ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r. com/ loca te / i j go

International Journal of Gynecology and Obstetrics (2008) 101, 166–171

CLINICAL ARTICLE

Heparanase expression increases throughout theendometrial hyperplasia–cancer sequenceJonathan Canaani a, Neta Ilanb, Stella Backc, Guy Gutmana,Israel Vlodavskyb, Dan Grisarua,⁎

a Department of Obstetrics–Gynecology, Tel-Aviv Sourasky Medical Center, Sackler Faulty of Medicine,Tel-Aviv University, Tel-Aviv, Israelb Cancer and Vascular and Biology Research Center, The Bruce Rappaport Faculty of Medicine, The Technion, Haifa, Israelc Department of Pathology, Tel-Aviv Sourasky Medical Center, Sackler Faulty of Medicine, Tel-Aviv University, Tel-Aviv, Israel

Received 25 September 2007; received in revised form 25 October 2007; accepted 25 October 2007

⁎ Corresponding author. DepartmentTel-Aviv Sourasky Medical Center, 6WeiIsrael. Tel.: +972 3 9625622; fax: +972

E-mail address: grisaro@post.tau.a

0020-7292/$ - see front matter © 2All rights reserved.doi:10.1016/j.ijgo.2007.10.019

Abstract

Objective: To assess the expression of heparanase in the different stages leading to endometrialcancer. Methods: The 38 examined specimens included adenocarcinoma, hyperplasia, andnormal endometrium specimens. Heparanase, estrogen, and progesterone receptor expressionswere analyzed immunohistochemically and the intensity was scored. Results: Secretory normalendometrium and simple hyperplasia specimens expressed the lowest mean values ofexpression (1.00 and 0.63, respectively); the complex hyperplasia specimens and G2endometrioid adenocarcinoma showed the highest values of expression (2.33 and 2.71,respectively). A linear trend (P=0.005) of heparanase expression was observed when comparingthe normal endometrium and simple hyperplasia group with the complex hyperplasia+G1carcinoma group and the G2+G3 carcinoma group. Evaluation of atrophic and inactiveendometrium compared with papillary serous carcinomas yielded no significant differences. Wefound no significant correlation between heparanase expression and estrogen receptor orprogesterone receptor expression. Conclusion: Heparanase expression was tightly regulated inendometrial tumorigenesis.© 2007 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd.All rights reserved.

KEYWORDSEndometrial cancer;Endometrial hyperplasia;Estrogen receptorHeparanase;Progesterone receptor

of Obstetrics and Gynecology,zmann Street, Tel-Aviv 64239,3 9625670.c.il (D. Grisaru).

007 International Federation of

1. Introduction

Endometrial carcinoma is the fourth most common malignancyin women of North America and Europe [1]. Endometrial carci-nomas are classified histologically and biologically into 2 types.Type I endometrial cancers are associated with long-term unop-posed estrogen exposure which gives rise to well-differentiated

Gynecology and Obstetrics. Published by Elsevier Ireland Ltd.

Figure 1 Heparanase expression in the endometrial cancerpathway (mean values). Note the increase in heparanase expres-sion with the progression from normal to hyperplasia and finallyto cancer. Abbreviations: NE, normal endometrium; SH, simplehyperplasia; CH, complex hyperplasia; G1, well-differentiatedadenocarcinoma; G2, intermediate differentiated adenocarci-noma; G3, poorly differentiated adenocarcinoma.

167Heparanase expression increases throughout the endometrial hyperplasia–cancer sequence

tumors througha sequenceof hyperplastic lesions. These tumorstend to have a favorable prognosis [2]. Earlier studies haveshown a progressive propensity of hyperplastic lesions to ad-vance tomalignancy, dependent upon specific histopathologicalcharacteristics [3]. In contrast, type II tumors are not estrogendependent and are associatedwith atrophic endometrium, theirhistology is nonendometrioid (i.e., serous papillary), and thetumors have aggressive biology [4]. The molecular mechanismsof carcinogenesis are still obscure, but abnormalities in KRAS2,PTEN, and β-catenin genes have been linked to type I tumors,while abnormalities in TP53 and ERBB2 (HER-2/neu) genes wereassociated with type II tumors [5].

