unni, T. C. Shives, R. G. Swee, and D.C. Dahlin, “Mesenchymal chondrosarcoma of bone and softtissue. A review of 111 cases,” Cancer, vol. 57, no. 12, pp. 2444–2453, 1986.

[62] A. H. Salvador, J. W. Beabout, and D. C. Dahlin, “Mesenchy-mal chondrosarcoma—observations on 30 new cases,” Cancer,vol. 28, no. 3, pp. 605–615, 1971.

[63] K. Devoe and N. Weidner, “Immunohistochemistry of smallround-cell tumors,” Seminars in Diagnostic Pathology, vol. 17,no. 3, pp. 216–224, 2000.

[64] J. C. Fanburg-Smith, A. Auerbach, J. S. Marwaha, Z. Wang,and E. J. Rushing, “Reappraisal of mesenchymal chondrosar-coma: novel morphologic observations of the hyaline cartilageand endochondral ossification and β-catenin, Sox9, andosteocalcin immunostaining of 22 cases,” Human Pathology,vol. 41, no. 5, pp. 653–662, 2010.

[65] S. Naumann, P. A. Krallman, K. Krishnan Unni, M. E. Fidler, J.R. Neff, and J. A. Bridge, “Translocation der(13;21)(q10;q10)in skeletal and extraskeletal mesenchymal chondrosarcoma,”Modern Pathology, vol. 15, no. 5, pp. 572–576, 2002.

[66] Y. K. Park, H. R. Park, S. G. Chi et al., “Overexpression of p53and rare genetic mutation in mesenchymal chondrosarcoma,”Oncology Reports, vol. 7, no. 5, pp. 1041–1047, 2000.

[67] J. Bjornsson, K. K. Unni, D. C. Dahlin, J. W. Beabout, and F.H. Sim, “Clear cell chondrosarcoma of bone: observations in47 cases,” American Journal of Surgical Pathology, vol. 8, no. 3,pp. 223–230, 1984.

[68] M. S. Collins, T. Koyama, R. G. Swee, and C. Y. Inwards, “Clearcell chondrosarcoma: radiographic, computed tomographicand magnetic resonance findings in 34 patients with patho-logic correlation,” Skeletal Radiology, vol. 32, no. 12, pp. 687–694, 2003.

[69] P. E. Swanson, “Clear cell tumors of bone,” Seminars inDiagnostic Pathology, vol. 14, no. 4, pp. 281–291, 1997.

[70] R. K. Kalil, C. Y. Inwards, K. K. Unni et al., “Dedifferentiatedclear cell chondrosarcoma,” American Journal of SurgicalPathology, vol. 24, no. 8, pp. 1079–1086, 2000.

[71] S. Soder, A. M. Oliveira, C. Y. Inwards, S. Muller, and T. Aigner,“Type II collagen, but not aggrecan expression, distinguishesclear cell chondrosarcoma and chondroblastoma,” Pathology,vol. 38, no. 1, pp. 35–38, 2006.

[72] Y. K. Park, H. R. Park, S. G. Chi, S. Ushigome, and K. K. Unni,“Overexpression of p53 and absent genetic mutation in clearcell chondrosarcoma,” International Journal of Oncology, vol.19, no. 2, pp. 353–357, 2001.

[73] J. Nishio, J. D. Reith, A. Ogose, G. Maale, J. R. Neff, and J. A.Bridge, “Cytogenetic findings in clear cell chondrosarcoma,”Cancer Genetics and Cytogenetics, vol. 162, no. 1, pp. 74–77,2005.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 953047, 6 pagesdoi:10.1155/2011/953047

Methodology Report

Technical Notes on Endoscopic Transnasal TranssphenoidalApproach for Clival Chondrosarcoma

Atsushi Kuge,1 Shinya Sato,1 Kaori Sakurada,1 Sunao Takemura,1 Zensho Kikuchi,1

Yuki Saito,1 and Takamasa Kayama1, 2

1 Department of Neurosurgery, Yamagata University Faculty of Medicine, 2-2-2 Iidanishi, Yamagata 990-9585, Japan2 National Cancer Center, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104-0045, Japan

Correspondence should be addressed to Atsushi Kuge, akuge@med.id.yamagata-u.ac.jp

Received 30 August 2010; Revised 11 January 2011; Accepted 17 January 2011

Academic Editor: Jose Casanova

Copyright © 2011 Atsushi Kuge et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Although there are various operative approaches for clival tumors, a transsphenoidal approach is one of choices when the maintumor extention is in an anterior-posterior direction with a slight lateral extension. However, this approach sometimes providesonly narrow and deep operative field. Recently, endoscopic transnasal transsphenoidal approach is quite an effective approach forclival tumors because of the improvement of surgical instruments, image guidance systems, and techniques and materials of woundclosure. In this paper, we describe the effectiveness, technical problems, and solution of this approach based on our experienceswith two clival chondrosarcomas that was removed by endoscopic transnasal transsphenoidal approach.

1. Introduction

Clival chondrosarcomas are rare group that are thoughtto originate from primitive mesenchymal cells or from theembryonal portion of the cartilaginous matrix of the cra-nium. And these tumors have aggressive features, infiltrativeneoplasms. Although it is generally agreed that surgery is thechoice of treatment for clival chondromas and chodrosar-comas, it is still controvertial whether it is preferable toattempt a radical resection, which implies wide exposure,extensive dissection, and the potential for significant surgicalmorbidity, or to determine the oncological features ofthe tumors and plan local control through more limitedapproaches. Transbasal, transmaxillary, transoral, transfacial,and subtemporal approaches have been used for clival lesion[1–5].

On the other hand, transsphenoidal approach [6, 7] wasused from 1960s [8], this approach was not widely spreadbecause of narrow and deep operative field using microscope,and difficulty of wound closure complicated cerebrospinalfluid leakage [7]. However, an endoscopic transsphenoid-al approach is now quite an effective approach for clival

tumors because of the improvement of surgical instruments,image guidance systems, and materials of wound closuresuch as fibrin glue. In this paper, we describe the effec-tiveness, technical problems and solution of this approachbased on our experiences with two clival chondrosarcomasthat was removed by endoscopic transnasal transsphenoidalapproach.

2. Patients and Methods

Consecutive two patients with a clival chondrosarcomatreated using an endonasal endoscopic approach were iden-tified [9]. In all cases, frameless neurosurgical navigation(VectorVision: Brain LAB Co., Ltd.) was used.

Case 1. A fifty-six-year-old male was suffering from diplopia.He showed left abducens nerve palsy, and MRI revealedclival enhanced mass lesion extending toward brainstem andcavernous sinus. First surgery was performed by craniotomy,and tumor was resected partially and get decompressionof brainstem. Immunohistochemically, vimentin was posi-tive, epithelial membrane antigen (EMA) and cytokeratin

2 Sarcoma

(a) (b)

(c) (d)

Figure 1: An original long navigation pointer (a) registered by the Universal Instrument Integration system ((b): VectorVision: BrainLABCo., Ltd.). Long and slim instruments we used for this operation, highspeed drill ((c): Primado: Nakanishi Co., Ltd.) and ultrasonic aspirator((d): SONOPET: M&M Co., Ltd.).

Figure 2: Case 1. Upper: MRI images 33 months after the first operation. The residual tumor is regrowing. Lower: MRI images after surgery,midline part of tumor was resected and decompressed brainstem.

(CAM5.2) were negative, and pathological diagnosis waschondrosarcoma. Thirty-three months after first surgery,tumor enlargement was confirmed. Second surgery was per-formed by endoscopic transnasal transsphenoidal approachbecause tumor was mainly extended toward the posterior

part. We used an endoscope with long axis (180 mm) andangled tip (0, 30, 70 degrees: HOPKINSII: KARL STORZCo., Ltd.). And we made special navigation probe to getmore correct information. It was registered by UniversalInstrument Integration System (VectorVision: BrainLAB

Sarcoma 3

R

RA

LAH

LA PH

AR

H

10 cm

10 cm

10 cm

Figure 3: Case 2. Upper: MRI images after the second operation. Lower: MRI after transsphenoidal surgery. Tumor was resected exceptsuprasellar part and decompressed brainstem.

CO., Ltd.) which could make favorite length of navigationprobe using materials you like. And instruments which longand slim curettes (Zonne Co., Ltd.), ultrasonic aspiratorprobe developed for abdominal surgery (SONOPET: M&MCo., Ltd.), high-speed drill (Primado: Nakanishi Co., Ltd.)which have curved handle did not to disturb operative fieldwithout the interference of each tools (Figure 1). We couldget resection of main part of tumor without cavernouslesions. Dural defect was observed, and it was repaired byfilling sphenoid sinus by autologous fat and fibrin glue. Aftersurgery, new neurological deficit and cerebrospinal fluidleakage were not seen. Residual tumor regrowth was notobserved for fifty months after surgery (Figure 2).

Case 2. A thirty-six-year-old male presented right motorweakness, diplopia, and facial sensory disturbance becauseof its oculomotor, trochlear, and trigeminal nerve impair-ment. CT and MRI showed enhanced mass lesion withcalcification located clivus, and it extended toward brain-stem and its lateral side. The first surgery was performedby craniotomy, and tumor was resected partially. Thirtymonths after surgery, residual tumor regrowth was detected.And a second transcranial surgery was performed. After

the second surgery, cranial nerve palsy gradually improved.But fifty months after the second surgery, the tumor wasgrowing toward brainstem and sphenoid sinus. The thirdsurgery was performed by endoscopic endonasal transsphe-noidal approach, because main part of tumor extensionwas midline. We performed the surgery with the samestrategy as Case 1, that was using instruments that donot disturb the operative field without the interference bytransphenoidal approach. We made suprasellar part of tumorremain intentionally, because suprasellar part was very hard,and we considered the possibility of the involvement ofperforators and injury of hypothalamus. We could performprecise manipulation without intraoperative complicationssuch as carotid artery and basilar artery, cranial nerves, andbrainstem injuries. This case was also seen dural defect, andit was repaired by patched with muscle fascia at defect andfilling sphenoid sinus by autologous fat and fibrin glue. Aftersurgery, there were no new deficit and cerebrospinal fluidleakage. Immunohistochemically, epithelial membrane anti-gen (EMA) and cytokeratin (CAM5.2) were negative, and thepathological diagnosis was chondrosarcoma. Residual tumorregrowth was not observed for thirty months after surgery(Figures 3, 4 and 5).

4 Sarcoma

Figure 4: Case 2. Intraoperative endoscopic view. Upper left: tumor capsule extending sphenoid sinus (asterisk), right: bilateral carotidprominence and expose carotid artery (triangle), lower left: lateral part of the lesion (star). Endoscope with long axis gave us clear operativeview but difficult to reach surgical instruments and intratumoral manipulation. Lower right: We could observe the surface of brainstem andbasilar artery through dural defect. basilar artery (square), brainstem (circle).

3. Discussion

Although there are various operative approaches for clivaltumors [1–5], a transsphenoidal approach [6, 7] is one ofchoices when the main tumor extention is in an anterior-posterior direction with a slight lateral extension.

Bouche et al. [8] reported transsphenoidal approachfor clival tumor in 1966, but in those days, there was thelimitation of removal of tumor which extend to lateraland deep part, because it is difficult to obtain sufficientmicroscopic operative view. As a result, this approach did notwidely spread.

Recently, development of neuroendoscope gives usimprovements of visibility of operative field. And innova-tion of surgical instruments that specialized in endoscopictranssphenoidal approach and image navigation systemsgives us safe and precise surgical procedures [10].

However, this approach sometimes provides only narrowand deep operative field. To remove the tumor whichrequires us to progress more into deeper and lateral parts,we have to use longer endoscope to observe details andinstruments could be reached to those parts. Rhoton Jr.[11] described dissection instruments have shafts at least

120 mm in length when reach intrasuprasellar lesion fortransspnenoidal surgery. When we consider the interferenceof bimanual manipulation and the surgical instrumentsinside and outside of the nasal cavity, we think that morelonger endoscope and surgical instruments seems to beneeded for clival lesion especially extend superior-posteriorpart.

In our case, we used endoscope with long axis (length:180 mm) and angled tip (0, 30, and 70 degrees, HOPKINSII:KARL STORZ Co., Ltd.) that was able to observe deepoperative field, superior and lateral part of lesion. Andinstruments which have long and slim curettes, ultrasonicaspirator probe for abdominal surgery, and high-speed drillwhich have curved handle were not to disturb operative fieldwithout the interference of each tool. These instruments gaveus easy and precise manipulation of tumor removal. Whenwe used neuronavigation system, we made our own longprobe to navigate the deep part. And this information mightmake us avoid intraoperative complications of parasellarcritical structures.

Cerebrospinal fluid leakage is a major complication oftranssphenoidal surgery, and it is an important factor toget good results of this surgery. Frank et al. [12] reported

Sarcoma 5

Figure 5

that they experienced 18% CSF leakages after endoscopictranssphenoidal surgery. And CSF leakages were controlledby fat packing. To prevent this complication, we do re-construction by fat packing, dural plasty using muscularfascia, pedicled nasoseptal flap, and fibrin glue. And thesetechniques were applied to parasellar lesion such as clivalchondrosarcoma.

