ass1 as a novel tumor suppressor gene in myxoﬁbrosarcomas: … · human cancer biology ass1 as a...

Human Cancer Biology

ASS1 as a Novel Tumor Suppressor Gene inMyxofibrosarcomas: Aberrant Loss via Epigenetic DNAMethylation Confers Aggressive Phenotypes, NegativePrognostic Impact, and Therapeutic Relevance

Hsuan-Ying Huang1, Wen-Ren Wu5, Yu-Hui Wang10, Jun-Wen Wang2, Fu-Min Fang3, Jen-Wei Tsai6,Shau-Hsuan Li4, Hsiao-Chin Hung1, Shih-Chen Yu1, Jui Lan1, Yow-Ling Shiue5, Chung-His Hsing12,Li-Tzong Chen7,11,14, and Chien-Feng Li5,8,9,13,14,15

AbstractPurpose: The principal goals were to identify and validate targetable metabolic drivers relevant to

myxofibrosarcoma pathogenesis using a published transcriptome.

Experimental Design: As the most significantly downregulated gene regulating amino acid meta-

bolism, argininosuccinate synthetase (ASS1) was selected for further analysis by methylation-specific

PCR, pyrosequencing, and immunohistochemistry of myxofibrosarcoma samples. The roles of ASS1 in

tumorigenesis and the therapeutic relevance of the arginine-depriving agent pegylated arginine deiminase

(ADI-PEG20) were elucidated in ASS1-deficient myxofibrosarcoma cell lines and xenografts with and

without stable ASS1 reexpression.

Results:ASS1 promoter hypermethylationwas detected inmyxofibrosarcoma samples and cell lines and

was strongly linked to ASS1 protein deficiency. The latter correlated with increased tumor grade and stage

and independently predicted a worse survival. ASS1-deficient cell lines were auxotrophic for arginine and

susceptible to ADI-PEG20 treatment, with dose-dependent reductions in cell viability and tumor growth

attributable to cell-cycle arrest in the S-phase. ASS1 expression was restored in 2 of 3 ASS1-deficient

myxofibrosarcoma cell lines by 5-aza-20-deoxycytidine, abrogating the inhibitory effect of ADI-PEG20.

Conditioned media following ASS1 reexpression attenuated HUVEC tube-forming capability, which was

associated with suppression of MMP-9 and an antiangiogenic effect in corresponding myxofibrosarcoma

xenografts. In addition to delayed wound closure and fewer invading cells in a Matrigel assay, ASS1

reexpression reduced tumor cell proliferation, induced G1-phase arrest, and downregulated cyclin E with

corresponding growth inhibition in soft agar and xenograft assays.

Conclusions:Our findings highlight ASS1 as a novel tumor suppressor inmyxofibrosarcomas, with loss

of expression linked to promoter methylation, clinical aggressiveness, and sensitivity to ADI-PEG20. Clin

Cancer Res; 19(11); 2861–72. �2013 AACR.

IntroductionMyxofibrosarcomas are characterized by increasedmetas-

tasis after relentless local recurrence (1–3). Such a recur-rence has been shown to parallel the cytogenetic complexityand eventually lead tometastasis (4, 5). The overall survival

rate in patients with a myxofibrosarcoma is approximately75% at 5 years, and several series have provided conflictingdata for clinicopathologic prognosticators (1, 3, 4, 6, 7).Weand others have addressed the importance of clear margins,which predicted better local control and translated into

Authors' Affiliations: Departments of 1Pathology, 2Orthopedic Sur-gery, 3Radiation Oncology, and 4Division of Oncology, Department ofInternal Medicine, Kaohsiung Chang Gung Memorial Hospital andChang Gung University College of Medicine; 5Institute of BiomedicalScience, National Sun Yat-Sen University; 6Department of Pathology,E-Da Hospital; Departments of 7Internal Medicine and Cancer Centerand 8Pathology, Kaohsiung Medical University Hospital, and 9Instituteof Clinical Medicine, Kaohsiung Medical University, Kaohsiung; Insti-tutes of 10Biosignal Transduction and 11Molecular Medicine, NationalCheng Kung University; Departments of 12Anesthesiology and13Pathology, Chi-Mei Medical Center; 14National Institute of CancerResearch, National Health Research Institutes; and Department of15Biotechnology, Southern Taiwan University of Science and Technol-ogy, Tainan, Taiwan

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

L.-T. Chen, and C.-F. Li contributed equally to this work.

Corresponding Authors: Li-Tzong Chen, National Institute of CancerResearch, National Health Research Institutes, Tainan, Taiwan. Phone:886-6-2083422, ext. 65110; Fax: 886-6-2083427; E-mail:[email protected] or Chien-Feng Li, Department of Pathology, Chi-MeiMedical Center, Tainan, Taiwan. Phone: 886-6-2812811, ext.53680; Fax:886-6-2511235; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-12-2641

�2013 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 2861

on January 12, 2021. © 2013 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst April 2, 2013; DOI: 10.1158/1078-0432.CCR-12-2641

http://clincancerres.aacrjournals.org/

survival benefits (1, 7). However, the molecular determi-nants of aggressiveness and therapeutic targets in myxofi-brosarcomas remain elusive.

Various genetic, epigenetic, and posttranslational aberra-tions ofmetabolism-associated enzymesmay alter signalingpathways in human cancers (8–12). Converging to adaptto core cell metabolism, these aberrations are essential formalignant transformation and sustained cell growth (8–12), making metabolic deregulation a cancer hallmark ofrenewed interest (8). Genome-wide approaches withderived targeted therapies have prompted efforts to char-acterize coordinately regulated gene expressions of specificpathways (13, 14). Because amino acids are building blocksin the anabolic process of cell proliferation, the develop-ment of compounds targeting the amino acidmetabolism isa promising cancer therapy (15). A successful example isasparaginase, which depletes asparagine in the plasma ofpatients with acute lymphoblastic leukemia devoid ofasparagine synthetase (16). To explore genes associatedwith tumorigenesis, we analyzed the published transcrip-tome of myxofibrosarcomas and identified argininosucci-nate synthetase (ASS1) as a prominently downregulatedgene regulating amino acid metabolism. ASS1 catalyzes thepenultimate step of arginine biosynthesis (17). Tumorsdevoid of ASS1, such as hepatocellular carcinomas andmelanomas, become auxotrophic for arginine (18, 19).This feature has been exploited as a biomarker to depriveplasma arginine using pegylated arginine deiminase (ADI-PEG20; refs. 20, 21).

Neither the functions of ASS1 in regulating tumorigenesisnor the clinical relevance and regulation of ASS1 expressionwere previously characterized in soft tissue sarcomas. Wehave shown that ASS1 is frequently methylated in myxofi-brosarcoma clinical samples and cell lines, resulting in

aberrant loss of ASS1 expression. TheASS1deficiency causesin vitro and in vivo susceptibility to ADI-PEG20 and con-tributes to the novel tumor suppressor role of ASS1, withvalidation in tumor samples, cell lines, and xenograftmodels.

Materials and MethodsAnalysis of published genomic and transcriptomicprofiling datasets

The complete datasets of our published genomic pro-filing study for myxofibrosarcoma tissues and cell lines(4) were available in Gene Expression Omnibus (GEO,GSE35483). To identify genes critical in myxofibrosar-coma pathogenesis, we reappraised expression profilingdatasets of myxofibrosarcomas versus nonneoplastic softtissues from GSE21122 deposited in GEO. The raw CELfiles of Affymetrix U133A microarray platform wereimported into Nexus Expression 3 software (BioDiscov-ery) to analyze all probe sets without preselection orfiltering. Supervised comparative analysis and functionalprofiling were conducted to identify significant differen-tially expressed genes, with special attention to aminoacid biosynthesis in Gene Ontology (GO: 0008652).Those with P less than 0.01 and log2-transformed expres-sion fold change more than 2 were chosen for furtheranalysis.

