IGFBP-1 is expressed specifically in ovarian clear cell adenocarcinoma

Transcription

1 Histopathology 2011, 58, DOI: /j x IGFBP-1 is expressed specifically in ovarian clear cell adenocarcinoma Shintaro Sugita, 1,2 Yukio Morishita, 1,2 Junko Kano, 1 Shuichiro Furuya, 2 Aya Shiba-Ishii 1 & Masayuki Noguchi 1,2 1 Department of Pathology, Institute of Basic Medical Science, Graduate School of Comprehensive Human Science, University of Tsukuba, and 2 Department of Pathology, Tsukuba University Hospital, Amakubo, Tsukuba, Ibaraki, Japan Date of submission 21 April 2010 Accepted for publication 12 August 2010 Sugita S, Morishita Y, Kano J, Furuya S, Shiba-Ishii A & Noguchi M (2011) Histopathology 58, IGFBP-1 is expressed specifically in ovarian clear cell adenocarcinoma Aims: To investigate the specific expression of insulinlike growth factor binding protein-1 (IGFBP-1) in ovarian clear cell adenocarcinoma (CCA). Methods and results: Immunohistochemistry and in situ hybridization for IGFBP-1 were performed in normal endometrium, placenta, and 100 surgically resected cases of ovarian cancer including 31 CCAs and 69 non- CCAs. Immunohistochemistry for hepatocyte nuclear factor-1 (HNF-1b) was also examined in all cases. Specific expression of IGFBP-1 was confirmed in secretory endometrium, decidua of placenta and Arias-Stella glands of miscarriage material. Among ovarian cancers, almost all cases of CCA showed expression of both IGFBP-1 protein and mrna, but non-cca hardly expressed IGFBP-1. There was a significant difference between CCA and non-cca in the expression of IGFBP- 1 protein and mrna. No correlation was found between the rate of IGFBP-1 expression and pathological T and N factors of the tumour node metastasis (TNM) classification. All CCA cases except for one exhibited expression of HNF-1b protein, whereas only 15.9% of non-ccas did so. Conclusion: The expression of IGFBP-1 in CCA is more specific than that of HNF-1b. IGFBP-1 shows expression by decidual endometrium and Arias-Stella glands, and CCA also exhibits characteristic expression. These results indicate that IGFBP-1 is a immunohistochemical marker for CCA. Keywords: clear cell adenocarcinoma, HNF-1b, IGFBP-1, immunohistochemistry, in situ hybridization Abbreviations: CCA, clear cell adenocarcinoma; EA, endometrioid adenocarcinoma; HNF-1b, hepatocyte nuclear factor-1beta; IGFBP-1, insulin-like growth factor binding protein-1; MA, mucinous adenucarcinoma; SA, serous adenocarcinoma Introduction Address for correspondence: M Noguchi, MD, Department of Pathology, Institute of Basic Medical Science, Graduate School of Comprehensive Human Science, University of Tsukuba, Tennodai, Tsukuba, Ibaraki , Japan. nmasayuk@md.tsukuba.ac.jp Clear cell adenocarcinoma (CCA) has the worst prognosis among the various subtypes of ovarian surface epithelial cancers, and is characterized histologically by glandular, papillary and solid proliferation of a large amount of glycogen-containing clear and hobnail tumour cells with a hyalinous fibrotic stroma. Middleaged and elderly women are affected predominantly, and the pathogenesis of CCA is associated closely with ovarian endometriosis. Using the suppression subtractive hybridization method, we have demonstrated previously that the gene expression of CS0DI066YJ10, as a full-length placental cdna clone, is reduced in cases of endometrial stromal sarcoma in comparison with stromal cells of endometrium in the normal secretory phase. 1 Homology retrieval revealed that CS0DI066YJ10 is identical to insulin-like growth factor binding protein 1 (IGFBP-1), which is one of the key members of the insulin-like growth factor (IGF) system. IGFBP regulates the biological effects of IGFs by binding to and Ó 2011 Blackwell Publishing Limited.

