{"paper_id":"8acae30f-73f4-4337-a930-f020e11c2b9a","body_text":"Expression of platelet-derived growth factors\nand their receptors in ovarian clear-cell\ncarcinoma and its putative precursors\nSohei Yamamoto1, Hitoshi Tsuda 1, Masashi Takano 2, Tsunekazu Kita 2, Kazuya Kudoh 2,\nKenichi Furuya 2, Seiichi Tamai 3 and Osamu Matsubara 1\n1Department of Basic Pathology, National Defense Medical College, Tokorozawa, Saitama, Japan;\n2Department of Obstetrics and Gynecology, National Defense Medical College, Tokorozawa, Saitama, Japan\nand 3Department of Laboratory Medicine, National Defense Medical College, Tokorozawa, Saitama, Japan\nRecent studies have shown that platelet-derived growth factors and their receptors are frequently co-expressed\nin ovarian cancers. Herein, we investigated the role of the platelet-derived growth factor pathway in the\ndevelopment of ovarian clear-cell adenocarcinoma, a highly chemoresistant form of ovarian cancer.\nImmunohistochemical expression of platelet-derived growth factor receptor- a and receptor- b, platelet-derived\ngrowth factor A-chain and B-chain was examined in 31 cases of clear-cell adenocarcinoma and 56 coexisting\nputative precursor lesions: 17 non-atypical and 19 atypical endometrioses, and 10 non-atypical and 10 atypical\nclear-cell adenofibroma components. Twenty-one solitary endometrioses were also examined. Vascular\nendothelial cells were always positive for all the markers examined, and were used as positive controls. The\nfrequencies of positivity for platelet-derived growth factor receptor- a and receptor- b, and platelet-derived\ngrowth factor A-chain increased in accordance with higher cytologic atypia in the putative precursors: 71, 47,\nand 59% in the 17 non-atypical endometrioses, 84, 73, and 84% in the 19 atypical endometrioses, 0% each in the\n10 non-atypical clear-cell adenofibromas, 100, 90, and 90% in the 10 atypical clear-cell adenofibromas, and 97,\n97, and 100% in the 31 clear-cell adenocarcinomas, respectively. Positivity for platelet-derived growth factor\nB-chain increased in accordance with increased atypia in clear-cell adenofibroma: 0% in non-atypical clear-cell\nadenofibromas, 30% in atypical clear-cell adenofibromas, and 60% in coexisting carcinomas. However, in\ncontrast, positivity for platelet-derived growth factor B-chain decreased in accordance with increased atypia in\nendometriosis coexisting with clear-cell adenocarcinomas: 35% in non-atypical endometrioses, 11% in atypical\nendometrioses, and 5% in coexisting carcinomas. Platelet-derived growth factor receptor- a and receptor-b, and\ntheir ligands A-chain and B-chain were positive in 14, 29, 19, and 62% of the solitary endometrioses,\nrespectively. These results indicate activation of the platelet-derived growth factor pathway in ovarian clear-cell\nadenocarcinomas and suggest biological differences between carcinomas that arise in association with clear-\ncell adenofibroma vs endometriosis.\nModern Pathology (2008) 21, 115–124; doi:10.1038/modpathol.3800984; published online 14 December 2007\nKeywords: ovarian clear-cell adenocarcinoma; PDGFR; autocrine/paracrine; endometriosis; clear-cell adeno-\nfibroma\nAmong the ovarian cancers, clear-cell adenocarci-\nnoma has been recognized as a distinct clinico-\npathological entity because of its characteristic\nhistology, frequent concurrence with endometriotic\nlesions, and highly chemoresistant nature resulting\nin an extremely poor prognosis when in high stage.1–3\nSurgical debulking of the tumor is the only\neffective treatment for improving the prognosis of\nclear-cell adenocarcinoma. Although chemothera-\npeutic regimens including platinum analogues,\ntaxanes, etoposide, and camptothecin have been\ndeveloped, mortality of patients with ovarian\nclear-cell adenocarcinoma has remained largely\nunchanged. Therefore, more effective treatment\noptions are needed, and these should be based\non an understanding of the pathways driving\nneoplastic transformation and tumor growth. Some\nof these strategies may involve targeting of growth\nfactors and growth factor receptors, which are\nimportant for the growth and development of\nspecific cancer types.\nReceived 19 June 2007; revised and accepted 1 October 2007;\npublished online 14 December 2007\nCorrespondence: Dr H Tsuda, MD, Department of Basic Pathology,\nNational Defense Medical College, 3-2 Namiki, Tokorozawa,\nSaitama 359-8513, Japan.\nE-mail: htsuda@ndmc.ac.jp\nModern Pathology (2008) 21, 115–124\n& 2008 USCAP , Inc All rights reserved 0893-3952/08 $30.00\nwww.modernpathology .org\n\nDespite a lack of data on cause–effect relation-\nships, histological and epidemiological observations\nhave consistently demonstrated a close association\nbetween endometriosis and ovarian clear-cell ade-\nnocarcinoma.1,4,5 In previous studies, loss of hetero-\nzygosity on chromosomes 9q, 10q, and 11q was\ndetected in both clear-cell adenocarcinoma and\ncoexisting endometriotic lesions. 