{"paper_id":"8880f49b-3b93-4254-bfa9-cab57ec60d85","body_text":"Accumulative copy number increase of MET\ndrives tumor development and histological\nprogression in a subset of ovarian clear-cell\nadenocarcinomas\nSohei Yamamoto1, Hitoshi Tsuda 2, Kosuke Miyai 1, Masashi Takano 3, Seiichi Tamai 4\nand Osamu Matsubara 1\n1Department of Basic Pathology, National Defense Medical College, Tokorozawa, Japan; 2Pathology and\nClinical Laboratory Division, National Cancer Center Hospital, Chuo-ku, Tokyo, Japan; 3Department of\nObstetrics and Gynecology, National Defense Medical College, Tokorozawa, Japan and 4Department of\nClinical Laboratory, National Defense Medical College Hospital, Tokorozawa, Japan\nOur previous study demonstrated that, among ovarian carcinomas, amplification of the MET gene and\noverexpression of MET specifically and commonly occur in clear-cell adenocarcinoma histology. This study\nwas conducted to address how these alterations contribute to development and progression of this highly\nchemoresistant form of ovarian cancer. We histologically reviewed 21 previously described MET amplification-\npositive clear-cell adenocarcinoma cases, and selected 11 tumors with synchronous endometriosis and\n2 tumors with adjacent clear-cell adenofibroma (CCAF) components. Using double in situ hybridization and\nimmunohistochemistry, copy number alterations of the MET gene and levels of MET protein expression were\nanalyzed in these putative precursor lesions and the corresponding invasive carcinoma components in this\nselected cohort. All of the non-atypical precursor lesions analyzed (ie, non-atypical endometrioses and the\nbenign CCAFs) were negative for MET gain. However, low-level ( Z3 MET copies in Z10% and Z4 MET copies in\n10–40% of tumor cells) gain of MET was detected in 4 (40%) of the 10 atypical endometrioses and 1 of the\n2 borderline CCAFs. Moreover, high-level ( Z4 MET copies in Z40% of tumor cells) gain of MET were detected in\nfive (50%) of the atypical endometrioses. In 4 (31%) of the 13 cases enrolled, intratumoral heterogeneity for MET\ngain was documented in invasive carcinoma components, wherein all the relatively differentiated carcinoma\ncomponents showed low-level gain of MET and all the corresponding poorly differentiated carcinomas showed\nhigh-level gain. The overall incidence of MET overexpression gradually increased from the precursors of\nnon-atypical form (0%), through those of atypical form (67%) and the relatively differentiated carcinoma\ncomponents (92%), to the poorly differentiated carcinoma components (100%). These results suggest\nthat accumulative MET gene copy number alterations causing MET overexpression are associated with\nhigher tumor grade and might drive the development and progression of the MET amplification-positive ovarian\nclear-cell adenocarcinoma.\nModern Pathology (2012) 25, 122–130; doi:10.1038/modpathol.2011.143; published online 7 October 2011\nKeywords:\nclear cell adenocarcinoma; c-Met; gene amplification; in situ hybridization; MET; ovary; progression\nPerhaps, the most important characteristic of any\ntumor is the combination of genetic alterations that\nunderlie its development and drive its progression.\nThe search for new target genes for anticancer\ntherapies has become a discrete field of translational\nresearch. MET, a unique member of the receptor\ntyrosine kinase family, has attracted much attention\nin the recent years, representing an intriguing target\nfor cancer therapy, although this has not yet been\nestablished in a clinical setting. The MET proto-\noncogene, located on chromosome 7q31, encodes\nthe MET kinase that is composed of three functional\ndomains, including the ligand-binding domain,\nReceived 2 May 2011; revised 26 July 2011; accepted 3 August\n2011; published online 7 October 2011\nCorrespondence: Dr H Tsuda, MD, Pathology and Clinical\nLaboratory Division, National Cancer Center Hospital, 5-1-1\nTsukiji, Chuo-ku, Tokyo 104-0045, Japan.\nE-mail: hstsuda@ncc.go.jp\nModern Pathology (2012) 25, 122–130\n122 & 2012 USCAP, Inc. All rights reserved 0893-3952/12 $32.00\nwww.modernpathology.org\n\nregulatory juxtamembrane domain, and the receptor\ntyrosine kinase domain. 1 Physiologically, when its\nligand, namely, hepatocyte growth factor (HGF),\nbinds to it, the MET receptor undergoes dimeriza-\ntion and autophosphorylation at specific tyrosine\nresidues within the cytoplasmic domain, creating\ndocking sites for intracellular signal transducers that\nactivate the Ras-mitogen-activated protein kinase\npathway, and the phosphatidylinositol 3-kinase-\nAKT-mTOR signaling pathway, as well as several\nother pathways.\n1,2\nIn contrast to the non-neoplastic cells, in the tumor\ncells, MET can be activated in a ligand-independent\nmanner through activating mutation, amplification,\nand overexpression of the MET gene.\n1–3 However, to\ndate, it is thought that somatic mutation of MET is a\nrare event in the sporadic primary carcinomas of\nadults, including ovarian carcinomas,\n4,5 with papil-\nlary carcinoma of the kidney being an exception. 