Microsatellite instability in Ewing tumor is not ...

J Cancer Res Clin Oncol DOI 10.1007/s00432-007-0220-2

ORIGINAL PAPER

Microsatellite instability in Ewing tumor is not associated with loss of mismatch repair protein expression I. Alldinger · K. L. Schaefer · D. Goedde · L. Ottaviano · U. Dirksen · A. Ranft · H. Juergens · H. E. Gabbert · W. T. Knoefel · C. Poremba

Received: 25 October 2006 / Accepted: 23 March 2007 © Springer-Verlag 2007

Abstract Only few clinical factors predict the prognosis of patients with Ewing tumors. Unfavorable outcome is associated with primary metastatic disease, age > 15 years, tumor volume above 200 ml, and the histological response to chemotherapy. The aim of this study was to elucidate the prevalence and clinical impact of microsatellite instability (MSI) together with the relation between MSI and mismatch repair protein expression in Ewing tumors. DNA from 61 primary Ewing tumors and 11 Ewing tumor cell lines was extracted and microsatellite analysis for the detection of instability or loss of heterozygosity was performed for the Wve markers

IA and KLS contributed equally to this study. I. Alldinger · W. T. Knoefel Department of General, Visceral and Pediatric Surgery, Heinrich-Heine-University of Düsseldorf, Düsseldorf, Germany K. L. Schaefer · D. Goedde · L. Ottaviano · H. E. Gabbert · C. Poremba Department of Pathology, University of Düsseldorf, Düsseldorf, Germany U. Dirksen · A. Ranft · H. Juergens Department of Pediatric Hematology and Oncology, University Children Hospital Münster, Münster, Germany A. Ranft Coordinating Centre for Clinical Trials, University Hospital Münster, Münster, Germany C. Poremba (&) Institute of Pathology, Heinrich-Heine-University Düsseldorf, Moorenstrasse 5, 40225 Düsseldorf, Germany e-mail: [email protected]

of the Bethesda panel BAT25, BAT26, D5S346, D2S123, and D17S250, which represents the established marker panel for the analysis of hereditary non-polyposis colorectal carcinoma (HNPCC) patients. In addition, single nucleotide repeat regions of the two tumor genes BAX and transforming growth factor receptor II (TGFBR2) were also included. All of the 61 samples were suitable for LOH analysis and 55 for the determination of MSIstatus. LOH of these microsatellite markers was detected in 9 of the 61 patients (14.8%). Over all, genetic instability, i.e. MSI and/or LOH, was detected in 17 tumors (27.9%). One out of the 11 tumor cell lines (STA ET1) was characterized by instability of all the Wve Bethesda markers, while from primary tumor samples, only one showed MSI in more than one microsatellite marker (D5S346 and D17S250, MSI-high). Eight of the Wfty-Wve patients (14.5%) showed instability of one microsatellite locus (MSI-low). No instability was detected in BAT26, D2S123, BAX and TGFBR2. There was no signiWcant correlation between MSI and loss of expression of mismatch repair proteins MLH1, MSH2, or MSH6. The impairment of the p53 signaling pathway (expression of TP53 and/or MDM2 by immunohistochemistry) was signiWcantly associated with reduced overall survival (15 of 49 patients (30.6%), P = 0.0410, log-rank test). We conclude that MSI is not prevalent in Ewing tumor and that the nature of instability diVers from the form observed in colorectal carcinoma, the model tumor of MSI. This is documented by the diVerent pattern of MSI (no BAT26 instability) in Ewing tumors and the lack of a strict correlation between MSI-high and loss of expression of MSH2, MSH6 and MLH1. Keywords Ewing sarcoma · Microsatellite instability · TP53 · MDM2

123

J Cancer Res Clin Oncol

Introduction Ewing tumors are bone-associated tumors aVecting predominantly children and adolescents. They are characterized by the fusion of the EWSR1 gene on chromosome 22q12 to the transcription factors FLI1 (85%), ERG (10%), or, in rare cases, to another member of the ets transcription factor family (Delattre et al. 1994). While the detection of this translocation either by RT-PCR or by FISH in routinely processed material is of high diagnostic value (DockhornDworniczak et al. 1994; Friedrichs et al. 2006), the clinical impact of the variability on the architecture of the fusion transcripts is still a matter of debate (Huang et al. 2005). Established factors associated with an unfavorable prognosis are primary disseminated disease, age > 15 years, large tumor volume, and axial primary site (Cotterill et al. 2000; Bacci et al. 2000). In the last decade, histological response to induction chemotherapy has been shown to be the strongest predictive factor (Picci et al. 1997). As not all patients receive surgery following induction chemotherapy, the latter is not available for all patients. Furthermore, several patients, lacking unfavorable predictive factors, show an early progress or relapse of their Ewing tumor. Therefore, additional predictive parameters are needed to identify these patients in order to provide a risk-adapted treatment strategy. Genetically, only few factors are known to be correlated with the prognosis of patients including impairment of the TP53 signaling pathway (Huang et al. 2005) and the number of secondary chromosomal alterations (Ozaki et al. 2001). In carcinomas and especially in colorectal carcinoma, structural, numerical chromosomal alterations (referred to as chromosomal instability) are associated with the impairment of p53 tumor suppressor signaling pathway and poor prognosis. On the other hand, microsatellite instability (MSI) as another form of genomic instability is correlated with prolonged survival of patients (Popat et al. 2005; Risques et al. 2003). MSI, which describes the insertion or deletion of a few nucleotides in short repetitive sequence elements, was described as a frequent genotype in tumors belonging to the hereditary non-polyposis colorectal carcinoma syndrome, HNPCC. In these tumors, the impairment of the mismatch repair system is caused by an inherited mutation in MLH1, MSH2 or other members of the family of repair genes. MSI also occurs in 10–15% of sporadic colorectal carcinoma. In these cases, transcriptional silencing by promoter hypermethylation of MLH1 rather than mutation is the mechanism leading to the functional inactivation of the mismatch repair complex. Since in HNPCC, the detection of MSI is used to ascertain the diagnosis of this familiar cancer syndrome, many eVorts went into Wnding a sensitive and speciWc set of marker satellites, leading to the establishment of the Bethesda panel that includes the

