Inhibin/Activin Subunits Are Immunohistochemically Expressed in Complete and Partial Hydatidiform Moles

ANTICANCER RESEARCH 27: 1995-2000 (2007)

Inhibin/Activin Subunits are Immunohistochemically Expressed in Complete and Partial Hydatidiform Moles

IOANNIS MYLONAS1, NAIM SHABANI1,2, JULIA VOGL1, JOSEF MAKOVITZKY3, SUSI KUNZE1, CHRISTINA KUHN1, SANDRA SCHULZE1, KLAUS FRIESE1 and UDO JESCHKE1

1Ludwig-Maximilians University Munich, First Department of Obstetrics and Gynaecology, 80337 Munich, Germany; 2Landes-Krankenhaus Schaerding, Abteilung für Frauenheilkunde und Geburtshilfe, A-4780 Schaerding, Austria; 3University of Rostock, Department of Obstetrics and Gynaecology, 18057 Rostock, Germany

Abstract. Background and aim: Inhibins are dimeric complete moles, suggesting the utilization of these antibodies glycoproteins, belonging to the transforming growth factor beta as diagnostic differentiation markers between complete and (TGF-‚) family, composed of an ·-subunit (INH-·) and one of partial moles. two possible ‚-subunits (‚A or ‚B). Additionally two further ‚- subunits (‚C and ‚E) have been cloned, although their function Inhibins are dimeric disulphide-linked glycoproteins and are remains still quite unclear. The detection by immuno- members of the transforming growth factor beta (TGF-‚) histochemistry of inhibin/activin subunits has been proposed as a family of cytokines; they were initially isolated from the useful marker of trophoblastic diseases. Interestingly, a complete gonads and identified as modulators of FSH production mole cannot be easily differentiated from a partial mole. from the anterior pituitary gland (1, 2). They are Therefore, the aim of this study was to determine expression heterodimers consisting of one ·-subunit and one of two changes of the five inhibin/activin subunits in partial and possible ‚-subunits (‚A- and ‚B-subunits). The ·-subunit complete moles. Materials and Methods: Histologically diagnosed can dimerize with either ‚A or ‚B to form inhibin-A (·-‚A) complete (n=6) and partial (n=3) hydatidiform moles were or -B (·-‚B), respectively. Activins are homodimers of ‚- immunohistochemical analyzed for INH-·, INH-‚A, INH-‚B, subunits linked by a disulfide bond. Depending on the INH-‚C and INH-‚E subunits. The immunohistochemical combination of the subunits, there are three isoforms of reaction in intermediate trophoblast was analyzed with a semi- activin, namely activin-A (‚A-‚A), activin-B (‚B-‚B) and quantitative score (IRS) and statistical analysis was performed. activin-AB (‚A-‚B) (1, 2). Recently, three additional ‚- Results: Immuno-histochemical reaction with INH-·, INH-‚A, subunits have been identified, determined as ‚C (3), ‚D INH-‚B, INH-‚C and INH-‚E subunits was demonstrated in (Xenopus only) (4) and ‚E (5), although their precise hydatidiform moles. The INH-‚A and INH-‚B expression was function remains unknown. significantly higher in complete compared to partial moles The expression of inhibin/activin subunits have been (p<0.05 each), while INH-·, INH-‚C and INH-‚E did not described in different female tissues, including normal and demonstrate any statistically significant differences. Conclusion: pathological human endometrium (6-9), and in normal and We demonstrated an immunohistochemical expression of all five pathological placenta (10-13), suggesting different important inhibin/activin subunits in partial and complete hydatidiform roles, such as paracrine modulators of reproductive function moles. The expression of INH-‚A and INH-‚B determined (1, 14). Inhibin/activin subunits are also expressed in immunohistochemically was significantly up-regulated in placental decidua, the syncytiotrophoblast (13) and in the trophoblast (10, 12), while recently inhibin/activin subunits were demonstrated in a hydatidiform mole with persistent polymorphic trophoblastic hyperplasia (15). Correspondence to: PD Dr. Udo Jeschke, Ludwig-Maximilians- Hydropic swelling of some or all villi and trophoblastic University Munich, First Department of Obstetrics and proliferation characterize hydatidiform mole, including two Gynaecology, Maistrasse 11, 80337 Munich, Germany. Tel: +49 89 distinct entities, complete and partial hydatidiform moles 5160 4266, Fax: +49 89 5160 4916, e-mail: [email protected] (16). Increased levels of circulating inhibin have previously muenchen.de been reported in molar pregnancy: it has been proposed Key Words: Inhibin/activin subunits, inhibin-·, inhibin-‚A, that serum inhibin concentrations may be of value in the inhibin-‚B, inhibin-‚C, inhibin-‚E, complete hydatidiform mole, management of trophoblastic disease (17) and increased partial hydatidiform mole. serological concentrations of inhibin can be found in