6\nOn the other hand, MET amplification was identified\nin 5–10% of gastric cancers, 7–9 4% of esophageal\ncancers,10 3–4% of lung cancers, 11,12 and 10% of\ncolorectal cancers.13\nOur recent study identified that MET overexpres-\nsion and gene amplification were commonly dete-\ncted in ovarian clear-cell adenocarcinoma, with\ntheir frequencies at 22 and 24%, respectively.\n14 On\nthe other hand, such alterations were extremely rare\nin the non-clear cell histological subtypes of ovarian\ncarcinoma (ie, serous, endometrioid, and mucinous\nadenocarcinomas).\n14 In that study, MET overexpres-\nsion was associated with a worse prognosis of the\nclear-cell adenocarcinoma patients, and MET gene\namplification was correlated with MET overexpres-\nsion and poorly differentiated histology of tumors. 14\nThese findings suggested that among ovarian carci-\nnomas, MET proto-oncogene is specifically involved\nin clear-cell adenocarcinoma by means of gene\namplification. However, it remains unclear how\nthese alterations contribute to tumor development\nand progression of this carcinoma type.\nClear-cell adenocarcinoma has been recognized to\nbe a highly chemo-resistant form of ovarian cancer.\n15–19\nAlthough little is known about the molecular genetic\nalterations underlying tumor development, a hypoth-\nesis of multi-step tumorigenesis of ovarian clear-cell\nadenocarcinoma, starting with histologically benign-\nappearing precursor lesions (ie, endometriosis and\nbenign clear-cell adenofibroma (CCAF)), progressing to\ntheir atypical counterparts (ie, atypical endometriosis\nand borderline CCAF), and ultimately to clear-\ncell adenocarcinoma has been proposed and widely\nconsidered.\n20–26\nIn this study, to unmask the timing of MET\nalterations in the development and progression of\novarian clear-cell adenocarcinoma, we selected 13\nclear-cell adenocarcinoma cases that were double\nin situ hybridization-positive cases, as determined\nby our previous series,\n14 and were in a group\nof carcinomas with adjacent putative precursor\nlesions. Using double in situ hybridization and\nimmunohistochemistry, the enrolled cases were\nanalyzed for MET gene copy number alterations\nand MET protein expression in terms of their intra-\ntumoral heterogeneity. Specifically, MET alterations\nwere examined in the benign-appearing putative\nprecursor lesions (ie, non-atypical endometriosis\nand benign CCAF), their atypical forms (atypical\nendometriosis and borderline CCAF), and the corre-\nsponding invasive carcinoma components. More-\nover, the invasive carcinoma components were\ndivided into the two histological categories\n(ie, carcinomas of poorly differentiated histology\nvs those of relatively differentiated histology), and\nthe heterogeneity regarding MET alterations in a\ncarcinoma was also assessed. It was anticipated that\nthis information would not only lead to a better\nunderstanding of the development of ovarian clear-\ncell adenocarcinoma, but also provide insight into\nthe potential treatment options for this highly\nchemo-resistant malignancy.\nMaterials and methods\nPatients and Tissue Samples\nHematoxylin- and eosin-stained sections from 21\nclear-cell adenocarcinoma cases, in which our pre-\nvious double in situ hybridization analysis revealed\nhigh-level polysomy or gene amplification of MET in\nthe invasive carcinoma components,\n14 were histolo-\ngically reviewed. From these 21 cases, based on the\nhistopathological criteria described previously, 21–24\n11 tumors with synchronous endometriotic lesions\n(endometriosis-associated carcinomas) and 2 tumors\nwith adjacent CCAF components (adenofibroma-\nassociated carcinomas) were identified. These 13\ncases were retrieved from the files of the Department\nof Clinical Laboratory, National Defense Medical\nCollege Hospital, Japan; all patients had undergone\nsurgical resection between 1987 and 2006, none had\nundergone chemotherapy or radiation therapy before\nsurgery, and all specimens were formalin-fixed and\nparaffin-embedded tissue sections. Clinical staging\nof disease was done according to the International\nFederation of Gynecology and Obstetrics system.\nOf the 11 cases with endometriosis-associated\ncarcinoma, 8 (73%) were stage I, 1 (9%) was stage\nII, 1 (9%) was stage III, and 1 (9%) was stage IV .\nOf the two cases with adenofibroma-associated\ncarcinoma, one was stage I, and another was stage\nII. The research protocol was approved by the ethics\ncommittee of the National Defense Medical College,\nTokorozawa, Japan.\nOn the basis of previously described histological\ncriteria for ‘atypical endometriosis’,\n21–23 endometriotic\nlesions adjacent to clear-cell adenocarcinoma were\nclassified into non-atypical and atypical forms. Of the\n11 endometriosis-associated cases, 9 had both\nnon-atypical and atypical endometriosis, 1 had non-\natypical endometriosis only, and 1 had atypical endo-\nmetriosis only. Therefore, 10 lesions of non-atypical\nMET in ovarian clear cell carcinogenesis\nS Yamamoto et al 123\nModern Pathology (2012) 25, 122–130\n\nendometriosis, 10 lesions of atypical endometriosis,\nand 11 clear-cell adenocarcinomas containing synchro-\nnous endometriosis were analyzed by double in situ\nhybridization and immunohistochemistry.\nBoth the adenofibroma-associated carcinomas had\ncomponents of benign CCAF and borderline CCAF .