123

mononucleotide STS’s BAT25 and BAT26 and the dinucleotide STS’s D2S123, D5S346, and D17S250. The instability of at least two of these markers is referred to as high MSI (MSI-H) associated with the loss of expression of at least one of the mismatch repair proteins and a better prognosis (Benatti et al. 2005). In other sporadic carcinomas the prevalence of MSI ranged from 0 to 37% of the cases (Samowitz et al. 2002; Yamamoto et al. 2001; Wu et al. 2004; Onda et al. 2001) and the clinical impact of MSI is by far not as clear as in colorectal cancer. Even less data are available for the molecular substructure, prevalence and prognostic role of MSI in sarcomas. Kawaguchi found MSI in 10 of the 40 soft tissue sarcomas, including the two MSI-H (Kawaguchi et al. 2005). In contrast to colorectal carcinoma, in this study a correlation between loss of MLH1/MSH2 expression and MSI could only be documented if MSI-L tumors were included. In one study on bone tumors by Tarkkanen et al. (1996), which included one Ewing tumor, no MSI was found. In contrast, Ewing tumors have been reported to harbor genomic instability in high frequency 13/23 patients (Ohali et al. 2004) using a panel of 17 microsatellites. Ebinger et al. (2005), on the other hand, analyzed a panel of eight mono- and dinucleotide microsatellites, which included the Bethesda panel of microsatellites, and found low frequency in 6% of the Ewing Tumors and no tumors displaying high-frequency MSI. Due to these divergent Wndings, we attempted to evaluate the prevalence of MSI in a collection of 61 primary Ewing tumor samples, and to analyze the correlation of MSI and loss of mismatch repair protein expression in this tumor entity together with the impact of MSI on the clinical outcome.

Patients and methods Patients We analyzed paraYn-embedded Ewing tumor specimens from 61 patients operated on Ewing tumor in diVerent hospitals in Germany. Thirty-eight patients were males (62.3%) and 23 were females. All specimens were analyzed for EWS translocations. Follow-up data were available for 49 patients registered in the EICESS92 or EUROE.W.I.N.G. 99 study. The histological diagnosis of Ewing tumor was conWrmed by a reference pathologist (C.P.) for each tumor. Cell lines The following 11 Ewing tumor cell lines, all characterized by an EWS gene rearrangement, were included: STA-ET-1,

J Cancer Res Clin Oncol

STA-ET-2.1, CADO, RM-82, TC-71, VH-64, WE-68, RDES, SK-N-MC, SK-ES, CHP-100 (Table 1). Cells were grown on gelatine coated culture Xask in RPMI supplemented with 1% Penicillin-Streptomycin (10,000 U/ml penicillin and 10 mg/ml streptomycin), 2 mM L-glutamine, and 10% FCS under standard conditions. DNA extraction A slice of each tumor was stained with hematoxylin/eosin to assess the percentage of tumor cells. DNA was extracted from archived, paraYn-embedded tumor samples. Three to six 10
m slices of the tumor sample were collected in a sterile, autoclaved test tube (1.5 ml). Tissue samples containing less than 80% of tumor cells as quantiWed by histopathological examination were microdissected prior to DNA extraction according to the suggestions of the German HNPCC Consortium (Muller et al. 2006). About 1 ml xylene was added, incubated at 45°C for 15 min and centrifuged at 12,000g for 5 min. Supernatant was discarded. This procedure was repeated twice. Then 1 ml 100% ethanol was added to the pellet, mixed and centrifuged at 12,000g for 5 min. This procedure was repeated twice. The pellet was air-dried for approximately 15 min; 500
l of a cell lysis solution were added, containing 50 mM Tris/HCl, pH 7.4, 25 mM EDTA, 500 mM NaCl, 0.1% Nonident P-40 and 1% SDS, and incubated at 65°C for 15 min; 20
l of proteinase K (20 mg/ml) were added and the sample incubated at 56°C overnight. In cases of incomplete cell solution another 10
l of proteinase K were added and the sample incubated for another hour. Then 1.2
l RNase A (10 mg/ml) was added and the sample incubated at 37°C for 1 h. DNA was isolated using the Qiagen DNA Extraction kit (Qiagen, Hilden, Germany) following the manufacturer’s recommendations. DNA concentration

Table 1 Ewing tumor cell lines Cell line

EWS rearrangement

TP53-status (Kovar et al. 1993)

CADO ES1

t(21;22)

Wild type

RD-ES

t(11;22)

Arg273Cys

RM 82

t(21;22)

Arg273His

SK ES1

t(11;22)

Cys176Phe

SKNMc

t(11;22)

Truncation

STA ET1

t(11;22)

Wild type

STA ET2.1

t(11;22)

Cys277Tyr

TC71

t(11;22)

Truncation

VH 64

t(11;22)

Wild type

WE 68

t(11;22)

Wild type

CHP 100

t(11;22)

