| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Research Resource |
Institut Cochin (B.B., F.M., T.-M.M., D.V., C.C.), Université Paris Descartes, Centre National de la Recherche Scientifique (Unité Mixte de Recherche 8104), and Institut National de la Santé et de la Recherche Médicale (B.B., F.M., T.-M.M., D.V., C.C.), Unité 567, 75014 Paris, France; Service de Gynécologie-Obstétrique 2 et Médecine de la Reproduction (B.B., C.C.), Centre Hospitalier Universitaire Cochin Saint-Vincent de Paul, Assistance Publique-Hôpitaux de Paris, 75674 Paris, France; and Department of Pathology (J.-C.N., I.F.), Erasme University Hospital, Free University of Brussels, B-1070 Brussels, Belgium
Address all correspondence and requests for reprints to: Dr. Bruno Borghese, Equipe 21, Département Génétique et Développement, Institut Cochin, 24 rue du Faubourg Saint-Jacques, 75014 Paris, France. E-mail: bruno.borghese{at}inserm.fr.
ABSTRACT
Endometriosis is a common gynecological disorder characterized by pain and infertility, where the lesions disseminate everywhere in the body with a preference for the pelvis. In that, it could be regarded as a benign metastatic disease, because its issue is not fatal. However, the molecular bases of this intriguing clinical condition are not well known. The objective of this study is to characterize the transcriptome differences between eutopic vs. ectopic endometrium with a special interest in pathways involved in cancerogenesis. We performed two hybridizations in technical replicate on highly specific long oligonucleotides microarrays (NimbleGen), with cDNA prepared from six-patients pools, where the same patient provided both eutopic and ectopic endometrium (endometriomas). To confirm the expression microarrays data, quantitative RT-PCR validation was performed on 12 individuals for 20 genes. Over 8000 transcripts were significantly modified (more than twice) in the lesions corresponding to 5600 down- or up-regulated genes. These were clustered through DAVID Bioinformatics Resources into 55 functional groups. The data are presented in a detailed and visual way on 24 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways implemented with induction ratios for each differentially expressed gene. An outstanding control of the cell cycle and a very specific modulation of the HOX genes were observed and provide some new evidence on why endometriosis only very rarely degenerates into cancer. The study constitutes a noteworthy update of gene profiling in endometriosis, by delivering the most complete and reliable list of dysregulated genes to date.
ENDOMETRIOSIS IS A COMMON gynecological disease affecting up to 10% of the women during their reproductive life (1). It is characterized by the ectopic localization of endometrial tissue at various sites in the general cavity of the body. The two main symptoms, including chronic pain and infertility, greatly deteriorate the quality of life of the patients. The ectopic endometrial tissue behaves similarly to eutopic endometrium. It displays the same response to endocrine stimuli, such as bleeding during menses, and explains why painful symptoms reemerge at this time. From the histological point of view, endometriotic lesions form an intricate network of glandular and stromal structures mixed with vascular, inflammatory, and fibromuscular elements (2).
Despite its high prevalence, the etiology of the disease is not well understood. The most commonly admitted mechanism by which endometriotic lesions arise is related to a retrograde flux of endometrial cells through the oviducts (3). In women with specific genetic backgrounds, complex networks of genes probably contribute to abnormal implantation of endometrial tissue at ectopic sites (4).
The most striking feature concerning deep endometriosis is its ability to invade the surrounding tissues and sometimes to metastasize in lymph nodes (5) and beyond the abdominal cavity (6). The lesions are clonal, originating from single cells implanting at abnormal locations (7). Accordingly, endometriosis closely resembles cancer. However, in endometriosis, the invasion is controlled and stops at a given time. Therefore, the issue of the disease is not fatal, which grants endometriosis the original and unique status of being a benign metastatic disease.
Understanding endometriosis is a perplexing challenge, considering the heterogeneity and complexity of the disease. Indeed, at least three major questions must be answered. 1) What are the molecular specificities of the lesions compared with the normal endometrium in a given patient? 2) What are the physiological and genetic differences between a patient that will develop endometriosis and a patient that will never do? And 3) what are the links between endometriosis and ovarian cancer and why endometriosis rarely degenerates into cancer (8)?
Facing the complexity of such a disease, high-throughput tools are presumably adequate to directly pinpoint pertinent gene targets. In the last few years, several studies have compared eutopic vs. ectopic endometrium using the whole tissue (9, 10, 11, 12) or laser-capture microdissection (LCM) to individualize the different cells contained in the lesion (13, 14). Up to now, the results of these studies were relatively discrepant, highlighting the lack of reproducibility of microarray technologies.
