Administration of βHCG leads to dose-dependent changes of gene expression signature of endometriotic stromal cells

Reproductive BioMedicine Online (2010) 20, 699– 706

www.sciencedirect.com www.rbmonline.com

ARTICLE

Administration of bHCG leads to dose-dependent changes of gene expression signature of endometriotic stromal cells ¨fler c, Gernot Hudelist a,b, Ambros Huber c, Michael Auer d, Martin Kno Leila Saleh c, Christian Hrachowitz e,f, Johannes C Huber g, Ernst Kubista Mahmood Manavi a, Martin Bilban h, Christian F Singer a,d,*

a,d

,

a

Division of Special Gynecology, Department of Obstetrics and Gynecology, Medical University of Vienna, Waehringer Guertel 18–20, A-1090 Vienna, Austria; b Department of OB/GYN, Center for Endometriosis, Villach General Hospital, Vienna, Austria; c Division of Obstetrics and Gynecology, Department of Obstetrics and Gynecology, Medical University of Vienna, Vienna, Austria; d Ludwig Boltzmann Institute for Clinical and Experimental Oncology, Vienna, Austria; e Department of Urology, Medical University of Vienna, Vienna, Austria; f Department of Obstetrics and Gynecology, Medical University of Vienna, Vienna, Austria; g Division of Endocrinology, Department of Obstetrics and Gynecology, Medical University of Vienna, Vienna, Austria; h Clinical Institute for Medical and Chemical Laboratory Diagnostics, Medical University of Vienna, Vienna, Austria * Corresponding author. E-mail address: [email protected] (CF Singer). Dr Singer is a board-certified gynaecologist and professor at the Medical University of Vienna. After graduation from Innsbruck University Medical School he successfully completed postdoctoral fellowships at Georgetown University (USA) and at the University of Leuven (Belgium), where he studied the influence of hormones on the development of malignant tumours. Professor Singer works and teaches at the Medical University of Vienna, where he has specialized in the treatment of breast cancer and the alleviation of climacteric symptoms. He was appointed Associated Professor of Obstetrics and Gynecology in 2004 and directs the clinic for hereditary breast and ovarian cancer at the Vienna General Hospital.

Abstract Preliminary studies have shown that systemic b-human chorionic gonadotrophin (bHCG) therapy alleviates endometriosis-

related chronic pelvic pain. The underlying mechanism, however, is completely unknown. This study has investigated the dosedependent alterations in the overall gene expression profile of endometriosis-derived stromal cells under increasing concentrations of bHCG by using the Affymetrix GeneChip U133 Set. It has been previously shown that bHCG concentrations of 0.1 U/ml and higher lead to a significant and dose-dependent increase in the expression of 68 genes. This study reports on a cluster analysis which identified three clusters of genes with a comparable expression pattern in response to increasing concentrations of bHCG. Most of the upregulated genes encoded proteins that are involved in cell adhesion, intercellular communication, extracellular matrix remodelling, apoptosis and inflammation. Stromal monocultures from eight patients, treated with and without 50 U/ml of bHCG, were then incubated and real-time polymerase chain reaction for the highly up-regulated genes PAI2, DUSP6, PLAU and MMP1 performed in order to validate the cDNA array findings in patients with endometriosis. Taken together, this study shows that bHCG induces dose-dependent

1472-6483/$ - see front matter ª 2010, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.rbmo.2010.01.010

700

G Hudelist et al.

characteristic response clusters in the gene expression profile of stromal cells obtained from endometriotic lesions which could explain the differential biological responses of bHCG in endometriosis. RBMOnline ª 2010, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. KEYWORDS: bHCG, cDNA array, endometriosis, gene expression signature