0250-7005/2007 $2.00+.40 1995 ANTICANCER RESEARCH 27: 1995-2000 (2007)

Table I. Antibodies used for immunohistochemical characterisation of complete and partial hydatidiform moles by immunohistochemistry.

Antibody Clone Isotype Dilution Source

Inhibin-· R1 mouse IgG2a 1:50 Serotec, Oxford, UK Inhibin-‚A E4 mouse IgG2b 1:50 Serotec, Oxford, UK Inhibin-‚B C5 mouse IgG2a 1:10 Serotec, Oxford, UK Inhibin-‚C Polyclonal Goat 1:50 R&D Systems, Inhibin-‚E Polyclonal rabbit 1:4000 In house antibody BioGenes, Berlin, Germany

hydatidiform mole (18). Recently, it was suggested that complex with the use of the mouse-IgG-Vectastain Elite ABC kit serum inhibin A and activin A measurements might be of (Vector Laboratories, Burlingame, CA, USA) for inhibin-·, -‚A value in diagnosis and short-term follow-up of molar and -‚B as well as goat-IgG-Vectastain Elite ABC kit (Vector pregnancy (19). Inhibin-· and -‚ subunits are consistently Laboratories) for inhibin-‚C and -‚E. The antibodies used are given in Table I. co-expressed immunohistochemically in syncytiotrophoblast Briefly, paraffin-fixed tissue sections were dewaxed using xylol in complete and partial moles, suggesting that these for 15 min, rehydrated in an ascending series of alcohol (70%, 96% glycoproteins might be useful tissue markers in the and 100%) and subjected to antigen retrieval on a high setting for differential diagnosis of trophoblastic lesions (20). 10 min in a pressure cooker in sodium citrate buffer (pH 6.0), Interestingly, complete moles can be reliably distinguished containing citrate acid 0.1M and sodium citrate 0.1M in distilled from non-molar pregnancy, but neither non-molar pregnancy water. After cooling, the slides were washed twice in PBS. nor complete moles are easily differentiated from partial Endogenous peroxidase activity was quenched by immersion in 3% hydrogen peroxide (Merck, Darmstadt, Germany) in methanol for mole (21, 22). However, limited data on histological 20 min. Non-specific binding of the primary antibodies was blocked expression of inhibin/activin subunits expression in these by incubating the sections with diluted normal serum (10 ml PBS two forms of hydatidiform moles exists. Therefore, the containing 150 Ìl horse serum; Vector Laboratories) for 20 min at aims of the present study were to determine expression room temperature. Sections were then incubated at room changes of the five inhibin/activin subunits in partial and temperature for 120 min with the primary antibodies. Inhibin-· was complete moles. diluted in PBS. After washing with PBS, the slides were incubated in diluted biotinylated serum (10 ml PBS containing 50 Ìl horse serum; Vector Laboratories) for another 30 min at room Materials and Methods temperature. After incubation with the avidin-biotin peroxidase complex for another 30 min and