Soft tissue sarcomas are a heterogenous group of neoplasms with various histological subtypes. Up to now, no individual causal molecular markers for.
Oncogene (1998) 16, 1183 ± 1185 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00
High prognostic signi®cance of Mdm2/p53 co-overexpression in soft tissue sarcomas of the extremities Peter WuÈrl1, Axel Meye2, Hannelore Schmidt2, Christine LautenschlaÈger3, Holger Kaltho4, Friedrich-Wilhelm Rath2 and Helge Taubert2 1
Clinic of General Surgery, 2Institute of Pathology, 3Institute of Medical Biometry and Informatics, 4University of HalleWittenberg, and Clinic of General Surgery, Christian Albrechts University of Kiel, Germany
Soft tissue sarcomas are a heterogenous group of neoplasms with various histological subtypes. Up to now, no individual causal molecular markers for prognosis and therapeutic success have been identi®ed. A tumorigenic connection between the oncogene product Mdm2 and tumor suppressor p53 is generally accepted, but their possible clinical relevance has not yet been investigated suciently in soft tissue sarcoma. In 86 primary soft tissue sarcoma of the extremities (ROresected, T1/2 N0 M0), Mdm2 and p53 overexpression were investigated by immunohistochemistry. The results were adjusted to clinico-pathological characteristics and evaluated for their prognostic relevance by multivariate analysis. In Cox's multivariate analysis with strati®cation of Mdm2 to p53 results, we determined four groups which had dierent prognostic values for relapse-free and overall survival (Mdm27/p5375Mdm27/p53+ 5Mdm2+/p5375Mdm2+/p53+). The most striking ®nding was a relative risk (rr) for overall survival of 18.77 (P=0.006) for patients with Mdm2/p53 cooverexpression (n=40). It is noticeably higher than the additive risk from both factors. Coincident Mdm2/p53 overexpression is an independent molecular marker with the highest prognostic relevance described for soft tissue sarcoma. Thus, a high risk sarcoma group has been de®ned which we believe requires alternative therapeutic approaches. Keywords: soft tissue sarcoma; Mdm2; p53; immunohistochemistry; multivariate prognostic analysis
Introduction Tumor suppressor p53 is essential for normal cell growth, development and dierentiation. A way to regulate p53 function is interaction with Mdm2, since Mdm2 can regulate stability of p53 via degradation (Haupt et al., 1997; Kubbutat et al., 1997). Thus, it can inhibit p53 dependent cell-cycle arrest or apoptosis (Haupt et al., 1993; Chen et al., 1996). In addition to direct Mdm2-p53 protein interactions via complexing, p53 controls transcriptional activity of mdm2, resulting in an autoregulative feedback loop (Oliner et al., 1993; Wu et al., 1993). Disturbances of the ®ne regulated balance between p53 and Mdm2 can result in uncontrolled cell growth and subsequent tumorigenesis.
Correspondence: H Taubert Received 25 June 1997; revised 9 October 1997; accepted 9 October 1997
Alterations of p53 and Mdm2 are usually described as alternative events in soft tissue sarcoma (STS) tumorigenesis (Leach et al., 1993; Florenes et al., 1994). In contrast to the prognostic relevance of p53 changes in STS (Cordon-Cardo et al., 1994; Taubert et al., 1996; WuÈrl et al., 1997), the clinical value of Mdm2 alterations has not been investigated suciently in STS up to now. This background prompted us to evaluate the possible prognostic relevance of simultaneous Mdm2 and p53 overexpression in STS by multivariate analysis. Results Out of 86 patients, 49 (56.9%) had tumors with p53 overexpression and 60 (69.7%) with Mdm2 overexpression, respectively. Forty (46.4%) tumors were positive for both antibodies, including seven of the eight p53 mutated cases (Taubert et al., 1996), and 17 (19.8%) tumors showed neither p53 nor Mdm2 positivity (Table 1). A total of 24 (27.9%) patients had a relapse; 18 of them died (15 with positivity for both antibodies). Among the 37 (43.0%) patients who died because of the disease, 28 (75.7%) had tumors with Mdm2/p53 co-overexpression. Univariate analysis of the relationships between staining and the attributes of adjustment in multivariate tests showed a signi®cant correlation (P=0.011) only for grading, but not for tumor entity and tumor localization (data not shown). In the multivariate Cox's analysis we found strong correlations between Mdm2 and p53 staining and overall survival as well as relapsefree survival (Table 1). In an unstrati®ed interpretation, Mdm2 overexpression has a relative risk (rr) of 6.51 (P=0.0049) and p53 of 3.95 (P=0.0015) for overall survival. In a strati®ed analysis describing Mdm2 results with regard to p53, we determined four groups with a dierent prognosis for overall survival and for relapsefree survival: (i) Mdm27/p537 (ii) Mdm27/p53+ (iii) Mdm2+/p537 and (iv) Mdm2+/p53+ (Figure 1). The most striking group is the one with Mdm2/p53 co-overexpression, where a relative risk of 18.77 (P=0.006) for overall survival and of 7.8 (P=0.006) for relapse-free survival were found. Discussion Our results ± based on relatively homogenous and highly standardized material connected with a long follow-up (56.1 months) ± are the ®rst which show the