Heparan sulfate proteoglycans (HSPGs) aremacromoleculespresent on the cell surface and extracellular matrix of everyeukaryotic cell. Heparan sulfate (HS) chains bindawide varietyof bioactive molecules, such as chemokines, lipoproteins,enzymes, and growth factors, thus influencing many physio-logical and pathological cellular and matrix processes, such ascell–cell, cell–matrix adhesion, cell proliferation, angiogen-esis, wound healing, tumorigenesis, andmetastasis [6]. Humanheparanase is an endo-β-glucuronidase that cleaves specificheparan sulfate side chains of HSPGs in a variety of normal andmalignant cells [7]. It has been implicated in a multitude ofbiological processes, such as tumor cell metastasis, angiogen-esis, inflammation, and wound healing by releasing HS-boundgrowth factors, such as basic fibroblast growth factor (bFGF)and vascular endothelial growth factor (VEGF) [8]. Increasedexpression of heparanase has been linked to progressivedysplasia and a worse prognosis in a myriad of cancers [9,10].Heparanase expression was found to be increased in correla-tion with clinical stage, tumor grade, lymph node metastasis,and decreased overall survival in endometrial cancer [11,12].One recent study suggested a possible role for estrogen ininducing heparanase overexpression in breast carcinoma[13].

In the present study we assessed the expression ofheparanase in the different stages leading to the 2 types ofendometrial cancer. In addition, we evaluated a possiblecorrelation between the expression of estrogen receptor (ER)and progesterone receptor (PR) and heparanase expression inthe endometrial carcinogenesis.

2. Materials and methods

2.1. Tissue samples

The files of patientswith endometrial cancerwhowere operated onat the Tel-Aviv Sourasky Medical Center, Israel, during 1999–2005were obtained from the Pathology Institute database. All patientshad surgical staging according to the FIGO classification [14]. Thespecimens were reviewed by a dedicated gynecological pathologist(S.B.) and only the ones with good histological architecture andsufficient tissue for analysis were chosen. Specimens from patientswith endometrial hyperplasia and normal endometrium were usedas controls. A total of 38 formalin-fixed and paraffin-embeddedspecimens (patients'median age 57 years; range, 35–94 years)wereselected, and they included 13 endometrioid adenocarcinomas, 3papillary serous carcinomas, 11 hyperplasias (8 simple and 3complex), and 11 normal endometrium specimens (atrophic,inactive, and secretory). Endometrioid adenocarcinoma specimensincluded 11 endometrioid cancers in stage I, and 2 in stage III. Allstudies were approved by the institutional review board.

2.2. Immunohistochemistry

Paraffin-embedded tissue sections (2 μm thick) were depar-affinized and rehydrated. The tissue was then denatured for20 min in a microwave oven in citrate buffer (0.01 mol/L,pH 6.0). Blocking steps included successive incubations in0.2% glycine, 3% hydrogen peroxide in methanol, and 10%goat serum. The first 2 steps were followed by 2 washes inphosphate-buffered saline (PBS). Sections were incubatedovernight at 4°C with a polyclonal mAb (733) antihumanheparanase antibody diluted 1:100 in PBS followed by incu-bation with horseradish peroxidase-conjugated goat–anti-mouse IgG+ IgM antibody (Jackson Laboratory, Bar Harbor, ME,USA). mAb 733 is directed against the c-terminus region of the50-kDa heparanase subunit. The preparation and specificity ofthis mAb were previously described and demonstrated [15].Color was developed using either Sigma Fast 3,3 diaminobenzi-dine tablet sets (Sigma Chemical Co, St. Louis, MO, USA) or aZymed aminoethyl carbazole (AEC) substrate kit (Zymed, SouthSan Francisco, CA, USA) for 10 min followed by counterstainwith Mayer's hematoxylin.

ER and PR expression were analyzed using the monoclonalantibody 6F11 and monoclonal antibody 312, respectively(Novocastra, Newcastle, UK). After deparaffinization, endogen-ous peroxidase activity was blocked with 3% hydrogen peroxidein absolute methanol. Microwave antigen retrieval was per-formed in a 10 mmol/L citrate buffer for 10 min at high power.

Antibody labeling was performed with the avidin–biotincomplex method and the 3, 3′-diaminobenzidine detection kiton a semi-automated instrument (Ventana ES; Ventana MedicalSystems, Tucson, AZ, USA). Antibody labeling took place at 37 °Cwith the sequential addition of the following reagents: inhibitorfor endogenous peroxidase, primary antibody, avidin and biotinblockers, biotinylated secondary antibody, avidin–biotin com-plexwith horseradish peroxidase, 3′-diaminobenzidine-hydrogenperoxide, copper enhancer, and hematoxylin. The primaryantibodies that were used were anti-PR (1:20 dilution withmicrowave antigen retrieval; Novocastra Laboratories, New-castle, UK) and anti-ER (Ventana Medical Systems, Tucson, AZ,USA) with the procedure carried out as specified by themanufacturer.