Almefty et al. [13] mentioned that chordoma andchondrosarcoma differed with regard to their origin andhistology and differed markedly with regard to outcome.That is, chordoma type (chordoma and chondroid chor-doma) demonstrated aggressive clinical course and pooroutcome. On the other hand, chondrosarcoma patients hada significantly better outcome compared with chordomapatients with regard to survival and recurrence-free survivalwithout high-dose radiotherapy. Almefty et al. mentionedthat radiotherapy may not be necessary in patients with low-grade chondrosarcoma. Our cases were diagnosed low-gradechondrosarcoma, and their features were consistent with theconcept of Almefty et al. group.

Recently, radiotherapy such as heavy ion irradiationdevelops as an additional treatment for clival tumors [9].While less invasive surgical approach would be recom-mended, it might be one of a choice as a strategy for clivaltumors, that is, a combination of subtotal resection andradiotherapy such as heavy ion irradiation for residual lesionto avoid complications of parasellar critical structures.

However, there are still insufficient measures for progres-sion into deep and lateral parts, even if using our surgicalinstruments as the above. Especially, we need developmentof surgical instrument to reach the lateral part of thelesion. Further development and improvement of specialinstruments for transsphenoidal approach will be needed forclival tumors such as chondrosarcoma.

References

[1] O. Al-Mefty and L. A. B. Borba, “Skull base chordomas: amanagement challenge,” Journal of Neurosurgery, vol. 86, no. 2,pp. 182–189, 1997.

[2] B. O. Colli and O. Al-Mefty, “Chordomas of the craniocervicaljunction: follow-up review and prognostic factors,” Journal ofNeurosurgery, vol. 95, no. 6, pp. 933–943, 2001.

[3] E. Gay, L. N. Sekhar, E. Rubinstein et al., “Chordomas andchondrosarcomas of the cranial base: results and follow-up of60 patients,” Neurosurgery, vol. 36, no. 5, pp. 887–897, 1995.

[4] M. N. Hadley, N. A. Martin, R. F. Spetzler, V. K. H. Sonntag,and P. C. Johnson, “Comparative transoral dural closuretechniques: a canine model,” Neurosurgery, vol. 22, no. 2,pp. 392–397, 1988.

[5] G. Lanzino, L. N. Sekhar, W. L. Hirsch, C. N. Sen, S. Pomonis,and C. H. Synderman, “Chondromas and chondrosarcomasinvolving the cavernous sinus: review of surgical treatmentand outcome in 31 patients,” Surgical Neurology, vol. 40,pp. 359–371, 1993.

6 Sarcoma

[6] E. De Divitiis, P. Cappabianca, L. M. Cavallo et al., “Endo-scopic transsphenoidal approach: adaptability of the proce-dure to different sellar lesions,” Neurosurgery, vol. 51, no. 3,pp. 699–707, 2002.

[7] G. Maira, R. Pallini, C. Anile et al., “Surgical treatment of clivalchordomas: the transsphenoidal approach revisited,” Journalof Neurosurgery, vol. 85, no. 5, pp. 784–792, 1996.

[8] J. Bouche, G. Guiot, J. Rougerie, and C. Freche, “The trans-sphenoidal route in the surgical approach to chordoma ofthe clivusLa voie trans-sphenoıdale dans l’abord chirurgicaldes chordomes du clivus,” Annales d’Oto-Laryngologie et deChirurgie Cervico-Faciale, vol. 83, no. 12, pp. 817–834, 1966.

[9] D. Schulz-Ertner, A. Nikoghosyan, C. Thilmann et al.,“Carbon ion radiotherapy for chordomas and low-gradechondrosarcomas of the skull base. Results in 67 patients,”Strahlentherapie und Onkologie, vol. 179, no. 9, pp. 598–605,2003.

[10] H. D. Jho, R. L. Carrau, M. R. McLaughlin, and S. C. Somaza,“Endoscopic transsphenoidal resection of a large chordomain the posterior fossa,” Acta Neurochirurgica, vol. 139, no. 4,pp. 343–348, 1997.

[11] A. L. Rhoton Jr., “The sellar region,” Neurosurgery, vol. 51,no. 4, pp. 335–374, 2002.

[12] G. Frank, V. Sciarretta, F. Calbucci, G. Farneti, D. Mazzatenta,and E. Pasquini, “The endoscopic transnasal transsphenoidalapproach for the treatment of cranial base chordomas andchondrosarcomas,” Neurosurgery, vol. 59, no. 1, pp. S50–S56,2006.

[13] K. Almefty, S. Pravdenkova, B. O. Colli, O. Al-Mefty, andM. Gokden, “Chordoma and chondrosarcoma: similar, butquite different, skull base tumors,” Cancer, vol. 110, no. 11,pp. 2457–2467, 2007.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 274281, 4 pagesdoi:10.1155/2011/274281

Review Article

Chondrosarcoma of the Mobile Spine and Sacrum

Ryan M. Stuckey and Rex A. W. Marco

Department of Orthopaedics, University of Texas Medical School at Houston, 6700 West Loop South, Suite 110,Bellaire, TX 77401, USA

Correspondence should be addressed to Rex A. W. Marco, rexmarco@gmail.com

Received 28 September 2010; Accepted 10 December 2010

Academic Editor: Charles Scoggins

Copyright © 2011 R. M. Stuckey and R. A. W. Marco. This is an open access article distributed under the Creative CommonsAttribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work isproperly cited.

Chondrosarcoma is a rare malignant tumor of bone. This family of tumors can be primary malignant tumors or a secondarymalignant transformation of an underlying benign cartilage tumor. Pain is often the initial presenting complaint whenchondrosarcoma involves the spine. In the mobile spine, chondrosarcoma commonly presents within the vertebral body and showsa predilection for the thoracic spine. Due to the resistance of chondrosarcoma to both radiation and chemotherapy, treatment isfocused on surgery. With en bloc excision of chondrosarcoma of the mobile spine and sacrum patients can have local recurrencerates as low as 20%.

1. Introduction

Chondrosarcoma is a rare malignant bone tumor thatproduces cartilage matrix. Chondrosarcoma can be subclas-sified as a primary malignant bone tumor or a secondarymalignant transformation of an underlying enchondroma orosteochondroma. The estimated annual incidence has beenreported as 1 in 200,000 [1]. The prevalence of chondrosar-coma in the mobile spine is reported between 6.5 and 10%,while approximately 5% are located within the sacrum [2–4].Chondrosarcoma typically presents in patients between the3rd and 7th decades.

2. Clinical Presentation

Patients with chondrosarcoma of the spine generally presentwith pain in the area of the lesion. The pain is often insidiousin nature and can be present for weeks to years. Boriani etal. noted a mass at presentation in 34% of patients in theirseries [2], and Shives et al. reported 40% of patients with amass at presentation [3]. The neurologic presentation can besomewhat mixed and can range from radicular pain to frankweakness. Radicular symptoms are seen in roughly 24% ofpatients [2, 5].

3. Location

Chondrosarcoma can occur within all regions of the spine.Roughly 20% of chondrosarcoma arises in the cervical spine,30% in the thoracic spine, 20% in the lumbar spine, and20% in the sacrum. Several series have shown a higherprevalence of chondrosarcoma in the thoracic spine. In theMD Anderson experience, 48% (10/21), 33% (7/21), and19% (4/21) were located in the thoracic, lumbar, and cervicalspines, respectively [5]. A similar breakdown was noted inthe series reported by Shives et al. [3]. Bergh et al. reportedon a series of 69 consecutive chondrosarcoma and found that16% (11/69) were located in the sacrum [6].

Within the vertebra, Chondrosarcoma can be isolatedto the body, the posterior elements, or a combination ofboth. Primary chondrosarcoma is predominately located inthe vertebral body, described as zones 10-3 according to theWBB staging system [2, 3]. In contrast, chondrosarcomathat arise from an underlying benign chondral lesion aretypically located in the posterior elements of the vertebra.Sacral chondrosarcoma are typically located eccentrically inthe upper portion of the sacrum (Figures 1(a), 1(b), and1(c)). This location can lead to local involvement of thesacroiliac joints [7].

2 Sarcoma

(a)

L

(b)

(c)

Figure 1: A 27-year-old female with a right sacroiliac secondary chondrosarcoma arising from a sacroiliac osteochondroma. The tumorinvolved half of the sacrum, and a sagittal resection through the midline of the sacrum was required to obtain wide margins. She developedscoliosis and sitting imbalance and underwent a subsequent posterior spinal fusion and instrumentation. She remains free of disease 48months after surgery.

4. Staging

The appropriate staging of chondrosarcoma of the spine isessential for prognosis and surgical treatment. Appropri-ate staging includes radiologic and histologic information.Imaging includes plain radiographs, CT scan, and MRI toevaluate the tumor. Chondrosarcoma presents radiographi-cally as a mixed lytic and blastic lesion. Low-grade lesionshave dense spicules of calcification and an eccentric, lobularappearance. High-grade lesions can have amorphous areasof calcifications and concentric growth of a soft tissuecomponent [8].

Additional imaging studies include a CT scan of thechest, abdomen, and pelvis, as well as an MRI of the spineto evaluate the rest of the spinal column. The appropriateimaging studies allow for a better understanding of the extentof the lesion as well as a differential diagnosis.

The histologic diagnosis is essential to the staging andultimate treatment of the spinal tumor. Prior excisional

biopsies can adversely affect the treatment and survival ofpatients with chondrosarcoma of the spine [9]. Thus a closed,image-guided core needle biopsy with a trocar providesa safe and effective method of obtaining a histologicaldiagnosis in patients with suspected chondrosarcoma of thespine. Histologic characteristics of chondrosarcoma includechondroid matrix, mitotic figures, hypercellularity, nuclearatypia, double-nucleated cells, and myxoid changes in thestroma [3, 4, 8, 10].

Two systems are commonly used to stage primary tumorsof the spine. Enneking developed a system based on thelocation of the tumor, the histologic grade of the lesion,and whether metastases are present or not [11]. Low-grade tumors are designated with the Roman numeral(I) and high-grade lesions (II); if metastases are present,the tumor is designated by the Roman numeral (III). Inaddition the tumor is described as intracompartmental (A)or extracompartmental (B). This classification system wasoriginally developed for tumors involving the appendicular

Sarcoma 3

skeleton and does have some limitations in the mobile spine.Weinstein (WBB) proposed a classification and surgicalstaging system developed specifically for spinal tumors [12].The WBB staging system divides the vertebra into zones inthe axial plane in a radiating pattern similar to a clock face.The numbers one through twelve designate the zones. Thesecond portion of this system classifies the tumor based onlayers designated by the letters A–E where (A) is prevertebralextraosseous, (B) intraosseous superficial, (C) intraosseousdeep, (D) extraosseous extradural, and (E) extraosseousintradural. The WBB staging system then allows for directionin the method of resection. If the tumor involves zones4–8 or 5–9, a vertebrectomy with a double approach isrecommended, if zones 2–5 or 7–11 are involved, then asagittal resection is recommended, and if the tumor involveszones 10-3, then a posterior arch resection is recommended.

5. Treatment

Chondrosarcoma are a family of slow-growing neoplasms. Ithas been well established in the literature that these tumorsare resistant to chemotherapy and radiation therapy and thusChondrosarcoma is a surgical disease [2–7, 10, 13]. En blocor radical resection of chondrosarcoma involving the longbones has led to 5-year survival rates as high as 54 to 78%[4, 6, 10].

Surgical treatment consisting of either curettage or enbloc resection has been described for both primary andrecurrent tumors [7, 13–16]. A description of the surgicalmargin is important for comparing results and the prognosisfollowing surgical treatment. Intralesional excision refers totumor cells present on the surface of the resected tissue.Marginal excision refers to a thin layer of tissue that isreactive, but without neoplastic tissue. Wide excision refersto a tumor-free cuff of normal tissue surrounding the lesion.

In the mobile spine and sacrum, the inherent intimaterelationship of the neurovascular structures and need forstructural stabilization can make en bloc resection difficult,and in some cases intralesional curettage is more appropriate.Boriani et al. proposed a set of criteria directing treatmenttoward curretage. These criteria include circumferentialspinal canal involvement, the need for spinal cord ligation tocomplete en bloc resection, and the potential for spinal cordischemia from ligation of the spinal segmental artery [2].

Unfortunately, results following curretage are poor. Intheir series of twenty-two patients with chondrosarcoma ofthe mobile spine, Boriani et al. reported at least one localrecurrence or progression of disease in 100% of patients(10/10) treated with curretage. 80% of these patients died ata mean of thirty-six months [2]. In contrast, 25% (3/12) ofpatients treated with en bloc marginal resection had a localrecurrence. The margins were classified as contaminatedor intralesional in two of the three patients with localrecurrence after en bloc excision. Shives et al. reportedon twenty patients with chondrosarcoma of the mobilespine. Similar to the previously cited study, 100% (11/11) ofpatients treated with intralesional excision had documenteddisease progression at a mean of 24.8 months. All of the

patients treated with an intralesional excision died [3].York et al. found a 69% recurrence rate in patients treatedintralesionally with a subtotal excision at a mean of 44.4months compared to a 20% recurrence rate in those treatedwith en bloc resection [5].

The role of en bloc resection in the treatment ofchondrosarcoma of the sacrum has also been well established[7, 13]. Fourney et al. reported on three patients withchondrosarcoma in their series on en bloc resection of sacraltumors. All three patients had local recurrence, but it isimportant to note that all three patients had previouslyundergone a subtotal excision prior to their planned en blocresection [7]. This further supports the important role ofnegative margins and en bloc resection in the successfultreatment of chondrosarcoma involving the spine.