Not analyzed

was determined measuring OD260 at pH 7.5 in TE buVer (Eppendorf Bio Photometer, Hamburg, Germany) while DNA purity was estimated using the OD260/OD280 ratio. Primers and PCR Microsatellite analysis was performed using the recommended reference panel for the detection of MSI in colorectal cancer. This panel is composed of two mononucleotide repeats (BAT25 and BAT26) and three dinucleotide repeats (D5S346, D2S123, and D17S250) (Boland et al. 1998). Additionally, the microsatellites in the BCL2-associated X protein gene (BAX, a G8-repeat on 19q13.3–4) and the transforming growth factor receptor II gene (TGFBR2, a A10-repeat on 3p22) were studied. Primers used to amplify microsatellite markers were as follows: BAT26 forward

TGACTACTTTTGACTTCAGCC

BAT26 reverse

AACCATTCAACATTTTTAACCC

BAT25 forward

TCGCCTCCAAGAATGTAAGT

BAT25 reverse

TCTGCATTTTAACTATGGCTC

D2S123 forward

AAACAGGATGCCTGCCTTTA

D2S123 reverse

GGACTTTCCACCTATGGGAC

D17S250 forward

GGAAGAATCAAATAGACAAT

D17S250 reverse

GCTGGCCATATATATATTTAAACC

D5S346 forward

ACTCACTCTAGTGATAAATCGGG

D5S346 reverse

AGCAGATAAGACAGTATTACAGTT

TGFBR2 forward

ATGACTTTATTCTGGAAGATGCTG

TGFBR2 reverse

CACATGAAGAAAGTCTCACCAGGC

BAX forward

TTCATCCAGGATCGAGCAGGGCGA

BAX reverse

CACTCGCTCAGCTTCTTGGTGGAC

BAT-26, D17S250, TGFBRII and BAX forward primers were labeled with the Xuorescent probe TET, BAT25 and D2S123 forward primers were labeled with the Xuorescent probe FAM, and D5S346 forward primer was labeled with HEX. PCR reactions were performed in a total volume of 20
l containing 20 ng of genomic DNA, 2
l 2 mM dNTP, 0.5 U Hotstart taq-Polymerase, 10 £ buVer (both QIAGEN, Hilden, Germany) and primers in optimized concentrations: BAT25, D2S123, D5S346 and D17S250 primer concentrations were 10
M, BAT-26 primer concentration 5
M and TGFBR2 and BAX primer concentration 2.5
M. Multiplex PCR was performed with BAT25, D2S123 and D5S346 in one PCR, BAT-26 and D17S250 in a second and TGFBR2 and BAX in a third PCR under the same conditions: denaturing at 95°C for 15 min, 35 cycles of denaturing at 94°C for 60 s, annealing at 52°C for 60 s and extension at 72°C for 90 s, followed by an extension step at 72°C for 15 min.

123

J Cancer Res Clin Oncol

PCR products were separated using an ABI Prism 310 single capillary genetic analyzer (Applied Biosystems, Foster City, CA, USA). MSI and LOH were assessed comparing the results with the peak pattern obtained from DNA from tumor-cell-free leukocytes. If one microsatellite was instable the tumor was considered MSI-L (low), in the case of two instable microsatellites it was considered MSI-H (high). MSI status was only reported if all the Wve markers showed suYcient PCR products.

external control. The loss of expression was recorded when nuclear staining was observed in normal tissue but not in adjacent malignant cells. For p16 (Han et al. 2001), p53 and MDM2 (Ralhan et al. 2000), a semiquantitative scoring on a four-point scale based on the percentage of positively staining tumor cells was employed according to the indicated references (50%: 3). For statistical analysis, groups of negative and week staining (score 0 and 1) or moderate and strong staining (scores 2 and 3) were combined.

LOH analysis Survival analysis An allelic imbalance was referred to as an LOH if the ratio between the two alleles of a heterozygous sample has shifted to greater than 1.5 compared to the corresponding tumor-free reference sample. Immunohistochemistry For each Ewing tumor, six sections (2–3
m thick) were cut and mounted on to glass slides covered with Histobond®. After dewaxing and rehydration of sections, antigenic site retrieval was accomplished by microwaving each slide for 5 min in 0.01 M citric acid buVer (pH 6.0). Endogenous peroxidase activity was blocked by incubation with 2% hydrogen peroxide for 20 min and non-speciWc binding prevented by incubation with 1% bovine serum albumin (BSA) in phosphate buVered saline (PBS). Sections were subsequently incubated with monoclonal antibodies against MSH2 (clone Ab-2, MERCK/Calbiochem, Darmstadt, Germany, dilution 1:80), MSH6 (clone 44, dilution 1:600) MLH1 (clone G168–728, dilution 1:240) (both BD Biosciences Pharmingen, Heidelberg, Germany), p53 (clone Ab-6, MERCK/Calbiochem, Darmstadt, Germany, dilution 1:500), p16 (clone Ab-7, NeoMarkers/Dunn, Asbach, Germany, dilution 1:50), or MDM2 (clone Ab-1, Oncogene Science/dianova, Hamburg, Germany, dilution 1:100) for 2 h at room temperature. Antibody binding was detected using the Elite Vectastain ABC kit together with the VIP peroxidase substrate kit for visualization (both Vector Laboratories Ltd, Peterborough, UK). Sections were counterstained with hematoxylin, then dehydrated and mounted.