6,7 So far, however,\nthere have been few data on the specific molecular\nor genetic alterations that are consistently involved\nin clear-cell adenocarcinoma, and it is still unclear\nwhether a specific gene or chromosomal alteration\nis essential for clear-cell adenocarcinoma develop-\nment.\nOn the other hand, clear-cell adenofibroma, a\nmajor form of benign or borderline ovarian clear-cell\ntumor, may be another type of clear-cell adenocarci-\nnoma precursor.8,9 Our previous study demonstrated\nthat clear-cell adenofibroma components coexisted\nin 21% of surgically resected clear-cell adenocarci-\nnomas.10 These clear-cell adenofibroma components\ncoexisting with clear-cell adenocarcinoma often\ncontain both apparently benign clear-cell adenofi-\nbroma (clear-cell adenofibroma without atypia) and\nclear-cell adenofibroma with cellular and structural\natypia (clear-cell adenofibroma with atypia or\nborderline clear-cell adenofibroma). 8–10 Moreover,\nin comparison with clear-cell adenocarcinoma with-\nout clear-cell adenofibroma components, clear-cell\nadenocarcinoma with clear-cell adenofibroma com-\nponents shows several distinct clinicopathologic\ncharacteristics, that is, a lower frequency of co-\nexisting endometriosis, a higher frequency of\nhistologically low-grade tumor with tubulocystic\nproliferative architecture, and lower cancer cell\nproliferative activity. 10 These data suggest that,\nbesides endometriosis, clear-cell adenofibroma\ncomponents may be another form of clear-cell\nadenocarcinoma precursor, although there is no\nevident molecular background to support this spec-\nulation.\nThe platelet-derived growth factor receptors\n(PDGFRs: PDGFR- a and PDGFR- b) are transmem-\nbrane receptor tyrosine kinases that are activated\nby platelet-derived growth factors (PDGFs). 11 The\nPDGFs are dimeric proteins composed of two closely\nrelated A-chain and B-chain polypeptides encoded\nby separate genes. The PDGFs selectively bind to\nreceptor subunits via specific epitopes. PDGF-AA\nbinds PDGFR- a, whereas PDGF-AB and PDGF-BB\nrecognize both PDGFR-a and PDGFR-b.11 On binding\ntheir ligands, the receptors dimerize and autopho-\nsphorylate specific tyrosine residues, resulting in\nactivation of a variety of intracellular signaling\nmolecules that control differentiation and cell\nproliferation.12 Tumor formation may result if these\nreceptors are persistently activated through func-\ntional mutation or amplification of their genes, or if\nPDGFRs and PDGFs constitute autocrine/paracrine\nstimulating loops. Although such mutations and\ngene amplification have not been described pre-\nviously in ovarian cancers, recent studies have\ndemonstrated that PDGFR- a, PDGFR- b, and PDGF-\nAB are commonly expressed in ovarian cancers at\nfrequencies as high as 87, 81, and 67%, respec-\ntively.13–16 Moreover, Henriksen et al17 have reported\nthat PDGFRs or PDGFs were not detected in any\nof the benign ovarian tumors or normal ovarian\nepithelium they examined. Experiments in vivo\nhave shown that PDGFRs activated by PDGFs\nmodulate Akt and MAPK phosphorylation and\nsignificantly influence ovarian cancer cell prolifera-\ntion and tumor expansion. 16 These observations\nsuggest that the PDGF–PDGFR system may play a\nfunctional role in the progression of ovarian cancers\nthrough autocrine or paracrine activation within the\ntumor tissues.\nIn the present study, we histologically reviewed a\nnumber of surgically resected cases of ovarian clear-\ncell adenocarcinoma, and selected endometriotic\nlesions and clear-cell adenofibroma components\nsynchronous with clear-cell adenocarcinoma. Using\nimmunohistochemistry, we examined the expres-\nsion of PDGFRs and PDGFs to clarify whether (1)\nPDGFRs and PDGFs are commonly expressed in\nclear-cell adenocarcinomas, (2) whether such ex-\npression is already evident in the two putative\nprecursor forms of clear-cell adenocarcinoma, and\n(3) whether the expression status of these molecules\nis related to morphological changes in the epithe-\nlium of these putative precursors. It was anticipated\nthat this information would not only lead to better\nunderstanding of the development of ovarian clear-\ncell adenocarcinoma, but also provide insight into\npotentially promising treatment options for this\nhighly chemoresistant malignancy.\nMaterials and methods\nCases\nThis study was performed with the approval of the\nInstitutional Internal Review Board on ethical\nissues, and informed consent was obtained from\nall patients. Sixty-seven cases of primary ovarian\nclear-cell adenocarcinoma and 21 solitary endome-\ntriotic lesions were identified from the files of\nthe Department of Laboratory Medicine, National\nDefense Medical College Hospital, Japan. These 67\ncases of clear-cell adenocarcinoma had been surgi-\ncally resected between the years 1987 and 2005, and\nthe patients had not undergone chemotherapy or\nradiation therapy before surgery. Cases of solitary\nendometriosis were obtained by salpingo-oophor-\nectomy with or without hysterectomy. All speci-\nmens were formalin-fixed and paraffin-embedded,\nand 4-mm thick sections were prepared for hematox-\nylin and eosin staining. All pathology specimens\nwere reviewed in our institution, and tumors were\nclassified according to the criteria of the World\nHealth Organization. 