Although en bloc resection can improve the clinicalresults of the surgical treatment of chondrosarcoma, theprocedures do not come without complications. In a reviewof 134 patients who underwent en bloc resection at a singleinstitution, complications were noted in 48 patients. A totalof 43 major complications were noted in 27 patients, and29 minor complications were noted in 28 patients. Majorcomplications included vascular injuries to the vena cava andaorta, myocardial infarction, pulmonary embolism, deepinfection requiring surgical debridement, transient renalfailure, ureteral injury, temporary paraplegia, and an exvacuo cerebral hematoma secondary to cerebrospinal fluidleak. Three patients died from complications related to theprocedure [17].

6. Conclusion

Chondrosarcoma is a malignancy of bone that can presentwithin the mobile spine and the sacrum. The effective stagingof chondrosarcoma plays an essential role in the diseasetreatment. This family of tumors is known to be resistant toboth chemotherapy and radiation therapy, thus treatment isdirected toward surgery. The prognosis of chondrosarcomainvolving the spine is related to the histological grade of thetumor, the age of the patient at the time of diagnosis, initialtreatment taking place in a primary tumor center, and mostimportantly the adequacy of the surgical margins. Patientswho undergo successful en bloc excision have recurrencerates as low as 20%.

References

[1] A. Y. Giuffrida, J. E. Burgueno, L. G. Koniaris, J. C. Gutierrez,R. Duncan, and S. P. Scully, “Chondrosarcoma in the UnitedStates (1973 to 2003): an analysis of 2890 cases from the SEERdatabase,” Journal of Bone and Joint Surgery. American, vol. 91,no. 5, pp. 1063–1072, 2009.

[2] S. Boriani, F. De Lure, S. Bandiera et al., “Chondrosarcoma ofthe mobile spine: report on 22 cases,” Spine, vol. 25, no. 7, pp.804–812, 2000.

[3] T. C. Shives, R. A. McLeod, K. K. Unni, and M. F. Schray,“Chondrosarcoma of the spine,” Journal of Bone and JointSurgery. American, vol. 71, no. 8, pp. 1158–1165, 1989.

4 Sarcoma

[4] S. Gitelis, F. Bertoni, P. Picci, and M. Campanacci, “Chon-drosarcoma of bone. The experience at the Istituto OrtopedicoRizzoli,” Journal of Bone and Joint Surgery. American, vol. 63,no. 8, pp. 1248–1257, 1981.

[5] J. E. York, R. H. Berk, G. N. Fuller et al., “Chondrosarcoma ofthe spine: 1954 to 1997,” Journal of Neurosurgery, vol. 90, no.1, pp. 73–78, 1999.

[6] P. Bergh, B. Gunterberg, J. M. Meis-Kindblom, and L. G.Kindblom, “Prognostic factors and outcome of pelvic, sacral,and spinal chondrosarcomas: a center-based study of 69 cases,”Cancer, vol. 91, no. 7, pp. 1201–1212, 2001.

[7] D. R. Fourney, L. D. Rhines, S. J. Hentschel et al., “En blocresection of primary sacral tumors: classification of surgicalapproaches and outcome,” Journal of Neurosurgery: Spine, vol.3, no. 2, pp. 111–122, 2005.

[8] R. A. Marco, S. Gitelis, G. T. Brebach, and J. H. Healey,“Cartilage tumors: evaluation and treatment,” The Journal ofthe American Academy of Orthopaedic Surgeons, vol. 8, no. 5,pp. 292–304, 2000.

[9] T. Yamazaki, G. S. McLoughlin, S. Patel, L. D. Rhines, andD. R. Fourney, “Feasibility and safety of en bloc resectionfor primary spine tumors: a systematic review by the SpineOncology Study Group,” Spine, vol. 34, no. 22, pp. S31–S38,2009.

[10] E. D. Henderson and D. C. Dahlin, “Chondrosarcoma ofbone—a study of two hundred and eighty-eight cases,” TheJournal of Bone and Joint Surgery. American, vol. 45, pp. 1450–1458, 1963.

[11] W. F. Enneking, “A system of staging musculoskeletal neo-plasms,” Clinical Orthopaedics and Related Research, vol. 204,pp. 9–24, 1986.

[12] S. Boriani, J. N. Weinstein, and R. Biagini, “Spine updateprimary bone tumors of the spine: terminology and surgicalstaging,” Spine, vol. 22, no. 9, pp. 1036–1044, 1997.

[13] P. C. Hsieh, R. Xu, D. M. Sciubba et al., “Long-term clinicaloutcomes following en bloc resections for sacral chordomasand chondrosarcomas: a series of twenty consecutive patients,”Spine, vol. 34, no. 20, pp. 2233–2239, 2009.

[14] B. Stener, “Total spondylectomy in chondrosarcoma arisingfrom the seventh thoracic vertebra,” Journal of Bone and JointSurgery. British, vol. 53, no. 2, pp. 288–295, 1971.

[15] K. Tomita, N. Kawahara, H. Baba, H. Tsuchiya, T. Fujita, andY. Toribatake, “Total en bloc spondylectomy: a new surgicaltechnique for primary malignant vertebral tumors,” Spine, vol.22, no. 3, pp. 324–333, 1997.

[16] N. Kawahara, K. Tomita, H. Murakami, S. Demura, K.Yoshioka, and T. Miyazaki, “Total excision of a recurrentchondrosarcoma of the thoracic spine: a case report of a seven-year-old boy with fifteen years follow-up,” Spine, vol. 35, no.11, pp. E481–E487, 2010.

[17] S. Bandiera, S. Boriani, R. Donthineni, L. Amendola, M.Cappuccio, and A. Gasbarrini, “Complications of en blocresections in the spine,” Orthopedic Clinics of North America,vol. 40, no. 1, pp. 125–131, 2009.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 342879, 7 pagesdoi:10.1155/2011/342879

Review Article

Chondrosarcoma of the Thorax

Philip A. Rascoe, Scott I. Reznik, and W. Roy Smythe

Scott & White Memorial Hospital and Clinic and Olin E. Teague Veterans’ Center,Texas A&M Health Science Center College of Medicine, Temple, TX 76508, USA

Correspondence should be addressed to Philip A. Rascoe, prascoe@swmail.sw.org

Received 9 September 2010; Revised 17 January 2011; Accepted 9 March 2011

Academic Editor: Beatrice Seddon

Copyright © 2011 Philip A. Rascoe et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Although a rare entity, chondrosarcoma is the most common malignant tumor of the chest wall. Most patients present with anenlarging, painful anterior chest wall mass arising from the costochondrosternal junction. CT scan with intravenous contrastis the gold standard radiographic study for diagnosis and operative planning. Contrary to previous dictum, resection may beperformed in an appropriate surgical candidate based on imaging characteristics or image-guided percutaneous biopsy results;incisional biopsy is rarely required. The keys to successful treatment are early recognition and radical excision with adequatemargins, as chondrosarcoma is relatively resistant to radiotherapy and conventional cytotoxic chemotherapy. Overall survivalis excellent in most surgical series from experienced centers. Complete excision with widely negative microscopic margins atthe initial operation is of the utmost importance, as local recurrence portends systemic metastasis and eventual tumor-relatedmortality. This paper summarizes data from relevant surgical series and thereupon draws conclusions regarding preoperative,intraoperative, and postoperative management of thoracic chondrosarcoma.

1. Introduction

The thoracic vertebrae, sternum, ribs, and costal cartilagesprovide the rigid structure of the thorax. The soft-tissueconstituents include skin, connective tissue, extrathoracicand intercostal musculature, and pleural mesothelium. Inaddition to providing protection for the underlying thoracicviscera, these structures function harmoniously to supportthe physiology of respiration.

Tumors of the chest wall encompass a wide varietyof benign and malignant conditions. The most commonentities are blood-borne rib metastases and direct chest wallinvasion from contiguous lung and breast carcinoma.

Primary chest wall tumors may arise from any of itssoft-tissue, bony, or cartilaginous constituents. These tumorsare quite rare, with approximately 500 new cases per yearin the United States. As such, most reports in the surgicalliterature consist of single-institutional studies with relativelyfew patients. Soft-tissue tumors account for roughly two-thirds of cases, while bony and cartilaginous masses account

for approximately one-third. In general, 50–80% of thesetumors are malignant, with increasing rates of malignancyfound as the proportion of soft-tissue tumors increaseswithin the series [1].

Chondrosarcoma is the most common malignant pri-mary tumor of both the bony thorax and, in fact, the entirechest wall [2]. It accounts for nearly one-third of all primarychest wall tumors. Nevertheless, it is exceedingly rare. Asonly 15% of chondrosarcomas arise from the chest wall, theannual predicted incidence of chest wall chondrosarcoma inthe United States is 60 cases [3].

2. Biology and Histology

Chondrosarcoma is a malignant cartilage-forming tumor ofbone. Interestingly, chondrosarcoma usually arises withinbones formed through ossification of a cartilaginous inter-mediate (endochondral ossification) and rarely arise frombones formed directly from fetal mesenchyme without a

2 Sarcoma

cartilaginous intermediate (membranous ossification). Assuch, chondrosarcomas develop most commonly in thoracic,pelvic, and appendicular long bones. Chondrosarcoma mostcommonly arises de novo within the medullary cavity ofbone (primary or central) but can result from malignanttransformation of the cartilage cap of a preexisting benigncartilaginous tumor such as enchondroma or osteochon-droma (secondary or peripheral) [4].

Histologically, chondrosarcomas are classified into 3grades, although grading is subject to interobserver vari-ability. Grade I tumors are composed predominantly ofextracellular hyaline matrix with chondrocytes displayingsmall, dense nuclei (Figure 1). Differentiation between gradeI chondrosarcoma and benign enchondroma can be difficult.Grade II tumors contain less chondroid matrix, are increas-ingly cellular, and may contain mitoses. Grade III tumorsare highly cellular, undifferentiated tumors demonstratingmarked pleomorphism, frequent mitoses, and a mucomyx-oid matrix [5]. Tumor grade is an important predictor oflocal recurrence, systemic metastasis, and survival [4, 5].

3. Clinical Presentation

Most malignant chest wall tumors present with symptoms.The vast majority of patients with thoracic chondrosarcomapresent with an enlarging, painful, anterior chest wall massarising from either the vicinity of the costochondral junctionor the sternum. Less commonly, posterior tumors present inthe paravertebral region, having arisen from the head of arib. Asymptomatic tumors detected incidentally by thoracicimaging are more likely to be benign. In fact, only 6.5% of96 patients in a retrospective series of patients seen at theMayo Clinic over 73 years were without physical findingsor symptoms [7]. There is a slight male predominance inmost series, with a median age near 50 years. Of 88 chon-drosarcomas treated at Memorial Sloan-Kettering CancerCenter, 43% arose in rib, 36% in scapula, 16% in sternum,and 5% in clavicle [3]. Roughly half of those originatingfrom rib present in the upper 4 ribs. Approximately 5–10%of patients present with synchronous systemic metastasis,most commonly to the lung, liver, and bone. As withextrathoracic bony and soft-tissue sarcoma, a history oftrauma in the location of the tumor is not uncommon.Although an association with prior trauma exists, causationhas proven difficult to establish. Similarly, chondrosarcomahas been reported to arise in previous radiation portals formetachronous malignancies such as lymphoma and breastcarcinoma, often many years after completion of therapy. Asstated, secondary chondrosarcoma can arise from malignantdegeneration of a benign chondroma or osteochondroma.

4. Diagnosis

Any patient with a suspected or confirmed chest wall massshould undergo an initial posterior-anterior (PA) and lateralchest radiograph. However, computed tomography (CT)scan of the lower neck, thorax, and upper abdomen withintravenous contrast is the gold standard radiographic study

Figure 1: Photomicrograph of grade I chondrosarcoma, demon-strating abundant extracellular hyaline matrix and scant cellularity(Courtesy of Robert S. Beissner MD, PhD, Department of Pathol-ogy, Scott & White Memorial Hospital and Clinic, Temple, TX,USA).

for both diagnosis and operative planning, as bone and soft-tissue windows are available to delineate tumor characteris-tics and extent of invasion, while lung windows evaluate forpulmonary metastasis. The characteristic CT appearance ofchondrosarcoma consists of a well-defined, lobulated soft-tissue mass with foci of chondroid matrix calcification [8]. Acharacteristic flocculent or “popcorn” pattern of calcificationhas been described for chondrosarcoma (Figure 2). Bonedestruction and invasion of overlying soft-tissue may alsoexist. Magnetic resonance imaging (MRI) is particularlyuseful for defining vascular or neural involvement andtherefore provides complementary information to the CTscan for tumors with mediastinal, paravertebral, or thoracicoutlet involvement. Otherwise, in contradistinction to chon-drosarcoma of the appendicular long bones, MRI offers noadvantage over CT and is not a necessary part of a routineradiologic evaluation. Positron emission tomography (PET)with fluorine-18 fluorodeoxyglucose (FDG) is performed torule out extrapulmonary metastases. PET has also demon-strated some utility in differentiating benign cartilaginoustumors from chondrosarcoma. Furthermore, preoperativePET imaging and comparison of standardized uptake value(SUV) with histologic grade may predict postsurgical patientoutcome. In a retrospective review of 31 patients withchondrosarcoma (25% with disease localized to the thorax),the combination of pretherapeutic SUV and histopathologictumor grade allowed identification of patients at high riskfor local relapse or metastatic disease [9]. Cross-sectionalbrain imaging is reserved for patients with recent onsetof neurologic symptoms as intracranial metastasis fromprimary chest wall neoplasms is rare.