The databases of two consecutive Ewing tumor trials of the German Society of Pediatric Oncology and Hematology (GPOH) (EICESS & Euro-E.W.I.N.G. 99) were frozen in November 2005. Overall survival (OS) time was calculated according to Kaplan and Meier from diagnosis to death or the last date seen where all living patients were censored. Univariate comparisons were made by using log-rank tests with an explorative character so no alpha-correction for multiple testing was done. Power of the tests to identify a reasonable diVerence in OS was somewhat limited due to small group sizes. Statistical analyses were carried out using the SAS statistical software package (Release 8.02, SAS Institute, Cary, North Carolina NC 27513, USA).

Results Patients The sixty-one patients in our study had a mean age of 20.7 years (3–57); 38 were males (62.3%), 23 females. A t(11;22) or t(21;22) chromosomal translocation was detected in tumor samples from 59 patients; 29 patients (47.5%) had an EWS/FLI1 gene fusion joining exon 7 to exon 6, 15 patients (24.5%) had an EWS/FLI1 transcript exon 7/exon 5. Other transcript variants were rare: EWS 7/ FLI1 8 (3 patients), EWS 10/ERG 6 (2 patients) EWS 10/ FLI1 5 (3 patients), EWS 10/FLI1 6 (3 patients), EWS 7/ ERG 6 (3 patients) and EWS 9/FLI1 7 (1 patient).

Immunohistochemistry scoring

Microsatellites

Staining results were scored independently by two pathologists (C.P. and D.G.). For all the six markers only nuclear staining was taken into account for the respective scoring system. According to the guidelines for the assessment of colorectal cancer, specimen were categorized positive if at least 10% of tumor cells show distinct nuclear staining of mismatch repair proteins. In the case of a missing internal control, normal colorectal tissue was used as a positive

DNA of suYcient quality to perform MSI analysis could be extracted from tissue specimens of 55 patients. Tumor specimen of only one patient was characterized by instability of two Bethesda markers (D5S346 and D17S346) and was therefore referred to as MSI-H. Tumors of eight patients (14.5%) showed instability of one microsatellite marker (D17S250 once, three times BAT25 and four times D5S346, Fig. 1) and were therefore called MSI-L. In none

123

J Cancer Res Clin Oncol Fig. 1 Detection of microsatellite instability and loss of heterozygosity by microsatellite analysis. a Analysis of primary tumor tissue. I Instability of satellite D17S250 in primary tumor specimen; above Ewing tumor tissue, below non-tumor control. II Loss of heterozygosity at D2S123; above Ewing tumor tissue, below non-tumor control. b Analysis of Ewing‘s tumor cell lines. STA ET1 (above) shows MSI in all markers analysed, e.g. I D17S250 or II BAT25, while for cell line WE68 (below) no additional peaks can be detected

of the 55 tumor tissue samples, the instability of BAT26 and D2S123-markers could be detected. PCR analysis for the mononucleotide repeats at the BAX and TGFBR2 genes could successfully be performed for 39 specimens showing no instability at all. Among the 11 Ewing tumor cell lines tested for MSI, all but one did not show any evidence of MSI, regardless of the missing “normal tissue”. In contrast, cell line STA ET1 was characterized by additional alleles in all the Wve Bethesda microsatellite markers and was therefore classiWed as MSI-H (Fig. 1). LOH LOH was detected in tumors of nine patients (16.4%), 6 at D17S250, 3 at D2S123 and 2 at D5S346. One tumor sample was characterized by LOH at two loci (D17S250 and D5S346). Only one tumor specimen was characterized by both MSI-L (BAT25) and LOH at D2S123 (Fig. 1). Overall 17 tumors (30.9%) showed genetic instability, i.e. LOH and/or MSI (Table 2).

Immunohistochemistry of mismatch repair proteins Tumor specimens of 59 patients could be successfully analyzed by means of immunohistochemistry. In 54 tumor samples we found a nuclear expression of MLH1, MSH2 and MSH6 (Fig. 2). In three specimen expressions, none of these proteins could be detected, with positive external control. Three reactions in two patients could not be assessed due to technical problems. One of the tumor tissues with no detectable expression of repair proteins was the patient with MSI-H. The other patients had stable microsatellites. The correlation between microsatellite status and expression of MSH2, MSH6 and MLH1 is shown in Table 2. Follow-up of patients Follow-up data were available for 49 patients. Median follow-up time was 4.28 years. A 3-year OS for all patients was 0.633 (95% CI § 0.135); 5-year OS was 0.591 (95% CI § 0.138).

123

J Cancer Res Clin Oncol Table 2 Correlation between microsatellite status and expression of mismatch repair proteins, p16, p53, and MDM2 Patient no.

EWS/FLI1–EWS/ERG

MS

LOH

MMR

p16

p53

MDM2

230

EWS7/FLI5

MSI-H

0

neg

0

0

0

253

n.d.

MSI-L

0

n.s

3

0

0

241

EWS10/ERG6

MSI-L

0

n.s.