\nThe histopathological criteria for ‘benign-’ and ‘border-\nline-’ CCAF were described in the previous report.24\nPoorly Differentiated Histology in Clear-Cell\nAdenocarcinoma\nTo further assess the intra-tumoral heterogeneity of\nMET alterations in invasive carcinoma components,\nwe established criteria for evaluating poorly differ-\nentiated carcinoma histology, and divided the carci-\nnoma components into two groups: poorly\ndifferentiated histology and relatively differentiated\nhistology. Details of the criteria for poorly differen-\ntiated histology of clear-cell adenocarcinoma were\ndescribed in our previous series.\n14 Briefly, when\ntumor cells grew as solid masses, cords, or individual\ntumor cells, infiltrating towards the surrounding\nstromal tissue without easily discernible glandular\ndifferentiation, these were defined as exhibiting\npoorly differentiated histology. On the other hand,\ntumor cells whose growth was noted to be typically\ntubulocystic, papillary, or a combination of these\n(tubulopapillary), were defined as having a relatively\ndifferentiated histology of clear-cell adenocarcinoma.\nOf the 13 cases enrolled, 10 cases had both the\nrelatively differentiated and the poorly differentiated\nhistological components, and other 3 had the rela-\ntively differentiated histological component only.\nTherefore, 13 lesions of the relatively differentiated\nclear-cell adenocarcinoma component and 10 lesions\nof the poorly differentiated clear-cell adenocarcinoma\ncomponent were analyzed by double in situ hybridi-\nzation and immunohistochemistry.\nBright-Field Double In Situ Hybridization\nA BenchMark XT automated slide processing system\n(V entana Medical Systems, Tucson, AZ, USA) was\nused for the optimization of the double in situ\nhybridization assay for copy number alteration of\nMET. Technical details for this assay were described\nin the previous reports.\n14,27 Briefly, for MET detec-\ntion, the INFORM MET DNA Probe (Ventana), a\ndinitrophenyl (DNP)-labeled probe, was applied to\nthe tissue sections, denatured at 95 1C, and hybri-\ndized at 52 1C for 6 h. After washing, the tissue\nsections were incubated with monoclonal rabbit\nanti-DNP antibody (Ventana) for 20 min, and then\nwith horseradish peroxidise-conjugated goat anti-\nrabbit antibody for 16 min at 37 1C. The metallic\nsilver deposit for MET ISH signal was developed\nusing the ultraView SISH Detection Kit (Ventana).\nFor CEN7 detection, the INFORM Chromosome 7\nProbe (V entana), a DNP-labelled oligoprobe, was\napplied to the tissue sections, denatured at 95 1C,\nand hybridized at 44 1C for 2 h. Then, after washing,\nthe tissue sections were incubated with monoclonal\nrabbit anti-DNP antibody for 20 min, and then with an\nalkaline phosphatase-conjugated goat anti-rabbit anti-\nbody for 12 min at 37 1C. The signal for CEN7 was\nvisualized with a Fast-Red and naphthol phosphate\nreaction, using an ultraView Red ISH Detection Kit.\nFinally, hematoxylin counterstaining was performed.\nImmunohistochemistry\nFor MET protein detection, a BenchMark XT\nautomated slide processing system was also used,\nas described previously.\n14 The primary antibody\nused was a rabbit monoclonal antibody against\nthe carboxyl region of the transmembrane human\nc-Met protein (CONFIRM anti-Total c-MET (SP44),\nready for use; Ventana). The immunoreaction was\nvisualized using an ultraView DAB Detection Kit\n(V entana) according to the manufacturer’s instruc-\ntions.\n27 Endothelial cells in the tumor tissues served\nas positive controls, and sections without the\nprimary antibody were used as negative controls.\nInterpretation of the Data\nFor double in situ hybridization analysis, the\nnumbers of dark brown and red dot signals,\ncorresponding to the copies of MET and those of\nCEN7, respectively, were counted in 80 inter-phase\ntumor cell nuclei by using a /C2 100 oil immersion\nobjective lens. All cases were arranged in three\ncategories as follows: no gain ( r2 MET copies in\n490% of cells), low-level gain ( Z3 MET copies in\nZ10% of cells and Z4 copies of MET in o40% of\ncells), and high-level gain (status of high polysomy\n(Z4 copies of MET in Z40% of cells) or gene\namplification (presence of tight gene clusters,\na MET/CEN7 ratio per cell of Z2, or Z15 copies of\nthe MET gene in Z10% of cells examined)), by using\nthe previously described criteria.\n14,28\nThe intensity of the immunoreaction was scored\nusing a four-tier system with the same criteria used for\nassessing HER2/neu immunoreactions in breast can-\ncer (the HercepTest criteria) as follows:29 negative, no\ndiscernible staining or background type staining; 1þ ,\ndefinite cytoplasmic staining and/or equivocal dis-\ncontinuous membrane staining; 2 þ , unequivocal\nmembrane staining with mild to moderate intensity;\n3 þ , strong and complete membrane staining. MET\noverexpression was defined as moderate (2 þ )o r\nstrong (3 þ ), when complete membrane staining was\nobserved in at least 10% of the cells of interest.\nStatistical Analysis\nStatistical analyses were performed using StatMate\nIII software (ATMS, Tokyo, Japan). Comparisons\nMET in ovarian clear cell carcinogenesis\n124 S Yamamoto et al\nModern Pathology (2012) 25, 122–130\n\nbetween parameters were computed by the w2-test.