Prior to entertaining major surgical resection, a completephysiologic evaluation is performed, including pulmonaryfunction tests (PFTs) with spirometry. This exam is partic-ularly important in patients in whom pulmonary resectionis anticipated, or in patients potentially requiring a large

Sarcoma 3

Figure 2: Typical CT appearance of an anterior chest wall chon-drosarcoma arising from the chondrosternal junction, demonstrat-ing prominent chondroid matrix mineralization resulting in acharacteristic flocculent or “popcorn” pattern of calcification [6].

anterior chest wall resection, as this tends to impede venti-latory mechanics postoperatively. Further interrogation byquantitative perfusion or exercise oximetry is required incases of questionable pulmonary reserve.

Previous teaching dictated that incisional biopsy beperformed routinely for all chest wall masses. However, this isno longer the recommendation. Small single rib lesions canbe removed as a therapeutic excisional biopsy. A lobulatedmass with flocculent calcification arising from the sternumor costochondral junction is a chondrosarcoma until provenotherwise and does not require preoperative diagnosis in asuitable operative candidate. Benign chondroma and low-grade chondrosarcoma are often histologically indistinguish-able by preoperative biopsy, and therefore both are treatedby wide excision [2]. If the diagnosis is in question, percu-taneous image-guided core needle biopsy and sophisticatedcytopathology techniques should yield a diagnosis in themajority of cases [10]. In rare cases in which percutaneouscore biopsy is unsuccessful, a carefully placed incisionalbiopsy is necessary. A small incision should be placed alongthe long axis of the tumor with careful consideration givenfor eventual removal of the area at definitive resection. Thewound should be closed in layers, including the capsule ofthe tumor, to prevent tumor spillage into the surroundingtissues.

5. Surgical Resection and Reconstruction

There is no indication for neoadjuvant radiation or chem-otherapy in a suitable surgical candidate with resectablethoracic chondrosarcoma. Wide resection of all thoracicdisease with appropriate margins, reconstruction of the bonychest wall to ensure preservation of respiratory mechanics,and soft-tissue coverage of reconstructive prostheses withhealthy, vascularized tissue is the treatment of choice. Pos-terior tumors in the paravertebral region may demonstrateinvolvement of vertebral bodies and/or neural foramina on

preoperative MRI. If so, preoperative neurosurgical consul-tation is obtained in anticipation of vertebrectomy. Similarly,if large volume or free soft-tissue transfer with microvascularanastomosis is anticipated, preoperative consultation with areconstructive surgeon may be of benefit.

An epidural catheter is inserted preoperatively unlessthere is concern for spinal involvement. Double-lumen oro-tracheal intubation is utilized in every patient to allowdeflation of the ipsilateral lung. Lung deflation facilitateswedge resection of adherent or involved lung and allowscareful palpation of uninvolved lung to identify unsuspectedmetastasis.

Patient positioning and incision placement are dictatedprimarily by the location of the lesion. Skin and soft-tissueinvolvement by local tumor extension, previous irradiation,or prior incisional biopsy must also be considered, as thesetissues should be resected en bloc and may indicate the needfor rotational or free tissue transfer. In general, patients withsternal tumors are placed supine, and patients with anteriortumors involving the costochondral junction are supine withthe ipsilateral thorax elevated 30–45◦. Patients with lateraland posterior tumors are placed in the lateral decubitusposition as for standard posterolateral thoracotomy.

If overlying skin and subcutaneous tissues are unin-volved, they are simply divided sharply to expose the under-lying chest wall musculature, which may be spared or dividedif uninvolved by tumor. If the deep muscular fascia isfelt to be adherent to the tumor, the overlying area ofmuscle should be resected en bloc with an appropriatemargin. If the extrathoracic musculature is to be spared forlater reconstruction, skin flaps may be raised to facilitatemobilization and retraction of the muscle for exposure. Theintercostal incision should be planned based on preoperativeimaging and palpation of the exterior surface of the mass,so that the pleural space is entered one uninvolved ribbelow the tumor. Upon entry into the pleural space, thesurgeon should palpate the tumor from within the chestto further plan the resection margin. The gross tumor-free margin should include at least a portion of a normalrib superiorly and inferiorly and at least 4 cm anteriorlyand posteriorly. 1 cm segments of each resected rib areobtained anteriorly and posteriorly, individually labeled,and submitted as final pathologic margins following decal-cification. Posteriorly located tumors frequently requiredisarticulation of the ribs from the transverse processes andvertebral bodies, with ligation of the neurovascular bundlesas they exit the neural foramina. En bloc vertebral resectionand reconstruction occasionally necessitates neurosurgicalassistance. The specimen is oriented for the pathologist. Anysuspicious soft-tissue margins can be submitted for frozensection analysis. The ipsilateral lung should be carefullypalpated for occult metastases, which should be resectedprior to commencement of reconstruction.

Chondrosarcoma of the sternum should also be resectedwith at least a 4 cm margin. This typically requires resectionof the entire body of the sternum. If the margin is adequate,part or all of the manubrium may be preserved and the stabil-ity of the thoracic inlet structures maintained, facilitating fullrange of motion and function at the shoulder. If resection of

4 Sarcoma

the manubrium is required to achieve an adequate margin,the clavicular heads are either disarticulated or resected enbloc along with the first costal cartilage. Sternal resection istypically performed through a vertical midline skin incision.The pectoralis major muscles are either reflected laterally orresected en bloc where adherent to tumor. A retrosternalplane is developed superiorly at the sternal notch (if totalsternectomy is anticipated) and inferiorly at the xiphoidprocess. Segments of costal cartilage are resected at eachlevel and submitted for pathologic analysis as describedabove. The intervening neurovascular bundles are ligatedand divided along with the intercostal muscles. The sternumis then elevated using a bone hook, and the resectionis completed by incision of the posterior perichondriumbilaterally.

Options for reconstruction of a bony thoracic defect arenumerous. Defects less than 5 cm in diameter which donot overly cardiac structures need not be reconstructed ifadequate soft-tissue coverage exists. In general, large anteriordefects negatively impact respiratory mechanics greaterthan posterolateral defects and therefore should be rigidlyreconstructed. Large defects at the posterior apex are coveredby the scapula and do not require reconstruction. However,if the inferior resection margin includes the fourth rib, itbecomes possible for the scapular tip to fall into the pleuralspace and catch under the superior edge of the fifth ribwhile in a normal postoperative anatomic position. Thiscreates a condition which is extremely painful and requiresreoperation to prevent further episodes. If this situationappears likely, prosthetic reconstruction of the defect shouldbe undertaken. Defects larger than 5 cm and those overlyingcardiac structures should be reconstructed. Both polypropy-lene (Prolene or Marlex) mesh and polytetrafluoroethy-lene (PTFE, Gore-Tex) have been successfully employedfor this purpose, and each material has its proponents.With the exception of chest wall resection accompanyingpneumonectomy (for which we employ PTFE to achievea watertight pleural seal), we prefer polypropylene mesh.The mesh provides a matrix for tissue ingrowth and has,on occasion, been retained even when infected. While itis less likely to engender an overlying seroma than PTFE,its interstices fill with fibrin and leakage of pleural fluidis rarely a problem. Large anterior defects require rigidreconstruction for preservation of respiratory mechanics andprotection of underlying cardiac structures. Rigidity, whenindicated, is provided by creating a methyl methacrylate“sandwich” consisting of a double sheet of polypropylenemesh with a thin layer of methyl methacrylate paste spreadbetween. Prior to setting, the prosthesis can be contouredappropriately by placing it over the patient’s thigh or flank.The polypropylene mesh is sewn to rib margins using heavypolypropylene suture. Placement of sutures and avoidanceof intercostal nerve entrapment is facilitated by creationof 1 mm holes in the ribs using a hand-held pneumaticor manual drill. The mesh should be pulled taught as thesutures are inserted and tied to ensure sturdy reconstruction.If overlying soft-tissue and musculature has been spared,meticulous layered closure over a Blake drain is all thatis required. If the overlying chest wall musculature was

resected en bloc, skin and adipose tissue alone should notbe closed over the prosthesis. Rather, the thoracic surgeon ora reconstructive surgical colleague should perform rotationof a pedicled muscle flap (latissimus, pectoralis, rectusabdominis, and serratus) for coverage of the prosthesis priorto skin closure. If full thickness chest wall, including skin,has been resected, rotation of a pedicled myocutaneous flapor free microvascular muscle transfer with skin graft isrequired. Figure 3 contains intraoperative photographs takenduring resection of an anterior thoracic chondrosarcoma andsubsequent reconstruction using polypropylene mesh andmethyl methacrylate.

Most patients can be extubated in the operating roomfollowing chest wall resection, and empirical prolongedintubation is not indicated. Properly reconstructed defectsshould not impair respiratory mechanics too severely. Ade-quate cough and avoidance of secretion retention are of theutmost importance to prevent pneumonia and respiratoryfailure. These are greatly facilitated by epidural analgesia.However, aggressive pulmonary physiotherapy and toiletbronchoscopy are sometimes required in the immediatepostoperative period.

Surgical Series. Due to their rarity, relatively few series ofprimary chest wall tumors have been reported. Chondrosar-coma is among the most frequently encountered tumortypes in most such series. A series of 110 patients treatedin Uniformed Services hospitals in Washington, DC wasreported in 1982. 46.4% had benign tumors, and 53.6%had malignant primary chest wall neoplasms, of whichchondrosarcoma was the second most common. Patientswith chondrosarcoma had the highest 5-year survival (89%)[11].

90 patients had chest wall resection for primary tumorsover a 20-year period at the Mayo Clinic. Approximately 2/3of tumors originated in soft-tissue, while 1/3 originated inbone or cartilage. Chondrosarcoma accounted for 24% of theentire series; moreover, it was the most common histologictype of primary chest wall bone tumors (58.6%). Greaterthan 50% of patients with malignant tumors developedrecurrence. 5-year recurrence-free survival was 56% inpatients with a 4 cm resection margin, as compared to only29% in patients with a 2 cm margin. 5-year survival forchondrosarcoma patients was 70%. Based on their findings,the authors recommended a resection margin of 4 cm ofnormal tissue for all primary malignant chest wall tumors[1].

In 1985, the Mayo group reported on their experiencewith 96 patients with primary chondrosarcoma of the chestwall. 81% arose in ribs and 19% in the sternum. 67%arose anteriorly near the costochondral or chondrosternaljunction, and 25% arose posteriorly from the head ofthe rib. Approximately 6% had metastatic disease at thetime of diagnosis. The 10-year disease-specific survival was60.4%. Retrospective review of operative reports and grossspecimens was used to separate patients into groups whounderwent wide resection (2–4 cm margins) and those whounderwent local excision only (minimal margin). 10-yearchondrosarcoma survival was 96.4% for those who had

Sarcoma 5

(a) (b)

(c)

Figure 3: Intraoperative photographs taken during resection of an anterior thoracic chondrosarcoma and subsequent reconstruction usingpolypropylene mesh and methyl methacrylate. The initial thoracotomy is created at least one rib below the level of palpable tumor (a).Surgical specimen following radical excision with adequate wide margins (b). Following reconstruction of the right anterior chest wall usingpolypropylene mesh and methyl methacrylate (c).

wide resection, compared with 65.4% for those who hadlocal excision only. Tumor diameter, grade, location, anddate of operation all significantly affected survival. Patientswith sternal tumors had improved 10-year survival (81.3%),compared with patients with rib tumors (55.6%). Recurrentchondrosarcoma developed in 52.1% of patients who under-went initial surgical resection at Mayo. 62% recurred locally,while 38% recurred both locally and systemically. Metastaseswere seen to develop up to 12 years following resection. Inter-estingly, no patient developed systemic metastasis withoutevidence of local recurrence. Indeed, local recurrence wasan ominous event, with 68% of patients recurring locallyultimately dying of chondrosarcoma [7].

In 1992, the group at Memorial Sloan-Kettering CancerCenter reported on their 40-year experience with pri-mary bony and cartilaginous chest wall malignancies. 10%of chondrosarcoma patients presented with synchronoussystemic metastases. Of 84 patients undergoing primaryresection for chondrosarcoma, 28% recurred locally as thefirst site of recurrence. Of 79 patient who presented withlocal disease only, 18% eventually developed lung metastasis.The overall 5-year survival for chondrosarcoma patients was64%. Local recurrence was seen to portend distant metastasis

in this series as well, with metastasis seen in only 4% ofpatients with no local recurrence compared with 29% whohad recurred locally. The authors concluded that the singlemost important factor predicting survival was completenessof first resection [3].

The group at M.D. Anderson Cancer Center presenteda 10-year review of 51 patients with primary chest wallsarcomas [10]. Of the 51 patients, 9 required incisionalbiopsy, with the remainder undergoing excisional biopsy,needle biopsy, or definitive resection based on radiologicappearance alone. 29% were chondrosarcomas. Histologywas the most significant predictor of survival. 5-year sur-vival for chondrosarcoma patients was 92.3%. Furthermore,patients with sternal tumors fared better than those withrib involvement, likely due to the higher proportion ofsternal chondrosarcomas. The authors stressed that theyavoid incisional and excisional biopsies, instead performingneedle biopsy if the diagnosis is in question or proceedingwith definitive resection if the radiographic diagnosis is clear.

Most recently, a report of 106 consecutive Swedishthoracic chondrosarcoma patients over a 22-year period waspublished by the Scandinavian Sarcoma Group. In this study,patients treated at orthopedic sarcoma centers were more

6 Sarcoma

likely to be resected with wide surgical margins than thosetreated at nonspecialty centers. Not surprisingly, this trans-lated to decreased local recurrence and increased 10-yearsurvival for patients treated at specialty centers [12]. Similarto the reported Mayo [7] and Memorial Sloan-Kettering[3] series, inadequate initial resection margins portendedlocal recurrence, which portended systemic metastasis andmortality.