1

0

0

239

EWS7/FLI5

MSI-L

0

pos

0

0

0

259

EWS7/ERG6

MSI-L

0

pos

2

1

1

201

EWS7/FLI6

MSI-L

1

pos

3

1

0

245

EWS7/FLI6

MSI-L

0

pos

3

1

1

235

EWS7/FLI6

MSI-L

0

pos

0

2

1

228

EWS7/FLI5

MSI-L

0

pos

3

2

3

236

EWS10/FLI5

MSS

0

neg

2

0

0

193

EWS7/FLI6

MSS

1

pos

3

0

1

216

EWS9/FLI7

MSS

0

pos

3

0

0

203

EWS7/FLI5

MSS

0

pos

3

0

0

220

EWS7/ERG6

MSS

0

pos

1

0

0

225

EWS7/FLI6

MSS

0

pos

3

0

0

232

EWS7/ERG6

MSS

0

pos

3

0

0

233

EWS7/FLI5

MSS

0

pos

3

0

0

238

EWS7/FLI5

MSS

1

pos

2

0

0

246

EWS10/FLI5

MSS

0

pos

3

0

0

250

EWS10/FLI6

MSS

0

pos

0

0

0

254

EWS7/FLI6

MSS

1

pos

3

0

0

229

EWS10/FLI6

MSS

0

posa

3

0

0

257

EWS7/FLI6

MSS

1

neg

3

1

0

218

EWS7/FLI8

MSS

0

pos

3

1

0

244

EWS7/FLI5

MSS

0

pos

0

1

0

252

EWS7/FLI6

MSS

0

pos

3

1

0

202

EWS7/FLI5

MSS

0

pos

3

1

1

207

EWS7/FLI5

MSS

0

pos

3

1

0

211

EWS7/FLI6

MSS

0

pos

3

1

0

213

EWS7/FLI6

MSS

0

pos

3

1

0

223

EWS7/FLI6

MSS

0

pos

3

1

0

224

EWS7/FLI6

MSS

0

pos

3

1

1

234

EWS7/FLI6

MSS

0

pos

3

1

0

226

EWS7/FLI6

MSS

0

pos

3

1

1

248

EWS7/FLI6

MSS

0

pos

3

1

0

251

EWS7/FLI5

MSS

0

pos

3

1

0

255

EWS7/FLI6

MSS

1

pos

1

1

1

200

EWS7/FLI6

MSS

0

pos

3

0

2

258

EWS7/FLI6

MSS

0

pos

3

0

3

196

n.d.

MSS

0

pos

0

1

2

215

EWS7/FLI6

MSS

0

pos

3

1

2

249

EWS7/FLI6

MSS

0

pos

3

1

3

209

EWS7/FLI5

MSS

0

n.s

2

2

0

219

EWS7/FLI6

MSS

1

pos

2

2

0

243

EWS7/FLI5

MSS

0

pos

3

2

1

247

EWS7/FLI6

MSS

0

pos

3

3

1

227

EWS7/FLI5

MSS

0

pos

1

2

2

123

J Cancer Res Clin Oncol Table 2 continued Patient no.

EWS/FLI1–EWS/ERG

MS

LOH

MMR

p16

p53

MDM2

231

EWS7/FLI6

MSS

0

pos

3

2

3

204

EWS7/FLI8

MSS

0

pos

3

3

2

212

EWS7/FLI5

MSS

1

pos

3

3

2

217

EWS7/FLI6

MSS

1

pos

0

3

2

242

EWS10/ERG6

MSS

0

pos

3

3

3

214

EWS7/FLI6

n.s.

0

n.s

3

0

0

195

EWS7/FLI5

n.s.

0

n.s

0

0

0

197

EWS7/FLI8

n.s.

0

pos

1

0

0

199

EWS7/FLI6

n.s.

0

pos

1

0

0

256

EWS10/FLI6

n.s.

0

pos

0

0

1

194

EWS10/FLI5

n.s.

0

pos

0

0

0

205

EWS7/FLI6

n.s.

0

pos

1

0

0

208

EWS7/FLI6

n.s.

0

pos

3

0

1

198

EWS7/FLI6

n.s.

0

posa

0

0

0

Patients were tested for EWS/FLI1 or EWS/ERG gene fusion as described elsewhere (Friedrichs et al. 2006) (n.d. not detectable). Microsatellite status (MS) was classiWed as instable (high or low, MSI-H or MSI-L), stable (MSS) or not suitable (n.s.). “1” indicates samples showing the loss of heterozygosity (LOH) for at least on microsatellite. Expression of mismatch repair proteins (MMR) was considered as either positive, negative or not suitable,while expression of p16, p53 and MDM2 was scored as negative, week, moderate, or strong (0, 1, 2, and 3, respectively) a

pos one marker not evaluable

Data analysis revealed no signiWcant diVerences between the OS of patients without (n = 35/42, 83.3%) and with MSI (at least one marker instable, n = 7, 16.7%, log-rank test, P = 0.429) or LOH (n = 11/49, 22.5%; P = 0.369). Figure 3 shows that also the combination of MSI and LOH to identify the tumors showing any kind of genetic instability, had no impact on the OS of patients (n = 14/36, 38.9%; P = 0.303). In addition to the direct analysis on genomic instability, i.e. MSI and LOH, we also investigated whether the p53 tumor suppressor protein itself and p16INK4A and MDM2, both involved in the TP53 signaling pathway, were aberrantly expressed in Ewing tumor samples. A Wrst pairwise analysis of p16, p53, and MDM2, respectively, revealed a signiWcant correlation between MDM2 and p53 positive samples (P = 0.0009, Fisher’s exact test) while no correlation could be observed between p16 and one of the other markers. In OS, no diVerence could be detected for patients without or low (n = 17/32, 53.1%) compared with moderate or strong expression of p16 INK4A (n = 15, 46.9%; P = 0.971). There was a trend toward decreased survival time for patients with moderate and strong expression of p53 (P = 0.09) but not for MDM2 (P = 0.31; for both: n = 11/ 49, 22.5%; no expression: n = 38, 77.5%). About 15 of the 49 tumor specimens (32.7%) showed expression either of p53 or of MDM2 or of both. OS time of these patients was considerably shorter than for the 34 patients (69.4%) without expression of these proteins with 3-year OS of 0.677 vs.