Introduction Endometriosis affects up to 15% of women in their reproductive years (Melis et al., 1994). Symptoms commonly associated with endometriosis include dysmenorrhoea, dyspareunia, dyschezia and infertility often confer a heavy social and economic burden (Ballard et al., 2006). The biological features of endometriotic cells resemble those characteristic for neoplastic cells: endometriotic tissue attaches to the peritoneal mesothelium and invades the extracellular matrix (ECM) associated with further infiltrative growth into adjacent anatomical structures (Chung et al., 2001). Laparoscopy and histological confirmation of disease is currently considered as the gold standard of diagnosis. Within this, a number of studies including randomized controlled trials provide evidence that excisional surgery is highly effective in treating endometriosis and associated symptoms such as chronic pelvic pain, infertility and subfertility (Jones and Sutton, 2003; Keckstein et al., 2003; Olive and Pritts, 2002). Nevertheless, several lines of evidence also show that endometriosis does recur following surgery in up to 20% of patients initially treated with surgical removal of endometriotic lesions (Kikuchi et al., 2006; Koga et al., 2006). In addition, persistence of symptoms following surgical removal of endometriosis may also indicate the presence of adenomyosis, rendering the majority of these patients unsuitable for surgery, i.e. hysterectomy, especially in cases where preservation of fertility is desired (Parker et al., 2006). Furthermore, a number of medical treatment strategies have been evaluated over the past decades in order to ameliorate endometriosis-associated symptoms. Gonadotrophin-releasing hormone (GnRH) agonists suppress ovarian oestrogen synthesis and are widely used in the treatment of pelvic pain but confer several side effects including depression, bone loss, decreased libido, sweats and infertility (Schroder et al., 2004). In contrast, oral contraceptives and gestagen-only preparations are generally well tolerated and do provide an efficient symptomatic treatment option for patients not aiming for pregnancy (Prentice et al., 2000). On the other hand, gestagens can also cause nausea, fluid retention and breakthrough-bleeding due to hypoestrogenaemia and may therefore be less well tolerated (ESHRE, 2008; Prentice et al., 2000). Physiological changes that confer a significant relief of symptoms are menopause and pregnancy. Whilst these effects can be explained by a physiological decrease in systemic oestrogens in menopausal patients, the reason for amelioration of symptoms during pregnancy in endometriosis patients is still unclear. It has been suggested that many of the beneficial effects of pregnancy are due to human chorionic gonadotrophin (bHCG). It has been recently demonstrated that human HCG has potent anti-endometriotic properties and therefore pain-reducing potential in vivo.

Injection of 1500–5000 IU HCG subcutaneously for 3–12 months reduced endometriosis-associated symptoms and improved quality of life in women who had been refractory to medical treatment strategies (Huber et al., 2004). While the clinical effects of parenteral bHCG preparations on endometriosis-related symptoms show a beneficial effect, the underlying biological mechanism of action is completely unknown. The consequences of bHCG treatment on endometriotic stromal cells on the RNA level have been previously reported and a number of genes have been identified that are up-regulated following the addition of bHCG to endometriotic fibroblasts (Huber et al., 2007). In order to further investigate the dose-dependent molecular changes that are associated with bHCG treatment, distinct clusters of genes have now been identified that show dose-dependent changes in expression levels by using Affymetrix GeneChips and validated these findings via real-time polymerase chain reaction (RT-PCR) for PAI2, DUSP6, PLAU, and MMP1 in nine patients with endometriosis.

Materials and methods Patients and samples Nine patients with histologically confirmed endometriosis were included in the present analysis. Patient characteristics and sample preparation has been described previously (Huber et al., 2007), and are shown here in Table 1.

CD56, CD45 and vimentin immunofluorescence The purity of fibroblast monocultures was confirmed by immunostaining for CD56 (mouse monoclonal antibody neural cell adhesion molecule 1 [NCAM-1], Ab-1 dilution 1:100; NeoMarkers, Fremont, CA, USA), CD45 (mouse monoclonal antibody CD45RB dilution 1:100; Dako Cytomation, Glostrup, Denmark) and vimentin (mouse monoclonal antibody M7020 dilution 1:100, Dako). A fraction of the cell suspension that was cultured for RNA isolation was plated on chamber slides (Greiner, Frickenhausen, Germany) and grown until sub-confluent. The cells were then washed with phosphate-buffered saline (PBS) and fixed in ice-cold 0.01% Triton X-100 (Sigma Chemical Co, USA) for 10 min at RT. After two additional washing steps with PBS, the primary antibodies (both diluted 1:100 in PBS containing 10 g/ml BSA fraction V) were added and incubated overnight at 4C. After three washing steps with PBS, the fluorescence-labelled secondary antibody (Alexa Fluor anti-mouse 488 for CD56 and CD45, and Alexa Fluor anti-mouse 546 for vimentin, 1:500 in both cases; Molecular Probes, Leiden, The Netherlands) was added and incubated for 30 min at RT. After three further washing steps with PBS, the specimens were

bHCG-induced alterations in gene expression profile in endometriosis Table 1

701

Patient characteristics.