In the present study, we used one of the most advanced gene expression profiling tools available to date to better understand what the molecular differences between endometriotic cells, endometrial cells, and neoplastic cells. The NimbleGen platform combines long isothermic oligonucleotides (60-mers) and permits screening of a large number of genes (47,633 transcripts), each being represented by eight oligonucleotides. The existence of several transcripts per gene makes it possible to discriminate expression levels of specific isoforms as well as to clearly distinguish between genes belonging to a given gene family. We reported here for the first time a systematic alteration of all HOX gene clusters expression in endometriosis. We noted also wide alterations of gene expression in 24 biological pathways, some of them being involved in oncogenesis.
RESULTS
The complete data set for these DNA microarrays has been deposited in the NCBI GEO database under accession number GSE12768. The reliability of the NimbleGen platform was confirmed by calculating the Pearson correlation coefficient for the data obtained from the two hybridizations (r = 0.94) and from the quantitative RT-PCR patterns of expression (r = 0.97) (supplemental Fig. 1, published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org).
We counted 2782 down-regulated genes (4174 transcripts) and 2823 up-regulated genes (3836 transcripts). The induction ratios ranged from 0.003–368. The 100 most differentially expressed transcripts are listed in supplemental Table 2 (published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org). The use of eight oligonucleotides per gene permitted us to distinguish specific isoforms and to analyze alternative splicing.
DAVID clusterized induced genes into 27 groups and repressed genes into 28 groups with an enrichment score above a simulation base calculated threshold (3.90 for the repressed genes and 3.60 for the induced genes) (supplemental Figs. 2 and 3). Among the most enriched clusters, the overexpressed extracellular matrix genes (570 genes including COL1A1 x11.5, THBS4 x8.0, ITGA11 x15.0, MMP26 ÷300.0, MMP23A x15.9, and MMP23B x20.6) and the down-regulated genes involved in cytoplasm motility (Golgi apparatus, tubulin and actin cytoskeleton, and protein transport) are the most remarkable.
Twenty-four Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways selected by DAVID for their statistical relevance have been implemented with induction ratios for each gene, providing a detailed and visual display of transcriptomic alterations in endometriosis (Fig. 1
, detail available as supplemental data). Some pathways are modified in a univocal way, such as steroid hormone metabolism (supplemental data) or cell cycle (Fig. 2). Others attest to a more complex regulation, such as Wnt signaling pathway (supplemental data).
|
|
). Genes nearby the clusters are not modified. In addition, inside the clusters, gene expression is linearly dysregulated according to the position of the HOX gene (clusters HOX-A and HOX-C) (Fig. 3).
|
This study makes available the most advanced and reliable data set of genetic alterations in endometriotic lesions. In particular, we reported two molecular characteristics of endometriomas that may contribute to protect the lesions from degenerating into ovarian cancer.
For the first time, we described a quasi-systematic alteration of all HOX genes clusters in endometriosis (Fig. 3
), whereas only HOX-A10 has been extensively studied (15, 16, 17, 18). HOX genes are evolutionarily conserved genes encoding transcriptional regulators that orchestrate embryonic development. HOX-A genes seem to be essential for adult cyclic endometrial development and for endometrial receptivity (19). Recently, HOX genes have been reported to be up-regulated in ovarian cancer (20, 21). In the present study, where long oligonucleotides permit us to differentiate unambiguously genes belonging to gene families, we found systematic alterations that were dissimilar according to the HOX cluster taken into consideration (down-regulation of HOX-A and HOX-B and up-regulation of HOX-C and, to a lesser extent, of HOX-D). These findings, opposite to those observed in ovarian cancer (20, 21), might therefore be associated with somewhat of a protection against malignant transformation of endometriotic cells.
Moreover, we observed a systematic down-regulation of genes involved in the cell cycle (Fig. 2). As a consequence, endometrioma seems to be quiescent, containing nondividing cells. This is consistent with two observations: 1) the up-regulation of cell adhesion controlling genes that could explain the extreme compactness of the lesions and 2) the down-regulation of intracytoplasmic traffic genes that could result in a setting at rest cellular metabolism. These conditions might decrease the risks of malignant transformation.
These findings are reinforced by an outstanding and original correlation between our data and those reported by the most extensive study to date, carried out on a GE Healthcare platform of 53,000 human transcripts (12). A total of 694 of 717 (96.7%) genes significantly modulated in this study were present in our NimbleGen array analysis (the others being lost due to defective annotation) with a consistency rate between the two experiments close to 81%. This rate is very striking because the two studies were performed in different laboratories on two different platforms and, of course, using a different set of patients. In our opinion, it gives much reliability to our results, which supply a list of over 5600 genes implicated in endometriosis, the largest one to date.