6. Postoperative Surveillance

Due to the possibility of late local and systemic recurrence,resected patients should undergo routine lifelong surveil-lance. Surveillance consists of physical examination andthoracic imaging with either PA/lateral radiograph or CTscan every 3–6 months for the first 5 years and annuallythereafter for a minimum of 10 years.

7. Local Recurrence

Despite wide local excision, local recurrence may occur inup to 50% of patients with thoracic chondrosarcoma. Hightumor grade, inadequate margins at initial resection, andresection at a nonspecialty center have been reported topredict local recurrence [7, 12]. In operable patients inwhom reexcision with margins can be achieved, subsequentresections are reasonable. Usually, local recurrence portendssystemic metastasis and poor outcome. However, in onereport, patients resected for local recurrence fared noworse than patients presenting for initial resection [10].Inoperable patients with local recurrence should be referredfor consideration of radiotherapy.

8. Radiotherapy

While chondrosarcoma is relatively radioresistant, conven-tional radiation therapy has demonstrated efficacy andshould be administered at sites of positive pathologicmargins (if a wider surgical margin cannot be achieved) orin cases of unresectability [5]. Particle therapy with protonsand carbon ions has proven efficacy in incompletely resectedchondrogenic tumors of the skull base [13, 14]. However,there are no reports describing utilization of these modalitiesfor thoracic chondrosarcoma.

9. Systemic Therapy

Unfortunately, at present, there is no effective systemiccytotoxic chemotherapy for metastatic chondrosarcoma.This resistance to therapy is most probably related toseveral intrinsic biologic characteristics of these neoplasms,among them abundance of chondroid matrix, low cellulargrade, and relative hypovascularity. Although prospectivedata are lacking, variant subtypes such as dedifferentiatedand mesenchymal chondrosarcoma may be sensitive tocombination chemotherapy and may be considered foradjuvant or palliative chemotherapy in the context of aclinical trial [5].

10. Lung Metastasis

As stated, approximately 5–10% of patients with primarychondrosarcoma of the chest wall present with synchronouspulmonary metastasis. Metachronous metastatic chon-drosarcoma may present as a slowly enlarging pulmonarynodule or as a pulmonary infiltrate which may be initiallyconfused as pneumonia [15]. Although pulmonary metasta-sectomy for sarcomatous neoplasms has proven therapeuticbenefit [16], there is a paucity of data regarding pulmonaryresection for metastatic chondrosarcoma in adults. Gen-eral principles of pulmonary metastasectomy are similarlyapplied irrespective of histologic origin. The primary tumorshould be controlled, and there should be no evidenceof extrathoracic metastatic disease. Furthermore, completeresection of all intrathoracic disease should be technicallyand physiologically possible, as debulking of pulmonarymetastases is of no benefit. Unilateral lesions are generallyapproached via axillary or posterolateral muscle-sparing tho-racotomy, which, as opposed to thoracoscopy, allow carefulpalpation of the entire ispilateral lung to evaluate for furthermetastases. Bilateral lesions may be approached via stagedthoracotomy, although median sternotomy or transversethoracosternotomy (clamshell incision) can be utilized insome circumstances. Margin-negative wedge resection is thepreferred surgical option. Lobectomy or pneumonectomycan be considered for deep lesions not amenable to wedgeresection.

11. Summary

Although a rare entity, chondrosarcoma is the most commonmalignant tumor of the chest wall. Most patients presentwith an enlarging, painful anterior chest wall mass. CT scanis the gold standard radiographic study for diagnosis andoperative planning. Contrary to previous dictum, resectionmay be performed in an appropriate surgical candidate basedon imaging characteristics or image-guided percutaneousbiopsy results; incisional biopsy is rarely required. The keysto successful treatment are early recognition and radicalexcision with adequate margins. Overall survival is excellentin most surgical series from experienced centers. Completeexcision with widely negative microscopic margins at theinitial operation is of the utmost importance, as localrecurrence portends systemic metastasis and eventual tumor-related mortality.

Acknowledgment

W. R. Smythe holds the Glen and Rita K. Roney EndowedChair in Surgery.

References

[1] R. M. King, P. C. Pairolero, and V. F. Trastek, “Primary chestwall tumors: factors affecting survival,” Annals of ThoracicSurgery, vol. 41, no. 6, pp. 597–601, 1986.

[2] J. Somers and L. P. Faber, “Chondroma and chondrosarcoma,”Seminars in thoracic and cardiovascular surgery, vol. 11, no. 3,pp. 270–277, 1999.

Sarcoma 7

[3] M. Burt, M. Fulton, S. Wessner-Dunlap et al., “Primary bonyand cartilaginous sarcomas of chest wall: results of therapy,”Annals of Thoracic Surgery, vol. 54, no. 2, pp. 226–232, 1992.

[4] J. V. M. G. Bovee, A. M. Cleton-Jansen, A. H. M. Taminiau,and P. C. W. Hogendoorn, “Emerging pathways in the devel-opment of chondrosarcoma of bone and implications fortargeted treatment,” Lancet Oncology, vol. 6, no. 8, pp. 599–607, 2005.

[5] H. Gelderblom, P. C. W. Hogendoorn, S. D. Dijkstra et al.,“The clinical approach towards chondrosarcoma,” Oncologist,vol. 13, no. 3, pp. 320–329, 2008.

[6] M. D. Murphey and E. A. Walker, “From the archives ofthe AFIP: imaging of primary chondrosarcoma: radiologic-pathologic correlation,” Radiographics, vol. 23, pp. 1245–1278,2003.

[7] M. K. McAfee, P. C. Pairolero, and E. J. Bergstralh, “Chon-drosarcoma of the chest wall: factors affecting survival,” Annalsof Thoracic Surgery, vol. 40, no. 6, pp. 535–541, 1985.

[8] P. O’Sullivan, H. O’Dwyer, J. Flint, P. L. Munk, and N.L. Muller, “Malignant chest wall neoplasms of bone andcartilage: a pictorial review of CT and MR findings,” BritishJournal of Radiology, vol. 80, no. 956, pp. 678–684, 2007.

[9] W. Brenner, E. U. Conrad, and J. F. Eary, “FDG PET imagingfor grading and prediction of outcome in chondrosarcomapatients,” European Journal of Nuclear Medicine and MolecularImaging, vol. 31, no. 2, pp. 189–195, 2004.

[10] G. L. Walsh, B. M. Davis, S. G. Swisher et al., “A single-institutional, multidisciplinary approach to primary sarcomasinvolving the chest wall requiring full-thickness resections,”Journal of Thoracic and Cardiovascular Surgery, vol. 121, no.1, pp. 48–60, 2001.

[11] G. M. Graeber, R. J. Snyder, and A. W. Fleming, “Initial andlong-term results in the management of primary chest wallneoplasms,” Annals of Thoracic Surgery, vol. 34, no. 6, pp. 664–673, 1982.

[12] B. Widhe and H. C. F. Bauer, “Surgical treatment is decisive foroutcome in chondrosarcoma of the chest wall: a population-based Scandinavian Sarcoma Group study of 106 patients,”Journal of Thoracic and Cardiovascular Surgery, vol. 137, no.3, pp. 610–614, 2009.

[13] M. Amichetti, D. Amelio, M. Cianchetti, R. Maurizi Enrici,and G. Minniti, “A systematic review of proton therapy in thetreatment of chondrosarcoma of the skull base,” NeurosurgicalReview, vol. 33, no. 2, pp. 155–165, 2010.

[14] A. V. Nikoghosyan, G. Rauch, M. W. Munter et al., “Ran-domised trial of proton vs. carbon ion radiation therapy inpatients with low and intermediate grade chondrosarcoma ofthe skull base, clinical phase III study,” BMC Cancer, vol. 10,article 606, 2010.

[15] C. Sergi, F. Willig et al., “Bronchopneumonia disguising lungmetastases of a painless central chondrosarcoma of pubis,”Pathology & Oncology Research, vol. 3, pp. 211–214, 1997.

[16] U. Pastorino, M. Buyse, G. Friedel et al., “Long-term resultsof lung metastasectomy: prognostic analyses based on 5206cases,” Journal of Thoracic and Cardiovascular Surgery, vol. 113,no. 1, pp. 37–49, 1997.

Hindawi Publishing CorporationSarcomaVolume 2011, Article ID 932451, 8 pagesdoi:10.1155/2011/932451

Review Article

The Bone Niche of Chondrosarcoma: A Sanctuary forDrug Resistance, Tumour Growth and also a Source ofNew Therapeutic Targets

E. David,1, 2 F. Blanchard,1, 2 M. F. Heymann,1, 2, 3 G. De Pinieux,4, 5 F. Gouin,1, 2, 3

F. Redini,1, 2 and D. Heymann1, 2, 3

1 INSERM, UMR 957, Physiopathologie de la Resorption Osseuse et Therapie des Tumeurs Osseuses Primitives, Faculte de Medecine,1 rue Gaston Veil, 44035 Nantes Cedex 1, 44035 Nantes, France

2 Universite de Nantes, Nantes Atlantique Universites, Laboratoire de Physiopathologie de la Resorption Osseuse et Therapie desTumeurs Osseuses Primitives, 44035 Nantes, France

3 University Hospital, Hotel Dieu, CHU de Nantes, 44035 Nantes, France4 EA3855, University Hospital, 2 bd Tonnelle, 37044 Tours Cedex, France5 University Hospital, Hopital Trousseau, CHRU de Tours, 37042 Tours Cedex, France

Correspondence should be addressed to D. Heymann, dominique.heymann@univ-nantes.fr

Received 25 November 2010; Revised 28 January 2011; Accepted 10 February 2011

Academic Editor: Ole Nielsen

Copyright © 2011 E. David et al. This is an open access article distributed under the Creative Commons Attribution License, whichpermits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Chondrosarcomas are malignant cartilage-forming tumours representing around 20% of malignant primary tumours of bone andaffect mainly adults in the third to sixth decade of life. Unfortunately, the molecular pathways controlling the genesis and thegrowth of chondrosarcoma cells are still not fully defined. It is well admitted that the invasion of bone by tumour cells affectsthe balance between early bone resorption and formation and induces an “inflammatory-like” environment which establishes adialogue between tumour cells and their environment. The bone tumour microenvironment is then described as a sanctuary thatcontributes to the drug resistance patterns and may control at least in part the tumour growth. The concept of “niche” defined as aspecialized microenvironment that can promote the emergence of tumour stem cells and provide all the required factors for theirdevelopment recently emerges in the literature. The present paper aims to summarize the main evidence sustaining the existenceof a specific bone niche in the pathogenesis of chondrosarcomas.

1. Introduction

Most chondrosarcomas (90%) are conventional chondrosar-comas which occur in the medullar cavity or at the bone sur-face. The fact that cartilaginous tumours are mainly observedin bones formed from endochondral ossification strengthensthe relationship between the differentiation of normal chon-drocytes and these neoplastic cells. Chondrosarcoma cellsare cytologically and phenotypically related to the differentchondrocyte subtypes observed in the growth plate, and allcell shapes can be observed in the tumour mass [1–4]. Thus,these similarities are in favour of a mesenchymal stem-cellorigin for chondrosarcoma cells [1, 5]. The development ofcancer cells in bone site responds to several biological mech-anisms potentially applicable to numerous other entities. For

instance, invasion of bone by a primary or metastatic tumourcell affects the balance between early bone resorption andbone formation. This dysregulation of osteoblast-osteoclastcoupling induces the release of factors initially trapped in thebone matrix, which in turn promote tumour cell prolifera-tion [6]. Thus, the bone tumour microenvironment controlsthe tumour growth and is also described as a sanctuary thatcontributes to drug resistance patterns [7]. The specific anddifferent bone sites in which the various sarcomas are ableto grow reinforce the prominence of the tumour micro-environment. Chondrosarcomas are also characterized bytheir chemo- and radioresistance leading to a therapeuticsurgical approach which remains the only available treatmentwith a 10-year survival between 30% and 80% dependingon the grade [8, 9]. Currently, surgical excision is the main

2 Sarcoma

treatment for all chondrosarcoma subtypes [10], and non-surgical treatments of their microenvironment are underinvestigation. In this context, a better understanding of thebone niche which interacts with chondrosarcoma is one ofthe future therapeutic options. The present paper aims todescribe the bone niche of chondrosarcoma, its role intumour growth and drug resistance, and its clinical interest asa therapeutic target.