Prognostic relevance of Mdm2/p53-overexpression in STS P WuÈrl et al
1184
Table 1 Immunohistochemical data and multivariate results of Cox's analysis for 86 STS patients (*reference group) Immunohistochemistry p53+ Mdm2+ Mdm27/p537* Mdm27/p53+ Mdm2+/p537 Mdm2+/p53+
Total 49 60 17 9 20 40
Overall survival Relative risk p 3.95 6.51 ± 2.89 6.38 18.77
Relapse free survival Relative risk p
0.0015 0.0049 ± 0.034 0.013 0.006
2.91 2.63 ± 2.4 4.19 7.8
0.021 0.029 ± 0.076 0.037 0.006
binding site. The extent of this cleavage correlates with induction of apoptosis (JaÈnicke et al., 1996). It would be of interest to investigate if Mdm2 binding may disturb cleavage of RB and succeeding induction of apoptosis. The present data suggest that, apart from the independent eects of Mdm2 and p53, there are much more than additive eects in the case of combined overexpression. The detection of Mdm2 and p53 cooverexpression has de®ned a clinical risk factor in STS with the highest prognostic relevance (rr=18.77) described until now. This ®nding constrains us to consider alternative therapeutic concepts which take the causal risk factors for these STS patients into consideration. A combined gene therapy with a p53 gene construct and mdm2 antisense has been suggested. Figure 1 Multivariate analysis of overall survival by p53/Mdm2 staining status
multivariate prognostic relevance of isolated Mdm2 and of coincident Mdm2/p53 overexpression for overall and relapse-free survival in STS. Two previous univariate analyses which examined p53 and Mdm2 separately had shown controversial results for clinical relevance (Cordon-Cardo et al., 1994; Reid et al., 1996). Further studies are required in order to establish whether the prognostic relevance of Mdm2/p53 cooverexpression is related to sarcoma development generally. These ®ndings are unexpected, because overexpression of Mdm2 can substantially decrease wild-type and mutant p53 expression (Haupt et al., 1997; Kubbutat et al., 1997). But modi®cation of either Mdm2 or p53, such as by phosphorylation, may protect from Mdm2induced p53 degradation (Kubbutat et al., 1997). In p53 the N-terminal region with a transcriptional activation domain is required for p53-dependent apoptosis (Chen et al., 1996). Mdm2 can bind to the N-terminus of p53 protein (Kussie et al., 1996) and inhibits G1 arrest and apoptosis function of p53 (Chen et al., 1996). Furthermore, apart from direct Mdm2/p53 interactions by highly speci®c binding (Kussie et al., 1996), both proteins can function side by side without complexing (Otto and Deppert, 1993). Thus Mdm2 can be oncogenic independently of p53, via activation of the E2F1/DP1 transcription factors (Martin et al., 1995) or via interaction with the RB protein (Xiao et al., 1995). Mdm2 protein binds to C-terminal fragment of RB (amino acids 792-928) and may perturb RB suppression of E2F function (Xiao et al., 1995). Furthermore, RB protein is cleaved by an ICE-like protease C-terminal (Asp886/Gly887) including Mdm2
Materials and methods We included 86 adult patients (44 men, 42 women) with primary STS of the extremities (22.1% arm, 77.9% leg) in stage T1 (24.4%) or T2 (75.6%) N0 M0 who had undergone primary radical surgical tumor resection (54.7% wide excision, 45.3% compartment resection) in our hospital (between 1986 and 1994). The 86 STS are composed of the six most common entities: 26 (30.2%) liposarcomas, 20 (23.3%) malignant ®brous histiocytomas, 18 (20.9%) ®brosarcomas, nine (10.5%) malignant peripheral nerve sheath tumors, seven (8.1%) rhabdomyosarcomas and six (7.0%) leiomyosarcomas. A total of 15 (17.4%) patients had had grade 1, 39 (45.4%) grade 2 and 32 (37.2%) grade 3 tumors. In eight out of the 86 (9.3%) tumors we had previously identi®ed p53 mutations (Taubert et al., 1996). The median follow-up for the living patients was 56.1 (31 ± 102) months and the mean age was 49 (19 ± 74) years. No patient had received pre-operative radiotherapy or chemotherapy. For all the patients and tumors, clinicopathologic data were assessed as described (WuÈrl et al., 1997). We applied a monoclonal Mdm2 antibody 19E3 and p53 antibody DO-1, using a previously published immunohistochemical procedure (Taubert et al., 1996; WuÈrl et al., 1997). The staining rate of N-terminal binding DO-1 antibody showed the best correlation to prognosis (WuÈrl et al., 1997). Tumors were considered positive (indicating overexpression) when more than 10% of the neoplastic cells were stained. Statistical analysis started with a univariate evaluation of prognostic factors in a log-rank test and was followed by a multivariate cox regression analysis adjusted to grading, tumor size, localization, entity and the kind of surgical resection.