Cell culture, transfection, and stable clonesThederivation andmaintenance of theOH931,NMFH-1,

and NMFH-2 myxofibrosarcoma cell lines and CCD966SKdermal fibroblasts were previously reported (22–25). Cul-ture methods of human umbilical venous endothelial cells(HUVEC) and the ASS1-expressing LPS510 cell line weredescribed in Supplementary Method S1, along with theauthentication of all cell lines by short-tandem repeatprofiling.

Full-length pASS1-DDK-Myc expression plasmid andempty pCMV6 vector (Origene) were validated by sequenc-ing. In a 24-well plate, 105 cells each of 3myxofibrosarcomacell lines (OH931, NMFH1, and NMFH2) were seeded andincubated with Lipofectamine 2000 (Invitrogene) for 4hours at 37�C to transfect various plasmids (1.5 mg). Tostandardize transfection efficiency, pEGFP (Promega) weretransfected in parallel. Afterwards, cells were cultured forfurther 24 hours at 37�C and lysed with Passive Lysis Buffer(Promega). Finally, myxofibrosarcoma cells stably expres-sing ASS1 or DDK-Myc tags alone were selected with neo-mycin. The transfected cells were analyzed for exogenousASS1 expression by Western blotting. To transfect SFGplasmid carrying the GFP-FFLuc gene (Clontech), 105

OH931 cells were incubatedwith lipofectamine 2000 (Invi-trogen) for 4 hours at 37�C. GFPþ OH931 cells were sortedby flow cytometry to isolate successfully transfected cells.The methods to establish stable EZH2-knockdown NMFH-1 cells and ASS1-knockdown LPS510 cells were essentiallyas previously reported (4), with modifications detailed inSupplementary Method S2.

Translational RelevanceIn the prognostication and treatment of myxofibro-

sarcomas, the molecular markers and therapeuticoptions are inadequate. Through data mining of themyxofibrosarcoma transcriptome, we identified ASS1gene as the most prominently downregulated candidateamong those regulating amino acid metabolism. AnASS1 protein deficiency was validated in myxofibrosar-coma samples and was associated with gene methyla-tion, adverse prognosticators, and worse outcomes. Invitro and in vivo, ASS1-deficientmyxofibrosarcomas weresusceptible to arginine-depriving ADI-PEG20 at thera-peutic doses, compared with ASS1-reexpressing counter-parts and fibroblasts. An ASS1 deficiency contributed toaggressive phenotypes, including tumor growth, migra-tion/invasion, and angiogenesis, which were attenuatedby forced ASS1 reexpression. We substantiate the clini-cal, biologic, and pharmacologic relevance of ASS1,highlighting its novel tumor suppressive role and ther-apeutic potential in myxofibrosarcomas.

Huang et al.

Clin Cancer Res; 19(11) June 1, 2013 Clinical Cancer Research2862




Quantification of transcripts of ASS1, EZH2, andangiogenesis-associated genesUsing the ABI StepOnePlus System, a real-time reverse-

transcription PCR (RT-PCR) was carried out on cell samplestomeasureASS1 and EZH2mRNA abundances and exploremediators of ASS1 in inhibiting angiogenesis with an RT-PCR expression array (PAHS-024, SABioscience) essentiallyas previously reported (4) and detailed in SupplementaryMethod S3.

Tumor characteristicsTissue procurement ofmyxofibrosarcomas was approved

by the Institutional review board (100-0895B). The criteriafor the diagnosis and parameter assessment were previouslyelaborated (1, 4, 26). Ninety formalin-fixed, primary myx-ofibrosarcoma samples (Table 1) receiving curative resec-tion without neoadjuvant radiation or chemotherapy wereanalyzed for an ASS1 methylation-specific PCR and ASS1immunostaining.

Evaluation of ASS1 gene methylationMethylation-specific PCR was carried out as described

(27) with pyrosequencing confirmation of all cell lines

and selected tissue samples detailed in SupplementaryMethod S4.

ImmunohistochemistryTissue microarray (TMA) sections were prepared for anti-

gen retrieval as described (4, 26, 28), encompassing all casessubjected tomethylation-specific PCR. The incubation withthe antibody against ASS1 (1:1,000, DesigneRx) and thecut-off value of less than 5% cytoplasmic reactivity to defineaberrant ASS1 loss were as previously reported (18). Wholesections from formalin-fixed xenografted specimens werestained with anti-CD31 (1:50, BD Pharmingen) and anti-Ki67 (1:200, Abcam).

Western blotting and enzyme-linked immunosorbentassays (ELISA) were as previously reported (4), withminor modifications detailed in Supplementary MethodsS5 and S6.

Pharmacologic agents and in vitro ADI-PEG20treatment

ADI-PEG20 and 3-Dznep (targeting EZH2-mediatedH3K27 histone methyltransferation) were provided byDesigneRx and Dr. Marquez (National Cancer Institute;Frederick, MD), respectively. UNC0638 (targeting G9a-mediated histone methyltransferation of H3K9), trichosta-tin (a histone deacetylase inhibitor), and 5-aza-20-deoxycy-tidine were obtained from Sigma. Pharmacokinetic charac-terization of the arginine-depriving efficiency of ADI-PEG20 was described in Supplementary Method S7, apartfrom ADI-PEG20 treatment in various cell lines.

Functional assaysTo elucidate functional alterations associated with ASS1

reexpression, ASS1 knockdown, and ADI-PEG20 treatment,various cancer phenotypes were evaluated using cell viabil-ity, bromodeoxyuridine (BrdUrd), electric cell-substrateimpedance sensing (ECIS), cell-cycle kinetic, soft-agar,HUVEC tube formation, and wound healing and invasionassays, as previously reported (4) with modificationsdetailed in Supplementary Methods-S8-S14.

Animal xenograftsThe protocol was approved (98121505) by the animal

use committee. OH931myxofibrosarcoma cells transfectedwith plasmids expressing GFP-FFLuc, ASS1, or the emptycontrol were inoculated into severe combined immunode-ficient (SCID) mice to analyze the in vivo effects of ADI-PEG20 treatment and ASS1 reexpression as described inSupplementary Method S15.

In vivo imaging system for monitoring ADI-PEG20treatment of xenografts

The successfully transfected, GFP-FFLuc-bearingOH931 cells were validated using the IVIS camera system(Xenogen) to ascertain the presence of bioluminesc-ence in complete media supplemented with 150 mg/mLD-luciferin. Real-time tumor growth and the therapeuticeffects of ADI-PEG20 were monitored in such OH931

Table 1. Associations of ASS1 expression withclinicopathologic parameters

ASS1 expressionP value

Deficient Intact

Gender 0.672Male 20 23Female 19 21

Age 0.666<60 years 18 15�60 years 28 29

Location 0.805Extremity 36 33Axial 10 11

Tumor depth 0.198Superficial 15 21Deep 30 23

FNCLCC grade 0.005a

Grade 1 14 25Grade 2 20 17Grade 3 12 2

AJCC stage 0.007a

Stage 1 4 15Stage 2 17 16Stage 3 22 11

ASS1 methylationb <0.001a

Not identified 2 17Present 21 4

aStatistically significant.bForty-four cases with methylation and immunohistochem-ical data.