Patient

Age (years)

Location

Dysmenorrhoea

Chronic pelvic pain

Prior treatment for endometriosis

Gravidity

Parity

Cycle phase

1 2 3 4 5 6 7 8 9

31 29 32 25 33 28 28 34 32

Peritoneal Peritoneal Peritoneal Peritoneal Peritoneal Peritoneal Peritoneal Peritoneal Peritoneal

Yes Yes Yes Yes Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes Yes Yes Yes Yes

0 0 0 0 0 0 0 0 0

1 Unknown 0 4 Unknown Unknown Unknown 2 0

1 Unknown 0 3 Unknown 1 0 1 0

Proliferative Proliferative Proliferative Proliferative Proliferative Proliferative Proliferative Proliferative Proliferative

Figure 1 Fluorescence immunohistochemistry of monocultures of stromal fibroblasts using (A) CD45 antibody, (C) CD56 antibody and (E) vimentin antibody. (B, D and F) Corresponding DAPI nuclear staining controls.

mounted in glycerol and covered with glass coverslips (Figure 1). Nuclei were stained with DAPI (Sigma). Control cells were incubated with replacement of the primary antibody by PBS (data not shown).

RNA extraction, amplification and GeneChip hybridization RNA extraction, amplification and GeneChip hybridization has been described previously (Huber et al., 2007). In brief,

double-stranded cDNA was synthesized from one patient by a chimeric oligonucleotide with an oligo-dT and a T7 RNA polymerase promoter at a concentration of 100 pmol/ll. Reverse transcription was performed as recommended by Affymetrix (Santa Clara, CA, USA) by using commercially available reagents (Invitrogen, UK). The reaction products were cleaned by phenol–chloroform extraction before biotin-labelling and an approximately 250-fold amplification were performed. 10 lg of each of the six labelled cRNA samples was then hybridized onto a separate Affymetrix U133

702 GeneChip set. The set consists of two GeneChip arrays and contains almost 45,000 probe sets representing more than 39,000 transcripts derived from approximately 33,000 well-substantiated human genes. The set design uses sequences selected from GenBank (www.ncbi.nlm.nih.gov, accessed 30 September 2009), dbEST (www.ncbi.nlm.nih. gov/dbEST, accessed 30 September 2009), and RefSeq (www.ncbi.nlm.nih.gov/RefSeq, accessed 30 September 2009). Prehybridization, hybridization, washing and staining with streptavidin–phycoerythrin were carried out according to the manufacturer’s recommendation. Antibody amplification was accomplished using a biotin-linked antistreptavidin antibody (Vector Laboratories, Buringame, CA, USA). The goat immunoglobulin G-blocking antibody (Sigma) was used for blocking non specific binding. The arrays were then stained and washed as recommended by Affymetrix, before being scanned on an Affymetrix GeneChip scanner (Agilent, Palo Alto, CA, USA). The raw, unnormalized signal intensity data were then analysed using the bioconductor software package (www.bioconductor.org, accessed 30 September 2009).

Statistical analysis Expression values were computed with the help of the robust-multiarray approach. This algorithm performs a background correction, a quantile normalization and finally computes normalized expression values (Irizarry et al., 2003). As replicate treatments were not available, fold changes were calculated. The untreated sample was chosen as reference. A fold change of two was chosen as cut-off value. This fold change has proved sufficient in identifying differentially expressed genes (Huynh et al., 2009). Eighty-six genes were identified and were then grouped with Ward’s hierarchical cluster algorithm. The Pearson correlation was used as distance measure. As demonstrated previously (Huber et al., 2007), six clusters were identified. Clusters 4, 5 and 6 were disregarded because they only contained eight genes and showed inconsistent results.

Real-time polymerase chain reaction Stromal cells from nine monocultures of stromal cells from peritoneal endometriotic lesions were isolated by collagenase treatment and grown in monocultures for 2–3 passages in the presence of FCS. Prior to the experiments, cells were washed and FCS replaced by serum-free media with or without 50 U/ml bHCG for 12 h. RNA was then extracted as previously described. Primers for PAI2 (HS00234032_m1), PLAU (HS00170182_m1), MMP1 (HS00899653_g1) and DUSP6 (HS00737962_m1) were purchased from Applied Biosystems (Foster City, CA, USA). The GAPDH primer set (HS99999905_m1) was used as internal control. First-strand cDNA was synthesized using 1 g of total RNA (DNase-treated) in a 20 ll reverse transcriptase reaction mixture. Serial 10fold dilutions (104 to 109 molecules) of GAPDH were used as a reference molecule for the standard curve calculation. All RT-PCR reactions were performed in a 25 ll mixture containing 1/20 volume of cDNA preparation (1 ll), 1· SYBR Green buffer (Applied Biosystems), 4 mmol/l MgCl2, 0.2 lmol/l of each primers, 0.2 mmol/l dNTP mix and 0.025 U of AmpliTaq Gold thermostable DNA polymerase