Table 1 Mean values of heparanase, estrogen receptor, andprogesterone receptor expression

No. Heparanase ER PR

Normal (n=11)Inactive 4 1.75 0.5 2Atrophic 3 2.67 0.33 2.33Secretory 4 1 1.75 2

Hyperplasia (n=11)Simple 8 0.63 2.25 3Complex 3 2.33 1.33 2

Endometrial carcinoma (n=13)G1 4 1.75 1 2.75G2 7 2.71 1.33 1.83G3 2 1.5 2.5 2.5

Papillary serous carcinoma (n=3) 2.33 0.33 0

Abbreviations: ER, estrogen receptor; PR, progesterone receptor.

168 J. Canaani et al.

2.3. Scoring methodology

Heparanase immunoreactivity in glandular tissue was determinedby categorizing staining intensity into 4 groups: (0) no staining; (1)weak staining, barely observed over nonspecific background

Figure 2 Immunohistochemical staining of heparanase, estrogen recein a panel of normal endometrium tissue samples (magnification ×20(column B); Secretory endometrium (column C). H&E staining for tis(magnification ×10). Note that heparanase is highly expressed in atroph

staining; (2) moderate staining observed as coarse aggregates,partially obscuring cytoplasm and counterstained chromatin; and(3) strong staining, observed as coarse membranous bandsobscuring the cell boundaries. ER and PR immunoreactivity wasscored as 0 (absent), +1 (5%–25%), +2 (25%–50%), and +3 (N50%).

2.4. Statistical analysis

Differences between the study groups were evaluated using theKruskal–Wallis test. The correlation between heparanaseexpression and ER and PR expression was determined using aSpearman's test. A one-way analysis of variance (ANOVA) wasused to determine a linear trend of heparanase expressionbetween the different groups. A P value b0.05 was taken as thelevel of significance for all tests.

3. Results

3.1. Heparanase expression in the endometrioidpathway

Heparanase expression was found to be significantly different(P=0.009) among the various stages in theendometrial pathwayto cancer (Fig. 1, Table 1). Secretory normal endometrium and

ptor (ER), and progesterone receptor (PR) (rows 2–4, respectively)). Atrophic endometrium (column A); Proliferative endometriumsue definition is shown for samples as outlined above (first row)ic tissue and absent in secretory and proliferative specimens.

169Heparanase expression increases throughout the endometrial hyperplasia–cancer sequence

simple hyperplasia specimens expressed the lowest meanvalues of expression (1.00 and 0.63, respectively) (Fig. 2),while the complex hyperplasia specimens and G2 specimensshowed the highest values of expression (2.33 and 2.71,respectively) (Figs. 3 and 4, Table 1).

Using the one-way ANOVA, we found a linear trend(P=0.005) of heparanase expression when comparing thenormal secretory endometrium and simple hyperplasia groupwith the complex hyperplasia+G1 carcinoma group and theG2+G3 carcinoma group (Fig. 1).

A comparison of endometrioid carcinomas (all grades) andcomplex hyperplasias yielded no difference among thegroups, with high mean expression values of 2.23 and 2.33,respectively.

3.2. Heparanase expression in the papillary serouspathway

Evaluation of atrophic and inactive endometrium comparedwith papillary serous carcinomas yielded no significantdifference between the groups, with high mean expressionvalues of 2.14 and 2.33, respectively (Table 1). In addition,

Figure 3 Immunohistochemical staining of heparanase, estro-gen receptor (ER), and progesterone (PR) (rows 2–4 respectively)in a panel of hyperplastic endometrium tissue samples (magni-fication ×20). Simple hyperplasia (column A); Complex hyper-plasia (Column B). H&E staining for tissue definition is shown forsamples as outlined above (first row) (magnification ×10). Notethat heparanase expression is absent in simple hyperplasia while ahigh level of heparanase is observed in complex hyperplasia tissue.

we found no significant difference between heparanaseexpression in papillary serous carcinomas and its expressionin G2+G3 endometrioid cancers, although both groupsexpressed high mean values (2.33 and 2.44, respectively).

3.3. Correlation between heparanase expression andER and PR expressions

ER and PR expressions were found among the various stagesin the endometrial pathway to cancer (Figs. 3 and 4, Table 1).A Spearman's correlation revealed no significant correlationbetween heparanase expression and ER expression (correla-tion coefficient=−0.232; P=0.16) nor any significant corre-lation between heparanase expression and PR expression(correlation coefficient=−0.198; P=0.23).