ASS1 Deficiency in Myxofibrosarcomas

www.aacrjournals.org Clin Cancer Res; 19(11) June 1, 2013 2863




xenografts. Mice were anesthetized with 3% isofluoraneafter intraperitoneal injection of 150 mg/kg body weightD-luciferin. Ten minutes after injection of the D-luciferin,images were acquired for 10 seconds to 2 minutes. Aphotograph was taken for projection of the pseudocolorimages representing the spatial distribution of photoncounts. For bioluminescence imaging plots, photon fluxwas calculated for each mouse using a square region ofinterest, encompassing the head of the mouse in a supineposition. This value was scaled to a comparable back-ground value (from a luciferin-injected mouse withouttumor cells) and normalized to the value obtained imme-diately after xenografting (day 0).

Statistical analysisAssociations of ASS1 immunoexpression with clinico-

pathologic factors and the ASS1 methylation status wereevaluated using a c2 or Fisher exact test as appropriate.Follow-up data were available in 89 patients, and themedian follow-up time was 30.6 (range, 2–229) months.The end points were metastasis-free and disease-specificsurvival. Univariate survival analyses were compared usingthe log-rank test. Amultivariatemodelwas carried out usingthe Cox regression, including parameters with univariateP < 0.05. Student t test was used to analyze the quantitativeRT-PCR results, and functional and pharmacologic assaysfor cell line and xenograft samples.

ResultsASS1 as a differentially downregulated gene regulatingamino-acid biosynthesis

From the dataset of 34 myxofibrosarcomas in the publictranscriptome, we focused on 30 probes covering 22 genesregulating amino-acid biosynthesis. Among statistically sig-nificant genes, ASS1 was the top-ranking differentiallyexpressed candidate (Fig. 1A and Supplementary TableS1), prompting us to characterize endogenous expressionsof the ASS1 transcript and protein inOH931, NMFH-1, andNMFH-2 myxofibrosarcoma cell lines. Unlike CCD966SKdermal fibroblasts, the real-time RT-PCR assay showedeither undetectable or barely detectable endogenous ASS1mRNA in all 3 myxofibrosarcoma cell lines (Fig. 1B, top),reflective of complete depletion of the ASS1 protein in theWestern blots (Fig. 1B, bottom).

Epigenetic methylation underlying the ASS1 deficiencyin myxofibrosarcoma cells

We speculated on whether DNA deletion might accountfor depletion of ASS1. Reappraisal of our published geno-mic profiling data of myxofibrosarcomas did not revealcopy number imbalances at the probes spanning ASS1 at9q34.11 in themajority of samples, except for one featuringa hemizygous deletion (Supplementary Fig. S1A). As dele-tion did not seem to be a predominant event underlying theASS1 deficiency in myxofibrosarcomas, we next considered

Nontumor soft tissueA

B C D

Myxofibrosarcoma

Myxofibrosarcoma

1.0

0.8

0.6

0.4

0.2

0 020

40

60

80

GC

Perc

enta

ge

CCD966 OH931

0.004

CCD966SK Fibro

blast

OH931

NMFH1

NMFH2

0.05

Undetect

ed

NMFH1 NMFH2

M U M U M U M U M U M U

CD966SK

CCD9661xTE M.Sssl ASS1

GAPDH

ASS1

GAPDH

OH931

(5-Aza)

OH931

OH931

0.06

0.05

0.77

0.004

0.35

Undetect

ed

NMFH1

NMFH1

NMFH1

(5-Aza)

NMFH2

NMFH2

NMFH2

(5-Aza)

133317094

CpG

0 200 400 600 800 1,000 (b.p.)

133376661Exon 1TSS

5′ Regulatory element of ASS1 gene

AS

S1

mR

NA

exp

ress

ion (

fold

)

1.0

0.8

0.6

0.4

0.2

0AS

S1

mR

NA

exp

ress

ion (

fold

)

ASS1MTHFD1ASNSBCAT1ENOPH1PYCR1CTHPSPHALDH18A1ASNSCTH

Figure 1. A, clustering analysis of myxofibrosarcoma samples revealed that ASS1 was the most prominently downregulated amino acid-regulating gene,followed byMTHFD1. B, ASS1 mRNA (top) and protein (bottom) expressions were diminished in 3 myxofibrosarcoma cell lines, compared with CCD966SKfibroblasts. C, methylation-specific PCR (MSP) targeting ASS1 intron 1 (top: TSS, transcription start site; blue rectangle, MSP; yellow rectangle,pyrosequencing) identified methylated bands in all ASS1-deficient myxofibrosarcoma cells but not in CCD966SK fibroblasts (bottom: TE, TE buffer; M.SssI,M.SssI methyltransferase-treated DNA). D, 5-Aza-20-deoxycytidine treatment (1 mmol/L, 4 days) restored expressions of ASS1 mRNA (top) and protein(bottom) in NMFH2 and OH931 but not in NMFH-1 cell lines, although all 3 showed methylation of ASS1 gene.

Huang et al.





whether this aberration resulted from epigenetic silencingsimilar to that seen in mesotheliomas and ovarian cancers(27, 29). Methylation-specific PCR showed ASS1 genemethylation in all 3 myxofibrosarcoma cell lines devoidof protein expression but not in ASS1-expressingCCD966SK dermal fibroblasts (Fig. 1C), consistent withthe pyrosequencing results (Supplementary Fig. S1B).Accordingly, myxofibrosarcoma cells were treated with5-aza-20-deoxycytidine, which restored ASS1 expressionwith comparable fold changes between transcripts andproteins in both OH931 and NMFH-2 cells, albeit beingmore prominent in NMFH-2 (Fig. 1D). However, 5-aza-20-deoxycytidine failed to rescue ASS1 expression inNMFH-1 cells. This difference prompted us to speculatethat deregulated modification of histone codes likelymediates ASS1 epigenetic silencing in this particular cellline. Regardless of the presence or absence of 5-aza-20-deoxycytidine, neither separate nor combined treatmentwith trichostatin, 3-Dznep, and/or UNC0638 restoredASS1 expression. This result remained identical whenshort-hairpin RNA (shRNA) against EZH2 was appliedas an alternative to 3-Dznep to suppress histone methyl-transferation of H3K27 by downregulating polycombrecessive complex 2-associated EZH2 that has intrinsicmethyltransferase activity (Supplementary Fig. S1C).

Associations of ASS1 immunoexpression withmethylation status, clinicopathologic features, andpatient survivalTMA-based ASS1 immunohistochemistry was conducted

on 90 primary myxofibrosarcomas linked to clinicopatho-logic characteristics listed in Table 1. Being identified in40 (44%) cases, aberrant ASS1 loss was strongly related tohypermethylation of ASS1 gene in 25 of 44 cases (57%)with informative data from methylation-specific PCR (P <0.001, Fig. 2A) andwas significantly associatedwith increas-ing FNCLCC grades (P¼ 0.005, Fig. 2B) and American JointCommittee on Cancer AJCC stages (P ¼ 0.007). The accu-racy of methylation-specific PCR was substantiated in 10selected tumor samples subjected to pyrosequencing(Supplementary Fig. S1B and Supplementary Table S2).The results indicated that a methylation-associated ASS1deficiency contributed to myxofibrosarcoma progression.Furthermore, an aberrant ASS1 deficiency univariately pre-dicted worse outcomes (Fig. 2C) for both disease-specific(P ¼ 0.0035) and metastasis-free survival (P ¼ 0.0052). Inthe multivariate comparison (Supplementary Table S3),this aberration also remained prognostically independent,along with higher grades for both disease-specific [P ¼0.0230, HR: 10.416] and metastasis-free survival (P ¼0.0213, HR: 4.237).