8\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n116\nModern Pathology (2008) 21, 115–124\n\nEndometriosis Synchronous with Clear-Cell\nAdenocarcinoma\nEndometriosis synchronous with clear-cell adeno-\ncarcinoma was defined as (1) endometriosis existing\nin histological continuity with, or adjacent to, the\nclear-cell adenocarcinoma, or (2) an endometriotic\ncyst in which clear-cell adenocarcinoma was\nobserved but histological continuity between\nthe carcinoma and endometriotic epithelium was\nlacking.\nWith reference to the histological criteria of\n‘atypical endometriosis’ described previously, cyto-\nlogical atypia was determined as present if, at most,\none of the following features was histologically\nevident in the endometriotic epithelium: large\nhyperchromatic or pale nuclei with moderate to\nmarked pleomorphism; an increased nuclear to\ncytoplasmic ratio; cellular crowding with stratifica-\ntion or tufting (Figure 1). 5\nConsequently, synchronous endometrioses were\nidentified in 21 (31%) of the 67 clear-cell adeno-\ncarcinomas. Of these 21, 15 (71%) had both\nendometriotic lesions with and without cytologic\natypia, 4 (19%) had only endometriotic lesions with\natypia, and 2 (10%) had only endometriotic lesions\nwithout atypia. Therefore, 17 endometriotic lesions\nwithout atypia, 19 lesions with atypia, and 21 clear-\ncell adenocarcinomas containing synchronous\nendometriosis were analyzed by immunohisto-\nchemistry.\nClear-Cell Adenofibroma Components Synchronous\nwith Clear-Cell Adenocarcinoma\nThe histological criteria of clear-cell adenofibroma\n(with or without atypia) have been described\npreviously.10 In short, clear-cell adenofibroma was\na surface epithelial–stromal tumor containing tubu-\nlocystic epithelial components embedded in a\nfibroma-like stroma (Figure 2a). The epithelial cells\nwere polygonal, hobnail or flat in shape, with clear,\nslightly granular or eosinophilic cytoplasm (Figure\n2b). Presence of atypia was determined by histo-\nlogical documentation based on the cellular (ie,\nnuclear pleomorphism and stratification of the\nepithelium) and structural (ie, size irregularity and\ncrowding of the tubulocystic architecture) features\nof each epithelial component (Figure 2c and d).\nConsequently, of the 67 clear-cell adenocarcino-\nmas, we identified 10 (15%) cases that had both\ncomponents of clear-cell adenofibroma with and\nwithout atypia. The designation ‘clear-cell adeno-\nfibroma with atypia’ is synonymous with, or has\nbeen referred to historically as clear-cell borderline\ntumor.8,9 Therefore, 10 lesions of clear-cell adeno-\nfibroma without atypia, 10 lesions of clear-cell\nadenofibroma with atypia, and 10 clear-cell adeno-\ncarcinomas containing clear-cell adenofibroma com-\nponents were analyzed by immunohistochemistry.\nAmong the 21 solitary endometrioses, reactive\nchanges were identified in 6 (29%) cases on the\nbasis of the histological features described by La\nGrenade and Silverberg, 18 that is, the epithelia had\nmildly hyperchromatic or pleomorphic nuclei, but\nlacked stratification, and were associated with\nmoderate to severe subepithelial inflammation that\nis rarely seen in atypical endometriosis.\nPresence or absence of atypia in endometriotic\nlesions and clear-cell adenofibroma components,\nand of reactive changes in solitary endometrioses\nwere evaluated by two of the authors (SY and HT),\nand discussed until a consensus was reached. None\nof the 21 clear-cell adenocarcinomas with synchro-\nnous endometriosis and the 10 clear-cell adenocar-\ncinomas with clear-cell adenofibroma components\noverlapped.\nImmunohistochemistry\nAll selected formalin-fixed and paraffin-embedded\nspecimens were cut into 4- mm thick serial sections\nand analyzed by immunohistochemistry. We used\nrabbit polyclonal antibodies raised against a peptide\nmapped at the C terminus of PDGFR- a (RB-16981-\nR7; prediluted; Neomarkers, CA, USA), PDGFR- b\n(sc-339; dilution 1/100; Santa Cruz Biotechnology,\nSanta Cruz, CA, USA), and PDGF-A (sc-128; dilution\n1/100; Santa Cruz Biotechnology), and against\namino acids 136–190 of PDGF-B of human origin\n(sc-7878; dilution 1/100; Santa Cruz Biotechnology).\nSections were deparaffinized and boiled in a\nmicrowave oven at 97 1C for 20 min in 0.01 mol/l\ncitrate buffer (pH 6.0), then allowed to cool at room\ntemperature. Endogenous peroxidase was blocked\nusing 5% hydrogen peroxide. The slides were\nincubated at 4 1C overnight with primary antibodies\nand then reacted with a dextran polymer reagent\ncombined with secondary antibodies and peroxi-\ndase (DAKO, Glostrup, Denmark) for 1 h at room\ntemperature. Specific antigen–antibody reactions\nwere visualized with 0.2% diaminobenzidine\nFigure 1 A case of endometriosis synchronous with clear-cell\nadenocarcinoma, showing cytologic aytpia. Nuclear enlargement\nand moderate pleomorphism are evident. H&E staining, original\nmagnification /C2200.