2 730 S Sugita et al. then being released from them at various sites in human organs. 2 IGFBP includes six members, IGFBP- 1 6, of which IGFBP-1 plays a particularly important role in gynaecological physiology. 3 IGFBP-1 closely regulates the menstrual cycle, decidualization of the endometrium, implantation, trophoblast invasion and fetal growth. 4 Strong and specific expression of IGFBP- 1 is observed ordinarily in the secretory endometrium, decidualized endometrial stromal cells and the decidua of the placenta. The expression of IGFBP-1 in such tissues has already been confirmed by immunohistochemistry and mrna in situ hybridization. 1 While the family of molecules of the IGF system are associated significantly with development of the organism and maintenance of a normal physiological state, they also play a role in tumour development in several organs including the lung, intestine, breast, prostate and gynaecological organs. 5,6 In the gynaecological field, they are involved in the development and progression of ovarian and uterine endometrial carcinomas. 7 Some studies have demonstrated a relationship between IGFBP-2,5 and the development and grade of ovarian epithelial cancers, including serous adenocarcinoma (SA), endometrioid adenocarcinoma (EA) and CCA. 8,9 However, no reported study has focused in detail upon the connection between IGFBP-1 and ovarian CCA. Hepatocyte nuclear factor-1 (HNF-1b) was recognized recently as a specific marker for ovarian CCA. 10 HNF-1b is one of the isoforms of HNF-1, which is a transcription factor expressed mainly in liver and kidney, where it regulates the promoters of several genes. On the other hand, HNF-1 has been reported to activate transcription of human IGFBP-1, since a binding site for HNF-1 is located upstream of the promoter region of IGFBP-1. These characteristics suggest that IGFBP-1 located downstream of HNF-1 would also be a good marker of CCA. 11,12 In the present study, the expression of IGFBP-1 was investigated by both immunohistochemistry and mrna in situ hybridization in the four major histological subtypes of ovarian surface epithelial cancers CCA, SA, EA and mucinous adenocarcinoma (MA) to determine whether or not the expression of IGFBP-1 is specific to CCA. We also addressed the role and importance of IGFBP-1 in the development of ovarian surface epithelial cancers, particularly CCA. Materials and methods sample selection The archival pathology files at the Department of Pathology, University Hospital of Tsukuba, Ibaraki, Japan, were searched for patients who had been diagnosed as having the various histological subtypes of ovarian surface epithelial carcinoma. Comprehensive informed consent from the patients was obtained before collection of the specimens. One hundred cases of primary ovarian carcinoma were selected for the study, comprising 31 cases of clear cell adenocarcinoma (CCA), 27 cases of serous adenocarcinoma (SA), 22 cases of endometrioid adenocarcinoma (EA) and 20 cases of mucinous adenocarcinoma (MA). Some nonneoplastic endometrial tissues were also selected for immunohistochemistry. These tissues included 10 cases of proliferative endometrium, secretory endometrium, decidua of the placenta and Arias-Stella glands in the endometrium of dilatation and curettage material following miscarriage, respectively. Five cases of endometrium with marked decidual change were also examined. pathological evaluation All haematoxylin and eosin (H&E) sections containing of the tumour were reviewed and the primary diagnosis of all cases was confirmed by two pathologists, S.S. and Y.M. Several tumours having a minor component of the other histological subtypes were reclassified into one of the four histological subtypes according to the predominant histology. Sections showing a representative histological picture were selected for immunohistochemistry and in situ hybridization in each case. immunohistochemistry Immunohistochemistry was performed using antibodies to IGFBP-1 and HNF-1b in accordance with the manufacturer s instructions and the procedure reported previously. 13 The antibodies used were a mouse monoclonal antibody to IGFBP-1 (sc-25257; lot no. G1508, E2907; dilution 1:100; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and a goat polyclonal antibody to HNF-1b (sc-7411; lot no. H1808; dilution 1:200; Santa Cruz). All formalin-fixed, paraffin-embedded specimens were cut into sections 3 lm thick. The slides were deparaffinized and rehydrated, and then for antigen retrieval they were pretreated by microwave heating at 98 C for 15 min (IGFBP-1) in TE buffer [10 mm Tris HCl with 1 mm ethylenediamine tetraacetic acid (EDTA), ph 8.0] and at 95 C for 20 min (HNF-1b) in sodium citrate buffer (10 mm sodium citrate monohydrate, ph 6.0). After the antigen retrieval procedure, endogenous peroxidase activity was blocked and then the slides were incubated with the primary antibodies at room temperature (RT) for