ADI-PEG20 attenuated cell viability in ASS1-deficientmyxofibrosarcoma cellsSince hepatocellular carcinomas respond to ADI-

PEG20 treatment because of auxotrophy for arginine intumor cells devoid of ASS1, we investigated whether ADI-PEG20 holds promise as targeted therapy in ASS1-defi-

cient myxofibrosarcomas (21). We characterized thepharmacokinetics of ADI-PEG20 in arginine deprivation,revealing a dose-dependent reduction in arginine in theculture medium of OH931 cells within a range of ADI-PEG20 concentrations up to 100 ng/mL (SupplementaryFig. S2A), at which point arginine was no longer detect-able. Myxofibrosarcoma cells and fibroblasts were incu-bated with the indicated concentrations of ADI-PEG20for 72 hours, drastically attenuating cell viability in allmyxofibrosarcoma cells at 100 ng/mL. However, ASS1-expressing CCD966SK fibroblasts showed resistancewithout cytotoxicity at a concentration of 500 ng/mL(Fig. 3A). We next evaluated how ADI-PEG20 affectedmyxofibrosarcoma cells engineered to stably reexpressASS1 (Fig. 3B) and found almost complete reversion ofthe inhibition of cell viability, substantiating that thisselective cytotoxicity to ADI-PEG20 was directly ascribedto the ASS1 deficiency in myxofibrosarcoma cells (Fig.3C). Because the methylated ASS1 gene could be reacti-vated with 5-aza-20-deoxycytidine in OH931 and NMFH-2 cells, we further explored whether this demethylatingtreatment compromised the sensitivity of these celllines to ADP-PEG20. After pretreatment with 1 mmol/L5-aza-20-deoxycytidine 4 days beforehand, incubationof ADP-PEG20 for a further 72 hours no longer dimin-ished cell viability in either OH931 or NMFH-2 cells(Fig. 3D), providing an argument against combined ther-apy with demethylating and arginine-depriving drugs inmyxofibrosarcomas.

Therapeutic effect of ADI-PEG20 in OH931 xenograftsWe examined further the therapeutic efficacy of ADI-

PEG20 in murine xenografts of OH931 cells transfectedwithGFP-FFLuc. Comparedwith phosphate-buffered saline(PBS)-treated mice, ADI-PEG20 led to significantly smallertumor volumes from day 14 posttreatment until day 28 atsacrifice (Fig. 4A). The kinetics of tumor growth inhibitionwith ADI-PEG20 treatment in mice were monitored bybioluminescence and quantified. Expressed as decreasedphoton flux, significantly attenuated bioluminescence sig-nals were seen in the ADI-PGE20–treated group at a dose aslow as 2.5 IU/kg (Fig. 4B). Upon sacrifice, obvious tumorshrinkage with regressing fibrosis was observed in ADI-PEG20–treated xenografts, in striking contrast to PBS con-trols (Supplementary Fig. S2B). On histologic examination(Fig. 4C), PBS controls displayed pleomorphic cells withfrequent mitoses in a fibromyxoid matrix, similar to a high-grade myxofibrosarcoma. However, ADI-PEG20–treatedxenografts showed quiescent-appearing tumor cells withremarkably fewer mitoses (P < 0.001) and increased col-lagenous matrix. Immunohistochemically (Fig. 4C), therewere fewer Ki67-labeled cells (P < 0.001) in the ADI-PEG–treated xenografts, and vice versa in the PBS controls,consistent with significantly less BrdUrd uptake in allASS1-reexpressingmyxofibrosarcoma cells (SupplementaryFig. S2C). To elucidate the basis of the antiproliferativeeffect of ADI-PEG20, we analyzed cell-cycle profiles of ADI-PEG20–treated versus PBS–treated OH931 cells by flow






cytometry, which revealed dose-dependent induction of theS-phase arrest by ADI-PEG20 (Supplementary Fig. S2C).

ASS1 as a novel tumor suppressor ofmyxofibrosarcomas

Apart from the therapeutic potential, the biologic effectsof ASS1 on various cancer subtypes remain unclear. Weused stable ASS1-reexpressing versus empty transfectantsto decipher how ASS1 modulates cellular phenotypes inmyxofibrosarcomas.

ASS1 reexpression inhibited tumoral angiogenesis. Intumorigenesis, it is compelling that an angiogenic switch isgoverned by counterbalancing molecules that induce orsuppress angiogenesis (8, 30). Because myxofibrosarcomasare characterized by an extensive vascular network, wehypothesized that ASS1 is involved in regulating angiogen-esis in myxofibrosarcomas. After incubation with condi-tionedmedia, theHUVEC tube-forming capability decreasedsignificantly to 64%, 63%, and 70% for OH931, NMFH-1,and NMFH-2 ASS1-reexpressing transfectants, respectively,

A

B

CDisease-specific survival

ASS1 intact

ASS1-deficient

ASS1 intact

ASS1-deficient

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 (d) 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 (d)

P = 0.0035 P = 0.0052

1.0

0.8

0.6

0.4

0.2

0.0

Cum

. surv

ival

1.0

0.8

0.6

0.4

0.2

0.0

Cum

. surv

ival

Metastasis-free survival

Low stageHigh stage

M U M U M U M U M U M U

1xTE M.Sssl Case 2 Case 34 Case 3 Case 68

Figure 2. A, methylated bands ofASS1 gene were seen inrepresentative high-stagecases but were absent in low-stage tumors. B, ASS1immunoexpression (bottom)was present in a grade 1myxofibrosarcoma (left) but waslacking in a grade 3 lesion (right). C,an ASS1 deficiency predicts worsedisease-specific (left) andmetastasis-free survival (right).

Huang et al.





compared with the corresponding controls (Fig. 5A). Thisresult signified that ASS1 deficiency conferred a proangio-genic function in myxofibrosarcomas, prompting us to

explore mediators of ASS1 in inhibiting angiogenesis withan RT-PCR expression array. According to the selectioncriteria described in Supplementary Method S3, mRNA

1.2

1.0

0.8

0.6

0.4

0.2

0.00 100 200 300

ADI-PEG20 (ng/mL)

0 200 400 600 800 1,000 1,200 1,400ADI-PEG20 (ng/mL)

0 200 400 600 800 1,000 1,200 1,400ADI-PEG20 (ng/mL)

0 200 400 600 800 1,000 1,200 1,400

ADI-PEG20 (ng/mL)

0 200 400 600 800 1,000 1,200 1,400 (ng/mL)

ADI-PEG20

400

CCD966SKNMFH1

NMFH1

NMFH1

NMFH2

NMFH2

NMFH2

ASS1

DDK (Tag)

Myc (Tag)

GAPDH

OH931

OH931 cell lineA B

C D

OH931 Control

Con

trol

4 DaysASS1

GAPDH

3 Days

ADI-PEG20

ADI-PEG20OH931 5-Aza

OH931+5-AzaOH931 wt

NMFH2+5-AzaNMFH2 wt

5-Aza

5-Aza (1 μmol/L)

NMFH2 Control

Con

trol

0 200 400 600 800 1,000 1,200 1,400 (ng/mL)

ADI-PEG20

4 DaysASS1

GAPDH

3 Days

ADI-PEG20

ADI-PEG20NMFH2 5-Aza

5-Aza

5-Aza (1 μmol/L)