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n117\nModern Pathology (2008) 21, 115–124\n\ntetrahydrochloride and hydrogen peroxide, and\ncounterstaining was performed using Mayer’s he-\nmatoxylin. Fallopian tube epithelium or endothe-\nlium of vessels included in the sections served as a\nbuilt-in-control for all four antibodies, and as a\nreference for normal expression of the proteins\nanalyzed.19,20 As negative controls, sections without\nthe primary antibody were used.\nCytoplasmic immunoreactivity for PDGFs (PDGF-\nA and PDGF-B) and both membranous and cyto-\nplasmic immunoreactivity for PDGFRs (PDGFR- a\nand PDGFR- b) were taken into account for the\nevaluation. Immunoreactivity in the clear-cell ade-\nnocarcinomas was scored according to intensity\nas negative, weak (1 þ ), moderate (2 þ , similar to\nthe control intensity), or strong (3 þ ). If 50% of\nepithelial components showed an immunoreactive\nintensity equal to or higher than that of fallopian\ntube epithelium, that is, 2 þ or 3 þ , the cases (or\nlesions) were regarded as positive. T wo observers\n(SY and HT) evaluated the results of the immuno-\nhistochemistry.\nStatistical Analyses\nStatistical analyses were performed using StatMate\nIII software (ATMS, Tokyo, Japan). The frequency of\nimmunopositivity for PDGFRs/PDGFs in the various\ncomponents was compared using Fisher’s exact test\nor w2-test. Differences at Po0.05 were considered\nstatistically significant.\nResults\nResults of immunohistochemistry are summarized\nin Tables 1 and 2.\nPDGFR/PDGF Expression in Clear-Cell\nAdenocarcinomas\nPDGFR-a expression was judged as positive in 30\n(97%) of the 31 clear-cell adenocarcinomas: 10\n(32%) cases scored 2 þ and 20 (65%) scored 3 þ\n(Figure 3a). PDGFR- b expression was positive in 30\nFigure 2 Histological features of clear-cell adenofibroma component without atypia ( a and b) and with atypia ( c and d). (a) A clear-cell\nadenofibroma component shows simple cystic glands separated by wide fascicles of stromal cells. ( b) Glands are delineated with\nmonolayers of cuboidal, flattened, or in part hobnail glandular cells with clear cytoplasm and uniform nuclei. ( c) Glands are variously\nsized and shaped, and focally crowded. ( d) Glands are delineated with one to three layers of cells with clear to eosinophilic cytoplasm\nand mildly pleomorphic, hyperchromatic nuclei. H&E staining, original magnification /C2100 for ( a and c); /C2400 for ( b and d).\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n118\nModern Pathology (2008) 21, 115–124\n\ncases (97%): 20 (65%) cases scored 2 þ and 10\n(32%) scored 3 þ (Figure 3b). PDGF-A expression\nwas positive in all 31 cases: 21 (68%) cases scored\n2 þ and 10 (32%) scored 3 þ (Figure 3c). PDGFR- a,\nPDGFR-b, and PDGF-A were positive in 95, 95, and\n100% of the 21 clear-cell adenocarcinomas asso-\nciated with endometriosis, and were always (100%)\npositive in the 10 clear-cell adenocarcinomas with\nclear-cell adenofibroma components (Table 1).\nPDGF-B was determined as positive in only 7\n(23%) cases, all of which scored 2 þ (Figure 3d) and\nshowed co-positivity for PDGFR- b. Positivity for\nPDGF-B in clear-cell adenocarcinomas with clear-\ncell adenofibroma components (6 of 10, 60%) was\nsignificantly higher than that in clear-cell adenocar-\ncinomas associated with endometriosis (1 of 21, 5%)\n(P ¼ 0.0017) (Table 2).\nPDGFR/PDGF Expression in Endometriosis\nSynchronous with Clear-Cell Adenocarcinomas\nPDGFR-a and PDGFR- b were positive in 12 (71%)\nand 8 (47%) of 17 endometriotic epithelia without\ncytologic atypia, respectively (Figure 4a and b).\nPDGF-A and PDGF-B were positive in 10 (59%) and\n6 (35%) of the 17 endometriotic lesions without\natypia, respectively (Figure 4c and d), and all of\nthese showed co-positivity with PDGFR- a and\nPDGFR-b, respectively (Table 2).\nAmong the 19 lesions showing cytologic atypia in\nthe endometriotic epithelium, PDGFR- a and\nPDGFR-b were positive in 16 (84%) and 14 (73%)\ncases, respectively (Figure 4e and f). PDGF-A and\nPDGF-B were positive in 16 (84%) and 2 (11%) of\nthese 19 lesions (Figure 4g and h). Moreover, 15\n(94%) of 16 PDGF-A-positive lesions and 2 PDGF-B-\npositive lesions showed co-positivity with PDGFR- a\nand PDGFR- b, respectively. Although PDGFRs\n(PDGFR-a and PDGFR- b) and PDGF-A tended to be\nmore frequently positive in the endometriotic\nlesions with cytologic atypia than in those without,\nthe difference was not statistically significant.\nTable 1 Immunoreactivity for PDGFRs and PDGFs in 31 cases of\novarian clear-cell adenocarcinoma\nMolecules Number of cases (%)\nImmunoreaction\nNegative Positive\nScore 0 1+ 2+ 3+\nPDGFR-a 0 1 (3) 10 (32) 20 (65)\nPDGFR-b 0 1 (3) 20 (65) 10 (32)\nPDGF-A 0 0 21 (68) 10 (32)\nPDGF-B 9 (29) 15 (48) 7 (23) 0\nPDGF , platelet-derived growth factor; PDGFR, platelet-derived growth\nfactor receptor.