4. Discussion

The present study is the first to assess heparanase expressionin the premalignant stages preceding endometrial carcinoma.We observed significantly different heparanase expressionthroughout the various stages, andmost importantly, a patternof increasedexpression throughout the hyperplasia–carcinomacontinuum.

Previous studies have shown that increased heparanaseexpression is correlated to tumor aggressiveness and meta-static potential of many human tumors, including bladder [16],colon [9], pancreas [10], and breast [17]. Only a few studiesinvestigated heparanase's expression in premalignant lesionsand, more specifically, its role in the hyperplasia–dysplasiasequence. Similar to our results, increased levels of heparanasemRNA were found in prostate cancer tissues compared withbenign prostate hyperplasia (BPH) samples [18]. In coloncancer, heparanase expression gradually increased along thenormal to themild and severe dysplasia to themalignant stages[9]. One earlier study postulated a possible role for heparanasein mouse mammary tumorigenesis through induction of gener-alized mammary hyperplasia [19]. Heparanase mRNA and pro-tein expression were increased progressively in sequentialpremalignant lesions in a mouse model of multistage pancrea-tic islet carcinogenesis [20]. In contrast, Stadlmann et al. [21]found that heparanase expression was decreased in malignantprostatic tissue comparedwith BPH tissue.Using the same anti-heparanase polyclonal antibody used by us in order to studyheparanase expression in colorectal cancer, Doviner et al. [22]observed heparanase upregulation in the progression fromnormal epithelia to moderate dysplasia, although expressionwas almost nonexistent in carcinomatous tissue. They sug-gested that heparanase may play different roles throughoutthe carcinogenesis process and thus may have a preferentialexpression pattern during that process.

Previous studies have shownaprogression paradigmdescrib-ing the natural history of endometrial hyperplasia in a limitednumber of patients [3]. Complementarymolecular research hasindicated a sequential progression model of genetic mutationsinvolving PTEN and KRAS mutations early in the process, i.e.,simple hyperplasia, with β-catenin mutations and microsatel-lite instability presenting at a later stage, atypical hyperplasia,and finally E-cadherin alteration with p53 mutations appearingin already neoplastic tissue [23]. Considering that eventhe most neoplastically prone lesion (i.e., atypical complex

Figure 4 Immunohistochemical staining of heparanase, estrogen receptor (ER), and progesterone (PR) (rows 2–4 respectively) in a panelof endometrial cancer tissue samples (magnification ×20). G1 endometrial carcinoma (A); G2 endometrial carcinoma (B,C); Papillary serousendometrial cancer (D). H&E staining for tissue definition is shown for samples as outlined above (first row) (magnification ×10). Note theweak heparanase staining in the G1 specimen compared with the intense staining pattern in the G2 and papillary cancer specimens.

170 J. Canaani et al.

hyperplasia) will progress to frank carcinoma in only 29% ofcases, there remains the question as to which lesions willindeed undergo malignant transformation. Our results suggesta potential novel marker to address this issue.

Classic diagnostic gyneco-pathology relies on subjectiveassessment of architectural and cytologic alterations to stratifypatients into high risk and low risk groups according to theWHO1994 classification system for endometrial hyperplasia. Thereis, however, considerable lack of reproducibility when usingthat classification, particularly in the diagnosis of cytologicatypia and, as a consequence, this has presumably led to over-treatment in terms of performing hysterectomies. As a result,an alternative classification system, endometrial intraepithe-lial neoplasia (EIN), was introduced in 2000 and is based oncomputerized analysis of morphometric features that reflectglandular volume, architectural complexity, and cytologic(nuclear) abnormality. Recent studies suggest that it has higheraccuracy in the detection of premalignant lesions [24]. EIN,however, requires selection of representative areas of focallesions at the risk of overlooking early lesions that have not yettransformed morphologically but have already undergone pre-neoplastic stages. Accordingly, we propose that heparanaseexpression levels could be used as a novel diagnostic tool,complementary to EIN, in risk stratification of endometrialhyperplasia.

Since the putative precursor lesion of papillary serouscarcinoma (SPC) is endometrial intraepithelial carcinoma(EIC), which has been known to develop in atrophic or restingendometrium [4], we tried to assess the expression pattern ofheparanase in atrophic endometrium and serous papillarycarcinoma. Our results suggest that heparanase expression isupregulated in both atrophic endometrium and SPC, thussupporting a poor prognosis in SPC and possibly implying thatatrophic endometrium is, by definition, more prone to carci-nogenesis. In contrast, a recent study by Kodama et al. [12]showed heparanase expression to be absent. The differencesbetween our study and theirs might be accounted for by thesmall number of samples tested by both groups.