repeated washing steps with PBS, Placental tissues diagnosed by a gynecological pathologist as visualisation was performed with substrate and the chromagen 3,3'- complete (n=6) and partial (n=3) hydatidiform moles were diaminobenzidine (DAB; Dako, Glostrup, Denmark) for 8-10 min. obtained from the First Department of Obstetrics and Gynaecology The slides were counterstained further with Mayer’s acidic of the LMU Munich. hematoxylin and washed in a series of alcohol (50-98%). After xylol treatment the slides were covered. Generation of a polyclonal activin-‚E peptide antibody. Anti-activin- Negative controls were performed by replacing the primary ‚E polyclonal antibodies were generated in rabbits against a antibody with normal horse serum (Vector Laboratories). polypeptide of 16 amino acids of activin-‚E (polypeptide-sequence: Positive cells showed a brownish color and the negative control, NH2-CRWGPRRRRQGSRTLL-COOH; amino acid position 144 as well as unstained cells, appeared blue. The standardisation, to 158; accession number: AAH05161). A primary dose of 200 Ìg dilution and optimisation of this protocol for inhibin-·, -‚A and activin-‚E polypeptide was emulsified in Freund’s complete adjuvant -‚B was primarily tested on normal premenopausal ovary tissue, (Sigma-Aldrich, Germany) and administered subcutaneously in while negative controls included postmenopausal ovarian tissue. rabbits. Three doses of the peptide emulsified in Freund’s For inhibin-‚C and inhibin-‚E normal liver tissue was used as incomplete adjuvant were administrated at intervals (6 weeks). After positive control. the third booster injection (14 days), blood was collected from the rabbit and the serum was separated. Antibodies were isolated using Immunohistochemical evaluation and statistical analysis. The column chromatography with a protein A column (Amersham intensity and distribution patterns of specific inhibin/activin subunit Pharmacia Biotech, Freiburg, Germany). immunohistochemical staining reaction was evaluated by two blinded, independent observers, including a gynecological Immunohistochemistry. Immunohistochemistry on paraffin sections pathologist (N. S.), using a semi-quantitative score (IRS score) as (7 Ìm) of the different placental tissue specimens was performed described elsewhere (23) and used in the evaluation of by incubating the slides in methanol/H2O2 (30 min) to inhibit inhibin/activin subunit expression in normal human endometrial endogenous peroxidase activity. Immunohistochemistry with specimens (6). The IRS score was calculated as follows: IRS=SI x inhibin-subunits was performed using a combination of pressure PP, where SI is the optical staining intensity (graded as: 0=no cooker heating and the standard streptavidin-biotin-peroxidase staining; 1=weak staining; 2=moderate staining and 3=strong