Contrary to certain studies (13, 14), we deliberately chose not to use LCM, although it could isolate specific cells by eliminating unwanted ones to give pure enriched cell populations. Some considerations persuaded us to reject LCM in this particular case: 1) following Eyster et al. (12), we estimated that all cell types in the endometriotic lesions contribute to the disease; 2) the study was designed to identify genes that are the most commonly modulated in endometriosis, without taking into account either cell types (no LCM) or individual specificities (use of DNA pools, same patient for case and control); 3) the LCM procedure has potential drawbacks such as affecting the RNA quality, resulting in bias in the gene profile (22); and 4) LCM is time consuming and too costly in the optics of a prognostic and individual application of microarrays in routine medical practice.
The data presented here provide a comprehensive and noteworthy update of gene profiling in endometriosis. The results are in favor of close but distinct features between endometriosis and cancer. Biologically, endometriosis resembles cancer less than it does clinically. Some control points seem to prevent the endometriotic cells from getting a more aggressive phenotype. This may be beneficial for research and therapy in the fields of endometriosis and cancer.
MATERIALS AND METHODS
Patients and Tissue Collection
We collected 12 samples of endometriomas and 12 samples of eutopic endometrium from 12 Caucasian patients in luteal phase, operated for painful stage IV endometriosis (23). The same patient provided both eutopic and ectopic endometrium to minimize the genetic heterogeneity. No patients were receiving hormone therapy at the time of surgery. All samples were histologically proven. The local ethics committee [Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale (CCPPRB) of Paris-Cochin] approved the study protocol, and all patients gave an informed consent.
RNA Extraction, cDNA Synthesis, cDNA Pooling, and Microarray Hybridizations
After surgical resection, the samples were immediately frozen in liquid nitrogen. Total RNA was extracted with Trizol (Invitrogen, Carlsbad, CA). RNA quality was checked by agarose gel electrophoresis and spectrophotometry. cDNA synthesis was performed with the SuperScript double-stranded cDNA synthesis kit (Invitrogen). The first cDNA pool was composed of six endometriomas, the second one by the endometria from the same patients and shipped to the NimbleGen platform in Reykjavik (Iceland). Labeling, hybridization, data collection, and normalization were carried out according to NimbleGen protocols. A same experiment was repeated with the six remaining samples. We decided to take into consideration genes modified more than 2-fold for further analysis.
Quantitative Real-Time PCR
A set of 20 genes, including nine induced, six repressed, and five nonmodified genes, was analyzed by quantitative RT-PCR using DNA from each individual patient. Primers for RT-PCR analysis (Supplemental Table 1) were chosen using the PRIMER3 software. Quantitative PCR was carried out on a Light Cycler thermocycler (Roche, Basel, Switzerland) using the amplification kit Platinum SYBR Green (Invitrogen). Results were analyzed with the LightCycler software using the second derivative method.
Functional Clustering
Two lists of genes (induced or repressed twice or more) were submitted separately to the DAVID database (http://david. abcc.ncifcrf.gov), which clusterizes genes according to a series of common keywords. The proportion of each keyword in the list is compared with the one in the whole genome, making it possible to compute P values and enrichment scores (geometric mean of the inverse log of each P value). We used a simulation to define minimal enrichment thresholds enabling us to consider gene clusters given by DAVID for further analysis (24). The detailed information of gene alterations was systematically reported on KEGG pathways (http://www.genome.ad.jp/kegg).
ACKNOWLEDGMENTS
We thank Dr. Jean Gogusev for critical reading of the manuscript.
FOOTNOTES
Disclosure Statement: All authors have nothing to declare.
First Published September 25, 2008
Abbreviations: KEGG, Kyoto Encyclopedia of Genes and Genomes; LCM, laser-capture microdissection.
Received for publication September 2, 2008. Accepted for publication September 9, 2008.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  E. M. C. Ohlsson Teague, C. G. Print, and M. L. Hull The role of microRNAs in endometriosis and associated reproductive conditions Hum. Reprod. Update, September 22, 2009; (2009) dmp034v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. C. Lin, M. Suzawa, R. D. Blind, S. C. Tobias, S. E. Bulun, T. S. Scanlan, and H. A. Ingraham Stimulating the GPR30 Estrogen Receptor with a Novel Tamoxifen Analogue Activates SF-1 and Promotes Endometrial Cell Proliferation Cancer Res., July 1, 2009; 69(13): 5415 - 5423. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