OH931

Wild-typeControl vector

ASS1 expression vector

Wild

-typ

e

Contr

ol (

Em

pty

) ve

ctor

AS

S1 e

xpre

ssio

n v

ect

or

Wild

-typ

e

Contr

ol (

Em

pty

) ve

ctor

AS

S1 e

xpre

ssio

n v

ect

or

Wild

-typ

e

Contr

ol (

Em

pty

) ve

ctor

AS

S1 e

xpre

ssio

n v

ect

or

Wild-type

Control vectorASS1 expression vector

Wild-type

Control vector

ASS1 expression vector

500 600

1.2

1.0

0.8

0.6

0.4

0.2

0.0

1.2

1.0

0.8

0.6

0.4

0.2

0.0

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Via

bili

ty

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Via

bili

ty

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Via

bili

ty

Via

bili

ty

Figure 3. ASS1 deficiency in myxofibrosarcomas accounts for susceptibility to ADI-PEG20. A, ADI-PEG20 (100 ng/mL) strongly decreased theviability of ASS1-deficient OH931, NMFH-1, and NMFH-2 myxofibrosarcoma cell lines, whereas CCD966SK fibroblasts showed no response. B, stableexogenous ASS1 reexpression was confirmed in OH931 (left), NMFH-1 (middle), and NMFH-2 (right) myxofibrosarcoma cell lines, showing the expectedbands with anti-ASS1, anti-DDK, and anti-Myc. C, unlike the sensitivity in the parent lines and empty controls, ASS1 reexpression led to resistance toADI-PEG20 inOH931 (top), NMFH-1 (middle), andNMFH-2 (bottom) cells. D, unlike the vehicle control, 5-aza-20-deoxycytidine rescuedASS1 expression butabrogated the inhibitory effect of ADI-PEG20 on OH931 (top) and NMFH-2 (bottom) cells.






expressions of 6 candidate proangiogenic genes were sig-nificantly and consistently downregulated in ASS1-reex-pressing transfectants; in contrast, there were no obviouslyupregulated genes (Supplementary Table S4). There was aconcomitant downregulation of VEGFR1, CXCL-9, andmatrix metalloproteinase (MMP)-9 proteins by Westernblotting (Fig. 5B), respectively, encoded by FLT1, CXCL9,and MMP9 genes, from all the myxofibrosarcoma ASS1-reexpressing transfectants. However, FGFR3, TGF-a, andCXCL10 protein levels decreased inconsistently in 1 or 2ASS1-reexpressing cell lines alone (Fig. 5B). Furthermore,ELISAs (Supplementary Fig. S3A) were conducted to quan-tify each protein in conditioned media from various trans-fectants ofmyxofibrosarcoma cell lines, whereas onlyMMP-9 was consistently and significantly downregulated in allASS1-reexpressing cell lines. These findings indicated thatASS1 may modulate the angiogenic switch of myxofibro-sarcomas through downregulating proangiogenic mole-cules, such as MMP-9.

ASS1 reexpression inhibited tumor growth by suppressingcell proliferation with the induction of G1 arrest. To deci-pher other biologic roles of ASS1 associated with its anti-tumor function, we compared the BrdUrd-incorporatingrate of myxofibrosarcoma cells, which was significantlyreduced in all ASS1-reexpressing myxofibrosarcoma cellline transfectants. Among these, we further evaluatedNMFH-2 cells for real-time cell proliferation by using anECIS assay, which revealed a significantly slower cell growthrate in ASS1-reexpressing transfectants (SupplementaryFig. S3B). These findings support the inhibition of ASS1 incell proliferation, prompting us to investigate how ASS1expression affects the cell-cycle redistribution and anchor-age-independent and in vivo tumor growth. Using flowcytometry, we observed that ASS1-reexpressing transfec-tants of OH931 and NMFH-2 cells displayed a cell-cyclearrest at the G1 phase. Especially prominent in NMFH-2cells was increased cells at the G1 phase with concomitant

lower percentages of cells in the S and G2 phases (Supple-mentary Fig. S4). We conducted Western blotting toevaluate alterations in G1- and G1–S–associated cyclins andcyclin-dependent kinases, which revealed consistentdownregulation of cyclin E expression across all myxofi-brosarcoma cell lines upon ASS1 reexpression. However,this cell-cycle arrest was not significantly induced inNMFH-1 cells by ASS1 reexpression, probably because of anextremely high fraction of cells in the G1 phase in the emptycontrol.

In a soft agar assay, significantly smaller and fewer cellcolonies appeared in ASS1-reexpressing transfectants inall myxofibrosarcoma cell lines, substantiating the roleof ASS1 in restraining anchorage-independent growth(Fig. 5C).

ASS1 reexpression inhibited cell migration and invasion.To investigate whether ASS1 is implicated in cell motilityand metastasis, a wound-healing assay was conducted,showing significantly slower wound closure in ASS1-re-expressing transfectants of all myxofibrosarcoma cell lines(Supplementary Fig. S5A). This indicated that ASS1 mightimpair themigration of myxofibrosarcoma cells. Moreover,OH931 and NMFH-1 cells stably transfected with ASS1 oran empty vector were seeded inmodified Boyden chambersto elucidate the capability for cell invasion. In 22 hours afterseeding, significantly fewer invading tumor cells werecounted for ASS1-reexpressing cells, suggesting an antiin-vasive attribute of ASS1 in myxofibrosarcomas (Supple-mentary Fig. S5B).

Stable ASS1 knockdown conferred proliferative and meta-static capabilities. As a surrogate for unavailable ASS1-expressing myxofibrosarcoma cells, LPS510 liposarcomacells were stably silenced against endogenous ASS1 tocross-validate its tumor-suppressive function. In Supple-mentary Fig. S6, the proliferation, viability, migration, andinvasion of tumor cells increased significantly in ASS1-knockdown LPS510 cells.

OH931 Xenograft

Control (PBS)

Control (8.7+/-0.5E+09) 5 IU (2.2+/-0.8E+07)

2.5 IU (5.4+/-0.7E+07) 10 IU (2.4+/-0.9E+07)

ADI-PEG20 2.5 IU/kgADI-PEG20 5 IU/kg

ADI-PEG20 10 IU/kg

ADI-PEG20Vehicle controlA B C

HE

Ki67

Baseline 1st Week 2nd Week 3rd Week

Time

4th Week

6,000

5,000

4,000

3,000

2,000

1,000

0

Tum

or

volu

me (

mm

3)

Figure 4. ADI-PEG20 inhibits tumor growth and cell proliferation in OH931 xenografts. A, the average tumor volume was larger in PBS-treatedxenografts than in ADI-PEG20–treated counterparts as plotted. B, using in vivo imaging system (IVIS) to measure the bioluminescence of the photonflux, growth kinetics of PBS- and ADI-PEG20–treated xenografts in SCID mice were monitored and photographed at day 28 for varioustreatment conditions. C, control xenografts (top left) displayed mitotically active pleomorphic cells within a myxoid stroma, whereas the ADI-PEG20–treated group (right top) showed regressing fibrotic changes with fewer mitoses. Ki-67 labeling was significantly higher (P < 0.001) in the control (leftbottom) than in the treated group (right bottom).

Huang et al.





In vivo tumor growth-inhibiting function of ASS1The role of ASS1 in tumor growth inhibition was also

examined in OH931 xenografts with and without ectopicASS1 reexpression. The ASS1-reexpressing group displayeda significantly smaller average tumor volume from day 11onward, and this trend continued until sacrifice on day 28(Fig. 5D, left). The excised xenograft tumor specimens were

obviously shrunken and lighter in weight compared withthe controls (Fig. 5D, right). On histologic examination(Fig. 5D, bottom), the control xenografts displayed spindle-shaped to pleomorphic cells in a fibromyxoid matrix,imparting a high-grade myxofibrosarcoma. However,ASS1-reexpressing xenografts showed reduced cellularitywith few, if any, mitoses (P < 0.001) in a myxoid matrix,

NMFH1A

C D

BNMFH2

NMFH1 NMFH2OH931

OH931

100%

80%

60%

40%

20%

0%

P = 0.001 P = 0.004 P = 0.001

OH931-contro

l

NMFH2-contro

l

NMFH2-ASS1

NMFH1-contro

l

OH931-ASS1

NMFH1-ASS1

Rela

tive

tube

form

ing a

bili

ty (

%)

Contr

ol

vecto

r

AS

S1

expre

ssio

n

Vector Vector Vector

FLT1

FGFR3

TGF-α

CXCL9

CXCL10

MMP9

GAPDH

ASS1 ASS1 ASS1

NMFH1-control

(empty vector)

Empty vector

HE

CD31

ASS1 reexpression

NMFH1 ASS1

expression

600

500

400

300

200

100

0

300

250

200

150

100

50

0

2,2002,0001,8001,6001,4001,2001,000

800600400200

00 5 10 15 20 25 30 (Day)

OH931 Control (n = 8)OH931 ASS1 Exp. (n = 8)

OH

931

Con

trol

OH

931

AS

S1

Exp

.