\nTable 2 Immunoreactivity for PDGFRs and PDGFs in clear-cell adenocarcinoma and in its putative precursor lesions\nNumber of positive cases (%)\nImmunoreaction positive\nTotal a: PDGFR- a b: PDGFR- b c: PDGF-A d: PDGF-B Both\na and c\nBoth\nb and d\nI: Immunoreaction in the putative precursors and clear-cell adenocarcinomas\nEndometriosis synchronous with clear-cell adenocarcinoma\nEndometriosis without atypia 17 12 (71) 8 (47) 10 (59) 6 (35) 10 (59) 6 (35)\nEndometriosis with atypia 19 16 (84) 14 (73) * 16 (84) * 2 (11) * 15 (79) * 2 (11)\nClear-cell adenocarcinoma 21 20 (95) 20 (95) 21 (100) 1 (5) 20 (95) 1 (5)\nClear-cell adenofibroma synchronous with clear-cell adenocarcinoma\nClear-cell adenofibroma without atypia 10 0\n*\n0\n*\n0\n*\n0\n*\n0\n*\n0\n*Clear-cell adenofibroma with atypia 10 10 (100) 9 (90) 9 (90) 3 (30) 9 (90) 3 (30)\nClear-cell adenocarcinoma 10 10 (100) 10 (100) 10 (100) 6 (60) 10 (100) 6 (60)\nSolitary endometriosis 21 3 (14) 6 (29) 4 (19) 13 (62) 1 (5) 5 (24)\nEndometriosis without reactive change 15 1 (7) 3 (20) 3 (20) 7 (47) * 0 2 (13)\nEndometriosis with reactive change 6 2 (33) 3 (50) 1 (17) 6 (100) 1 (17) 3 (50)\nII: Comparison of the PDGFRs and PDGF expression between solitary endometriosis and endometrioses synchronous with clear-cell\nadenocarcinomas\nSolitary endometriosis 21 3 (14) * 6 (29) 4 (19) * 13 (62) 1 (5) * 5 (24)\nEndometriosis synchronous with clear-cell\nadenocarcinomasa\n17 12 (71) 8 (47) 10 (59) 6 (35) 10 (59) 6 (35)\nPDGF , platelet-derived growth factor; PDGFR, platelet-derived growth factor receptor.\n*Po0.05.\naEndometriosis without cytologic atypia.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n119\nModern Pathology (2008) 21, 115–124\n\nImmunopositivity for PDGFR- a, PDGFR- b, and\nPDGF-A in endometriotic lesions without atypia\nwas lower than that in clear-cell adenocarcinomas:\n71% (12 of 17) vs 95% (20 of 21), 47% (8 of 17) vs\n95% (20 of 21), and 59% (10 of 17) vs 100% (21 of\n21), respectively. In contrast, PDGF-B was expressed\nmore frequently in synchronous endometriotic\nlesions without atypia than in clear-cell adenocarci-\nnomas: 35% (6 of 17) vs 5% (1 of 21), respectively\n(Table 2I). Consequently, comparison between en-\ndometriotic lesions without atypia and coexisting\nclear-cell adenocarcinomas revealed significant dif-\nferences with regard to PDGFR- b (P ¼ 0.0011),\nPDGF-A ( P ¼ 0.0015), and PDGF-B ( P ¼ 0.022).\nPDGFR/PDGF Expression in the Clear-Cell\nAdenofibromas Synchronous with Clear-Cell\nAdenocarcinomas\nPDGFRs (PDGFR- a and PDGFR- b) and PDGFs\n(PDGF-A and PDGF-B) were negative in the epithe-\nlia of clear-cell adenofibroma components without\natypia (Figure 4i–l) (Table 2).\nOf the 10 clear-cell adenofibroma components\nwith aytpia, PDGFR- a and PDGFR-b were positive in\n10 (100%) and 9 (90%) lesions, respectively (Figure\n4m and n). PDGF-A and PDGF-B were positive in 9\n(90%) and 3 (30%) of the 10 clear-cell adenofibroma\ncomponents with atypia, respectively (Figure 4o and\np). Moreover, all 9 (100%) of the PDGF-A-positive\nlesions and all 3 (100%) of the PDGF-B-positive\nlesions showed co-positivity with PDGFR- a and\nPDGFR-b, respectively.\nPDGFR/PDGF Expression in Solitary Endometriotic\nEpithelium\nAmong the 21 cases with solitary endometriosis,\nPDGFR-a, PDGFR- b, PDGF-A, and PDGF-B were\npositive in 3 (14%), 6 (29%), 4 (19%), and 13\n(62%), respectively. One (25%) of the 4 PDGF-A-\npositive cases and 5 (42%) of the 13 PDGF-B-\nFigure 3 Expression of PDGFR- a, PDGFR-b, PDGF-A, and PDGF-B in ovarian clear-cell adenocarcinomas. Positive immunoreactions for\n(a) PDGFR- a (scored 3 þ ), ( b) PDGFR- b (scored 3 þ ), ( c) PDGF-A (scored 3 þ ), and ( d) PDGF-B (scored 2 þ ). Arrowheads indicate\nthe capillary endothelium showing moderate immunoreactivity with each primary antibody. Immunoperoxidase staining, original\nmagnification /C2400.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n120\nModern Pathology (2008) 21, 115–124\n\npositive cases showed co-positivity with PDGFR- a\nand PDGFR- b, respectively (Table 2).\nOf the six lesions showing reactive changes,\nPDGFR-a, PDGFR- b, PDGF-A, and PDGF-B were\npositive in two (33%), three (50%), one (17%), and\nsix (100%), respectively (Figure 5a and b) (Table 2I).\nTherefore, PDGFRs and PDGF-B were more fre-\nquently positive in solitary endometrioses showing\nreactive changes than in those without, and there\nwas a significant difference with regard to PDGF-B-\npositivity (47% (7 of 15) vs 100% (6 of 6), P ¼ 0.032)\n(Table 2I).\nWhen the solitary endometrioses and the endo-\nmetrioses synchronous with clear-cell adenocarci-\nnomas were compared, PDGFRs and PDGF-A were\nmore frequently positive in the latter, and significant\ndifferences were evident with regard to PDGFR- a\nand PDGF-A positivity (PDGFR- a, 14% (3 of 21) in\nthe former vs 71% (12 of 17) in the latter, Po0.001;\nPDGF-A, 19% (4 of 21) in the former vs 59% (10 of\n17) in the latter, P ¼ 0.014) (Table 2II). On the other\nhand, PDGF-B was more frequently positive in\nsolitary endometriosis than in endometriosis syn-\nchronous with clear-cell adenocarcinoma, but not to\na significant degree (62% (13 of 21) vs 35% (6 of 17),\nP ¼ 0.096) (Table 2II).