2. The Bone Niche Is Composed ofHeterogeneous Cell Types withCoupled Activities

In 2003, two research laboratories demonstrated that oste-oblasts formed an osteoblastic niche to sustain hemopoiesis[11, 12]. Osteoblasts establish an “epithelial-like” tissuewhich physically interacts with hemopoietic stem cells andcontributes to their maintenance in a quiescent stage throughthe interaction between Tie-2 and angiopoietin-1 [13]. Nils-son et al. showed that primitive hematopoietic cells residedclose to the bone surface [14]. From these observations, theconcept of bone niche has strongly evolved and has beenapplied to cancer stem cells [15]. Indeed, the “niche” isa functional microenvironment able to promote the emer-gence of cancer stem cells and to provide all factors requiredfor their development. Naturally, this concept is well rec-ognized in the context of hematologic malignancies such asmultiple myeloma [16] or leukemia [17], and these diseasesappear as a stem-cell disease with a hierarchy analogousto normal hematopoietic development. However, the boneniche is not limited to osteoblasts and during skeletal remod-elling, numerous cell types (preosteoclasts, preosteoblasts,endothelial cells, macrophages, etc.) are closely located inthe bone matrix and their functional coordination is a pre-requisite to maintain the bone and the bone niche microar-chitecture. Using three-dimensional visualizations, Andersenet al. clearly demonstrated the functional relevance of thesecellular interactions in the bone niche [18]. In physiologicalconditions in which bone resorption and bone formationare coupled, the bone surface is always covered by canopycomposed by flat cells expressing osteoblastic markers andassociated with sinusoidal vessels [18]. Disruption of thiscanopy results in the dysregulation of the coupled bone-formation bone-resorption process and leads to a bonedeficiency [18]. These very elegant observations revealedthat the bone niche is composed of multiple cell entities.Macrophages also contribute to the bone niche as shown byChang et al. [19]. Indeed, a discrete population of residentmacrophages has been identified between bone lining cellswithin endosteum and periosteum. These osteal tissuemacrophages are involved in bone dynamics by controllingosteoblast functions and, more specifically, are required forefficient osteoblast mineralization [19]. Into the bone niche,self-renewal and differentiation activity are clearly balancedas shown for hemopoietic stem cells [16, 17], and thisbalance is being controlled by the level of hypoxia, whichmodulates the interactions between tumour cells and thecomponents of bone niche. The proliferation stage of stem

cells is predominant with increased levels of oxygen andhypoxia resulting in opposite effects [20, 21].

The concept of bone niche is also currently discussed forsolid tumours and strengthens the very modern theory of“seed and soils” proposed by Paget in 1887 in which tumourcells (“seeds”) would colonize receptive foci (“soils”) [22].This data is supported by the fact that specific molecules(e.g., cadherin and osteopontin) contribute to the stabi-lization of cancer cells in bone niches mimicking the cellinteractions which take place during hemopoiesis [23, 24].Such interactions have been identified in the premetastaticniche of breast carcinoma, where carcinoma cells grow avidlyin bone which stores a variety of cytokines and growthfactors and thus provide an extremely fertile environmentfor growing cells [25, 26]. The seed and soil theory can bealso envisaged for the primary bone tumours. In a recentstudy, we reported an unexpected local osteosarcoma relapsewhich occurred at the exact site of autologous fat grafts ina patient who did not present any predictive factor of localrecurrence [27]. Moreover, we showed that tumour growthwas promoted by fat injection using a human osteosarcomamodel induced in athymic nude mice. We then demonstratedthat the mesenchymal stem cells isolated from adiposetissue induced exactly the same effect, probably reactivatingquiescent tumour cells locally deposited into the bone tissue[27]. A recent study reinforces this theory by presenting 8cases of osteosarcoma development several years after benignbone tumour treatment by curettage associated with bonegraft. To explain the development of “de novo” sarcomasin these patients, an attraction mechanism of mesenchymalstem cells by the scaffold has been hypothesized [28].Although mechanisms by which cancer stem cells could drivethe tumour growth are still unknown, modulation of themicroenvironment by mesenchymal stem cells may interferewith the biological behavior of this cell subpopulation.Similarly, inflammatory process associated with surgery maybe also responsible for the reactivation of dormant tumourcells [29, 30]. Thus, a disturbance of the microenvironmentand the bone niche modifies the proliferation/differentiationprogram of the tumour cells.

3. The Bone Niche of Chondrosarcoma

The key role of bone microenvironment in chondrosarcomadevelopment has been suspected many years ago. Indeed,a rat intraosseous model simulating the progression ofhuman chondrosarcoma has been set up to assess the inter-actions between bone environment and chondrosarcoma[31]. Transplantation of swarm rat chondrosarcoma withinbone marrow or in close contact to the bone with inducedperiosteal lesions led to extensive bone remodelling withtrabecular bone rarefaction and periosteal apposition asso-ciated with tumour growth. In contrast with these results,transplantation in close contact to the bone but withoutany periosteal lesion had no effect on bone, suggesting thatbone healing factors interact with tumour development. Thetumours which developed in intramedullary environmentpresented different foci with various gradings confirming

Sarcoma 3

that bone environment is an important factor in the patho-genesis of chondrosarcoma [31]. Histological examinationof conventional chondrosarcoma reveals the presence ofnumerous cells types in close contact to the cartilaginoustumour cells (Figures 1(a) and 2). The morphology of car-tilaginous tumour cells depends on the grading of thetumour and associated cartilage-like tissue composed bytumour chondrocytes with heterogeneous shapes (Figures1(b)–1(e)) and tumour cell types with mesenchymal aspect(Figure 1(e)). The tumour mass is characterized by lobularfoci separated by vascularized soft tissue, which establishesa continuum with bone marrow or with the surroundingtissues (Figure 1). When chondrosarcoma develops in themedullary space (central or primary chondrosarcoma), thetumour cells induce the dysregulation of the balance betweenosteoblasts and osteoclasts, degrading the trabecular bone,perturbing the bone marrow environment. When chon-drosarcoma develops from the bone surface (peripheral orsecondary chondrosarcoma), tumour mass exhibits a similarlobular morphology associated with a periosteal reaction[31]. These peripheral chondrosarcoma develop on preex-isting osteochondroma defined as the most common benignbone tumours and characterized by a cartilage-capped exo-phytic lesion that arises from the bone cortex. Nevertheless,the limit between osteochondroma and chondrosarcoma isstill unclear, especially with low-grade chondrosarcoma thatis closely related to osteochondroma. These tumours interactwith periosteum mimicking the “bone niche”. Periosteumis a continuous membrane intimately linked covering thebone, well vascularized and containing osteoprogenitor cellsincluding mesenchymal stem cells [32–34]. Thereby, periph-eral and central chondrosarcoma can interact with the samekind of bone microenvironment. The permeation of tumourcells into the bone tissue is associated with the activation ofbone resorption through the induction of osteoclast forma-tion (Figures 1(f) and 2). In fact, the bone niche of chon-drosarcoma includes all cell types described in the other neo-plastic bone diseases. The narrow relationship between chon-drosarcoma cells, soft tissue, vessels, and bone cells strength-ens the relevance of a specific bone niche able to sustain tum-our growth.

Can we suspect the existence of cancer stem cells in thisbone niche which could be at the origin of chondrosarcomaand become quiescent in specific circumstances? Expressionof SOX9 in human chondrosarcomas suggests that chon-drosarcomas originate from a multipotent stem cell commit-ted to differentiation along the chondrogenic pathway [35].Moreover, the results of the cDNA array analyses emphasizethe heterogeneous nature of chondrosarcoma. Using similarapproaches, Boeuf et al. [36] proposed a new classificationof chondrosarcoma in two clusters: a prechondrogenicphenotype with immature cells and a chondrogenic pheno-type composed of more mature cells. Primary conventionalcentral chondrosarcoma cells could be then grouped into twomain clusters with distinctive marker expression signatures:one group clustering together with mesenchymal stem cells(CD49b-high/CD10-low/CD221-high) and a second groupclustering close to fibroblasts (CD49b-low/CD10-high/CD221-low) [37]. These data strongly suggest the existence

of cancer stem cells possibly with mesenchymal stem cellsor fibroblast markers. Although most of the literature onchondrosarcoma has confirmed that adequate surgery is themainstay of treatment for local tumour control, which itselfconstitutes a risk factor for survival, an additional featureof chondrosarcoma is also the high level of local recidiveeven in case of adequate surgery [38–40]. This feature isalso in favour of the existence of cancer stem cells in thebone marrow which may remain dormant until some yetunknown signals promote their growth or/and metastasisformation in bone tissue.

Hypoxia is a signal resulting in a large number of adaptivechanges aimed at surviving in the hypoxic environment aswell as correcting the oxygen deficit. Hypoxia inducing factor(HIF)-1 is a dimeric transcription factor composed of HIF-1 alpha and beta subunits. HIF-1 protein levels increase as aresult of decreased degradation of the oxygen sensitive sub-unit HIF-1α. HIF-1 modulates changes in gene expressionduring hypoxia. Although the angiogenesis compound ofcartilage tumours is heterogenous [41], hypoxia modulatesthe proliferation of chondrosarcoma cells similarly to theother solid tumour types and hemopoietic neoplasia. Thus,there is a significant relationship between the expressionof HIF-1α, the microvessel density and the proliferatingcell nuclear antigen [42]. Several authors demonstrated thatmalignant chondrocytes increased HIF-1α expression in anoxygen concentration-dependent manner and increased V-EGF expression in response to hypoxia [43–46] which isclosely related to the potential malignancy of chondrosar-coma [47, 48]. Hypoxia is also known to increase chemokinereceptor expression such as CXCR4 in numerous cell types[49] and CXCR4/SDF1 also indirectly promotes the prolifer-ation and migration of tumour cells and enhances tumour-associated angiogenesis [50]. CXCR4 expressed by tumourcells contributes to their migration into the premetastaticniche [51]. Interestingly, chondrosarcoma cell invasion isincreased by hypoxia-induced expression of CXCR4 andMMP1, a process mediated by HIF1α and ERK [52], andCXCL12, also called SDF-1, increases the invasiveness ofchondrosarcoma cells [53]. Other chemokine/chemokinereceptors couples are also involved in chondrosarcomaprogression. Thus, the interaction of CCL5 (RANTES), aproduct of activated T cells present in bone environmentduring the tumour process with CCR5 expressed on the cellmembrane enhances the migration of chondrosarcoma cellsthrough the increase of MMP-3 production [54]. Overall,these data point out the similarities between the behaviourof chondrosarcoma cells and the invasion of leukaemia cellsin the bone niche [51]. Osteopontin is also a typical exampleof these similarities. Indeed, osteopontin could mediatethe anchoring of cancer cells in osteoblastic niches in amanner that mimics the mechanisms used by osteoblastto retain hematopoietic stem cells in these niches and tonegatively regulate stem-cell pool size [55]. Osteopontin alsoinfluence the behaviour of carcinoma cells (proliferation,invasiveness, etc.) [56]. Similarly, osteopontin located in thebone matrix increases the migration and MMP expressionin human chondrosarcoma and contributes to the patho-genesis of chondrosarcoma in its bone niche [57]. More

4 Sarcoma

(a)

(b)

(c) (d)

(e) (f)

Bone tissue

∗

∗

∗

∗

▲

▲▲

▲

▲

▲

Figure 1: The bone niche of chondrosarcoma is composed by various cellular entities. Chondrosarcoma tissue shows heterogeneous cellmorphology (a–d) with chondrocyte-like (b–d) and mesenchymal features (f). Chondrosarcoma bone niche is associated with several celltypes including osteoclasts (e), endothelial cells vascularized soft tissue (f). HES staining, original magnification (×20, a and b; ×40: c–e).Tumour cells: arrow head, asterix: blood vessels, and arrow: osteoclast.

recently, Vincourt et al. [58] demonstrated not only that therespective levels of C-propeptides of procollagens I and IIin chondrogenic tumours but also that the interactions ofchondrosarcoma cells with the surrounding extracellularmatrix may modulate tumour progression, angiogenesis, andmetastasis. C-propeptides of procollagen I favor angiogenesisand tumour progression, whereas C-propeptides of pro-collagen II exert antitumour and antangiogenic propertiesthrough apoptosis induction when they are immobilized,and progression and metastasis when they are soluble [58].Endostatin derived from collagen XVIII, a potent endoge-nous antiangiogenic factor that induces regression of varioustumours of epithelial origin, prevents the chondrosarcomagrowth via its potential activity on endothelial cells [59].These results demonstrate that bone microenvironment andextracellular matrix establish a very complex bone nicheadapted to the tumour progression.

The interactions between the extracellular matrix of boneniche and chondrosarcoma cells are tightly controlled bycytokines and growth factors produced by the environmentalcells (osteoblasts, endothelial cells, macrophages, lympho-cytes, etc.) and also by tumour cells themselves [60]. Proin-flammatory cytokines are particularly associated with thepathogenesis of chondrosarcoma. Interleukin (IL)-1 regu-lates the expression of a disintegrin and metalloproteinasewith thrombospondin motifs 1 (ADAMTS1) and VEGF bychondrosarcoma cells, then contributing to a strong positiveimpact of IL-1 on vascularization and tumour progression[61]. TNF-α, another proinflammatory cytokine, inducedMMP-12 expression in chondrosarcoma cells when chondro-cytes undergo malignant transformation [62] and increasedalso MMP-13 [63]. Members of TGF-β superfamily playalso a crucial role in migration and metastasis of humanchondrosarcoma. For instance, TGF-β1 and BMP-2 increase

Sarcoma 5

(a) (b)

(c)

Bone tissueBone tissue

Bone tissue

∗∗

Figure 2: Chondrosarcoma growth is strongly linked to the bone tissue. Relationship between bone tissue and chondrosarcoma cells (a–c). Infiltration of chondrosarcoma cells into the bone tissue (permeation) (a–c). Chondrosarcoma development is associated with boneresorption foci (b). HES staining, original magnification (×20, a and b; ×40: c). Arrow: bone resorption area, arrow head: necrosis ofchondrosarcoma tissue, and ∗: viable tumour component.

motility of human chondrosarcoma via the PI3K/Akt path-way [64, 65]. Oncostatin (OSM), a member of IL-6 cytokinefamily, induces a hypertrophic differentiation, with reducedSOX9 and induced Cbfa1, Coll10, MMP13, VEGF, andRANKL expression in chondrosarcoma cells. RANKL being apro-osteoclastogenesis factor and then a proresorptive factor,OSM enhances osteoclast formation at the tumour/boneinterface and reduces the ectopic bone neoformation [66].