G Hudelist et al. (Applied Biosystems). Real-time quantifications were performed using RT-PCR performed in an ABI PRISM 7000 Sequence Detection System thermal cycler (Applied Biosystems) using the Taqman Universal PCR Master Mix protocol (Applied Biosystems).

Results CD56, CD45 and vimentin immunofluorescence Figure 1 demonstrates the results of the negative (CD45, CD56) and positive (vimentin) immunofluorescence staining of monocultures of stromal fibroblasts using a CD45, CD56 antibody and a vimentin antibody, with corresponding DAPI nuclear staining controls thereby demonstrating the purity of fibroblast monocultures.

Changes in gene expression pattern in response to increasing bHCG concentrations Fibroblasts obtained from peritoneal endometriotic lesions were incubated with increasing concentrations (0, 0.1, 1, 10, 50 and 100 IU/ml) of bHCG for 12 h. After extraction of total RNA, a separate analysis of the gene expression signature of each of the six conditions was performed by using the Affymetrix U133 GeneChip. A previous study identified 68 genes that showed a consistent and dose-dependent up-regulation of gene expression in response to increasing doses of bHCG (Huber et al., 2007). In order to further evaluate the dose-dependent effects of bHCG treatment, the previous analysis was expanded by hierarchical clustering which allowed the identification of three clusters of gene expression which were characterized by comparable changes in mRNA transcription in response to increasing doses of bHCG (Figure 2). A total of 23 genes exhibited a dose-dependent increase in their expression that was beginning to become evident at concentrations of 0.1 U/ml bHCG (cluster 1). These included genes such as those for epithelial membrane protein 1 (EMP1) and matrix metalloproteinase 1 (MMP1), tissue factor pathway inhibitor 2 (TFPI2) and coagulation factor II (thrombin) receptor-like 1 (F2RL1) which are known to be involved in ECM remodelling and cell adhesion processes. A further increase in bHCG concentrations 1 U/ml and higher resulted in the identification of 34 genes that showed a consistent and dose-dependent increase in their expression (cluster 2). Genes associated with cluster 2 included those for podocalyxin-like protein (PODXL), tissue plasminogen activator (PLAT) urokinase plasminogen activator (PLAU), urokinase plasminogen activator receptor (PLAUR), EPH receptor B6 (EPHB6) and endothelial cell-specific molecule 1 (ESM1) which again have been shown to play important roles in cell adhesion and attachment. Up-regulated genes involved in cell cycle control according to cluster 2 included CDC6 (cell division cycle 6 homologue), cyclin E2 (CCNE2) and Ras-induced senescence 1 (RIS1). Furthermore, the genes reacted to anti-inflammatory proteins pentaxin-related gene (PTX3) and interleukin 13 receptor alpha 2 (IL13RA2) exhibited elevations under bHCG treatment with concentrations 1 U/ml and higher (cluster 2). Finally, a cluster of 14 genes (cluster 3) was also identified with a statistically significant and dose-dependent

bHCG-induced alterations in gene expression profile in endometriosis

703

Figure 2 (Left panel) Identification of clusters of genes with a similar change in mRNA expression in response to increasing concentrations of HCG after 12 h of exposure: lane S_1a, 0 U/ml; S_2a, 0.1 U/ml; S_3a, 1 U/ml; S_4a, 10 U/ml; S_5a, 50 U/ml; S_6, 100 U/ml. (Right panel) Corresponding logarithmic representation of gene expression: lane 1, 0 U/ml; 2, 0.1 U/ml; 3, 1 U/ml; 4, 10 U/ml; 5, 50 U/ml; 6, 100 U/ml. Each line represents mRNA expression of one gene. Genes depicted in red represent downregulated genes, while genes depicted in blue represent up-regulated genes. Some genes, such as EMP1, are represented in both clusters 1 and 3 because, in a small number of cases, the analysis software attributes expression levels that show a borderline expression pattern into two categories. Some genes appear more than once in the same cluster due to slightly different results from replicate measurements causing the software to print each of the slightly different arrays.