3 IGFBP-1 expression in ovarian CCA min (IGFBP-1) or at 4 C overnight (HNF-1b). The immunohistochemical reaction was visualized using Envision (IGFBP-1) or the ABC (HNF-1b) method with diaminobenzidine as the chromogen. Normal decidual cells of human placenta and a typical case of CCA were used as the positive control for IGFBP-1 and HNF-1b, respectively. Expression of IGFBP-1 was defined as negative (score 0; no staining), weakly positive (score 1+; staining in <5% of tumour cells), moderately positive (score 2+; staining in 5 20% of tumour cells) and strongly positive (score 3+; staining in >20% of tumour cells). Expression of HNF-1b was assessed as negative (score 0; no staining), weakly positive (score 1+; staining in <10% of tumour cells), moderately positive (score 2+; staining in 10 50% of tumor cells) and strongly positive (score 3+; staining in >50% of tumour cells). igfbp-1 in situ hybridization In situ hybridization was performed as reported previously. 14 To prepare the crna probes for in situ hybridization of IGFBP-1, total RNA was extracted from frozen decidual tissue of placenta material without any maternal disease. T7 RNA polymerase promoterattached primers for IGFBP-1 were designed using the primer designing tool of Primer 3 ( edu/primer3/) on the basis of the sequence in the GenBank database (Accession no. NM ). IGFBP-1 mrna was amplified by reverse transcriptase polymerase chain reaction (RT PCR) with T7 RNA polymerase promoter-attached primers (forward, 5 -CTTAATACGACTCACTATAGGGCTATGATGGCTCG- AAGGCTC-3 ; and reverse, 5 -CTTAATACGACTCACT- ATAGGGGAGACCCAGGGATCCTCTTC-3 ), and then the PCR products were transcribed to antisense or sense crna probes labelled with digoxigenin (DIG) with a DIG RNA Labeling kit (Roche Diagnostic GmbH, Penzberg, Germany) in accordance with the manufacturer s instructions. Through this process, we established antisense and sense crna probes of about 260 base pairs (bp) for in situ hybridization (Figure 1). Formalin-fixed, paraffin-embedded ovarian cancer tissues were cut into sections 3 lm thick and mounted on silane-coated slides (Matsunami Glass Ind., Osaka, Japan). Hybridization was performed at 50 C for 16 h with a hybridization mixture containing DIG-labelled antisense and sense probes. Detection of the hybridized crna probes was performed with horseradish peroxidase-conjugated rabbit anti-dig antibody (DakoCytomation, Kyoto, Japan), and the signals were amplified using a GenPointTM Tyramide Signal Amplification System (Dako North America, Carpinteria, CA, USA). Signal detection was performed by immunohistochemical reaction with diaminobenzidine as a chromogen. Expression of IGFBP-1 mrna was defined as negative (score 0; no staining), weakly positive (score 1+; staining in <5% of tumour cells), moderately positive (score 2+; staining in 5 20% of tumour cells) and strongly positive (score 3+; staining in >20% of tumour cells). statistical analysis All statistical analysis was carried out by v 2 test using Dr SSPS II software (SSPS, Chicago, IL, USA). For all analysis, differences at P < 0.05 were considered statistically significant. Results clinicopathological findings in 31 cases of cca The patients ranged in age from 25 to 83 years (mean, 50.4 years). In patients with CCA, the right ovary was affected in 18 cases (58.1%), the left ovary in 11 (35.5%) and the bilateral ovaries in two (6.5%). Twenty-four of the 31 CCA cases were classified as pt1 (77.4%) by the pathological tumour node metastasis (TNM) classification, and seven as pt2 or 3 (22.6%). Dissection of regional lymph nodes was performed in 25 patients with CCA, of whom three exhibited regional lymph node metastasis (12%). immunohistochemistry of non-neoplastic endometrial tissue First, we examined the expression of IGFBP-1 protein in normal gynaecological materials including endometrial tissue in the proliferative and secretory phases and in miscarriage material and placenta. Endometrial tissues showing marked decidual change were also examined. All samples of endometrial stromal cells in the proliferative phase were negative for IGFBP-1. Strong expression of IGFBP-1 was observed in decidualized stromal cells in secretory endometrium and markedly decidualized endometrium. Secretory endometrial epithelium, especially in mid- to late secretory phase, expressed IGFBP-1 weakly, but the intensity was lower than that of Arias-Stella glands. Decidua of the placenta also expressed IGFBP-1 consistently in all cases. Interestingly, all samples of Arias-Stella glands exhibited both cytoplasmic IGFBP-1 and nuclear HNF-1b expression (Figure 2).