4. The Bone Niche: A Sanctuary forthe Drug Resistance and a Source ofNew Therapeutic Targets

Although bone niche represents an adequate microenviron-ment for the survival/proliferation of cancer stem cells andhas been identified as a major parameter regulating themetastatic process [67], recent studies also described thetumour microenvironment as a sanctuary contributing tothe phenomenon of drug resistance [68]. The process ofdrug resistance has been shown to be mediated through(i) soluble factors such as cytokines or adhesion moleculesconstituting de novo drug resistances or (ii) acquired drugresistance linked to resistance mechanisms caused by selec-tive pressure of chemotherapy or other therapeutic drugs[68]. Chondrosarcomas are poorly vascularized in correla-tion with resistance to systemic chemotherapy and exhibitpoor metastatic potential. However, although this poorvascularization represents a first explanation for the drug

resistance, the bone niche also contributes to this resistanceas observed for other tumour entities. In this context, abetter definition of bone niche leads to the identificationof relevant drug targets to improve the efficiency of thecurrent treatment. This concept has been already validatedin leukemia [69]. In sarcomas, similar approaches havebeen also envisaged [70]. Targeting of angiogenesis hasbeen assessed in combination of chemotherapy and inducedtumour necrosis [71]. Cyclooxygnease-2 (COX-2), a media-tor of angiogenesis, is expressed in malignant cartilaginoustumours [72]. In chondrosarcoma, the use of celecoxib,a COX-2 inhibitor, first results in a decrease in tumourvolume followed unfortunately by a relapsed tumour growthafter 6 weeks [73]. Higher doses of COX-2 may be used,or a combinatory therapy based on this concept may bedesigned. HDAC4 represses VEGF expression and associatedangiogenesis in chondrosarcoma [74]. Similarly, a therapeu-tic approach of chondrosarcoma based on HDAC inhibitoradministration may be interesting [75, 76]. Bisphosphonatesand rapamycin and its derivatives have been originallydeveloped, respectively, as antiresorptive and antifungalagents [77, 78]. However, in vitro and in vivo experimentsdemonstrated that these compounds are multifunctionalmolecules exerting their effects not only on bone remodellingbut also on tumour cell growth. mTOR targeting hasbeen envisaged for numerous cancer types including malig-nant primary bone tumours [78–80], and a very impres-sive response of myxoid chondrosarcoma has been obtained

6 Sarcoma

in combination with cyclophosphamide [81]. The maintargets of bisphosphonates are bone-resorbing osteoclasts[82] which contribute to the hemopoietic and tumour boneniche [82]. Bisphosphonates also reduce the proliferationand invasion of chondrosarcoma [83, 84]. In preclinicalmodel of chondrosarcoma, zoledronic acid slows down ratprimary development and recurrent tumour progressionafter intralesional curettage and increases overall survival[85]. Thus, osteoclasts targeting may be used in prevention ofchondrosarcoma recurrence. Cytokinic treatment representsanother relevant therapeutic approach of chondrosaroma[2]. Oncostatin M, a member of the IL-6 cytokine familymainly produced by macrophages, neutrophils, and T lym-phocytes, is a cytostatic factor for chondrosarcomas invitro and in vivo [66]. This growth inhibitory effect isalso observed with two other cytokines of the same familyable to reduce chondrosarcoma expansion but with a lowerefficiency: IL-6 in association with its soluble receptor andIL-27 [66]. This list is not exhaustive but gives some evidenceof the interest to target or to modulate the bone niche com-ponents to improve chondrosarcoma treatment.

5. Conclusion

The treatment of chondrosarcoma is currently based onsurgery, radiotherapy, and chemotherapy being occasionallyused for metastatic tumours. However, a recent concept hasemerged based on the key role played by the tumour micro-environment in the tumour invasiveness and in the drug-resistance phenomenon. This bone niche allows to identifynew therapeutic targets for chondrosarcoma, and it appearsclearly that a better understanding of the chondrosarcomabone niche will open nonsurgical therapeutic options forchondrosarcoma which could also be combined with surgery.

Acknowledgments

This work was supported by La Ligue Contre Le Cancer.E. David received a fellowship from the Ministere de laRecherche.

References

[1] S. Boeuf, P. Kunz, T. Hennig et al., “A chondrogenic geneexpression signature in mesenchymal stem cells is a classifier ofconventional central chondrosarcoma,” Journal of Pathology,vol. 216, no. 2, pp. 158–166, 2008.

[2] J. V. M. G. Bovee, A. M. Cleton-Jansen, A. H. M. Taminiau,and P. C. W. Hogendoorn, “Emerging pathways in thedevelopment of chondrosarcoma of bone and implications fortargeted treatment,” Lancet Oncology, vol. 6, no. 8, pp. 599–607, 2005.

[3] T. Aigner, “Towards a new understanding and classification ofchondrogenic neoplasias of the skeleton—biochemistry andcell biology of chondrosarcoma and its variants,” VirchowsArchiv, vol. 441, no. 3, pp. 219–230, 2002.

[4] H. Gelderblom, P. C. W. Hogendoorn, S. D. Dijkstra et al.,“The clinical approach towards chondrosarcoma,” Oncologist,vol. 13, no. 3, pp. 320–329, 2008.

[5] G. De Pinieux and C. Bouvier, “Recent advances in the biologyof bone tumors and new diagnostic tools,” in Bone Cancer, D.Heymann, Ed., chapter 19, pp. 225–234, Academic Press, NewYork, NY, USA, 2010.

[6] Y. Wittrant, S. Theoleyre, C. Chipoy et al., “RANKL/RANK/OPG: new therapeutic targets in bone tumours and associatedosteolysis,” Biochimica et Biophysica Acta, vol. 1704, no. 2, pp.49–57, 2004.

[7] M. B. Meads, L. A. Hazlehurst, and W. S. Dalton, “The bonemarrow microenvironment as a tumor sanctuary and contrib-utor to drug resistance,” Clinical Cancer Research, vol. 14, no.9, pp. 2519–2526, 2008.

[8] M. Campanacci, Bone and Soft Tissue Tumors, Springer, NewYork, NY, USA, 2nd edition, 1999.

[9] F. Y. Lee, H. J. Mankin, G. Fondren et al., “Chondrosarcomaof bone: an assessment of outcome,” Journal of Bone and JointSurgery. Series A, vol. 81, no. 3, pp. 326–338, 1999.

[10] E.L. Staals, E. Palmerini, S. Ferrari, and M. Mercuri, “Non-surgical treatment of chondrosarcoma: current concepts andfuture perspectives,” in Bone Cancer, D. Heymann, Ed., chap-ter 31, pp. 375–383, Academic Press, New York, NY, USA,2010.

[11] J. Zhang, C. Niu, L. Ye et al., “Identification of the haematopoi-etic stem cell niche and control of the niche size,” Nature, vol.425, no. 6960, pp. 836–841, 2003.

[12] L. M. Calvi, G. B. Adams, K. W. Weibrecht et al., “Osteoblasticcells regulate the haematopoietic stem cell niche,” Nature, vol.425, no. 6960, pp. 841–846, 2003.

[13] F. Arai, A. Hirao, M. Ohmura et al., “Tie2/angiopoietin-1signaling regulates hematopoietic stem cell quiescence in thebone marrow niche,” Cell, vol. 118, no. 2, pp. 149–161, 2004.

[14] S. K. Nilsson, H. M. Johnston, and J. A. Coverdale, “Spatiallocalization of transplanted hemopoietic stem cells: inferencesfor the localization of stem cell niches,” Blood, vol. 97, no. 8,pp. 2293–2299, 2001.

[15] H. Iwasaki and T. Suda, “Cancer stem cells and their niche,”Cancer Science, vol. 100, no. 7, pp. 1166–1172, 2009.

[16] G. W. Basak, A. S. Srivastava, R. Malhotra, and E. Carrier,“Multiple myeloma bone marrow niche,” Current Pharmaceu-tical Biotechnology, vol. 10, no. 3, pp. 335–346, 2009.

[17] S. W. Lane, D. T. Scadden, and D. G. Gilliland, “Theleukemic stem cell niche: current concepts and therapeuticopportunities,” Blood, vol. 114, no. 6, pp. 1150–1157, 2009.

[18] T. L. Andersen, T. E. Sondergaard, K. E. Skorzynska etal., “A physical mechanism for coupling bone resorptionand formation in adult human bone,” American Journal ofPathology, vol. 174, no. 1, pp. 239–247, 2009.

[19] M. K. Chang, L. J. Raggatt, K. A. Alexander et al., “Osteal tissuemacrophages are intercalated throughout human and mousebone lining tissues and regulate osteoblast function in vitroand in vivo,” Journal of Immunology, vol. 181, no. 2, pp. 1232–1244, 2008.

[20] K. Ito, A. Hirao, F. Arai et al., “Reactive oxygen species actthrough p38 MAPK to limit the lifespan of hematopoietic stemcells,” Nature Medicine, vol. 12, no. 4, pp. 446–451, 2006.

[21] K. Hosokawa, F. Arai, H. Yoshihara et al., “Function of oxida-tive stress in the regulation of hematopoietic stem cell-nicheinteraction,” Biochemical and Biophysical Research Communi-cations, vol. 363, no. 3, pp. 578–583, 2007.

[22] S. Paget, “The distribution of secondary growths in cancer ofthe breast,” The Lancet, vol. 133, no. 3421, pp. 571–573, 1889.

[23] T. Yin and L. Li, “The stem cell niches in bone,” Journal ofClinical Investigation, vol. 116, no. 5, pp. 1195–1201, 2006.

Sarcoma 7

[24] K. Hosokawa, F. Arai, H. Yoshihara et al., “Knockdown of N-cadherin suppresses the long-term engraftment of hematopoi-etic stem cells,” Blood, vol. 116, no. 4, pp. 554–563, 2010.

[25] P. Clezardin, “Early bone metastasis-associated molecular andcellular events,” in Bone Cancer, D. Heymann, Ed., chapter 3,pp. 41–45, Academic Press, New York, NY, USA, 2010.

[26] C. Coghlin and G. I. Murray, “Current and emerging conceptsin tumour metastasis,” Journal of Pathology, vol. 222, no. 1, pp.1–15, 2010.

[27] P. Perrot, J. Rousseau, A. -L. Bouffaut et al., “Safety concernbetween autologous fat graft, mesenchymal stem cell andosteosarcoma recurrence,” PLoS One, vol. 5, no. 6, Article IDe10999, 2010.

[28] P. Picci, G. Sieberova, M. Alberghini et al., “Late sarcomadevelopment after curettage and bone grafting of benign bonetumors,” European Journal of Radiology, vol. 77, no. 1, pp. 19–25, 2011.

[29] R. Demicheli, M. W. Retsky, W. J. M. Hrushesky, and M.Baum, “Tumor dormancy and surgery-driven interruption ofdormancy in breast cancer: learning from failures,” NatureClinical Practice Oncology, vol. 4, no. 12, pp. 699–710, 2007.

[30] R. Demicheli, M. W. Retsky, W. J. M. Hrushesky, M. Baum,and I. D. Gukas, “The effects of surgery on tumor growth: acentury of investigations,” Annals of Oncology, vol. 19, no. 11,pp. 1821–1828, 2008.

[31] E. Grimaud, C. Damiens, A. V. Rousselle, N. Passuti, D.Heymann, and F. Gouin, “Bone remodelling and tumourgrade modifications induced by interactions between boneand Swarm rat chondrosarcoma,” Histology and Histopathol-ogy, vol. 17, no. 4, pp. 1103–1111, 2002.

[32] M. Chanavaz, “The periosteum: the “umbilical cord” ofbone. Quantification of the blood supply of cortical boneof periosteal origin,” Revue de Stomatologie et de ChirurgieMaxillo-Faciale, vol. 96, no. 4, pp. 262–267, 1995.

[33] C. A. Squier, S. Ghoneim, and C. R. Kremenak, “Ultrastruc-ture of the periosteum from membrane bone,” Journal ofAnatomy, vol. 171, pp. 233–239, 1990.

[34] X. Zhang, H. A. Awad, R. J. O’Keefe, R. E. Guldberg,and E. M. Schwarz, “A perspective: engineering periosteumfor structural bone graft healing,” Clinical Orthopaedics andRelated Research, vol. 466, no. 8, pp. 1777–1787, 2008.

[35] M. Soderstrom, T. Bohling, T. Ekfors, L. Nelimarkka, H. T.Aro, and E. Vuorio, “Molecular profiling of human chon-drosarcomas for matrix production and cancer markers,”International Journal of Cancer, vol. 100, no. 2, pp. 144–151,2002.

[36] S. Boeuf, P. Kunz, T. Hennig et al., “A chondrogenic geneexpression signature in mesenchymal stem cells is a classifier ofconventional central chondrosarcoma,” Journal of Pathology,vol. 216, no. 2, pp. 158–166, 2008.

[37] J. Diaz-Romero, S. Romeo, J. V. M. G. Bovee, P. C. W.Hogendoorn, P. F. Heini, and P. Mainil-Varlet, “Hierarchicalclustering of flow cytometry data for the study of conventionalcentral chondrosarcoma,” Journal of Cellular Physiology, vol.225, no. 2, pp. 601–611, 2010.