Tu

mo

r vo

lum

e (

mm

3)

1,800

1,500

1,200

900

600

300

0

NMFH1 OH931 NMFH2

EV ASS1 Exp. EV ASS1 Exp. EV ASS1 Exp.

Nu

mb

er

of

colo

nie

s/w

ell

Nu

mb

er

of

colo

nie

s/w

ell

Nu

mb

er

of

colo

nie

s/w

ell

OH931 ASS1

expression

NMFH2 ASS1

expression

OH931-control

(empty vector)

NMFH2-control

(empty vector)

Figure 5. Tumor suppressor role of ASS1 in governing tumor angiogenesis and growth. A, a less well-formed HUVEC vascular network can be seenwhen exposed to conditioned media from three ASS1-reexpressing transfectants (top) by calculating the tube-forming ability plotted in the histograms(bottom). B, validation of differentially expressed genes in the RT-PCR array showed that only protein abundances of VEGFR1, CXCL-9, and MMP-9were consistently downregulated in all 3 ASS1-reexpressing transfectants. C, more and larger cell colonies were formed on soft agar in thecorresponding controls (top) than in the ASS1-reexpressing transfectants (middle) as plotted (bottom). D, left, the average tumor volume of emptyvector-transfected OH931 xenografts was significantly larger than that of ASS1-reexpressing counterparts from day 11 onward. Right, after mice weresacrificed on day 28, control xenografts were obviously larger on macroscopic examination. Bottom, control xenografts (left column) displayedpleomorphic sarcoma cells in a fibromyxoid stroma, whereas the ASS1-reexpressing group (right column) showed slightly atypical tumor cells with asignificantly lower vessel density based on CD31 staining.






reminiscent of a low-grade myxofibrosarcoma. Immuno-histochemically (Fig. 5D, bottom), ASS1-reexpressing xeno-grafts displayed significantly fewer intratumoral vessels, asevidenced by staining with the CD31 endothelial marker.

DiscussionSoft tissue sarcomas are mesenchymal malignancies of

great histologic and biologic heterogeneity, and there areinadequate molecular markers and therapeutic optionsfor clinical management (13, 31, 32). By reappraisingdifferentially expressed genes in a published transcrip-tome, we provided evidence for ASS1 as a novel moleculartarget of myxofibrosarcomas associated with amino-acidmetabolism implicated in tumor progression and thera-peutic strategies. For the first time among soft tissuesarcomas, ASS1, a rate-limiting enzyme that convertscitrulline into arginine in the urea cycle (17), was foundto have been aberrantly lost in a considerable percentageof myxofibrosarcomas. This protein deficiency correlatedwith ASS1 gene silencing secondary to promoter meth-ylation and conferred preferential susceptibility to argi-nine-deprivation with ADI-PEG20. The antitumor efficacyof ADI-PEG20 was dose-dependent and substantiated inboth in vitro and in vivo xenograft models, opening a newavenue in myxofibrosarcoma management for whichcomplete resection remains the mainstay (7, 33). This istherapeutically relevant given the relentless recurrenceand unsatisfactory chemoradiotherapeutic responses ofmyxofibrosarcomas (34). We also provide multiple linesof evidence reinforcing the novel tumor-suppressive roleof ASS1 in myxofibrosarcoma pathogenesis, including itsantiangiogenic, antiproliferative, and antimigratory/inva-sive properties.

Arginine is a nonessential amino acid in mammaliancells because of the ubiquitous presence of ASS1 (17).However, its expression differs in abundance in varioustissues and is mostly regulated at the transcriptional levelby extracellular hormones, nutrients, cytokines, and tran-scription factors (17). Compared with normal tissuecounterparts, a wide spectrum of tumor types displayremarkable variations in ASS1 expression levels, whichare elevated in gastric and colorectal cancers but frequent-ly downregulated in hepatocellular carcinomas, melano-mas, and mesotheliomas, among others (18, 20, 21, 27).However, the mechanistic underpinning responsible forsuppression or reexpression of ASS1 is yet to be eluci-dated, especially in cancer tissues. Until recently, DNAmethylation was identified as an inactivating mechanismof ASS1 expression only in pleural mesotheliomas, lym-phomas, and ovarian carcinomas (27, 29, 35). Neverthe-less, data for ASS1 methylation in hepatocellular carci-nomas or melanomas are limited, despite being 2 of thebest-known neoplasms devoid of ASS1 expression, withauxotrophy for arginine and sensitivity to ADI-PEG20(20). ASS1 was silenced by DNA methylation in allmyxofibrosarcoma cell lines by methylation-specific PCRand pyrosequencing. This epigenetic aberration was also

detected in over a half of successfully analyzed primarymyxofibrosarcomas with a strong association with pro-tein deficiency. Immunohistochemically, the ASS1 defi-ciency in primary myxofibrosarcomas correlated withincreasing tumor grade and stage and independentlypredicted a worse outcome, especially metastasis-freeand disease-specific survival. This clinical relevanceprompted us to address the functional attributes of ASS1associated with various cancer hallmarks.

In all myxofibrosarcoma cell lines, stable ASS1 reexpres-sion led to decreased BrdUrd uptake, signifying attenuatedDNA synthesis, which was reconciled with significantlyreduced proliferation of ASS1-reexpressing NMFH-2 cellsby real-time ECIS measurements. Furthermore, anchorage-independent growth was impaired in all ASS1-reexpressingtransfectants with significantly fewer and smaller coloniesin the soft-agar assays. These findings highlight the anti-proliferative function of ASS1 in suppressing the growth ofmyxofibrosarcomas, which was associated with the induc-tion of G1 arrest, most likely resulting fromdownregulationof cyclin E. Distant dissemination is the leading cause ofmortality in sarcomas, and is a great hindrance to treatment(31, 32). In osteosarcomas of the bone, reduced ASS1immunoexpression was reported to be a biomarker ofpulmonarymetastasis after neoadjuvant chemotherapy andcurative resection, whereas there was no mechanistic elu-cidation of the ASS1 deficiency with enhanced tumormetastasis (36). Herein, we further reinforced the role ofASS1 in suppressingmetastasis. In stable ASS1 transfectantsof myxofibrosarcoma cell lines, we observed delayed clo-sure in wound-healing assays and decreased tumor cellinvasion using Matrigel assays. These results are consistentwith the ASS1 deficiency as a harbinger of worsemetastasis-free survival in myxofibrosarcomas. Moreover, we alsovalidated the antiproliferative and antimetastatic propertiesof ASS1 using RNA interference in the ASS1-expressingLPS510 cell line

In cancer cells, angiogenesis is necessary to deliver nutri-ents and oxygen and eliminate metabolic wastes (8). Theangiogenic switch is characterized by deregulated predom-inance of proangiogenic over antiangiogenic factors (8, 30).ASS1 reexpression was substantiated to exhibit an antian-giogenic effect in all 3 myxofibrosarcoma cell lines withsignificantly attenuated HUVEC tube-forming capability.The downregulating effect of ASS1 on MMP-9 expressionwas significant and consistent in 3 different myxofibrosar-coma cell lines, as validated by a real-time RT-PCR,Westernblotting, andELISA.Among the secretedMMPs,MMP-2 andMMP-9 have attracted particular attraction because of theirfrequent overexpression and association with aggres-siveness in various cancer types (37). Also unique withintheir catalytic domains is the presence of fibronectin type IIrepeats that bind to stromal collagens whereMMP-9 cleavescollagen IV, unmasks cryptic sites, and mobilizes the vas-cular endothelial growth factor (VEGF), which is importantfor angiogenesis (37, 38). Regardless of whether it is secret-ed from tumor cells or inflammatory and stromal cells in thetumor microenvironment, the classical role of activated

Huang et al.