4 732 S Sugita et al. A B C D Figure 1. Insulin-like growth factor binding protein-1 (IGFBP-1) protein and mrna expression in decidua cells of human placenta tissue. A, Decidua cells of placenta tissue (haematoxylin and eosin). B, Immunohistochemistry of IGFBP-1 protein. Decidua cells show expression of IGFBP-1 protein. C,D, In situ hybridization of IGFBP-1 mrna. C, Decidua cells exhibit reactivity for IGFBP-1 mrna with anti-sense crna probe of IGFBP-1. D, No reactivity was found with a sense-crna probe of IGFBP-1 mrna. expression of igfbp-1 in ovarian cancers demonstrated by immunohistochemistry and in situ hybridization Twenty-eight of the 31 cases of CCA (90.3%) were immunoreactive for IGFBP-1 by immunohistochemistry (Table 1 and Figure 3). The immunoreactive cases comprised five cases with a score of 3+ (17.9%), six with a score of 2+ (21.4%) and 17 with a score of 1+ (60.7%). Three cases of CCA were negative for IGFBP-1 protein expression. In contrast, four of the 69 non-cca cases (5.8%) showed expression of IGFBP-1 protein with a score of 1+; these comprised two cases of SA and EA, respectively; the two cases of EA did not show typical secretory change (Figure 4). Expression of IGFBP-1 mrna was also observed in 28 of 31 CCA cases (90.3%) by in situ hybridization (Table 2 and Figure 3). The cases positive for IGFBP-1 mrna comprised nine with a score of 3+ (32.1%), eight with a score of 2+ (28.6%) and 11 with a score of 1+ (39.3%). Three cases of CCA demonstrated no expression of IGFBP-1 mrna, and these cases were the same as those lacking expression of IGFBP-1 protein. Conversely, two of 17 non-cca cases (11.8%) expressed mrna for IGFBP-1 in a minority of the tumour cells with a score of 1+. These two cases were SA and EA, respectively. In summary, almost all cases of CCA expressed both IGFBP-1 protein and mrna at least in limited areas, whereas the majority of non-ccas lacked expression of either IGFBP-1 protein or mrna. In

5 IGFBP-1 expression in ovarian CCA 733 A B C D E F G H I Figure 2. Immunohistochemistry of insulin-like growth factor binding protein-1 (IGFBP-1) and hepatocyte nuclear factor-1beta (HNF-1b) in non-neoplastic endometrial tissue. A, Proliferative endometrium. B, Stromal cells in proliferative endometrium were negative for IGFBP-1. C, Decidualized secretory endometrium. D, Decidualized stromal cells in secretory endometrium express IGFBP-1 protein. E, Markedly decidualized secretory endometrium. F, Decidualized stromal cells are positive for IGFBP-1. G, Arias-Stella endometrial glands. H, Cytoplasmic reactivity of IGFBP-1 in Arias-Stella endometrial glands of dilatation and curettage (DC) material. I, Nuclear expression of HNF-1b in Arias-Stella endometrial glands of DC material. terms of the difference in expression score between IGFBP-1 protein and IGFBP-1 mrna, 17 of the 31 CCAs showed the same score, whereas in the other 14 the scores obtained by immunohistochemistry and in situ hybridization differed (Table 3). Ten of these 14 cases showed higher expression of IGFBP-1 mrna than the protein. However, four of the 14 cases showed lower expression of IGFBP-1 mrna than the protein (Table 3). Statistically, a