ER expression had been previously shown to be decreased inhigh-grade endometrial tumors compared with low-gradetumors [25]. Elkin et al. [13] suggested that estrogen mayinduce heparanase overexpression in breast cancer via stim-ulation of transcriptional activity of the heparanase promoter.Our results, however, did not show ER expression to be corre-lated with heparanase expression and thus they do not supportthis potential crosstalk between estrogen and heparanasein endometrial cancer. Estrogen mediates its biologicaleffects in target tissues by binding to specific intracellularreceptors, estrogen receptor-α (ER α) and ER β. Although theuterus is thought to predominantly express ER α, increased cell

171Heparanase expression increases throughout the endometrial hyperplasia–cancer sequence

proliferation and exaggerated response to estrogen in ER βknockoutmice indicates that ERβmightmodulate ERα functionin the uterus and have an antiproliferative function. Therefore,an imbalance in ERα and ERβ expression could be a crucial stepin estrogen-dependent tumorigenesis. Tamoxifen is the mostprescribed antineoplastic drugworldwide, and it is used to treatpatients with all stages of hormone-responsive breast cancer.Despite having a robust anti-estrogenic function in the breast,tamoxifen shows only partial estrogenic effects in other targettissues. These partial estrogenic actions have beneficial effectson bones and the cardiovascular system in postmenopausalwomen, but they are also associated with an increasedincidence of cancer in the endometrium. This may explain thedifferent correlation of heparanase and ERexpressions betweenthe breast and endometrium tissues.

Atypical hyperplasia and endometrial cancer are usuallytreated with hysterectomy, however 2%–14% of these patientsare under 40 years of age and wish to maintain fertility. Thesewomen are currently treated with progestin therapy and havegood response rates in atypical complex hyperplasia (83%–94%)andmoderate response rates in endometrial cancer (57%–75%).Nevertheless, there is a 13% recurrence rate in atypical hyper-plasia and an up to 50% recurrence rate in endometrial cancerpatients [26], indicating that additional therapeuticmodalitiesshould be considered. Consequently, heparanase inhibitiontherapy with agents such as PI-88might serve as future adjuncttherapy for endometrial cancer and hyperplasia.

In summary, the results of the present study showheparanase expression to be tightly regulated in endometrialtumorigenesis and indicate that heparanase may prove to bea novel marker and prognostic factor in the management ofpatients with endometrial hyperplasia and cancer. Futurestudies may demonstrate a therapeutic role for heparanaseusing targeted biologic therapy.

Acknowledgment

This work was performed in partial fulfillment of the M.D.thesis requirements of the Sackler Faculty ofMedicine, Tel-AvivUniversity (J.C.). The study was supported by the ResearchFund of the Department of Obstetrics and Gynecology at theTel-Aviv Sourasky Medical Center.

References

[1] 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.

[2] Bokhman JV. Two pathogenetic types of endometrial carcinoma.Gynecol Oncol 1983;15:10—7.

[3] Kurman RJ, Kaminski PF, Norris HJ. The behavior of endometrialhyperplasia. A long-term study of “untreated” hyperplasia in170 patients. Cancer 1985;56:403—12.

[4] Sherman ME. Theories of endometrial carcinogenesis: a multi-disciplinary approach. Mod Pathol 2000;13:295—308.

[5] Amant F, Moerman P, Neven P, Timmerman D, Van Limbergen E,Vergote I. Endometrial cancer. Lancet 2005;366:491—505.

[6] Sasisekharan R, Shriver Z, Venkataraman G, Narayanasami U.Roles of heparin-sulphate glycosaminoglycans in cancer. NatRev Cancer 2002;2:521—8.

[7] Vlodavsky I, Friedmann Y, Elkin M, Aingorn H, Atzmon R, Ishai-Michaeli R, et al. Mammalian heparanase: gene cloning, expres-sion and function in tumor progression and metastasis. Nat Med1999;5:793—802.

[8] Zcharia E, Metzger S, Chajek-Shaul T, Aingorn H, Elkin M,Friedmann Y, et al. Transgenic expression of mammalian hepa-ranase uncovers physiological functions of heparan sulfate intissue morphogenesis, vascularization, and feeding behavior.FASEB J 2004;18:252—63.