\nDifferences at Po0.05 were considered statistically\nsignificant.\nResults\nThe results of double in situ hybridization and\nimmunohistochemistry for detection of MET altera-\ntions are given in Table 1.\nCopy Number Alterations of the MET Gene and\nOverexpression of MET Frequently Occur in the\nPutative Precursor Lesions, Especially in their\nAtypical Forms\nAll the 10 non-atypical endometrioses examined\nwere found to exhibit no gain of MET by double\nin situ hybridization assay, and all showed weak\nimmunoreactions for MET, with scores of 1 þ\n(Figure 1). Of the 10 atypical endometrioses,\n1 (10%), 4 (40%), and 5 (50%) lesions were defined\nas exhibiting no gain, low-level gain, and high-level\ngain of MET, respectively (Figure 1). All the atypical\nendometriosis showing high-level gain of the MET\ngene and three of the four atypical endometrioses\nshowing low-level gain were defined as exhibiting\nMET overexpression by immunohistochemistry\n(Figure 1). The remaining one atypical endometrio-\nsis showed no gain of MET, with a weak (score 1 þ )\nimmunoreaction for MET.\nBoth of benign CCAFs showed no gain ofMET and a\nweak (score 1 þ ) immunoreaction for MET (Figure 2).\nOne borderline CCAF showed low-level gain of MET\n(Figure 2), and another borderline CCAF showed no\ngain. Weak immunoreaction for MET was noted in\nthese two borderline CCAF components, but MET\noverexpression was not observed (Figure 2).\nHeterogeneity of MET alterations in the invasive\ncarcinoma components is common (Table 1).\nOf the five endometriosis-associated cases where-\nin adjacent atypical endometrioses harbored a\nhigh-level gain of MET (case numbers 2, 4, 8, 9,\nand 10), all the corresponding carcinoma compo-\nnents examined, regardless of the extent of tumor\ndifferentiation (relatively differentiated vs poorly\ndifferentiated components), also showed high-level\ngain of MET and overexpression of MET. In case\nnumber 9, although both the relatively differentiated\nand poorly differentiated carcinoma components\nwere defined as exhibiting a high-level gain, only\nthe latter component showed true amplification of\nMET (Figure 3).\nOf the five tumors in which the adjacent pre-\ncursors (ie, atypical endometriosis and borderline\nCCAF) showed low-level gain of MET (case numbers\n1, 3, 5, 6, and 13), intra-tumoral heterogeneity of\nthe copy number alterations of the MET gene was\ndetected in two carcinomas (case numbers 5 and 6).\nIn these two cases, carcinoma components of\nrelatively differentiated histology were defined\nas exhibiting low-level gain of MET, whereas those\nof poorly differentiated histology exhibited high-\nlevel gain.\nOf the other three tumors in which double in situ\nhybridization analysis did not reveal copy number\nalterations of MET in the adjacent precursor lesions\n(case numbers 7, 11, and 12), intra-tumoral hetero-\ngeneity of the copy number alterations of MET was\ndetected in two carcinomas (case numbers 7 and 11).\nTable 1 MET alterations in ovarian clear-cell adenocarcinomas and their co-existing putative precursor lesions, as detected by\nimmunohistochemistry and double in situ hybridization\nCase no.\n(adjacent\nprecursor lesion)\nClinical stage\nof disease a\nPrecursor lesion\nwithout atypia\nPrecursor lesion\nwith atypia\nCarcinoma\nRelatively\ndifferentiated\nPoorly\ndifferentiated\nIHC DISH IHC DISH IHC DISH IHC DISH\n1 (Endometriosis) Ic 1+ No gain 3+ Low gain 3+ High gain — —\n2 (Endometriosis) Ic 1+ No gain 2+ High gain 3+ High gain 2+ High gain\n3 (Endometriosis) Ia 1+ No gain 1+ Low gain 2+ High gain — —\n4 (Endometriosis) IV 1+ No gain 2+ High gain 2+ High gain 2+ High gain\n5 (Endometriosis) IIIc 1+ No gain 2+ Low gain 2+ Low gain 3+ High gain\n6 (Endometriosis) Ic 1+ No gain 2+ Low gain 3+ Low gain 3+ High gain\n7 (Endometriosis) Ia 1+ No gain 1+ No gain 2+ Low gain 3+ Amplified\n8 (Endometriosis) Ia 1+ No gain 2+ High gain 2+ High gain 3+ High gain\n9 (Endometriosis) Ia 1+ No gain 2+ High gain 3+ High gain 3+ Amplified\n10 (Endometriosis) Ic — — 2+ High gain 2+ High gain — —\n11 (Endometriosis) IIc 1+ No gain — — 1+ Low gain 2+ High gain\n12 (CCAF) Ic 1+ No gain 1+ No gain 2+ High gain 2+ High gain\n13 (CCAF) IIc 1+ No gain 1+ Low gain 2+ High gain 3+ High gain\nAbbreviations: CCAF, clear-cell adenofibroma; DISH, double in situ hybridization; IHC, immunohistochemistry.\naClinical stages of disease defined by International Federation of Gynecology and Obstetrics.\nMET in ovarian clear cell carcinogenesis\nS Yamamoto et al 125\nModern Pathology (2012) 25, 122–130\n\nIn these two cases, carcinoma components of\nrelatively differentiated histology were defined as\nexhibiting low-level gain of MET, whereas those of\npoorly differentiated histology exhibited high-level\ngain. In case number 7, the poorly differentiated\ncarcinoma component showed true amplification of\nthe MET gene. Both the relatively differentiated and\npoorly differentiated carcinoma components in the\nremaining one case (case number 12) showed a high-\nlevel gain of the MET gene.