0.533, and 5-year OS of 0.646 vs. 0.467 (P = 0.041, see Fig. 4).

Discussion To estimate the impact of MSI on the biology of Ewing tumors and to get information on putative underlying genetic mechanisms, we screened a series of paraYnembedded tumor samples from 61 patients for MSI and correlated these data with the loss of immunohistochemical MLH1, MSH2, or MSH6 expression. Applying the Bethesda microsatellite marker panel, which was established to characterize HNPCC, 14.5% of tumors were categorized as MSI-L while only one tumor showed two instable microsatellites and was characterized as MSI-H. It is important to note that the pattern of instable microsatellite markers clearly diVers from the pattern found in colorectal carcinoma. Of the two mononucleotid marker loci which are frequently aVected in MSI-H colorectal carcinoma (Popat et al. 2005), instability of BAT25 was detected only once and instability of BAT26 in none of the Ewing tumor samples. These results are comparable with a previous study of 17 tumor samples (Ebinger et al. 2005). An interesting exception was found in the Ewing tumor cell line STA ET1, which is instable for all the Wve Bethesda markers. Due to the unavailability of the primary tumor we do not know if this genetic phenotype was already present in the original tumor tissue or was acquired during in vitro culturing.

123

J Cancer Res Clin Oncol Fig. 2 Expression of marker proteins in Ewing tumor tissue detected by immunohistochemistry (ampliWcation £200)

Fig. 3 Overall-survival (OS) MSI or LOH positive vs. both low and negative (n = 49)

Another remarkable diVerence between colorectal carcinoma and Ewing tumor exists in the coherence of the expression of mismatch repair proteins and MSI-H: In colorectal carcinoma, MSI and the loss of either MLH1 or MSH2 shows convincing correlation (Ward et al. 2005). In

123

Fig. 4 Overall-survival (OS) p53 or MDM2 positive vs. both low and negative (N = 49)

contrast, in our collection of primary Ewing tumor samples, either no or all (MLH1, MSH2, MSH6) repair proteins were lost and this loss was not strictly correlated with MSI or even MSI-H. Using a set of 13 microsatellite markers (only D17S250 and BAT26 from the Bethesda panel), Ohali et al. (2004)

J Cancer Res Clin Oncol

detected MSI in 48% of tumor samples from a series of 23 Ewing tumor patients (4 MSI-H, 7 MSI-L). The diVering rates of MSI tumors in this study and our investigations may be due to the diVerent microsatellite markers employed. Since the aim of our study was to investigate the impact of MSI, which is related to mismatch repair deWciency as observed in colorectal carcinoma, we focused on the Bethesda panel of microsatellite markers. The correlation between mismatch repair protein deWciency and MSI in tumor entities other than colorectal carcinoma is quite heterogeneous: For familial and sporadic breast carcinoma, both MSI and loss of MLH1/MSH2 expression represent an extremely rare event (Adem et al. 2003). On the other hand, in ovarian carcinoma (Geisler et al. 2003) and endometrial carcinoma (Stefansson et al. 2002) pathological expression of mismatch repair proteins and MSI-high are signiWcantly correlated. In adult acute myeloid leukemia MSI was found in 19.6% of the patients, but without correlating to the expression of MLH1, MSH2, or MSH6 as measured by RT-PCR. In 10 out of the 40 soft tissue sarcomas, MSI was detectable; however, a correlation between loss of MLH1/MSH2 expression and MSI could only be documented if MSI-L tumors were included (Kawaguchi et al. 2005). Finally, for the rare malignant melanoma of soft parts (MMSP, synonym: clear cell sarcoma of soft tissue), which are also characterized by a chromosomal translocation involving the EWS gene, a recent study has indicated a dramatically reduced prevalence of MSI compared with dermal malignant melanoma and, at the same time, could exclude loss of MLH1/MSH2 expression for all MMSP samples analyzed (n = 9). Taken together, these examples show that for both epithelial and mesenchymal tumors, there is no general rule in terms of prevalence of MSI and correlation to mismatch repair protein expression, and therefore, these features have to be determined separately for each tumor entity. For our collective of Ewing’s tumors, there was no signiWcant correlation between MSI-status and patients survival. One possible explanation for the better prognosis of MSI-H colorectal carcinomas (Popat et al. 2005; Han et al. 2001) is the high amount of mutated proteins, which in turn are leading to the activation of the immune system as evidenced by elevated numbers of activated cytotoxic T cells inWltrating the tumor tissue (Banerjea et al. 2004; Phillips et al. 2004). This mechanism was not observed in colorectal carcinoma showing comparable low levels of MSI (MSI-L) like in our Ewing tumors. This may be an explanation that no favorable eVect was observed for these sporadic MSI + tumors. While MSI represents only one manifestation of genetic instability, in former studies, others and we were able to demonstrate that chromosomal instability is associated with tumor progression in Ewing tumor (Hattinger et al. 2002;