significant difference was recognized between CCA cases and non-cca cases in the expression of IGFBP-1 protein and mrna (P < ) (Tables 1 and 2). Clinicopathologically, no significant difference was found, neither the sensitivity of IGFBP-1 expression and the extent of primary tumour (pt) (Table 4) nor regional lymph node metastasis (pn) (data not shown). expression of hnf-1 b in ovarian cancers demonstrated by immunohistochemistry Thirty of the 31 CCA cases (96.8%) were positive for HNF-1b, comprising 15 cases with a score of 3+ (50%), eight with a score of 2+ (26.7%) and seven with a score of 1+ (23.3%) (Table 5 and Figure 3). Only one case

6 734 S Sugita et al. Table 1. Expression of IGFBP-1 protein in ovarian adenocarcinoma by immunohistochemistry Histological type Total Negative (score 0) (score 1+ to 3+) P Score 1+ Score 2+ Score 3+ CCA 31 (100) 3 (8.7) 28 (90.3) < (60.7) 6 (21.4) 5 (17.9) Non-CCA 69 (100) 65 (94.2) 4 (5.8) 4* (100) 0 (0) 0 (0) IGFBP1, Insulin-like growth factor binding protein-1; CCA, clear cell adenocarcinoma. Score 0, negative; score 1+, <5%; score 2+, 5 20%; score 3+, >20%. Non-CCA includes serous adenocarcinoma, endometrioid adenocarcinoma and mucinous adenocarcinoma. Sensitivity, 90.3%; specificity, 94.2%; positive predictive value, 87.5%; negative predictive value, 95.6%. *Four score one cases of non-cca were two serous adenocarcinomas and two endometrioid adenocarcinomas. A B C D Figure 3. Clear cell adenocarcinoma (CCA) case showing high expression of insulin-like growth factor binding protein-1 (IGFBP-1) and hepatocyte nuclear factor-1beta (HNF-1b). A, CCA consisting of papillary and glandular structures of clear tumour cells (haematoxylin and eosin. B, Tumour cells express IGFBP-1 protein by immunohistochemistry. C, Tumour cells express IGFBP-1 mrna by in situ hybridization. D, Tumour cells are immunoreactive for HNF-1b.

7 IGFBP-1 expression in ovarian CCA 735 A B C D Figure 4. Non-clear cell adenocarcinoma (CCA) case [endometrioid adenocarcinoma (EA)] showing low expression of insulin-like growth factor binding protein-1 (IGFBP-1). A, EA consists of papillary proliferation of columnar and polyhedral tumour cells [haematoxylin and eosin (H&E)]. B, A few tumour cells are positive for IGFBP-1 protein by immunohistochemistry. C, The other EA case with glandular structures (H&E). D, A few tumour cells are positive for IGFBP-1 protein by immunohistochemistry. Table 2. Expression of IGFBP-1 mrna in ovarian adenocarcinoma by in situ hybridization Histological type Total Negative (score 0) (score 1+ to 3+) P Score 1+ Score 2+ Score 3+ CCA 31 (100) 3 (9.7) 28 (90.3) < (39.3) 8 (28.6) 9 (32.1) Non-CCA 17 (100) 15 (88.2) 2 (11.8) 2* (100) 0 (0) 0 (0) Score 0, negative; score 1+, <5%; score 2+, 5 20%; score 3+, >20%. CCA, Clear cell adenocarcinoma; IGFBP-1, insulin-like growth factor binding protein-1. Non-CCA includes serous adenocarcinoma, endometrioid adenocarcinoma, and mucinous adenocarcinoma. Sensitivity, 90.3%; specificity, 88.2%; positive predictive value, 93.3%; negative predictive value, 83.3%. *IGFBP-1-positive cases of non-cca were serous adenocarcinoma and endometrioid adenocarcinoma.