\nThe overall incidence of high-level gain of the\nMET gene was found to gradually increase from the\nprecursors of non-atypical form (0%), through those\nof atypical form (42%) and the relatively differen-\ntiated carcinoma components (69%), to the poorly\ndifferentiated carcinoma components (100%). Con-\nsequently, statistical analyses revealed significant\ndifferences in the frequency between the non-\natypical and atypical precursors ( P ¼ 0.019), and\nbetween the atypical precursors and poorly differ-\nentiated carcinoma components ( P ¼ 0.0046). More-\nover, the overall incidence of MET overexpression\nwas also observed to gradually increase from the\nprecursors of non-atypical form (0%), through those\nof atypical form (67%) and the relatively differen-\ntiated carcinoma components (92%), to the poorly\ndifferentiated carcinoma components (100%). Con-\nsequently, statistical analyses revealed significant\ndifferences in the frequency between the non-\natypical and atypical precursors ( P ¼ 0.0007) or\ninvasive carcinoma components ( Po0.0001).\nDiscussion\nThe main findings of our investigation can be\nsummarized as follows: (1) MET overexpression\nand copy number alterations were not detected in\nthe non-atypical forms of precursor lesions, such\nas non-atypical endometriosis and benign CCAF;\n(2) these alterations were frequently detected in the\natypical forms of precursors (ie, atypical endome-\ntriosis and borderline CCAF) in the MET amplifica-\ntion-positive ovarian clear-cell adenocarcinoma, and\nsome of these lesions already harbor a high-level\ngain of MET; (3) intra-tumoral heterogeneity of MET\nalterations in invasive carcinoma components is\ncommon; and (4) rates of the MET alterations were\nFigure 1 MET alterations in non-atypical endometriosis ( a–c) and atypical endometriosis ( d–f) synchronous with ovarian clear-cell\nadenocarcinoma. ( a) Representative histological features of non-atypical endometriosis. The endometriotic epithelia lack evident\ncytological or structural atypia. ( b) Weak but definite cytoplasmic immunoreaction is noted in the endometriotic epithelium. According\nto the described criteria, this lesion was classified as weak (score 1 þ ) for MET immunoreaction. ( c) Most of the epithelial cells in this\nphotomicrographs show one to two dark brown (corresponding to the MET gene) and red (corresponding to the centromeric DNA region\non chromosome 7 (CEN7)) signals. According to the described criteria, this lesion was defined as exhibiting no gain of the MET gene.\n(d) Representative histological features of an atypical endometriosis. Endometriotic epithelia show moderate degrees of cytological atypia\nand cellular stratification. ( e) Continuous membrane immunoreaction with moderate intensity is noted in this atypical endometriosis.\nAccording to the described criteria, this lesion was classified as exhibiting MET overexpression (score 2 þ ). ( f) Atypical endometriosis\nclassified as showing high-level gain of the MET gene. About half of the epithelial cells in this photomicrograph show four pairs of MET\nand CEN7 signals. ( a and d) HE stain, original magnification /C2 400 for ( a) and /C2 200 for ( d). (b and e) Immunoperoxidase stain, original\nmagnification /C2 400 for both. ( c and f) Double in situ hybridization assays.\nMET in ovarian clear cell carcinogenesis\n126 S Yamamoto et al\nModern Pathology (2012) 25, 122–130\n\nhigher in the poorly differentiated carcinoma histol-\nogy than in the relatively differentiated histology of\novarian clear-cell adenocarcinomas. These findings\nsuggest that MET alterations occur as an early event\nin carcinogenesis of the MET amplification-positive\nclear-cell adenocarcinoma of ovary, and that, these\nalterations might drive the development and\nprogression in a subset of ovarian clear-cell adeno-\ncarcinoma. This is the first report to demonstrate\nthe evidence of MET copy number abnormalities\nand MET overexpression in the putative precursor\nlesions of ovarian clear-cell adenocarcinoma.\nAlthough our analysis failed to reveal MET over-\nexpression in the non-atypical precursor lesions, all\nthese lesions showed weak levels of MET-associated\nimmunoreactions. Small amounts of MET protein\nexpression were detected in the normal Mullerian\nepithelium (ie, endocervical glands, endometrium,\nand ovarian surface epithelium). 30 Moreover, the\npathophysiology is still elusive. It has been gener-\nally believed that HGF retains a multifunctional\nrole (ie, functions in mitogenesis, cell migration,\nangiogenesis, and morphogenesis) in the pathogen-\nesis of the pelvic solitary endometriosis, in combi-\nnation with peritoneal macrophages and ovarian\nsteroids.31 Therefore, it could be suggested that in\n(non-atypical) endometrioses and benign CCAFs,\nlow-level expression of MET (ie, expression level\nscored as 1 þ in the present study), in combination\nwith the locally secreted HGF , might be related\nto cellular differentiation, for example, to the\nMullerian epithelial differentiation, and not be\ndirectly associated with cancer development and\nprogression.