Schaefer et al. 2004). It is well established that the loss of p53 eYciency is directly linked to both, tolerance of gross chromosomal replication errors as well as to an adverse eVect on patient’s prognosis (Huang et al. 2005; de Alava et al. 2000). Moreover, several studies show a signiWcant inverse correlation between MSI/mismatch repair impairment and positive staining for p53 in colon cancer (Ogino et al. 2006; Samowitz et al. 2001) and the same was also found for endometrial carcinoma: the endometroid subtype is characterized by MSI, near diploid karyotype, and lower incidence of p53 impairment while the non-endometroid subtype shows more often p53 mutation, aneuploidy but low prevalence of MSI (Cerezo et al. 2006). As in this study on Ewing’s sarcoma, the signiWcant coexpression of p53 and MDM2 is reported also for other tumor entities, colorectal carcinoma, (Broll et al. 1999), renal carcinoma, (Haitel et al. 2000), endometroid carcinoma (Ambros et al. 1996), supporting the existence of regulatory mechanism involving these to proteins. Using immunohistochemical positive staining of p53 and/or MDM2 as a surrogate marker of p53 eYciency, we could underline the high impact of this signaling pathway on patient outcome in contrast to MSI status: about one-third of patients was scored p53/MDM2 positive and was found to show a signiWcantly reduced survival time. On the basis of these Wndings, we conclude the kind of MSI found in Ewing tumors remarkably diVers from MSI found in colorectal carcinoma with respect to frequency, the pattern of microsatellite markers aVected, and the relationship to the mismatch repair proteins MLH1 and MSH2. The conWrmation that p53 pathway impairment in Ewing’s tumors is associated with a worse prognosis of patients supports the claim to include this parameter in the routine staging analysis of this tumor entity. Acknowledgments This work was supported by grants from the “Forschungskommission der Medizinischen Fakultät Düsseldorf” and EuroBoNet (6th Framework Network of Excellence of the European Union).

ConXict of interest statement

None declared.

References Adem C, Soderberg CL, Cunningham JM, Reynolds C, Sebo TJ, Thibodeau SN et al (2003) Microsatellite instability in hereditary and sporadic breast cancers. Int J Cancer 107(4):580–582 Ambros RA, Sheehan CE, Kallakury BV, Ross JS, Malfetano J, Paunovich E et al (1996) MDM2 and p53 protein expression in the histologic subtypes of endometrial carcinoma. Mod Pathol 9(12):1165–1169 Bacci G, Ferrari S, Comandone A, Zanone A, Ruggieri P, Longhi A et al (2000) Neoadjuvant chemotherapy for Ewing’s sarcoma of bone in patients older than thirty-nine years. Acta Oncol 39(1):111–116

123

J Cancer Res Clin Oncol Banerjea A, Ahmed S, Hands RE, Huang F, Han X, Shaw PM et al (2004) Colorectal cancers with microsatellite instability display mRNA expression signatures characteristic of increased immunogenicity. Mol Cancer 3:21 Benatti P, Gafa R, Barana D, Marino M, Scarselli A, Pedroni M et al (2005) Microsatellite instability and colorectal cancer prognosis. Clin Cancer Res 11(23):8332–8340 Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW et al (1998) A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58(22):5248–5257 Broll R, Stark A, Windhovel U, Best R, Strik MW, Schimmelpenning H et al (1999) Expression of p53 and mdm2 mRNA and protein in colorectal carcinomas. Eur J Cancer 35(7):1083–1088 Cerezo L, Cardenes H, Michael H (2006) Molecular alterations in the pathogenesis of endometrial adenocarcinoma: therapeutic implications. Clin Transl Oncol 8(4):231–241 Cotterill SJ, Ahrens S, Paulussen M, Jurgens HF, Voute PA, Gadner H et al (2000) Prognostic factors in Ewing’s tumor of bone: analysis of 975 patients from the European Intergroup Cooperative Ewing’s Sarcoma Study Group. J Clin Oncol 18(17): 3108–3114 de Alava E, Antonescu CR, Panizo A, Leung D, Meyers PA, Huvos AG et al (2000) Prognostic impact of P53 status in Ewing sarcoma. Cancer 89(4):783–792 Delattre O, Zucman J, Melot T, Garau XS, Zucker JM, Lenoir GM et al (1994) The Ewing family of tumors: a subgroup of small-roundcell tumors deWned by speciWc chimeric transcripts. N Engl J Med 331(5):294–299 Dockhorn-Dworniczak B, Schafer KL, Dantcheva R, Blasius S, Winkelmann W, Strehl S et al (1994) Diagnostic value of the molecular genetic detection of the t(11;22) translocation in Ewing’s tumours. Virchows Arch 425(2):107–112 Ebinger M, Bock T, Kandolf R, Sotlar K, Bultmann BD, Greil J (2005) Standard mono- and dinucleotide repeats do not appear to be sensitive markers of microsatellite instability in the Ewing family of tumors. Cancer Genet Cytogenet 157(2):189– 190 Friedrichs N, Kriegl L, Poremba C, Schaefer KL, Gabbert HE, Shimomura A et al (2006) Pitfalls in the detection of t(11;22) translocation by Xuorescence in situ hybridization and RT-PCR: a single-blinded study. Diagn Mol Pathol 15(2):83–89 Geisler JP, Goodheart MJ, Sood AK, Holmes RJ, Hatterman-Zogg MA, Buller RE (2003) Mismatch repair gene expression defects contribute to microsatellite instability in ovarian carcinoma. Cancer 98(10):2199–2206 Haitel A, Wiener HG, Baethge U, Marberger M, Susani M (2000) mdm2 expression as a prognostic indicator in clear cell renal cell carcinoma: comparison with p53 overexpression and clinicopathological parameters. Clin Cancer Res 6(5):1840–1844 Han S, Ahn SH, Park K, Bae BN, Kim KH, Kim HJ et al (2001) P16INK4a protein expression is associated with poor survival of the breast cancer patients after CMF chemotherapy. Breast Cancer Res Treat 70(3):205–212 Hattinger CM, Potschger U, Tarkkanen M, Squire J, Zielenska M, Kiuru-Kuhlefelt S et al (2002) Prognostic impact of chromosomal aberrations in Ewing tumours. Br J Cancer 86(11):1763– 1769 Huang HY, Illei PB, Zhao Z, Mazumdar M, Huvos AG, Healey JH et al (2005) Ewing sarcomas with p53 mutation or p16/p14ARF homozygous deletion: a highly lethal subset associated with poor chemoresponse. J Clin Oncol 23(3):548–558 Kawaguchi K, Oda Y, Takahira T, Saito T, Yamamoto H, Kobayashi C et al (2005) Microsatellite instability and hMLH1 and hMSH2