8 736 S Sugita et al. Table 3. Comparison of expression score between IGFBP-1 protein and IGFBP-1 mrna IGFBP-1 mrna expression Score 0 Score 1+ Score 2+ Score 3+ Total IGFBP-1 protein expression Score Score Score Score Total IGFBP-1, Insulin-like growth factor binding protein-1. showed no expression of HNF-1b protein, and this case was also negative for IGFBP-1 protein and mrna. In contrast, 11 of the 69 non-cca cases (15.9%) were immunoreactive for HNF-1b, but the majority of these positive cases showed weak expression of HNF-1b with a score of 1+. A statistically significant difference was recognized between CCA cases and non-cca cases with regard to the expression of HNF-1b protein (P < ) (Table 5). One non-cca case positive for IGFBP-1 protein was included among the 11 that were positive for HNF-1b. Discussion We have demonstrated that the immunohistochemical expression of IGFBP-1 protein is specific to cases of CCA, although its expression is limited to relatively small areas in many cases. We have also confirmed the specific expression of IGFBP-1 at the mrna level. IGFBP-1 mrna expression was detected specifically in CCA cases by in situ hybridization, although the distribution of the sensitivity score differed between the protein level and the mrna level. Stronger and wider expression of IGFBP-1 was observed at the mrna level than at the protein level, although it is not meaningful to compare the levels of protein and mrna expression determined by these two methods. These results suggest that CCAs probably express IGFBP-1 more strongly at the mrna level than at the protein level. We suspected that IGFBP-1 protein might show a tendency to disappear rapidly from the cytoplasm after its synthesis, because generally a large proportion of IGFBP is secreted into the blood and Table 4. Association of pathological T factor and sensitivity of IGFBP-1 protein expression in ovarian clear cell adenocarcinoma Sensitivity of IGFBP-1 protein expression Pathological T factor (pt) Total Low (score 0 or 1+) High (score 2+ or 3+) P pt1 24 (100) 15 (62.5) 9 (37.5) >0.05 pt2 or 3 7 (100) 5 (71.4) 2 (28.6) pt1, pathological T factor of tumour node metastasis (TNM) classification. IGFBP-1, insulin-like growth factor binding protein-1. Table 5. Expression of HNF-1b protein in ovarian adenocarcinoma by immunohistochemistry Histological type Total Negative (score 0) (score 1+ to 3+) P Score 1+ Score 2+ Score 3+ CCA 31 (100) 1 (3.2) 30 (96.8) < (23.3) 8 (26.7) 15 (50) Non-CCA 69 (100) 58 (84.1) 11 (15.9) 10 (90.9) 1 (9.1) 0 (0) Score 0, negative; score 1+, <10%; score 2+, 10 50%; score 3+, >50% CCA, Clear cell adenocarcinoma; HNF-1b, hepatocyte nuclear factor-1beta. Non-CCA includes serous adenocarcinoma, endometrioid adenocarcinoma, and mucinous adenocarcinoma. Sensitivity, 96.8%; specificity, 84.1%; positive predictive value, 73.1%; negative predictive value, 98.3%.

9 IGFBP-1 expression in ovarian CCA 737 circulates throughout the body by binding to insulation growth factors (IGFs). 2 Conversely, the localization of IGFBP-1 protein expression sometimes differed from that of IGFBP-1 mrna expression in the same specimen. As the preservation of the specimens was similar in each case, we concluded that mrna expression and protein expression were not usually simultaneous in individual tumour cells, and showed heterogeneity of distribution in individual tumours. In contrast, non- CCA cases hardly expressed IGFBP-1 at either the mrna or protein level. These results indicated that IGFBP-1 expression is a feature of ovarian CCA and might represent a characteristic phenotype. No previously reported study has focused in detail upon IGFBP-1 expression in ovarian adenocarcinoma using human histological specimens. In general, IGFBP-1 plays an important role in human female reproductive physiology, 3,4 and we have confirmed its specific expression in secretory endometrial stromal cells and decidua of the placenta by in situ hybridization. 1 IGFBP-1 was recognized formerly as placenta-specific tissue protein 12 (PP12), and Inaba et al. 15 have investigated PP12 expression in specimens of ovarian cancer. They discussed the expression of pregnancy-specific protein 1 (SP1), PP5, PP10, PP11 and PP12 in cases of mucinous and serous cystadenocarcinoma by immunohistochemistry, and concluded that these proteins, especially PP5 and PP11, might be useful markers for monitoring patients with ovarian adenocarcinoma and for early diagnosis. However, no CCA cases were included in their study, and they did not mention PP12 expression in cases of CCA. Conversely, some IGFBPs, except for IGFBP-1, have been discussed in relation to their association with the development of, and susceptibility to, ovarian cancer. 