[38] F. Fiorenza, A. Abudu, R. J. Grimer et al., “Risk factors forsurvival and local control in chondrosarcoma of bone,” Journalof Bone and Joint Surgery. Series B, vol. 84, no. 1, pp. 93–99,2002.

[39] F. Y. Lee, H. J. Mankin, G. Fondren et al., “Chondrosarcomaof bone: An assessment of outcome,” Journal of Bone and JointSurgery. Series A, vol. 81, no. 3, pp. 326–338, 1999.

[40] S. Gitelis, F. Bertoni, P. Picci, and M. Campanacci, “Chon-drosarcoma of bone. The experience at the Istituto Ortopedico

Rizzoli,” Journal of Bone and Joint Surgery. Series A, vol. 63, no.8, pp. 1248–1257, 1981.

[41] T. Kalinski, S. Sel, I. Kouznetsova, M. Ropke, and A. Roessner,“Heterogeneity of angiogenesis and blood vessel maturationin cartilage tumors,” Pathology Research and Practice, vol. 205,no. 5, pp. 339–345, 2009.

[42] T. Kubo, T. Sugita, S. Shimose, T. Matsuo, K. Arihiro, andM. Ochi, “Expression of hypoxia-inducible factor-1α and itsrelationship to tumour angiogenesis and cell proliferation incartilage tumours,” Journal of Bone and Joint Surgery. Series B,vol. 90, no. 3, pp. 364–370, 2008.

[43] C. Lin, R. McGough, B. Aswad, J. A. Block, and R. Terek,“Hypoxia induces HIF-1α and VEGF expression in chon-drosarcoma cells and chondrocytes,” Journal of OrthopaedicResearch, vol. 22, no. 6, pp. 1175–1181, 2004.

[44] G. Ayala, C. Liu, R. Nicosia, S. Horowitz, and R. Lackman,“Microvasculature and VEGF expression in cartilaginoustumors,” Human Pathology, vol. 31, no. 3, pp. 341–346, 2000.

[45] R. L. McGough, C. Lin, P. Meitner, B. I. Aswad, and R. M.Terek, “Angiogenic cytokines in cartilage tumors,” ClinicalOrthopaedics and Related Research, no. 397, pp. 62–69, 2002.

[46] T. Furumatsu, K. Nishida, A. Kawai, M. Namba, H. Inoue,and Y. Ninomiya, “Human chondrosarcoma secretes vascularendothelial growth factor to induce tumor angiogenesis andstores basic fibroblast growth factor for regulation of its owngrowth,” International Journal of Cancer, vol. 97, no. 3, pp.313–322, 2002.

[47] S. A. Nakagawa, A. Lopes, A. L. De Carvalho et al., “Nitricoxide synthases, cyclooxygenase-2, nitrotyrosine, and angio-genesis in chondrosarcoma and their relation to prognosis,”Journal of Bone and Joint Surgery. Series A, vol. 92, no. 8, pp.1738–1746, 2010.

[48] S. Boeuf, J. V. M. G. Bovee, B. Lehner, P. C. W. Hogendoorn,and W. Richter, “Correlation of hypoxic signalling to histolog-ical grade and outcome in cartilage tumours,” Histopathology,vol. 56, no. 5, pp. 641–651, 2010.

[49] T. Schioppa, B. Uranchimeg, A. Saccani et al., “Regulationof the chemokine receptor CXCR4 by hypoxia,” Journal ofExperimental Medicine, vol. 198, no. 9, pp. 1391–1402, 2003.

[50] J. Wang, J. Wang, Y. Sun et al., “Diverse signaling pathwaysthrough the SDF-1/CXCR4 chemokine axis in prostate cancercell lines leads to altered patterns of cytokine secretion andangiogenesis,” Cellular Signalling, vol. 17, no. 12, pp. 1578–1592, 2005.

[51] R. N. Kaplan, R. D. Riba, S. Zacharoulis et al., “VEGFR1-positive haematopoietic bone marrow progenitors initiate thepre-metastatic niche,” Nature, vol. 438, no. 7069, pp. 820–827,2005.

[52] X. Sun, L. Wei, Q. Chen, and R. M. Terek, “CXCR4/SDF1mediate hypoxia induced chondrosarcoma cell invasionthrough ERK signaling and increased MMP1 expression,”Molecular Cancer, vol. 9, article no. 17, 2010.

[53] T. H. Lai, Y. C. Fong, W. M. Fu, R. S. Yang, and C. H.Tang, “Stromal cell-derived factor-1 increase αvβ3 integrinexpression and invasion in human chondrosarcoma cells,”Journal of Cellular Physiology, vol. 218, no. 2, pp. 334–342,2009.

[54] C. H. Tang, A. Yamamoto, Y. T. Lin, YI. C. Fong, and T. W. Tan,“Involvement of matrix metalloproteinase-3 in CCL5/CCR5pathway of chondrosarcomas metastasis,” Biochemical Phar-macology, vol. 79, no. 2, pp. 209–217, 2010.

[55] S. Stier, Y. Ko, R. Forkert et al., “Osteopontin is a hematopoi-etic stem cell niche component that negatively regulates stem

8 Sarcoma

cell pool size,” Journal of Experimental Medicine, vol. 201, no.11, pp. 1781–1791, 2005.

[56] J. C. Reichert, V. M. C. Quent, L. J. Burke, S. H. Stansfield,J. A. Clements, and D. W. Hutmacher, “Mineralized humanprimary osteoblast matrices as a model system to analyseinteractions of prostate cancer cells with the bone microenvi-ronment,” Biomaterials, vol. 31, no. 31, pp. 7928–7936, 2010.

[57] Y. J. Chen, Y. Y. Wei, H. T. Chen et al., “Osteopontinincreases migration and MMP-9 up-regulation via αvβ3integrin, FAK, ERK, and NF-κB-dependent pathway in humanchondrosarcoma cells,” Journal of Cellular Physiology, vol. 221,no. 1, pp. 98–108, 2009.

[58] J. B. Vincourt, S. Etienne, J. Cottet et al., “C-propeptidesof procollagens Iα1 and II that differentially accumulate inenchondromas versus chondrosarcomas regulate tumor cellsurvival and migration,” Cancer Research, vol. 70, no. 11, pp.4739–4748, 2010.

[59] T. Furumatsu, N. Yamaguchi, K. Nishida et al., “Endostatininhibits adhesion of endothelial cells to collagen I via αβintegrin, a possible cause of prevention of chondrosarcomagrowth,” Journal of Biochemistry, vol. 131, no. 4, pp. 619–626,2002.

[60] P. Rutkowski, J. Kaminska, M. Kowalska, W. Ruka, and J.Steffen, “Cytokine and cytokine receptor serum levels in adultbone sarcoma patients: correlations with local tumor extentand prognosis,” Journal of Surgical Oncology, vol. 84, no. 3, pp.151–159, 2003.

[61] T. Kalinski, S. Krueger, S. Sel, K. Werner, M. Ropke, and A.Roessner, “ADAMTS1 is regulated by interleukin-1β, not byhypoxia, in chondrosarcoma,” Human Pathology, vol. 38, no.1, pp. 86–94, 2007.

[62] E. KerkelA, T. Bohling, R. Herva, J. A. Uria, and U. Saarialho-Kere, “Human macrophage metalloelastase (MMP-12) expres-sion is induced in chondrocytes during fetal development andmalignant transformation,” Bone, vol. 29, no. 5, pp. 487–493,2001.

[63] S. W. Yoon, J. S. Chun, M. H. Sung, J. Y. Kim, and H. Poo,“α-MSH inhibits TNF-α-induced matrix metalloproteinase-13 expression by modulating p38 kinase and nuclear fac-tor κB signaling in human chondrosarcoma HTB-94 cells,”Osteoarthritis and Cartilage, vol. 16, no. 1, pp. 115–124, 2008.

[64] Y. Y. Yeh, C. C. Chiao, W. Y. Kuo et al., “TGF-β1 increasesmotility and αvβ3 integrin up-regulation via PI3K, Akt andNF-κB-dependent pathway in human chondrosarcoma cells,”Biochemical Pharmacology, vol. 75, no. 6, pp. 1292–1301, 2008.

[65] Y. C. Fong, T. M. Li, C. M. Wu et al., “BMP-2 increasesmigration of human chondrosarcoma cells via PI3K/Aktpathway,” Journal of Cellular Physiology, vol. 217, no. 3, pp.846–855, 2008.

[66] E. David, P. Guihard, B. Brounais et al., “Direct anti-cancereffect of oncostatin M on chondrosarcoma,” InternationalJournal of Cancer, vol. 128, no. 8, pp. 1822–1835, 2011.

[67] J. A. Joyce and J. W. Pollard, “Microenvironmental regulationof metastasis,” Nature Reviews Cancer, vol. 9, no. 4, pp. 239–252, 2009.

[68] M. B. Meads, L. A. Hazlehurst, and W. S. Dalton, “Thebone marrow microenvironment as a tumor sanctuary andcontributor to drug resistance,” Clinical Cancer Research, vol.14, no. 9, pp. 2519–2526, 2008.

[69] M. Konopleva, Y. Tabe, Z. Zeng, and M. Andreeff, “Therapeu-tic targeting of microenvironmental interactions in leukemia:mechanisms and approaches,” Drug Resistance Updates, vol.12, no. 4-5, pp. 103–113, 2009.

[70] K. Scotlandi, P. Picci, and H. Kovar, “Targeted therapies inbone sarcomas,” Current Cancer Drug Targets, vol. 9, no. 7, pp.843–853, 2009.

[71] H. Morioka, L. Weissbach, T. Vogel et al., “Antiangiogenesistreatment combined with chemotherapy produces chon-drosarcoma necrosis,” Clinical Cancer Research, vol. 9, no. 3,pp. 1211–1217, 2003.

[72] K. M. Sutton, M. Wright, G. Fondren, C. A. Towle, and H. J.Mankin, “Cyclooxygenase-2 expression in chondrosarcoma,”Oncology, vol. 66, no. 4, pp. 275–280, 2004.

[73] Y. M. Schrage, I. Machado, D. Meijer et al., “COX-2 expressionin chondrosarcoma: a role for celecoxib treatment?” EuropeanJournal of Cancer, vol. 46, no. 3, pp. 616–624, 2010.

[74] X. Sun, L. Wei, Q. Chen, and R. M. Terek, “HDAC4represses vascular endothelial growth factor expression inchondrosarcoma by modulating RUNX2 activity,” Journal ofBiological Chemistry, vol. 284, no. 33, pp. 21881–21890, 2009.

[75] F. Blanchard and C. Chipoy, “Histone deacetylase inhibitors:new drugs for the treatment of inflammatory diseases?” DrugDiscovery Today, vol. 10, no. 3, pp. 197–204, 2005.

[76] R. Sakimura, K. Tanaka, S. Yamamoto et al., “The effects ofhistone deacetylase inhibitors on the induction of differenti-ation in chondrosarcoma cells,” Clinical Cancer Research, vol.13, no. 1, pp. 275–282, 2007.

[77] B. Ory, G. Moriceau, F. Redini, and D. Heymann, “mTORinhibitors (Rapamycin and its derivatives) and nitrogencontaining bisphosphonates: bi-functional compounds for thetreatment of bone tumours,” Current Medicinal Chemistry,vol. 14, no. 13, pp. 1381–1387, 2007.

[78] D. Heymann, B. Ory, F. Gouin, J. R. Green, and F. Redini,“Bisphosphonates: new therapeutic agents for the treatmentof bone tumors,” Trends in Molecular Medicine, vol. 10, no. 7,pp. 337–343, 2004.

[79] R. E. Brown, “Brief communication: morphoproteomic por-trait of the mTOR pathway in mesenchymal chondrosarcoma,”Annals of Clinical and Laboratory Science, vol. 34, no. 4, pp.397–399, 2004.

[80] G. Moriceau, B. Ory, L. Mitrofan et al., “Zoledronic acidpotentiates mTOR inhibition and abolishes the resistance ofosteosarcoma cells to RAD001 (everolimus): pivotal role ofthe prenylation process,” Cancer Research, vol. 70, no. 24, pp.10329–10339, 2010.

[81] O. Merimsky, R. Bernstein-Molho, and R. Sagi-Eisenberg,“Targeting the mammalian target of rapamycin in myxoidchondrosarcoma,” Anti-Cancer Drugs, vol. 19, no. 10, pp.1019–1021, 2008.

[82] O. Kollet, A. Dar, S. Shivtiel et al., “Osteoclasts degrade endo-steal components and promote mobilization of hematopoieticprogenitor cells,” Nature Medicine, vol. 12, no. 6, pp. 657–664,2006.

[83] T. J. Lai, S. F. Hsu, T. M. Li et al., “Alendronate inhibits cellinvasion and MMP-2 secretion in human chondrosarcoma cellline,” Acta Pharmacologica Sinica, vol. 28, no. 8, pp. 1231–1235, 2007.

[84] T. Kubo, S. Shimose, T. Matsuo et al., “Inhibitory effects ofa new bisphosphonate, minodronate, on proliferation andinvasion of a variety of malignant bone tumor cells,” Journalof Orthopaedic Research, vol. 24, no. 6, pp. 1138–1144, 2006.

[85] F. Gouin, B. Ory, F. Redini, and D. Heymann, “Zoledronicacid slows down rat primary chondrosarcoma development,recurrent tumor progression after intralesional curretage andincreases overall survival,” International Journal of Cancer, vol.119, no. 5, pp. 980–984, 2006.