704 up-regulation of gene expression at bHCG concentrations of 10 U/ml and more (P = 0.0242, Student’s t-test). Cluster 3 included genes such as proapoptotic bone morphogenic protein (BMP2), integrin b5 (ITGB5) and anti-inflammatory interleukin 11 (IL-11). Furthermore, genes of cluster 3 implicated in down-stream receptor signalling were E3 ubiquitin ligase SMURF2 (SMURF2) and dual specificity phosphatase 6 (DUSP6). Percentage distribution of gene clusters according to their cellular function are depicted in Figure 3.

RT-PCR validation experiments In order to validate the gene expression pattern obtained by cDNA arrays in other patients, four significantly up-regulated transcripts were chosen for RT-PCR: PLAU (P = 0.0057, Student’s t-test), MMP1 (P = 0.0138), PAI2 (P = 0.0301) and DUSP6 (P = 0.0086). Nine samples of total RNA from stromal cells that had or had not been treated with 50 U/ml bHCG were tested. The expression was normalized compared with the housekeeping gene GAPDH. The results of the RT-PCR are shown in Figure 4. While mRNA for both proteins was detected in all samples, the data showed a clear up-regulation of PLAU under the influence of bHCG in eight of the nine patients (Figure 4C). In the case of MMP1, gene expression was also increased in the stromal cells from six of the nine patients investigated (Figure 4D). The PAI2 gene expression was similarly stimulated by HCG in seven of the nine patients (Figure 4A), and DUSP6 also clearly increased in six of the nine patients (Figure 4B).

Discussion The present work further demonstrates that treatment of endometriotic fibroblasts does indeed cause several changes in the gene expression pattern of these cells which may also reflect in vivo alterations of certain genes in response to

Figure 3 Bar graph reflecting percentage distribution (Y-axis) of cluster 1 (light grey bar), cluster 2 (dark grey bar), cluster 3 (white bar) according to their function (X-axis) in cell adhesion and communication (1), extracellular matrix remodelling (2), proapoptotic mechanisms (3), anti-inflammatory processes (4), down-stream receptor signalling (5), cell cycle control (6) and cell growth and transcriptional processes (7).

G Hudelist et al. bHCG administration in patients with endometriosis. Interestingly, bHCG treatment was associated with a dosedependent increase of certain gene transcripts which further allowed the identification of three clusters of gene sets. Concentrations of 0.1 U/ml of bHCG led to the up-regulation of genes that have been implicated in cell signalling and adhesion such as EMP1 and MMP1. In relation to this, transfection of human tumour cell lines with, and subsequent expression of EMP1 results in a direct inhibition of cell proliferation, but is also associated with an increase in other gene transcripts that are mostly related to cell signalling, cell communication and particularly to adhesion (Wang et al., 2003). The detection of ECM-degrading proteases such as MMP1 in endometriotic lesions has led to the suggestion that these enzymes are associated with the development and pathogenesis of endometriosis (Gottschalk et al., 2000; Hudelist et al., 2005). However, proteases also have an important role in physiological tissue turnover and wound healing (Pilcher et al., 1998). Furthermore, the selective and tightly regulated expression of MMP1 in the human endometrium during the premenstrual phase results in demarcation and sequestration of functional endometrium during menstrual bleeding (Kokorine et al., 1996). In the peritoneum, MMP1 and its inhibitor are involved in the repair of serosal tissues in response to pelvic surgery, and the tissue plasminogen activator PLAT provides the potential to resolve post-operative fibrinous collections in peritoneal lesions – even at sites at which the mesothelial cells have been injured, removed or destroyed (Bruse et al., 2004; Chegini et al., 2001). Cluster analysis of genes progressively up-regulated at concentrations of 1 U/ml bHCG and higher (cluster 2) included gene transcripts of receptor subtype B6 (EPHB6), PLAU, its receptor PLAUR and the antiinflammatory genes such as for interleukin 13 receptor antagonist 13 (IL13RA2). EPH receptors and ephedrins have been shown to act as regulators of repulsion and adhesion of cellular organization (Bruse et al., 2004). The EPH receptor subtype B6 gene (EPHB6) is physiologically expressed in normal human tissues, but is down-regulated in invasive tumours, thereby presumably adding to their invasive phenotype (Fox et al., 2006). These observations nicely fit the finding of increased stromal EPHR6 in response to bHCG and suggest a causal role of bHCG in inhibition of ECM degradation by endometriotic cells. Similarly, the beneficial effect of bHCG could further be explained by the induction of the anti-inflammatory genes such as IL13RA2. IL13RA2 has been demonstrated to neutralize the pro-inflammatory action of IL13, a pivotal mediator of cytokine-dependent diseases (Mentink-Kane and Wynn, 2004). A further increase in bHCG finally led to identification of a third cluster of genes including transcripts of apoptosis-associated genes such as BMP and integrin 5. Both proteins have been implicated in inhibitory mechanisms of cell growth and induction of apoptosis (Kawamura et al., 2000; Yin et al., 1999) which might thereby explain growth-inhibitory effects of bHCG on endometriotic implants. In conclusion, the present work demonstrates that administration of bHCG confers alteration of expression levels of a number of genes involved in mechanisms of apoptosis, reorganization of the ECM and inflammation. These changes appear to be dose-dependent reflected by the cluster of genes identified and confirmed by PCR analysis. As a