\nDiscussion\nIn the present study, most (97%, 30 of 31) of the\novarian clear-cell adenocarcinomas analyzed were\npositive for either PDGFR- a or PDGFR- b, and all\nof the 30 PDGFR- a-positive carcinomas showed\nco-positivity for the PDGF-A chain, which is a\ncomponent of PDGF-AA or PDGF-AB. Although\nseveral reports have documented the frequent\nFigure 4 Expression of PDGFR- a, PDGFR-b, PDGF-A, and PDGF-B in putative precursor lesions for ovarian clear-cell adenocarcinomas.\nImmunohistochemistry for PDGFR-a (a, e, i, and m), PDGFR-b (b, f, j, and n), PDGF-A ( c, g, k, and o), and PDGF-B ( d, h, l, and p) in clear-\ncell adenocarcinoma-associated endometriosis without cytologic atypia ( a–d) and with cytologic atypia ( e–h), clear-cell adenocarcinoma-\nassociated clear-cell adenofibroma components without atypia ( i–l), and with atypia ( m–p). PDGFR-a, PDGFR-b, and PDGF-A are positive\nin the epithelial components of endometriosis without atypia ( a–c) and with atypia ( e, f–g), and in clear-cell adenofibroma with atypia\n(m–o). PDGF-B is positive in epithelial components of endometriosis without atypia ( d) and clear-cell adenofibroma with atypia ( p),\nbut negative in endometriosis with atypia ( h), and clear-cell adenofibroma without atypia ( l). Immunoperoxidase staining, original\nmagnification /C2200.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n121\nModern Pathology (2008) 21, 115–124\n\nexpression of PDGFRs in ovarian carcinomas, the\nnumber of clear-cell adenocarcinoma cases included\nin those studies was generally small. 13–15,17 The\nlargest series by Matei et al 16 demonstrated that 17\n(90%) and 18 (95%) of 19 clear-cell adenocarcino-\nmas were immunohistochemically positive for\nPDGFR-a and PDGF-AB, respectively, the former\nbeing almost concordant with our data.\nIn contrast, the frequency of positivity for the\nPDGF-B chain, which is a component of PDGF-AB\nor -BB, was 23% of clear-cell adenocarcinomas in\nthe present study, being much lower than the\nfrequency of PDGF-AB expression (95%) reported\nby Matei et al.16 This discrepancy may have resulted\nfrom the differences in the antibody used, protocols\nand evaluation of the immunohistochemistry.\nPDGFR-b, which is a receptor for PDGF-AB or\nPDGF-BB, was mostly (97%) positive in the present\nseries of clear-cell adenocarcinomas. However,\nPDGFs including PDGF-BB are natural ligands\npresent in stromal cells such as fibroblasts, and the\nsmooth muscle and endothelium of blood vessels,\nwhich constitute the major proportion of the stroma\nnecessary for tumor growth and invasion, and these\nwere constantly positive for PDGF-A and PDGF-B in\nthe present study. 19,20 These observations appear to\nsupport the hypothesis that not only an autocrine\nloop but also a paracrine loop of PDGFRs/PDGFs\ncontributes to tumor progression of clear-cell ade-\nnocarcinoma. Although not studied in the present\nseries, PDGF-D, which has been recently cloned,\nbinds PDGFR- b and shows mitogenic properties. 21\nThe expression of PDGF-D in clear-cell adenocarci-\nnoma would therefore be of interest.\nIn the endometriotic lesions, synchronous with\nclear-cell adenocarcinomas, the frequencies of\nPDGFR (PDGFR- a and PDGFR- b) and PDGF-A\nexpression increased in accordance with the acqui-\nsition of cytologic atypia by the epithelium, and a\nsimilar trend was also evident in solitary endome-\ntriotic lesions. Moreover, except for PDGF-B, ex-\npression of PDGFs and PDGFRs was more frequent\nin endometriotic lesions synchronous with clear-\ncell adenocarcinomas than in solitary endome-\ntrioses. These findings suggest that the establish-\nment of autocrine/paracrine loops of PDGFRs/\nPDGFs, especially those mediated by PDGFR- a,\nmight be an early event in the development of\nendometriosis-associated clear-cell adenocarcino-\nmas, and that activation of PDGFR has already\noccurred by the stage of solitary endometriosis\nwithout cellular atypia.\nIn contrast to the endometriotic lesions, PDGFs\nand PDGFRs were not detected in the epithelium of\napparently benign clear-cell adenofibroma compo-\nnents (clear-cell adenofibroma without atypia).\nHowever, similarly to endometriotic epithelium,\nexpression of PDGFs and PDGFRs was frequently\ndetected in clear-cell adenofibroma components\nwhen cytologic atypia was present. These findings\nsuggest that establishment of autocrine/paracrine\nloops for PDGFR activation is not essential for\nthe formation of clear-cell adenofibroma without\natypia, but is highly associated with the acquisi-\ntion of cytologic atypia by the epithelium of\nclear-cell adenofibroma and the development\nof clear-cell adenofibroma-associated clear-cell\nadenocarcinomas.