123

expression analysis in soft tissue sarcomas. Oncol Rep 13(2):241–246 Kovar H, Auinger A, Jug G, Aryee D, Zoubek A, Salzer-Kuntschik M, Gadner H (1993) Narrow spectrum of infrequent p53 mutations and absence of MDM2 ampliWcation in Ewing tumours. Oncogene 8(10):2683–2690 Muller A, Beckmann C, Westphal G, Bocker Edmonston T, Friedrichs N, Dietmaier W et al (2006) Prevalence of the mismatch-repairdeWcient phenotype in colonic adenomas arising in HNPCC patients: results of a 5-year follow-up study. Int J Colorectal Dis 21(7):632–641 Ogino S, Brahmandam M, Kawasaki T, Kirkner GJ, Loda M, Fuchs CS (2006) Combined analysis of COX-2 and p53 expressions reveals synergistic inverse correlations with microsatellite instability and CpG island methylator phenotype in colorectal cancer. Neoplasia 8(6):458–464 Ohali A, Avigad S, Cohen IJ, Meller I, Kollender Y, Issakov J et al (2004) High frequency of genomic instability in Ewing family of tumors. Cancer Genet Cytogenet 150(1):50–56 Onda M, Nakamura I, Suzuki S, Takenoshita S, Brogren CH, Stampanoni S et al (2001) Microsatellite instability in thyroid cancer: hot spots, clinicopathological implications, and prognostic signiWcance. Clin Cancer Res 7(11):3444–3449 Ozaki T, Paulussen M, Poremba C, Brinkschmidt C, Rerin J, Ahrens S et al (2001) Genetic imbalances revealed by comparative genomic hybridization in Ewing tumors. Genes Chromosomes Cancer 32(2):164–171 Phillips SM, Banerjea A, Feakins R, Li SR, Bustin SA, Dorudi S (2004) Tumour-inWltrating lymphocytes in colorectal cancer with microsatellite instability are activated and cytotoxic. Br J Surg 91(4):469–475 Picci P, Bohling T, Bacci G, Ferrari S, Sangiorgi L, Mercuri M et al (1997) Chemotherapy-induced tumor necrosis as a prognostic factor in localized Ewing’s sarcoma of the extremities. J Clin Oncol 15(4):1553–1559 Popat S, Hubner R, Houlston RS (2005) Systematic review of microsatellite instability and colorectal cancer prognosis. J Clin Oncol 23(3):609–618 Ralhan R, Sandhya A, Meera M, Bohdan W, Nootan SK (2000) Induction of MDM2-P2 transcripts correlates with stabilized wild-type p53 in betel- and tobacco-related human oral cancer. Am J Pathol 157(2):587–596 Risques RA, Moreno V, Ribas M, Marcuello E, Capella G, Peinado MA (2003) Genetic pathways and genome-wide determinants of clinical outcome in colorectal cancer. Cancer Res 63(21):7206– 7214 Samowitz WS, Holden JA, Curtin K, Edwards SL, Walker AR, Lin HA et al (2001) Inverse relationship between microsatellite instability and K-ras and p53 gene alterations in colon cancer. Am J Pathol 158(4):1517–1524 Samowitz WS, Curtin K, Neuhausen S, SchaVer D, Slattery ML (2002) Prognostic implications of BAX and TGFBRII mutations in colon cancers with microsatellite instability. Genes Chromosomes Cancer 35(4):368–371 Schaefer KL, Brachwitz K, Wai DH, Braun Y, Diallo R, Korsching E et al (2004) Expression proWling of t(12;22) positive clear cell sarcoma of soft tissue cell lines reveals characteristic up-regulation of potential new marker genes including ERBB3. Cancer Res 64(10):3395–3405 Stefansson I, Akslen LA, MacDonald N, Ryan A, Das S, Jacobs IJ et al (2002) Loss of hMSH2 and hMSH6 expression is frequent in sporadic endometrial carcinomas with microsatellite instability: a population-based study. Clin Cancer Res 8(1):138–143 Tarkkanen M, Aaltonen LA, Bohling T, Kivioja A, Karaharju E, Elomaa I, et al (1996) No evidence of microsatellite instability in bone tumours. Br J Cancer 74(3):453–455

J Cancer Res Clin Oncol Ward RL, Turner J, Williams R, Pekarsky B, Packham D, Velickovic M et al (2005) Routine testing for mismatch repair deWciency in sporadic colorectal cancer is justiWed. J Pathol 207(4):377–384 Wu M, Semba S, Oue N, Ikehara N, Yasui W, Yokozaki H (2004) BRAF/K-ras mutation, microsatellite instability, and promoter

hypermethylation of hMLH1/MGMT in human gastric carcinomas. Gastric Cancer 7(4):246–253 Yamamoto H, Itoh F, Nakamura H, Fukushima H, Sasaki S, Perucho M et al (2001) Genetic and clinical features of human pancreatic ductal adenocarcinomas with widespread microsatellite instability. Cancer Res 61(7):3139–3144

123