\nIn our previous report, although a strong correla-\ntion between MET overexpression and double in situ\nhybridization positivity was statistically supported,\nthere were some discrepancies between the results\nof immunohistochemistry and double in situ hybri-\ndization analyses. However, in the present study, all\ncarcinoma components and atypical precursor\nlesions harboring a high-level gain of MET demon-\nstrated MET overexpression. These differences in\nthe results obtained were probably because of the\ndifferent sample types used for analyses; whole-\nsection samples were used in the present series,\nwhereas two tissue microarray cores per case were\nused in the previous report, and the latter might\nonly represent small parts of the carcinoma. In\naddition, six (67%) of the nine lesions with low-\nlevel gain of MET showed MET overexpression, and\nnone of the lesions with no gain of MET showed\nMET overexpression. Therefore, it is highly likely\nthat MET gene amplification is indeed an important\nFigure 2 MET alterations in the benign clear-cell adenofibroma (CCAF) (a–c), and the borderline CCAF ( d–f) adjacent to ovarian clear-cell\nadenocarcinoma. (a) A case of benign CCAF , lacking cytological or structural atypia. ( b) Weak but definite cytoplasmic immunoreaction\nis noted in the epithelial component. This lesion was assigned a score of 1 þ for MET immunoreaction. ( c) The epithelial cells of CCAF\nshowing one to two pairs of MET (dark brown) and CEN7 (red) signals. This lesion was defined as exhibiting no gain of the MET gene.\n(d) A case of borderline CCAF . Epithelial cells show a moderate degree of cytological atypia. ( e) Weak (score 1 þ ) immunoreactions were\nnoted in the epithelial components. ( f) The epithelial cells of borderline CCAF showing three to four pairs of MET and CEN7 signals.\nAccording to the described criteria, this component was defined as exhibiting a low-level gain of MET.( a and d) Hematoxylin and eosin\n(HE) stain, original magnification /C2 400 for both. ( b and e) Immunoperoxidase stain, original magnification /C2 400 for both. ( c and f)\nDouble in situ hybridization assays.\nMET in ovarian clear cell carcinogenesis\nS Yamamoto et al 127\nModern Pathology (2012) 25, 122–130\n\nmechanism for MET overexpression in ovarian clear-\ncell carcinogenesis.\nWhat are the potential implications of these\nfindings with regard to treatment of ovarian clear-\ncell adenocarcinoma with MET inhibitors? In recent\nyears, the evidence accumulated suggests that MET\nis an exciting and novel drug target for the treatment\nof MET-overexpressing ovarian cancers, because of\nthe success observed in vitro and in vivo.32,33 Various\nparameters have been suggested to be the predictors\nof the response to MET kinase inhibitors, including\nstrong expression, as seen, for example, in lung\ncancers, and gene amplification, as seen in lung and\ngastric cancers. 12,34–37 Taken together with our pre-\nsent data, MET appears to be an important therapeu-\ntic target in the treatment of ovarian clear-cell\nadenocarcinomas with MET alterations, justifying\nthe exploration of anti-MET treatment strategies.\nMoreover, as HGF is the only known ligand for\nthe MET receptor, neutralization of HGF using\nribozymes38 or antagonistic fragments such as\nNK4,39 and neutralizing antibodies 40 may also be\nthe potentially attractive compounds for inhibiting\nHGF/MET signaling.\nIdentification of the molecular mechanisms respon-\nsible for aggressive tumor behavior is also important\nfor the development of potential new treatment\nstrategies. Although, until recently, no histological\nfeatures predictive of the clinical outcome in ovarian\nclear-cell adenocarcinoma patients have been agreed\nupon, we recently reported that clear-cell adeno-\ncarcinomas with poorly differentiated histology,\nFigure 3 Intra-tumoral heterogeneity of MET gain in an ovarian clear-cell adenocarcinoma. ( a) Representative histology of a relatively\ndifferentiated carcinoma component in case number 9. Carcinoma cells growing infiltrating tubular architectures composed of medium-\nsize and uniformly rounded glands of tumor cells. Some glands of tumor cells are combined with small papillary architectures,\nexhibiting a ‘tubulopapillary’ feature. ( b) In this microphotograph, the carcinoma cells of the relatively differentiated histology show\nseven or four pairs of MET (dark brown) and CEP7 (red) signals. This lesion was defined as exhibiting a high gain of MET, but not true\namplification. (c) A poorly differentiated histological component in case number. Tumor cells growing by solid sheets of carcinoma cells,\nwithout distinct glandular formations. ( d) True gene amplification in tumor cells of poorly differentiated histology in case number. In the\nfocused tumor cell, MET signals form gene clusters, and the ratio of MET:CEN7 per cell is clearly more than 2. ( a and c) Hematoxylin and\neosin (HE) stain, original magnification /C2 100 for ( a) and /C2 200 for ( c). ( b and d) Double in situ hybridization assays.\nMET in ovarian clear cell carcinogenesis\n128 S Yamamoto et al\nModern Pathology (2012) 25, 122–130\n\naccounting for 33% of the cases studied, had\nsignificantly worse outcomes than those with well\n(not poorly)-differentiated histology, both in early-\nstage and advanced-stage diseases. 