7,16 Using tissue microarray analysis, Wang et al. 8 demonstrated that the expression of IGFBP-2 and IGFBP-5 differed among high-grade ovarian SA, CCA and MA, and concluded that both IGFBPs are important for the development of high-grade SA cases alone. Some studies have shown that serum IGFBP-2 levels are elevated in early and advanced ovarian cancers, regardless of histological type, and that increased expression of IGFBP-2 mrna can be identified in ovarian cancer at any stage. 9 A recent study by Martin et al. suggested for the first time that insulin-like growth factor 2 mrna-binding protein 3 (IGF2BP3, also known as IMP3) could be a prognostic biomarker of ovarian CCA. 17 In the present study, there was no relationship between the sensitivity of IGFBP-1 expression and pathological T and N factors of the TNM classification in CCA patients. HNF-1b has been established recently as a specific histological marker of CCA, as first reported by Tsuchiya et al., and its specific expression has already been confirmed by many researchers. 10,13,18 20 HNF-1 is a transcription factor expressed mainly in the liver and kidney, and regulates the promoters of several genes. HNF-1 consists of two isoforms: HNF-1a and HNF-1b. It has been revealed that HNF-1a binds to the promoter region of the IGFBP-1 gene and regulates its expression. 11,12 In the present study, we found that all the IGFBP-1-positive cases of CCA also expressed HNF-1b, another isoform of HNF-1, and we speculate that both HNF-1b and IGFBP-1 might represent a characteristic phenotype of CCA. Indeed, the sensitivity of IGFBP-1 expression was lower than that of HNF-1b in many CCA cases, but we conclude that IGFBP-1 might also be a more specific marker of CCA than HNF-1b. We also performed immunohistochemistry for IG- FBP-1 and HNF-1b in Arias-Stella glands in miscarriage material, as the histology of Arias-Stella glands mimics that of CCA. The Arias-Stella reaction has been recognized classically as a frequent histological sign of pregnancy in endometrial glands, and is often noticed in routine DC materials from abortion cases. The most characteristic histological feature of Arias-Stella glands is conspicuous clear cell change with hypersecretory features, and the nucleus sometimes exhibits notable atypia. Some studies have emphasized the importance of differential diagnosis, including benign clear cell lesions and malignant clear cell neoplasms, because of the histological similarity of gynaecological clear cell malignancy and glands showing the Arias-Stella reaction. 21,22 Such clear cell lesions can sometimes present diagnostic difficulty if the Arias-Stella reaction occurs in the absence of pregnancy in older women. We have demonstrated that Arias-Stella glands show both cytoplasmic expression of IGFBP-1 and nuclear expression of HNF-1b, and that the expression pattern of the two markers is similar to that in CCA. These results indicate that CCA and Arias-Stella endometrial glands share common features, not only morphologically but also immunophenotypically. Placental decidua and decidualized endometrial stromal cells also express IGFBP-1 protein strongly. In some situations, the distinction between clear cell adenocarcinoma and endometrioid adenocacinoma showing secretory change may be difficult because both of the tumour cells have clear cytoplasm. The secretory variant of endometrioid adenocarcinoma often consists entirely of tumour cells with clear subnuclear or supranuclear vacuoles, similar to endometrial epithelium in the early secretory phase. Among 22 ovarian endometrioid adenocarcinoma