MMP-9 is to degrade the extracellular matrix to enhancetumor invasion and metastasis (37, 39). Therefore, theantimigratory and anti-invasive capabilities of ASS1 re-expression may be partly mediated by modulating MMP-9 expression and activity in myxofibrosarcomas.As a pegylated arginine deiminase, ADI-PEG20 has a

long half-life in the plasma to deprive arginine. Promisingbenefits have been obtained in clinical trials of advancedhepatocellular carcinomas and melanomas (20), and thenumber of other cancer types known to lack ASS1 expres-sion is increasing (18). Among soft tissue sarcomas,in vitro and in vivo growth inhibition was first achievedby ADI-PEG20 treatment in myxofibrosarcoma at thera-peutic doses. It attenuated cell proliferation, induced aS-phase arrest, and led to a profound loss of tumorige-nicity in OH931 xenografts. This potent effect indicatedthe therapeutic potential of ADI-PEG20 in myxofibro-sarcomas, warranting a systematic assessment of ASS1immunoexpression for prospective clinical trials. How-ever, 5-aza-20-deoxycytidine pretreatment recovered ASS1expression in 2 of 3 ASS1-methylated myxofibrosarcomacells at the cost of losing the original susceptibility toADI-PEG20. In contrast, in the ASS1-methylated NMFH-1cells, 5-aza-20-deoxycytidine pretreatment failed to regainASS1 expression irrespective of combined or separateuse of EZH2 knockdown and various inhibitors againstformation of repressive histone codes. This finding inNMFH-1 myxofibrosarcoma cells was not in keeping withthe series of ASS1-methylated T-cell lymphoma cell lines,in which the ASS1 enzyme was readily reexpressed bydemethylating agents (35). Our results indicated that acombination therapy of ADI-PEG20 and demethylatingagents appears without effect in myxofibrosarcomas, asubset of which may use additional mechanism(s) otherthan DNA methylation and/or deregulated histone mod-ifications to abolish ASS1 expression.In conclusion, ASS1 deficiency is associated with pro-

moter methylation and has a negative impact on prognosisin the clinic. ASS1 has been substantiated in vitro and in vivo

as a tumor suppressor in myxofibrosarcoma via antiangio-genic, antiproliferative, and antimigratory/anti-invasiveproperties. ASS1 deficiency renders ADI-PEG20 as a prom-ising strategy in the management of myxofibrosarcomas.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: H.-Y. Huang, Y-H. Wang, F.-M. Fang, Li-T. Chen,C.-F. LiDevelopment of methodology: H.-Y. Huang, Y-H. Wang, S.-C. Yu, Y.-L.Shiue, C.-F. LiAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): W.-R. Wu, Y-H. Wang, Y.-L. Shiue, C.-H.Hsing, Li-T. Chen, C.-F. LiAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.):H.-Y.Huang, J.-W.Wang, J.-W. Tsai, S.-H. Li,H.-C.Hung, S.-C. Yu, J. Lan, Li-T. Chen, C.-F. LiWriting, review, and/or revision of the manuscript: H.-Y. Huang, Li-T.Chen, C.-F. LiAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): H.-Y. Huang, J. Lan, C.-H. Hsing,Li-T. Chen, C.-F. LiStudy supervision: H.-Y. Huang, F.-M. Fang, Li-T. Chen

AcknowledgmentsThe authors thank Dr. Bridge and Dr. Ogose for providing myxofibro-

sarcoma cells, Dr. Fletcher for LPS510 cells, Polaris for ADI-PEG20, Dr.Marquez for 3-Dznep, Dr. Szlosarek for critical review and editing of themanuscript, and Chang Gung genomic core laboratory for technical assis-tance (CMRPG880251).

Grant SupportThis work was sponsored by National Science Council [NSC96-2320-B-

182A-008-MY3 (to H-Y. Huang); NSC99-2628-B-182-003-MY3 (to H.-Y.Huang); NSC99-2320-B-384-001-MY2 (to C-F. Li); NSC 101-2320-B-384-001-MY3 (to C-F Li)], Department of Health [DOH100-TMP-009-3 (to C-F.Li) and DOH101-TD-C-111-004 (to C-F. Li and L.-T. Chen)], and ChangGung Hospital (CMRPG870753, CMRPG8A03212; to H-Y. Huang) ofTaiwan.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received August 9, 2012; revised March 18, 2013; accepted March 24,2013; published OnlineFirst April 2, 2013.

References1. Lin CN, Chou SC, Li CF, Tsai KB, Chen WC, Hsiung CY, et al.

Prognostic factors of myxofibrosarcomas: implications of marginstatus, tumor necrosis, and mitotic rate on survival. J Surg Oncol2006;93:294–303.

2. Mentzel T, Calonje E, Wadden C, Camplejohn RS, Beham A, SmithMA, et al. Myxofibrosarcoma. Clinicopathologic analysis of 75 caseswith emphasis on the low-grade variant. Am J Surg Pathol1996;20:391–405.

3. Mutter RW, Singer S, Zhang Z, Brennan MF, Alektiar KM. Theenigma of myxofibrosarcoma of the extremity. Cancer 2012;118:518–27.

4. Li CF, Wang JM, Kang HY, Huang CK, Wang JW, Fang FM, et al.Characterization of gene amplification-driven SKP2 overexpression inmyxofibrosarcoma: potential implications in tumor progression andtherapeutics. Clin Cancer Res 2012;18:1598–610.

5. Willems SM, Debiec-Rychter M, Szuhai K, Hogendoorn PC, SciotR. Local recurrence of myxofibrosarcoma is associated withincrease in tumour grade and cytogenetic aberrations, suggesting

a multistep tumour progression model. Mod Pathol 2006;19:407–16.

6. Oda Y, Takahira T, Kawaguchi K, Yamamoto H, Tamiya S, Matsuda S,et al. Altered expression of cell cycle regulators in myxofibrosarcoma,with special emphasis on their prognostic implications. Hum Pathol2003;34:1035–42.

7. Sanfilippo R, Miceli R, Grosso F, Fiore M, Puma E, Pennacchioli E, et al.Myxofibrosarcoma:prognostic factorsandsurvival inaseriesofpatientstreated at a single institution. Ann Surg Oncol 2011;18:720–5.

8. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation.Cell 2011;144:646–74.

9. Cairns RA, Harris IS, Mak TW. Regulation of cancer cell metabolism.Nat Rev Cancer 2011;11:85–95.

10. Gatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis?Nat Rev Cancer 2004;4:891–9.

11. Liu X, Wang X, Zhang J, Lam EK, Shin VY, Cheng AS, et al. Warburgeffect revisited: an epigenetic link between glycolysis and gastriccarcinogenesis. Oncogene 2010;29:442–50.






12. Smerc A, Sodja E, Legisa M. Posttranslational modification of 6-phosphofructo-1-kinase as an important feature of cancer metabo-lism. PLoS ONE 2011;6:e19645.

13. Barretina J, Taylor BS, Banerji S, Ramos AH, Lagos-Quintana M,Decarolis PL, et al. Subtype-specific genomic alterations definenew targets for soft-tissue sarcoma therapy. Nat Genet 2010;42:715–21.