41 Moreover, the\npresence of poorly differentiated histology was highly\nassociated with resistance of the tumor to post-\noperative platinum-based chemotherapy. 41 These\nfindings suggested that tumors with poorly differ-\nentiated histology can be regarded as a high-grade\nsubtype of ovarian clear-cell adenocarcinoma. In the\ncurrent study, intratumoral heterogeneity for MET\nalterations were found to be common in invasive\ncarcinoma components, wherein MET alterations\nwere usually more evident in the poorly differen-\ntiated histology than in the relatively differentiated\ncarcinoma components, suggesting that MET altera-\ntions may be associated with the histological progres-\nsion of MET amplification-positive ovarian clear-cell\nadenocarcinoma. Therefore, MET gains may provide\na carcinoma component with evolutionary advan-\ntages, favoring its transformation into higher grade\nand aggressive subtypes of this carcinoma type.\nConsidering the overall data available, it can be\nstated that MET should be examined as a potential\ntarget for treatment of ovarian clear-cell adenocarci-\nnoma, especially those of the aggressive, high-grade,\nand chemo-resistant subgroup.\nIn conclusion, our data suggest that copy number\nalterations of MET and MET overexpression are\ncritical steps in the early development of MET-\namplified and high-grade ovarian clear-cell adeno-\ncarcinomas, and that MET amplification, in parti-\ncular, may prove to be an excellent biomarker of\nhistological progression of this carcinoma. These\nresults contribute to the understanding of the\npathogenesis of clear-cell adenocarcinoma and sup-\nport the development of targeted therapies that\ninhibit MET activation.\nAcknowledgements\nThis work was supported in part by a grant-in-aid\nfor cancer research from the Ministry of Health,\nLabor, and Welfare, Japan (HT), and by a grant from\nthe Foundation for Promotion of Cancer Research\n(SY and HT). We are grateful to Eiko Munechika,\nMT, Roche Diagnostics, Tokyo, Japan, for technical\nassistance.\nDisclosure/conflict of interest\nThe authors declare no conflict of interest.\nReferences\n1 Ma PC, Maulik G, Christensen J, et al. c-Met: structure,\nfunctions and potential for therapeutic inhibition.\nCancer Metastasis Rev 2003;22:309–325.\n2 Peruzzi B, Bottaro DP . Targeting the c-Met signaling\npathway in cancer. Clin Cancer Res 2006;12:3657–3660.\n3 Birchmeier C, Birchmeier W, Gherardi E, et al. Met,\nmetastasis, motility and more. Nat Rev Mol Cell Biol\n2003;4:915–925.\n4 Lorenzato A, Olivero M, Patane ` S, et al. Novel somatic\nmutations of the MET oncogene in human carcinoma\nmetastases activating cell motility and invasion.\nCancer Res 2002;62:7025–7030.\n5 Tanyi J, Tory K, Rigo ´ Jr J, et al. Evaluation of the\ntyrosine kinase domain of the Met proto-oncogene in\nsporadic ovarian carcinomas. Pathol Oncol Res 1999;\n5:187–191.\n6 Schmidt L, Junker K, Nakaigawa N, et al. Novel\nmutations of the MET proto-oncogene in papillary\nrenal carcinomas. Oncogene 1999;18:2343–2350.\n7 Kuniyasu H, Yasui W, Kitadai Y, et al. Frequent\namplification of the c-met gene in scirrhous type\nstomach cancer. Biochem Biophys Res Commun 1992;\n189:227–232.\n8 Tsujimoto H, Sugihara H, Hagiwara A, et al. Amplifi-\ncation of growth factor receptor genes and DNA ploidy\npattern in the progression of gastric cancer. Virchows\nArch 1997;431:383–389.\n9 Hara T, Ooi A, Kobayashi M, et al. Amplification of\nc-myc, K-sam, and c-met in gastric cancers: detection\nby fluorescence in situ hybridization. Lab Invest\n1998;78:1143–1153.\n10 Miller CT, Lin L, Casper AM, et al. Genomic amplifica-\ntion of MET with boundaries within fragile site FRA7G\nand upregulation of MET pathways in esophageal\nadenocarcinoma. Oncogene 2006;25:409–418.\n11 Zhao X, Weir BA, LaFramboise T, et al. Homozygous\ndeletions and chromosome amplifications in human\nlung carcinomas revealed by single nucleotide polymor-\nphism array analysis. Cancer Res 2005;65:5561–5570.\n12 Bean J, Brennan C, Shih JY, et al. MET amplification\noccurs with or without T790 M mutations in EGFR\nmutant lung tumors with acquired resistance to\ngefitinib or erlotinib. Proc Natl Acad Sci USA 2007;\n104:20932–20937.\n13 Di Renzo MF , Olivero M, Giacomini A, et al. Over-\nexpression and amplification of the met/HGF receptor\ngene during the progression of colorectal cancer. Clin\nCancer Res 1995;1:147–154.\n14 Yamamoto S, Tsuda H, Miyai K, et al. Gene amplifica-\ntion and protein overexpression of MET are common\nevents in ovarian clear-cell adenocarcinoma: Their\nroles in tumor progression and prognostication of the\npatient. Mod Pathol 2011;24:1146–1155.\n15 Lee KR, Tavassoli FA, Part J, et al . Surface epithelial-\nstromal tumours. In: Tavassoli FA, Devilee P (eds).\nWorld Health Organization Classification of Tumours.\nPathology and Genetics of Tumours of the Breast and\nFemale Genital Organs. IARC Press: Lyon, 2003, pp\n117–145.\n16 Chan JK, Teoh D, Hu JM, et al. Do clear cell ovarian\ncarcinomas have poorer prognosis compared to other\nepithelial cell types? A study of 1411 clear cell ovarian\ncancers. Gynecol Oncol 2008;109:370–376.\n17 Mizuno M, Kikkawa F , Shibata K, et al. Long-term\nfollow-up and prognostic factor analysis in clear cell\nadenocarcinoma of the ovary. J Surg Oncol 2006;\n94:138–143.