1996 Mylonas et al: Inhibin and Activin in Hydatidiform Moles staining) and PP the percentage of positively-stained cells. PP was chemically detectable inhibin-· subunit in placental tissue estimated by counting ~200 cells (defined as 0=no staining; was mainly localized within the syncytiotrophoblast with 1=<10% staining; 2=11-50% staining; 3=51-80% staining and positive staining of the decidua (10). Production of inhibin 4=>81% staining). Digital images were obtained with a digital by these cells may account for raised serum levels during camera system (Olympus, Tokyo, Japan) and were saved on computer. The Mann-Whitney rank-sum test was used to compare pregnancy. Inhibin can also be demonstrated in the means of the different IRS scores (SPSS; Chicago, IL, USA). choriocarcinoma and in nongestational trophoblastic tissue, Significance of differences of the means was assumed at p≤0.05. including in hydatidiform mole with persistent polymorphic trophoblastic hyperplasia (15). The detection by Results immunohistochemistry of inhibin/activin subunits has been recently proposed, in association with beta-hCG, as a useful Immunohistochemical staining was performed using an marker of trophoblastic neoplasia (25, 26). It has been appropriate positive control comprising ovaries containing suggested that measurements of serum inhibin A and activin follicular cysts or normal liver. Inhibin-· stained positive A might be more useful in diagnosing and following up with ovarian granulosa cells and theca interna cells, while molar pregnancies than hCG (19). However, other research inhibin-‚A and -‚B subunits also stained positively with groups demonstrated that serum molecular forms of human ovarian tissue as previously described (6). Inhibin- inhibins (inhibins A and B) might not be of relevance in the ‚C and -‚E were immunohistochemically expressed in liver biological survey of patients with gestational trophoblastic and ovarian tissue as described elsewhere (4, 5). diseases (27, 28). Additional immunohistochemical and All five inhibin/activin subunits were detected in partial serological studies are still needed to evaluate the and complete hydatidiform moles. Inhibin-· was primarily measurement of inhibin/activin subunits in the clinical expressed in syncytiotrophoblast cells of partial (Figure 1A) setting regarding trophoblastic lesions. and complete moles (Figure 1B). The ‚A-subunit was also Using immunhistochemistry we demonstrated the detected in molar syncytiotrophoblast cells but with a expression of all five inhibin/activin-subunits in complete weaker intensity in partial (Figure 1C) compared to and partial hydatidiform moles. We observed a significant complete moles (Figure 1D). The syncytiotrophoblast in lower expression of the inhibin-‚A and -‚B subunit in partial and complete hydatidiform moles also demonstrated partial compared to complete moles. Interstingly, expression a positive immunohistochemical staining of the inhibin-‚B of the inhibin-·, -‚C and -‚E subunits did not show any subunit. Interestingly, the staining reaction was less in the statistical significance. We therefore assume that the higher partial (Figure 1E) than in the complete moles (Figure 1F). ‚A and ‚B subunit expression results in a higher activin A Inhibin-‚C and -‚E also demonstrated a positive staining and activin B secretion. This assumption might explain the reaction, albeit weaker than the other inhibin/activin higher activin A concentration in serum in gestational subunits (Figures 1G-J). trophoblastic diseases, while serological inhibin The immunoreactive score for inhibin-· demonstrated concentration (inhibins A and B) might not be of relevance no significant differences between partial and complete in the survey of patients with hydatidiform moles. However, moles (Figure 2), while the staining intensity of inhibin- other possibilities like the formation of inhibin C or E as ‚A and -‚B subunit was significantly higher in complete well as different activin molecules cannot be excluded and hydatidiform moles than in partial ones (p<0.05 each) warrent further studies. (Figure 2). The immunoreactive score for inhibin-‚C and -‚E did not show any significant differences between Conclusion either hydatidiform mole entities (Figure 2). All five inhibin/activin subunits were observed in Discussion complete and partial hydatidiform moles using immunohistochemistry. Inhibin/activin subunits were distributed While inhibins/activins were initially characterised as on the cell surface, playing a significant role in the endocrine and paracrine hormonal regulators of the pathogenesis of gestational trophoblastic diseases. The hypothalamic-pituitary-gonadal axis, it is now clear that they immunohistochemical reaction with inhibin-· was similar are expressed in a wide range of tissues, including normal in complete and partial moles. The inhibin-‚A and -‚B and pathological placenta (10-13, 24). The human placenta expression was significantly higher in complete compared expresses inhibin/activin mRNA (13) and inhibin/activin to partial moles (p<0.05 each), while inhibin-‚C and -‚E subunit protein throughout pregnancy, being the primary did not demonstrate any statistically significant source of maternal circulating inhibin and activin levels (24) differences. Therefore, the ‚A- and ‚B-antibodies can be Inhibin/activin subunits are also localized in trophoblast used as diagnostic differentiation markers between (10) and trophoblastic tumours (25, 26). The immunohisto- complete and partial moles. Although the precise role of