bHCG-induced alterations in gene expression profile in endometriosis

705

Figure 4 Real-time polymerase chain reaction for (A) PAI2, (B) DUSP6, (C) PLAU and (D) MMP1 performed on total RNA extracted from nine monocultures of stromal cells obtained from peritoneal endometriotic lesions which were incubated without (h) or with (n) 50 U/ml bHCG for 12 h.

consequence, this study suggests that the clinical effects of bHCG in patients with endometriosis, either in pregnancy or following the therapeutic administration of bHCG could partly be explained by these alterations at the transcriptional level.

Acknowledgements The authors acknowledge the technical support provided by Barbara Weidinger and Ernst Ru ¨cklinger for support with statistical analysis. This work has been supported by a grant by AMGEN and by a grant by the Jubila ¨umsfonds des Bu ¨rgermeisters der Bundeshauptstadt Wien.

References Ballard, K., Lowton, K., Wright, J., 2006. What’s the delay? A qualitative study of women’s experiences of reaching a diagnosis of endometriosis. Fertil. Steril. 86, 1296–1301. Bruse, C., Radu, D., Bergqvist, A., 2004. In situ localization of mRNA for the fibrinolytic factors uPA, PAI-1 and uPAR in endometriotic and endometrial tissue. Mol. Hum. Reprod. 10, 159–166. Chegini, N., Kotseos, K., Zhao, Y., et al., 2001. Expression of matrix metalloproteinase (MMP1) and tissue inhibitor of MMP in serosal tissue of intraperitoneal organs and adhesions. Fertil. Steril. 76, 1212–1219. Chung, H.W., Wen, Y., Chun, S.H., Nezhat, C., Woo, B.H., Lake Polan, M., 2001. Matrix metalloproteinase-9 and tissue inhibitor of metalloproteinase-3 mRNA expression in ectopic and eutopic endometrium in women with endometriosis: a rationale for endometriotic invasiveness. Fertil. Steril. 75, 152–159. ESHRE, 2008. The ESHRE Guideline on Endometriosis 2008. Fox, B.P., Tabone, C.J., Kandpal, R.P., 2006. Potential clinical relevance of Eph receptors and ephrin ligands expressed in