\nOur previous study demonstrated that clear-cell\nadenocarcinomas with clear-cell adenofibroma com-\nponents were more frequently low-grade histologi-\ncally, showed a tubulocystic architectural pattern,\nand had lower cancer cell proliferation activity than\nclear-cell adenocarcinomas without clear-cell ade-\nnofibroma, which were mostly (68%) endometrio-\nsis-associated.10 In the present study, the incidence\nof PDGF-B positivity was higher in clear-cell\nadenofibroma-associated clear-cell adenocarcino-\nmas (60%) than in endometriosis-associated\nclear-cell adenocarcinomas (5%). All non-atypical\nclear-cell adenofibroma components studied were\nnegative for PDGF-B. In contrast, PDGF-B positivity\nwas relatively high in both solitary endometrioses\n(62%) and non-atypical endometrioses synchronous\nwith clear-cell adenocarcinoma (35%). Therefore, it\nFigure 5 A case of solitary endometriosis showing positive\nimmunoreactivity for ( a) PDGFR- b and ( b) PDGF-B. Immuno-\nperoxidase staining, original magnification /C2200.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n122\nModern Pathology (2008) 21, 115–124\n\ncan be speculated that the role of PDGFR-b and PDGF-\nBB (or PDGF-AB) expression may be associated with\ncell differentiation rather than cell proliferation or\nneoplastic transformation, although how the differ-\nence in PDGF-B expression status between these\nclear-cell adenocarcinomas contributes to the bio-\nlogical nature of the tumors is largely unknown.\nOvarian carcinoma is one of the human epithelial\nmalignancies showing histological heterogeneity,\nand clear-cell adenocarcinoma is one of its repre-\nsentative variations. 1 Therefore, the possibility that\nareas of apparently benign endometriosis-like or\nclear-cell adenofibroma-like lesions represent well\ndifferentiated or ‘mature’ components of the carci-\nnoma cannot be ruled out, although previous reports\nhave indicated that stromal features of clear-cell\nadenofibroma (fibroma-like) could be an aid for\ndifferentiating clear-cell adenofibroma from clear-\ncell adenocarcinoma with a prominent tubulocystic\npattern.9,10 In the present study, expression of PDGF-\nA and PDGFRs was detected less frequently in the\nendometriotic epithelium than in clear-cell adeno-\ncarcinomas, and was not detected in apparently\nbenign clear-cell adenofibroma components. These\nobservations suggest that the apparently benign\nareas coexisting with clear-cell adenocarcinomas,\nthat is, non-atypical endometriosis and clear-cell\nadenofibroma without atypia, may be true precur-\nsors of ovarian clear-cell adenocarcinomas. More-\nover, the striking differences in the frequency of\nexpression of PDGFRs/PDGFs between endometrio-\ntic lesions and non-atypical clear-cell adenofibro-\nmas support the idea that clear-cell adenofibroma\nand endometriosis are two distinct forms of pre-\ncursor for ovarian clear-cell adenocarcinoma.\nThe incidence of frequent high-intensity expres-\nsion of PDGFRs in clear-cell adenocarcinomas\ndemonstrated in this series appears to throw light\non the molecular background involved in clear-cell\nadenocarcinoma formation, and also provides new\ninsights into possible molecular-based therapies for\nthis highly chemoresistant malignancy. PDGFRs, as\nwell as KIT, belong to the type III receptor tyrosine\nkinases, and can be specifically treated by tyrosine\nkinase inhibitors such as imatinib mesylate\n(STI571).22 A recent study has demonstrated that\nimatinib mesylate inhibits ovarian cancer cell pro-\nliferation and PDGF-induced S-phase entry through\nPDGFR-a and Akt inactivation. 23 Although a phase II\ntrial of imatinib mesylate as a single agent failed to\nproduce satisfactory efficacy against recurrent and\nplatinum-resistant ovarian cancer, 24 cases with\nclear-cell adenocarcinoma were not enrolled and,\ntherefore, the effect of imatinib or other tyrosine\nkinase inhibitors on ovarian clear-cell adenocarci-\nnoma remains undetermined.\nIn the present study, we have demonstrated that\nPDGFRs and the PDGF-A chain are frequently\nexpressed in ovarian clear-cell adenocarcinomas.\nEpithelial–epithelial and epithelial–stromal inter-\nactions via autocrine and/or paracrine activation of\nPDGFRs might drive the development of clear-cell\nadenocarcinoma during the process of multistage\ncarcinogenesis. The present results also suggest\nbiological differences between clear-cell adenocar-\ncinomas that arise in association with clear-cell\nadenofibroma vs endometriosis.\nAcknowledgements\nThis work was supported in part by a grant-in-aid\nfor special research from National Defense Medical\nCollege, and by a grant-in-aid for cancer research\nfrom the Ministry of Health, Labor, and Welfare,\nJapan. We are thankful to Dr David Douglas\n(Douglas’ Scientific Editorial Services, Tokorozawa,\nJapan) for his professional review of the final\nversion of the manuscript.\nDisclosure/conflict of interest\nWe indicate no potential conflicts of interest.