1997 ANTICANCER RESEARCH 27: 1995-2000 (2007)

Figure 1. continued these inhibin/activin subunits in human complete and Acknowledgements partial hydatidiform moles is still unclear, they could be involved in autocrine/paracrine signalling, contributing to We would like to thank the nurses, medical doctors and laboratory several aspects such as angiogenesis and tissue staff for obtaining the endometrial material. We also express our remodelling. gratitude to Mrs. I. Wiest and Mrs. S. Kunze for their excellent

1998 Mylonas et al: Inhibin and Activin in Hydatidiform Moles

Figure 1. Immunohistochemical staining reaction of inhibin/activin subunits in in complete and partial hydatidiform moles. Magnification: x100.

Figure 2. The immunoreactive score for inhibin-· demonstrated no significant differences between partial and complete moles, while the staining intensity of inhibin-‚A and -‚B subunits increased significantly between partial and complete hydatidiform moles *(p<0.05). The immunoreactive score for inhibin-‚C and -‚E did not show any significant differences between either hydatidiform mole entities. Data represent mean±SEM. *Significance was assumed at p<0.05.

1999 ANTICANCER RESEARCH 27: 1995-2000 (2007) work with the endometrial samples. This study was partially 14 Welt CK: The physiology and pathophysiology of inhibin, supported by FöFoLe program, the Friedrich-Baur-Institute and activin and follistatin in female reproduction. Curr Opin Obstet the Weigland Stipendium Program of the Ludwig-Maximilians Gynecol 14: 317-323, 2002. University Munich for I. Mylonas. 15 Mylonas I, Makovitzky J, Vogel M, Shabani N, Jeschke U and Friese K: Expression of inhibin/activin subunits, sialyl-Lewis A References (CA 19-9, sLea) and sialyl-Lewis X (sLex) carbohydrate antigens in a hydatidiform mole with persistent polymorphic 1 de Kretser DM, Hedger MP, Loveland KL and Phillips DJ: trophoblastic hyperplasia. Anticancer Res 25: 1725-1730, 2005. Inhibins, activins and follistatin in reproduction. Hum Reprod 16 Vassilakos P and Kajii T: Letter: Hydatidiform mole: two Update 8: 529-541, 2002. entities. Lancet 1: 259, 1976. 2 Vale W, Rivier C, Hsueh A, Campen C, Meunier H, Bicsak T, 17 Yohkaichiya T, Fukaya T, Hoshiai H, Yajima A and de Kretser Vaughan J, Corrigan A, Bardin W and Sawchenko P: Chemical DM: Inhibin: a new circulating marker of hydatidiform mole? and biological characterization of the inhibin family of protein BMJ 298: 1684-1686, 1989. hormones. Recent Prog Horm Res 44: 1-34, 1988. 18 Badonnel Y, Barbe F, Legagneur H, Poncelet E and Schweitzer 3 Hötten G, Neidhardt H, Schneider C and Pohl J: Cloning of a M: Inhibin as a marker for hydatidiform mole: a comparative new member of the TGF-beta family: a putative new activin beta study with the determinations of intact human chorionic C chain. Biochem Biophys Res Commun 206: 608-613, 1995. gonadotrophin and its free beta-subunit. Clin Endocrinol (Oxf) 4 Oda S, Nishimatsu S, Murakami K and Ueno N: Molecular 41: 155-162, 1994. cloning and functional analysis of a new activin beta subunit: a 19 Florio P, Severi FM, Cobellis L, Danero S, Bome A, Luisi S and dorsal mesoderm-inducing activity in Xenopus. Biochem Petraglia F: Serum activin A and inhibin A. New clinical Biophys Res Commun 210: 581-588, 1995. markers for hydatidiform mole. Cancer 94: 2618-2622, 2002. 