prostate carcinoma cell lines. Biochem. Biophys. Res. Commun. 342, 1263–1272. Gottschalk, C., Malberg, K., Arndt, M., et al., 2000. Matrix metalloproteinases and TACE play a role in the pathogenesis of endometriosis. Adv. Exp. Med. Biol. 477, 483–486. Huber, A.V., Huber, J.C., Kolbus, A., et al., 2004. Systemic HCG treatment in patients with endometriosis: a new perspective for a painful disease. Wien. Klin. Wochenschr. 116, 839–843. Huber, A., Hudelist, G., Knofler, M., Saleh, L., Huber, J.C., Singer, C.F., 2007. Effect of highly purified human chorionic gonadotropin preparations on the gene expression signature of stromal cells derived from endometriotic lesions: potential mechanisms for the therapeutic effect of human chorionic gonadotropin in vivo. Fertil. Steril. 88, 1232–1239. Hudelist, G., Lass, H., Keckstein, J., et al., 2005. Interleukin 1alpha and tissue-lytic matrix metalloproteinase-1 are elevated in ectopic endometrium of patients with endometriosis. Hum. Reprod. 20, 1695–1701. Huynh, K.M., Kim, G., Kim, D.J., et al., 2009. Gene expression analysis of terminal differentiation of human melanoma cells highlights global reductions in cell cycle-associated genes. Gene 433, 32–39. Irizarry, R.A., Hobbs, B., Collin, F., et al., 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4, 249–264. Jones, K.D., Sutton, C., 2003. Patient satisfaction and changes in pain scores after ablative laparoscopic surgery for stage III–IV endometriosis and endometriotic cysts. Fertil. Steril. 79, 1086–1090. Kawamura, C., Kizaki, M., Yamato, K., et al., 2000. Bone morphogenetic protein-2 induces apoptosis in human myeloma cells with modulation of STAT3. Blood 96, 2005–2011. Keckstein, J., Ulrich, U., Kandolf, O., Wiesinger, H., Wustlich, M., 2003. Laparoscopic therapy of intestinal endometriosis and the ranking of drug treatment. Zentralbl. Gynakol. 125, 259–266. Kikuchi, I., Takeuchi, H., Kitade, M., Shimanuki, H., Kumakiri, J., Kinoshita, K., 2006. Recurrence rate of endometriomas following a laparoscopic cystectomy. Acta Obstet. Gynecol. Scand. 85, 1120–1124.

706 Koga, K., Takemura, Y., Osuga, Y., Yoshino, O., Hirota, Y., Hirata, T., Morimoto, C., Harada, M., Yano, T., Taketani, Y., 2006. Recurrence of ovarian endometrioma after laparoscopic excision. Hum. Reprod. 21, 2171–2174. Kokorine, I., Marbaix, E., Henriet, P., Okada, Y., Donnez, J., Eeckhout, Y., Courtoy, P.J., 1996. Focal cellular origin and regulation of interstitial collagenase (matrix metalloproteinase1) are related to menstrual breakdown in the human endometrium. J. Cell Sci. 109, 2151–2160. Melis, G.B., Ajossa, S., Guerriero, S., Paoletti, A.M., Angiolucci, M., Piras, B., Caffiero, A., Mais, V., 1994. Epidemiology and diagnosis of endometriosis. Ann. NY Acad. Sci. 734, 352– 357. Mentink-Kane, M.M., Wynn, T.A., 2004. Opposing roles for IL-13 and IL-13 receptor alpha 2 in health and disease. Immunol. Rev. 202, 191–202. Olive, D.L., Pritts, E.A., 2002. The treatment of endometriosis: a review of the evidence. Ann. NY Acad. Sci. 955, 360–372, discussion 389–93, 396–406. Parker, J.D., Leondires, M., Sinaii, N., Premkumar, A., Nieman, L.K., Stratton, P., 2006. Persistence of dysmenorrhea and nonmenstrual pain after optimal endometriosis surgery may indicate adenomyosis. Fertil. Steril. 86, 711–715.

G Hudelist et al. Pilcher, B.K., Sudbeck, B.D., Dumin, J.A., Welgus, H.G., Parks, W.C., 1998. Collagenase-1 and collagen in epidermal repair. Arch. Dermatol. Res. 290 (Suppl.), S37–S46. Prentice, A., Deary, A.J., Bland, E., 2000. Progestagens and antiprogestagens for pain associated with endometriosis. Cochrane Database Syst. Rev., CD002122. Schroder, A.K., Diedrich, K., Ludwig, M., 2004. Medical management of endometriosis: a systematic review. IDrugs 7, 451–463. Wang, H.T., Kong, J.P., Ding, F., Wang, X.Q., Wang, M.R., Liu, L.X., Wu, M., Liu, Z.H., 2003. Analysis of gene expression profile induced by EMP-1 in esophageal cancer cells using cDNA microarray. World J. Gastroenterol. 9, 392–398. Yin, L., Fu, S., Zhao, X., Garcia-Sanchez, F., Deisseroth, A.B., 1999. Correlation of cell proliferation inhibition and apoptosis induction with expression of human beta 5 integrin on hematopoietic cells. Chin. Med. J. (Engl.) 112, 659–664. Declaration: The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported. Received 11 December 2008; refereed 29 January 2009; accepted 10 December 2009.