[9] Friedmann Y, Vlodavsky I, Aingorn H, Aviv A, Peretz T, Pecker I,et al. Expression of heparanase in normal, dysplastic, and neo-plastic human colonic mucosa and stroma. Evidence for its rolein colonic tumorigenesis. Am J Pathol 2000;157:1167—75.

[10] Koliopanos A, Friess H, Kleeff J, Shi X, Liao Q, Pecker I, et al.Heparanase expression in primary and metastatic pancreaticcancer. Cancer Res 2001;61:4655—9.

[11] Watanabe M, Aoki Y, Kase H, Tanaka K. Heparanase expressionand angiogenesis in endometrial cancer. Gynecol ObstetInvestig 2003;56:77—82.

[12] Kodama J, Kusumoto T, Seki N, Matsuo T, Ojima Y, Nakamura K,et al. Heparanase expression in both normal endometrium andendometrial cancer. Int J Gynecol Cancer 2006;16:1401—6.

[13] Elkin M, Cohen I, Zcharia E, Orgel A, Guatta-Rangini Z, Peretz T,et al. Regulation of heparanase gene expression by estrogen inbreast cancer. Cancer Res 2003;63:8821−6.

[14] Creasman WT, Odicino F, Maisonneuve P, Quinn MA, Beller U,Benedet JL, et al. Carcinoma of the corpus uteri. Int J GynecolObstet 2006;95(Suppl 1):S105—43.

[15] Zetser A, Levy-Adam F, Kaplan V, Gingis-Velitski S, Bashenko Y,Schubert S, et al. Processing and activation of latentheparanase occurs in lysosomes. J Cell Sci 2004;117:2249—58.

[16] Gohji K, Hirano H, Okamoto M, Kitazawa S, Toyoshima M, DongJ, et al. Expression of three extracellular matrix degradativeenzymes in bladder cancer. Int J Cancer 2001;95:295—301.

[17] Maxhimer JB,Quiros RM, Stewart R,Dowlatshahi K,Gattuso P, FanM, et al. Heparanase-1 expression is associated with the metas-tatic potential of breast cancer. Surgery 2002;132:326—33.

[18] OgishimaT, ShiinaH, Breault JE, Tabatabai L, BassettWW, EnokidaH, Li LC, et al. Increased heparanase expression is caused bypromoter hypomethylation and up-regulation of transcriptionalfactor early growth response-1 in human prostate cancer. ClinCancer Res 2005;11:1028—36.

[19] Zcharia E, Metzger S, Chajek-Shaul T, Friedmann Y, Pappo O, AvivA, et al. Molecular properties and involvement of heparanasein cancer progression and mammary gland morphogenesis.J Mammary Gland Biol Neoplasia 2001;6:311—22.

[20] Joyce JA, Freeman C, Meyer-Morse N, Parish CR, Hanahan D. Afunctional heparan sulfate mimetic implicates both heparanaseand heparan sulfate in tumor angiogenesis and invasion in amouse model of multistage cancer. Oncogene 2005;24:4037—51.

[21] Stadlmann S, Moser PL, Pollheimer J, Steiner P, Krugmann J,Dirnhofer S, et al. Heparanase-1 gene expression in normal,hyperplastic andneoplastic prostatic tissue. Eur J Cancer 2003;39:2229—33.

[22] Doviner V, Maly B, Kaplan V, Gingis-Velitski S, Ilan N, VlodavskyI, et al. Spatial and temporal heparanase expression in colonmucosa throughout the adenoma-carcinoma sequence. ModPathol 2006;19:878—88.

[23] Lax SF. Molecular genetic pathways in various types ofendometrial carcinoma: from a phenotypical to a molecular-based classification. Virchows Arch 2004;444:213—23.

[24] Baak JP, Orbo A, van Diest PJ, Jiwa M, de Bruin P, Broeckaert M,et al. Prospective multicenter evaluation of the morphometricD-score for prediction of the outcome of endometrial hyper-plasias. Am J Surg Pathol 2001;25:930—5.

[25] RiceLW, StoneRL,XuM,GalganoM, StolerMH,Everett EN, JazaeriAA. Biologic targets for therapeutic intervention in endometrioidendometrial adenocarcinoma and malignant mixed mülleriantumors. Am J Obstet Gynecol 2006;194:1119−26.

[26] Jadoul P, Donnez J. Conservative treatment may be beneficialfor young women with atypical endometrial hyperplasia orendometrial adenocarcinoma. Fertil Steril 2003;80:1315—24.