5 Fang J, Yin W, Smiley E, Wang SQ and Bonadio J: Molecular 20 Kommoss F, Schmidt D, Coerdt W, Olert J and Muntefering H: cloning of the mouse activin beta E subunit gene. Biochem Immunohistochemical expression analysis of inhibin-alpha and Biophys Res Commun 228: 669-674, 1996. -beta subunits in partial and complete moles, trophoblastic 6 Mylonas I, Jeschke U, Wiest I, Hoeing A, Vogl J, Shabani N, tumors, and endometrial decidua. Int J Gynecol Pathol 20: 380- Kuhn C, Schulze S, Kupka MS and Friese K: Inhibin/activin 385, 2001. subunits alpha, beta-A and beta-B are differentially expressed 21 Howat AJ, Beck S, Fox H, Harris SC, Hill AS, Nicholson CM in normal human endometrium throughout the menstrual cycle. and Williams RA: Can histopathologists reliably diagnose molar Histochem Cell Biol 122: 461-471, 2004. pregnancy? J Clin Pathol 46: 599-602, 1993. 7 Mylonas I, Makovitzky J, Fernow A, Richter DU, Jeschke U, 22 Javey H, Borazjani G, Behmard S and Langley FA: Briese V, Gerber B and Friese K: Expression of the Discrepancies in the histological diagnosis of hydatidiform inhibin/activin subunits alpha (alpha), beta-A (betaA) and beta- mole. Br J Obstet Gynaecol 86: 480-483, 1979. B (betaB) in benign human endometrial polyps and tamoxifen- 23 Mylonas I, Speer R, Makovitzky J, Richter DU, Briese V, associated polyps. Arch Gynecol Obstet 272: 59-66, 2005. Jeschke U and Friese K: Immunohistochemical analysis of 8 Mylonas I, Makovitzky J, Hoeing A, Richter DU, Vogl J, steroid receptors and glycodelin A (PP14) in isolated glandular Schulze S, Jeschke U, Briese V and Friese K: Inhibin/activin epithelial cells of normal human endometrium. Histochem Cell subunits beta-A (-betaA) and beta-B (-betaB) are differentially Biol 114: 405-411, 2000. localised in normal, hyperplastic and malignant human 24 Florio P, Cobellis L, Luisi S, Ciarmela P, Severi FM, Bocchi C endometrial tissue. Acta Histochem 108: 1-11, 2006. and Petraglia F: Changes in inhibins and activin secretion in 9 Mylonas I, Makovitzky J, Richter DU, Jeschke U, Briese V and healthy and pathological pregnancies. Mol Cell Endocrinol 180: Friese K: Expression of the inhibin-alpha subunit in normal, 123-130, 2001. hyperplastic and malignant endometrial tissue: an 25 Pelkey TJ, Frierson HF, Jr, Mills SE and Stoler MH: Detection immunohistochemical analysis. Gynecol Oncol 93: 92-97, 2004. of the alpha-subunit of inhibin in trophoblastic neoplasia. Hum 10 McCluggage WG, Ashe P, McBride H, Maxwell P and Sloan Pathol 30: 26-31, 1999. JM: Localization of the cellular expression of inhibin in 26 Shih IM and Kurman RJ: Immunohistochemical localization of trophoblastic tissue. Histopathology 32: 252-256, 1998. inhibin-alpha in the placenta and gestational trophoblastic 11 Mylonas I, Schiessl B, Jeschke U, Vogl J, Makrigiannakis A, lesions. Int J Gynecol Pathol 18: 144-150, 1999. Kuhn C, Schulze S, Kainer F and Friese K: Expression of 27 Kato T, Seki K, Matsui H and Sekiya S: Circulating inhibin inhibin/activin subunits alpha (-a), betaA (-‚A) and betaB (-‚B) forms in patients with hydatidiform mole. Gynecol Obstet in placental tissue of normal, preeclamptic and HELLP Invest 54: 114-117, 2002. pregnancies. Endocr Pathol 17(1): 19-33, 2006. 28 Pautier P, Ghione S, Brailly-Tabard S, Lhomme C, Morice P 12 Mylonas I, Schiessl B, Jeschke U, Vogl J, Makrigiannakis A, and Bidart JM: Are serum inhibin concentrations new markers Kuhn C, Schulze S, Kainer F and Friese K: Expression of of placental tumours in the course of chemotherapy? Hum inhibin/activin subunits alpha, betaA and betaB in placental Reprod 16: 2434-2437, 2001. tissue of normal and intrauterine growth restricted (IUGR) pregnancies. J Mol Histol 37(1-2): 43-52, 2006. 13 Petraglia F, Garuti GC, Calza L, Roberts V, Giardino L, Genazzani AR, Vale W and Meunier H: Inhibin subunits in Received December 13, 2006 human placenta: localization and messenger ribonucleic acid Revised February 28, 2007 levels during pregnancy. Am J Obstet Gynecol 165: 750-758, 1991. Accepted March 2, 2007

2000