\n18 O’Brien ME, Schofield JB, T an S, et al. Clear cell\nepithelial ovarian cancer (mesonephroid): bad prognosis\nonly in early stages. Gynecol Oncol 1993;49:250–254.\nMET in ovarian clear cell carcinogenesis\nS Yamamoto et al 129\nModern Pathology (2012) 25, 122–130\n\n19 Takano M, Kikuchi Y, Yaegashi N, et al. Clear cell\ncarcinoma of the ovary: a retrospective multicentre\nexperience of 254 patients with complete surgical\nstaging. Br J Cancer 2006;94:1369–1374.\n20 Prowse AH, Manek S, Varma R, et al. Molecular\ngenetic evidence that endometriosis is a precursor of\novarian cancer. Int J Cancer 2006;119:556–562.\n21 Fukunaga M, Nomura K, Ishikawa E, et al. Ovarian\natypical endometriosis: its close association with\nmalignant epithelial tumours. Histopathology 1997;\n30:249–255.\n22 Yamamoto S, Tsuda H, Takano M, et al. Expression of\nplatelet-derived growth factors and their receptors in\novarian clear-cell carcinoma and its putative precur-\nsors. Mod Pathol 2008;21:115–124.\n23 Yamamoto S, Tsuda H, Miyai K, et al. Cumulative\nalterations of p27-related cell-cycle regulators in the\ndevelopment of endometriosis-associated ovarian clear\ncell adenocarcinoma. Histopathology 2010;56:740–749.\n24 Yamamoto S, Tsuda H, Takano M, et al. Clear-cell\nadenofibroma can be a clonal precursor for clear-cell\nadenocarcinoma of the ovary: a possible alternative\novarian clear-cell carcinogenic pathway. J Pathol\n2008;216:103–110.\n25 Kurman RJ, Shih IeM. The origin and pathogenesis of\nepithelial ovarian cancer: a proposed unifying theory.\nAm J Surg Pathol 2010;34:433–443.\n26 Tan DS, Kaye S. Ovarian clear cell adenocarcinoma: a\ncontinuing enigma. J Clin Pathol 2007;60:355–360.\n27 Nitta H, Hauss-Wegrzyniak B, Lehrkamp M, et al.\nDevelopment of automated brightfield double in situ\nhybridization (BDISH) application for HER2 gene and\nchromosome 17 centromere (CEN 17) for breast\ncarcinomas and an assay performance comparison to\nmanual dual color HER2 fluorescence in situ hybridi-\nzation (FISH). Diagn Pathol 2008;3:41.\n28 Cappuzzo F , Hirsch FR, Rossi E, et al. Epidermal\ngrowth factor receptor gene and protein and gefitinib\nsensitivity in non-small-cell lung cancer. J Natl Cancer\nInst 2005;97:643–655.\n29 Pauletti G, Dandekar S, Rong H, et al. Assessment of\nmethods for tissue-based detection of the HER-2/neu\nalteration in human breast cancer: a direct comparison\nof fluorescence in situ hybridization and immunohis-\ntochemistry. J Clin Oncol 2000;18:3651–3664.\n30 Huntsman D, Resau JH, Klineberg E, et al. Comparison\nof c-met expression in ovarian epithelial tumors and\nnormal epithelia of the female reproductive tract by\nquantitative laser scan microscopy. Am J Pathol\n1999;155:343–348.\n31 Khan KN, Kitajima M, Hiraki K, et al. Immunopatho-\ngenesis of pelvic endometriosis: role of hepatocyte\ngrowth factor, macrophages and ovarian steroids. Am J\nReprod Immunol 2008;60:383–404.\n32 Sawada K, Radjabi AR, Shinomiya N, et al. c-Met\noverexpression is a prognostic factor in ovarian cancer\nand an effective target for inhibition of peritoneal\ndissemination and invasion. Cancer Res 2007;67:\n1670–1679.\n33 Zillhardt M, Christensen JG, Lengyel E. An orally\navailable small-molecule inhibitor of c-Met, PF-\n2341066, reduces tumor burden and metastasis in a\npreclinical model of ovarian cancer metastasis. Neo-\nplasia 2010;12:1–10.\n34 Ma PC, Jagadeeswaran R, Jagadeesh S, et al. Functional\nexpression and mutations of c-Met and its therapeutic\ninhibition with SU11274 and small interfering RNA\nin non-small cell lung cancer. Cancer Res 2005;\n65:1479–1488.\n35 Ma PC, Tretiakova MS, MacKinnon AC, et al. Expres-\nsion and mutational analysis of MET in human\nsolid cancers. Genes Chromosomes Cancer 2008;47:\n1025–1037.\n36 Engelman JA, Zejnullahu K, Mitsudomi T, et al. MET\namplification leads to gefitinib resistance in lung\ncancer by activating ERBB3 signaling. Science 2007;\n316:1039–1043.\n37 Smolen GA, Sordella R, Muir B, et al. Amplification of\nMET may identify a subset of cancers with extreme\nsensitivity to the selective tyrosine kinase inhibitor\nPHA-665752. Proc Natl Acad Sci USA 2006;103:\n2316–2321.\n38 Abounader R, Ranganathan S, Lal B, et al.\nReversion of\nhuman glioblastoma malignancy by U1 small nuclear\nRNA/ribozyme targeting of scatter factor/hepatocyte\ngrowth factor and c-met expression. J Natl Cancer Inst\n1999;91:1548–1556.\n39 Matsumoto K, Nakamura T. NK4 (HGF-antagonist/\nangiogenesis inhibitor) in cancer biology and thera-\npeutics. Cancer Sci 2003;94:321–327.\n40 Jun HT, Sun J, Rex K, et al. AMG 102, a fully human\nanti-hepatocyte growth factor/scatter factor neutraliz-\ning antibody, enhances the efficacy of temozolomide or\ndocetaxel in U-87 MG cells and xenografts. Clin Cancer\nRes 2007;13:6735–6742.\n41 Yamamoto S, Tsuda H, Shimazaki H, et al. Clear cell\nadenocarcinoma with a component of poorly differ-\nentiated histology: a poor prognostic subgroup of\novarian clear cell adenocarcinoma. Int J Gynecol\nPathol 2011;30:431–441.\nMET in ovarian clear cell carcinogenesis\n130 S Yamamoto et al\nModern Pathology (2012) 25, 122–130","source_license":"public-domain-us","license_restricted":false}