14. Snijders AM, Nowak N, Segraves R, BlackwoodS, BrownN, Conroy J,et al. Assembly of microarrays for genome-wide measurement of DNAcopy number. Nat Genet 2001;29:263–4.

15. Tennant DA, Duran RV, Gottlieb E. Targetingmetabolic transformationfor cancer therapy. Nat Rev Cancer 2010;10:267–77.

16. Pieters R, Hunger SP, Boos J, Rizzari C, Silverman L, Baruchel A, et al.L-asparaginase treatment in acute lymphoblastic leukemia: a focus onErwinia asparaginase. Cancer 2011;117:238–49.

17. Husson A, Brasse-Lagnel C, Fairand A, Renouf S, Lavoinne A. Argi-ninosuccinate synthetase from the urea cycle to the citrulline-NOcycle. Eur J Biochem 2003;270:1887–99.

18. Dillon BJ, Prieto VG, Curley SA, Ensor CM, Holtsberg FW, BomalaskiJS, et al. Incidence and distribution of argininosuccinate synthetasedeficiency in human cancers: a method for identifying cancers sensi-tive to arginine deprivation. Cancer 2004;100:826–33.

19. Wheatley DN. Arginine deprivation and metabolomics: importantaspects of intermediary metabolism in relation to the differentialsensitivity of normal and tumour cells. Semin Cancer Biol 2005;15:247–53.

20. Delage B, Fennell DA, Nicholson L, McNeish I, Lemoine NR, Crook T,et al. Arginine deprivation and argininosuccinate synthetase expres-sion in the treatment of cancer. Int J Cancer 2010;126:2762–72.

21. Yang TS, Lu SN, Chao Y, Sheen IS, Lin CC, Wang TE, et al. Arandomised phase II study of pegylated arginine deiminase (ADI-PEG20) in Asian advanced hepatocellular carcinoma patients. Br J Cancer2010;103:954–60.

22. Kawashima H, Ogose A, Gu W, Nishio J, Kudo N, Kondo N, et al.Establishment and characterization of a novel myxofibrosarcoma cellline. Cancer Genet Cytogenet 2005;161:28–35.

23. Krause AK, Hinrichs SH, Orndal C, DeBoer J, Neff JR, Bridge JA.Characterization of a human myxoid malignant fibrous histiocytomacell line, OH931. Cancer Genet Cytogenet 1997;94:138–43.

24. Lee JC, Li CF, Fang FM, Wang JW, Jeng YM, Yu SC, et al. Prognosticimplication of MET overexpression in myxofibrosarcomas: an integra-tive array comparative genomic hybridization, real-time quantitativePCR, immunoblotting, and immunohistochemical analysis. ModPathol 2010;23:1379–92.

25. Ariizumi T, Ogose A, Kawashima H, Hotta T, Umezu H, Endo N.Multinucleation followed by an acytokinetic cell division in myxofibro-sarcoma with giant cell proliferation. J Exp Clin Cancer Res 2009;28:44.

26. Huang HY, Li CF, Fang FM, Tsai JW, Li SH, Lee YT, et al. Prognosticimplication of ezrin overexpression in myxofibrosarcomas. Ann SurgOncol 2010;17:3212–9.

27. Szlosarek PW, Klabatsa A, Pallaska A, Sheaff M, Smith P, Crook T,et al. In vivo loss of expression of argininosuccinate synthetase inmalignant pleural mesothelioma is a biomarker for susceptibility toarginine depletion. Clin Cancer Res 2006;12:7126–31.

28. Huang HY, Kang HY, Li CF, Eng HL, Chou SC, Lin CN, et al. Skp2overexpression is highly representative of intrinsic biological aggres-siveness and independently associatedwith poor prognosis in primarylocalized myxofibrosarcomas. Clin Cancer Res 2006;12:487–98.

29. NicholsonLJ,SmithPR,Hiller L, SzlosarekPW,KimberleyC,Sehouli J,et al. Epigenetic silencing of argininosuccinate synthetase confersresistance to platinum-induced cell death but collateral sensitivity toarginine auxotrophy in ovarian cancer. Int JCancer 2009;125:1454–63.

30. Nyberg P, Xie L, Kalluri R. Endogenous inhibitors of angiogenesis.Cancer Res 2005;65:3967–79.

31. Coindre JM, Terrier P, Guillou L, Le Doussal V, Collin F, Ranchere D,et al. Predictive value of grade for metastasis development in the mainhistologic types of adult soft tissue sarcomas: a study of 1240 patientsfrom theFrenchFederationofCancerCentersSarcomaGroup.Cancer2001;91:1914–26.

32. Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, et al.Cancer statistics, 2005. CA Cancer J Clin 2005;55:10–30.

33. Huang HY, Lal P, Qin J, Brennan MF, Antonescu CR. Low-grademyxofibrosarcoma: a clinicopathologic analysis of 49 cases treatedat a single institution with simultaneous assessment of the efficacy of3-tier and 4-tier grading systems. Hum Pathol 2004;35:612–21.

34. Haglund KE, Raut CP, Nascimento AF, Wang Q, George S, Baldini EH.Recurrence patterns and survival for patients with intermediate- andhigh-grade myxofibrosarcoma. Int J Radiat Oncol Biol Phys 2012;82:361–7.

35. Delage B, Luong P, Maharaj L, O'Riain C, Syed N, Crook T, et al.Promoter methylation of argininosuccinate synthetase-1 sensitiseslymphomas to arginine deiminase treatment, autophagy and cas-pase-dependent apoptosis. Cell Death Dis 2012;3:e342.

36. Kobayashi E, Masuda M, Nakayama R, Ichikawa H, Satow R, Shita-shige M, et al. Reduced argininosuccinate synthetase is a predictivebiomarker for the development of pulmonary metastasis in patientswith osteosarcoma. Mol Cancer Ther 2010;9:535–44.

37. BauvoisB.New facets ofmatrixmetalloproteinasesMMP-2andMMP-9 as cell surface transducers: outside-in signaling and relationship totumor progression. Biochim Biophys Acta 2012;1825:29–36.

38. Xu J, Rodriguez D, Petitclerc E, Kim JJ, Hangai M, Moon YS, et al.Proteolytic exposure of a cryptic site within collagen type IV is requiredfor angiogenesis and tumorgrowth in vivo. JCellBiol 2001;154:1069–79.

39. EgebladM,Werb Z. New functions for thematrixmetalloproteinases incancer progression. Nat Rev Cancer 2002;2:161–74.

Huang et al.





2013;19:2861-2872. Published OnlineFirst April 2, 2013.Clin Cancer Res Hsuan-Ying Huang, Wen-Ren Wu, Yu-Hui Wang, et al. RelevancePhenotypes, Negative Prognostic Impact, and TherapeuticAberrant Loss via Epigenetic DNA Methylation Confers Aggressive

as a Novel Tumor Suppressor Gene in Myxofibrosarcomas:ASS1

Updated version

10.1158/1078-0432.CCR-12-2641doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2013/03/29/1078-0432.CCR-12-2641.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/19/11/2861.full#ref-list-1

This article cites 39 articles, 6 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/19/11/2861.full#related-urls

This article has been cited by 10 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/19/11/2861To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-12-2641

http://clincancerres.aacrjournals.org/content/suppl/2013/03/29/1078-0432.CCR-12-2641.DC1

http://clincancerres.aacrjournals.org/content/19/11/2861.full#ref-list-1

http://clincancerres.aacrjournals.org/content/19/11/2861.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/19/11/2861


ass1 as a novel tumor suppressor gene in myxoﬁbrosarcomas: … · human cancer biology ass1 as a...

Documents