10 738 S Sugita et al. cases examined in this study, there was no endometrioid adenocarcinoma showing secretory change. We performed immunohistochemistry for IGFBP-1 in several cases of uterine endometrial adenocarcinoma with secretory change, including the secretory variant of endometrioid adenocarcinoma. The results showed that no reactivity of IGFBP-1 was observed in the tumours (data not shown). In conclusion, the expression of IGFBP-1 is more specific in ovarian CCA than in other histological types of ovarian cancer, in comparison with HNF-1b. The sensitivity of IGFBP-1 is obviously lower than that of HNF-1b in CCA cases, and no clinicopathological significance of IGFBP-1 expression in CCA patients was shown. IGFBP-1 may represent a characteristic phenotypic component of CCA, and is a potentially specific marker of this cancer. References 1. Yamada K, Kano J, Tsunoda H et al. Phenotypic characterization of endometrial stromal sarcoma of the uterus. Cancer Sci. 2006; 97; Jones JI, Clemmons DR. Insulin-like growth factors and their binding proteins: biological actions. Endocr. Rev. 1995; 16; Druckmann R, Rohr UD. IGF-1 in gynaecology and obstetrics: update Maturitas 2002; 41(Suppl. 1); S65 S Fowler DJ, Nicolaides KH, Miell JP. Insulin-like growth factor binding protein-1 (IGFBP-1): a multifunctional role in the human female reproductive tract. Hum. Reprod. Update 2000; 6; LeRoith D, Roberts CT Jr. The insulin-like growth factor system and cancer. Cancer Lett. 2003; 195; Yu H, Rohan T. Role of the insulin-like growth factor family in cancer development and progression. J. Natl. Cancer Inst. 2000; 92; Hirano S, Ito N, Takahashi S, Tamaya T. Clinical implications of insulin-like growth factors through the presence of their binding proteins and receptors expressed in gynecological cancers. Eur. J. Gynaecol. Oncol. 2004; 25; Wang H, Rosen DG, Wang H, Fuller GN, Zhang W, Liu J. Insulinlike growth factor-binding proteins 2 and 5 are differentially regulated in ovarian cancer of different histologic types. Mod. Pathol. 2006; 19; Lancaster JM, Sayer RA, Blanchette C et al. High expression of insulin-like growth factor binding protein-2 messenger RNA in epithelial ovarian cancers produces elevated preoperative serum levels. Int. J. Gynecol. Cancer 2006; 16; Tsuchiya A, Sakamoto M, Yasuda J et al. Expression profiling in ovarian clear cell carcinoma: identification of hepatocyte nuclear factor-1 beta as a molecular marker and a possible molecular target for therapy of ovarian clear cell carcinoma. Am. J. Pathol. 2003; 163; Powell DR, Suwanichkul A. HNF1 activates transcription of the human gene for insulin-like growth factor binding protein-1. DNA Cell Biol. 1993; 12; Suh DS, Rechler MM. Hepatocyte nuclear factor 1 and the glucocorticoid receptor synergistically activate transcription of the rat insulin-like growth factor binding protein-1 gene. Mol. Endocrinol. (Balt.) 1997; 11; Kato N, Sasou S, Motoyama T. Expression of hepatocyte nuclear factor-1beta (HNF-1beta) in clear cell tumors and endometriosis of the ovary. Mod. Pathol. 2006; 19; Pei Y, Kano J, Iijima T, Morishita Y, Inadome Y, Noguchi M. Overexpression of Dickkopf 3 in hepatoblastomas and hepatocellular carcinomas. Virchows Arch. 2009; 454; Inaba N, Ishige H, Ijichi M et al. Immunohistochemical detection of pregnancy-specific protein (SP1) and placenta-specific tissue proteins (PP5, PP10, PP11 and PP12) in ovarian adenocarcinomas. Oncodev. Biol. Med. 1982; 3; Walker G, MacLeod K, Williams AR, Cameron DA, Smyth JF, Langdon SP. Insulin-like growth factor binding proteins IG- FBP3, IGFBP4, and IGFBP5 predict endocrine responsiveness in patients with ovarian cancer. Clin. Cancer Res. 2007; 13; Kobel M, Xu H, Bourne PA et al. IGF2BP3 (IMP3) expression is a marker of unfavorable prognosis in ovarian carcinoma of clear cell subtype. Mod. Pathol. 2009; 22; Kato N, Motoyama T. Overexpression of osteopontin in clear cell carcinoma of the ovary: close association with HNF-1beta expression. Histopathology 2008; 52; Higashiguchi A, Yamada T, Susumu N et al. Specific expression of hepatocyte nuclear factor-1beta in the ovarian clear cell adenocarcinoma and its application to cytological diagnosis. Cancer Sci. 2007; 98; Yamamoto S, Tsuda H, Aida S, Shimazaki H, Tamai S, Matsubara O. Immunohistochemical detection of hepatocyte nuclear factor 1beta in ovarian and endometrial clear-cell adenocarcinomas and non-neoplastic endometrium. Hum. Pathol. 2007; 38; Vang R, Barner R, Wheeler DT, Strauss BL. Immunohistochemical staining for Ki-67 and p53 helps distinguish endometrial Arias-Stella reaction from high-grade carcinoma, including clear cell carcinoma. Int. J. Gynecol. Pathol. 2004; 23; Nucci MR, Young RH. Arias-Stella reaction of the endocervix: a report of 18 cases with emphasis on its varied histology and differential diagnosis. Am. J. Surg. Pathol. 2004; 2;