\nReferences\n1 Seidman JD, Russell P , Kurman RJ. Surface epithelial\ntumors of the ovary. In: Kurman RJ (ed). Blaustein’s\nPathology of the Female Genital Tract, 5th edn.\nSpringer-V erlag: New York, 2001, pp 791–904.\n2 Sugiyama T, Kamura T, Kigawa J, et al. Clinical\ncharacteristics of clear cell carcinoma of the ovary:\na distinct histologic type with poor prognosis and\nresistance to platinum-based chemotherapy. Cancer\n2000;88:2584–2589.\n3 Ikeda K, Sakai K, Yamamoto R, et al. Multivariate\nanalysis for prognostic significance of histologic sub-\ntype, GST-pi, MDR-1, and p53 in stages II-IV ovarian\ncancer. Int J Gynecol Cancer 2003;13:776–784.\n4 Vercellini P , Parazzini F , Bolis G, et al. Endometriosis\nand ovarian cancer. Am J Obstet Gynecol 1993;169:\n181–182.\n5 Fukunaga M, Nomura K, Ishikawa E, et al. Ovarian\natypical endometriosis: its close association with\nmalignant epithelial tumours. Histopathology 1997;30:\n249–255.\n6 Jiang X, Morland SJ, Hitchcock A, et al. Allelotyping of\nendometriosis with adjacent ovarian carcinoma reveals\nevidence of a common lineage. Cancer Res 1998;58:\n1707–1712.\n7 Sato N, Tsunoda H, Nishida M, et al. Loss of\nheterozygosity on 10q23.3 and mutation of the tumor\nsuppressor gene PTEN in benign endometrial cyst of\nthe ovary: possible sequence progression from benign\nendometrial cyst to endometrioid carcinoma and clear\ncell carcinoma of the ovary. Cancer Res 2000;60:\n7052–7056.\n8 Tavassoli FA, Devilee P , (eds). World Health Organiza-\ntion classification of tumours. Pathology and Genetics\nof Tumours of the Breast and Female Genital Organs.\nIARC Press: Lyon, 2003, pp 218–228.\n9 Bell DA, Scully RE. Benign and borderline clear\ncell adenofibromas of the ovary. Cancer 1985;56:\n2922–2931.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n123\nModern Pathology (2008) 21, 115–124\n\n10 Yamamoto S, Tsuda H, Yoshikawa T, et al. Clear cell\nadenocarcinoma associated with clear cell adenofibro-\nmatous components: a subgroup of ovarian clear cell\nadenocarcinoma with distinct clinicopathological\ncharacteristics. Am J Surg Pathol 2007;31:999–1006.\n11 Hart CE, Forstrom JW, Kelly JD, et al. T wo classes of\nPDGF receptor recognize different isoforms of PDGF .\nScience 1988;240:1529–1531.\n12 Heldin CH, Ostman A, Ronnstrand L. Signal trans-\nduction via platelet-derived growth factor receptors.\nBiochim Biophys Acta 1998;1378:F79–F113.\n13 Schmandt RE, Broaddus R, Lu KH, et al. Expression of\nc-ABL, c-KIT, and platelet-derived growth factor\nreceptor-beta in ovarian serous carcinoma and\nnormal ovarian surface epithelium. Cancer 2003;98:\n758–764.\n14 Lassus H, Sihto H, Leminen A, et al. Genetic altera-\ntions and protein expression of KIT and PDGFRA in\nserous ovarian carcinoma. Br J Cancer 2004;91:\n2048–2055.\n15 Wilczynski SP , Chen YY, Chen W, et al. Expression and\nmutational analysis of tyrosine kinase receptors c-kit,\nPDGFRalpha, and PDGFRbeta in ovarian cancers. Hum\nPathol 2005;36:242–249.\n16 Matei D, Emerson RE, Lai YC, et al. Autocrine\nactivation of PDGFRalpha promotes the progression\nof ovarian cancer. Oncogene 2006;25:2060–2069.\n17 Henriksen R, Funa K, Wilander E, et al. Expression and\nprognostic significance of platelet-derived growth\nfactor and its receptors in epithelial ovarian neo-\nplasms. Cancer Res 1993;53:4550–4554.\n18 La Grenade A, Silverberg SG. Ovarian tumors asso-\nciated with atypical endometriosis. Hum Pathol\n1988;19:1080–1084.\n19 Siegbahn A, Hammacher A, Westermark B, et al.\nDifferential effects of the various isoforms of platelet-\nderived growth factor on chemotaxis of fibroblasts,\nmonocytes, and granulocytes. J Clin Invest 1990;85:\n916–920.\n20 Tallquist MD, Soriano P , Klinghoffer RA. Growth factor\nsignaling pathways in vascular development. Onco-\ngene 1999;18:7917–7932.\n21 Lokker NA, Sullivan CM, Hollenbach SJ, et al. Platelet-\nderived growth factor (PDGF) autocrine signaling\nregulates survival and mitogenic pathways in glioblas-\ntoma cells: evidence that the novel PDGF-C and PDGF-\nD ligands may play a role in the development of brain\ntumors. Cancer Res 2002;62:3729–3735.\n22 Heinrich MC, Blanke CD, Druker BJ, et al. Inhibition\nof KIT tyrosine kinase activity: a novel molecular\napproach to the treatment of KIT-positive malignan-\ncies. J Clin Oncol 2002;20:1692–1703.\n23 Matei D, Chang DD, Jeng MH. Imatinib mesylate\n(Gleevec) inhibits ovarian cancer cell growth through\na mechanism dependent on platelet-derived growth\nfactor receptor alpha and Akt inactivation. Clin Cancer\nRes 2004;10:681–690.\n24 Coleman RL, Broaddus RR, Bodurka DC, et al. Phase II\ntrial of imatinib mesylate in patients with recurrent\nplatinum- and taxane-resistant epithelial ovarian\nand primary peritoneal cancers. Gynecol Oncol 2006;\n101:126–131.\nPDGF/PDGFR in ovarian clear-cell adenocarcinomas\nS Yamamoto et al\n124\nModern Pathology (2008) 21, 115–124","source_license":"public-domain-us","license_restricted":false}