Acknowledgements We are thankful to our colleagues from the Institute of Pathology from the University Halle for continuous support. AM was supported by a grant of the `Marianneund-Dr.-Fritz-Walter-Fischer-Stiftung im Stifterverband fuÈr die Deutsche Wissenschaft'.
Prognostic relevance of Mdm2/p53-overexpression in STS P WuÈrl et al
1185
References Chen J, Wu X, Lin J and Levine AJ. (1996). Mol. Cell. Biol., 16, 2445 ± 2452. Cordon-Cardo C, Latres E, Drobnjak M, Oliva M, Pollack D, Woodru JM, Marechal V, Chen J, Brennan MF and Levine A. (1994). Cancer Res., 54, 794 ± 799. Florenes VA, Maelandsmo GM, Forus A, Andreassen A, Myklebost O and Fodstad O. (1994). J. Natl. Cancer Inst., 86, 1297 ± 1302. Haupt Y, Maya R, Kazaz A and Oren M. (1997). Nature, 387, 296 ± 299. Haupt Y, Barak Y and Oren M. (1993). EMBO J., 15, 1596 ± 1606. JaÈnicke RU, Walker PA, Lin XY and Porter AG. (1996). EMBO J., 15, 6969 ± 6978. Kubbutat MHG, Jones SN and Vousden KH. (1997). Nature, 387, 299 ± 303. Kussie PH, Gorina S, Marechal V, Elenbaas B, Moreau J, Levine AJ and Pavletich NP. (1996). Science, 274, 948 ± 953. Leach FS, Tokino T, Meltzer P, Burrell M, Oliner JD, Smith S, Hill DE, Sidransky D, Kinzler KW and Vogelstein B. (1993). Cancer Res., 53, 2231 ± 2234.
Lin J, Wu X, Chen J, Chang A and Levine AJ. (1995). Functions of the p53 protein in growth regulation and tumor suppression. Cold Spring Harbor Symposia on Quantitative Biology LIX, 215 ± 223. Martin K, Trouche D, Hagemeier C, Sorensen TS, La Thangue NB and Kouzarides T. (1995). Nature, 375, 691 ± 694. Oliner JD, Pietenpol JA, Thiagalingam S, Gyuris J, Kinzler KW and Vogelstein B. (1993). Nature, 362, 857 ± 860. Otto A and Deppert W. (1993). Oncogene, 8, 2591 ± 2603. Reid AH, Reid AH, Tsai MS, Venzon DJ, Wright CF, Lack EE and O'Leary TJ. (1996). Diagn. Mol. Pathol., 5, 65 ± 73. Taubert H, Meye A and WuÈrl P. (1996). Cancer Res., 56, 4134 ± 4136. Wu XW, Bayle JH, Olson D and Levine AJ. (1993). Genes Dev., 7, 1126 ± 1132. WuÈrl P, Taubert H, Meye A, Berger D, LautenschlaÈger C, Holzhausen H-J, Schmidt H, Rath F-W and Dralle H. J. Cancer Res. Clin. Oncol., 123, 502 ± 508. Xiao ZX, Chen J, Levine AJ, Modjtahedi N, Xing J, Sellers R and Livingston DM. (1995). Nature, 375, 694 ± 698.
