Leiomyoma and Myometrial Gene Expression Profiles and Their Responses to Gonadotropin-Releasing Hormone Analog Therapy
http://www.100md.com
内分泌学杂志 2005年第3期
Department of Obstetrics and Gynecology, University of Florida, Gainesville, Florida 32610
Address all correspondence and requests for reprints to: Dr. Nasser Chegini, Department of Obstetrics and Gynecology, University of Florida, Box 100294, Gainesville, Florida 32610. E-mail: cheginin@obgyn.ufl.edu.
Abstract
Gene microarray was used to characterize the molecular environment of leiomyoma and matched myometrium during growth and in response to GnRH analog (GnRHa) therapy as well as GnRHa direct action on primary cultures of leiomyoma and myometrial smooth muscle cells (LSMC and MSMC). Unsupervised and supervised analysis of gene expression values and statistical analysis in R programming with a false discovery rate of P 0.02 resulted in identification of 153 and 122 differentially expressed genes in leiomyoma and myometrium in untreated and GnRHa-treated cohorts, respectively. The expression of 170 and 164 genes was affected by GnRHa therapy in these tissues compared with their respective untreated group. GnRHa (0.1 μM), in a time-dependent manner (2, 6, and 12 h), targeted the expression of 281 genes (P 0.005) in LSMC and MSMC, 48 of which genes were found in common with GnRHa-treated tissues. Functional annotations assigned these genes as key regulators of processes involving transcription, translational, signal transduction, structural activities, and apoptosis. We validated the expression of IL-11, early growth response 3, TGF-?-induced factor, TGF-?-inducible early gene response, CITED2 (cAMP response element binding protein-binding protein/p300-interacting transactivator with ED-rich tail), Nur77, growth arrest-specific 1, p27, p57, and G protein-coupled receptor kinase 5, representing cytokine, common transcription factors, cell cycle regulators, and signal transduction, at tissue levels and in LSMC and MSMC in response to GnRHa time-dependent action using real-time PCR, Western blotting, and immunohistochemistry. In conclusion, using different, complementary approaches, we characterized leiomyoma and myometrium molecular fingerprints and identified several previously unrecognized genes as targets of GnRHa action, implying that local expression and activation of these genes may represent features differentiating leiomyoma and myometrial environments during growth and GnRHa-induced regression.
Introduction
LEIOMYOMAS ARE BENIGN uterine tumors originating from the transformation of myometrial smooth muscle cells and/or connective tissue fibroblasts during the reproductive years. Despite recent progress in our understanding of leiomyoma’s molecular environment, the identity of a molecules(s) responsible for initiating the transformation of these cells into leiomyoma remains unknown. However, ovarian steroids are essential for leiomyoma growth, and creating a hypoestrogenic condition with GnRH analog (GnRHa) therapy is often used for their medical management (1, 2, 3, 4, 5, 6, 7). Hypoestrogenic conditions created by GnRHa therapy affect both leiomyoma and myometrium; however, clinical observations indicate a difference in their responses to changes in the hormonal milieu (7). In addition to GnRHa therapy, clinical and preclinical assessments of selective estrogen and progesterone receptor modulators, either alone or in combination with GnRHa, have shown efficacy in leiomyoma regression (4, 5, 6).
With respect to the leiomyoma molecular environment, several genome-wide allele-typing studies have evaluated the association between genomic instability and the pathogenesis of leiomyoma (for review, see Ref.8). These studies have led to the identification of several candidate genes; however, in the majority of cases evidence of genomic instability is either lacking or inconsistent, implying that the existence of unrecognized pathways leads to the development of leiomyoma. Additional studies have provided support for various autocrine/paracrine regulators in the pathogenesis of leiomyoma, including local estrogen production, growth factors, cytokines, chemokines, and their receptors, whose expression are regulated by ovarian steroids (1, 2). These studies in many instances demonstrated altered expression of these factors and/or their receptors in leiomyoma compared with normal myometrium. In recent years, cDNA microarray has been used as a high throughput method to identify a large number of differentially expressed and regulated genes in various tissues and cells. Using this approach, several recent studies, including our own, have also assisted in fingerprinting the gene expression profile of leiomyoma and myometrium during the menstrual cycle (9, 10, 11, 12, 13, 14). However, only the expression of a few of these newly identified genes has been validated, and their regulation and correlation with pathogenesis of leiomyoma remain to be investigated.
With respect to GnRHa therapeutic action, it is traditionally believed to act primarily at the level of the pituitary-gonadal axis, and by suppressing ovarian steroid production, it causes leiomyoma regression. However, the identification of GnRH and GnRH receptor expression in several peripheral tissues, including the uterus, has indicated an autocrine/paracrine role for GnRH and additional sites of action for GnRHa therapy (15, 16, 17, 18, 19, 20). We provided support for this concept by demonstrating the expression of GnRH as well as GnRH I and II receptors mRNA in leiomyoma and myometrium and their isolated smooth muscle cells (15, 16). Several in vitro studies, including our own, have also demonstrated GnRHa direct action on various cell types derived from peripheral tissues resulting in alteration of cell growth, apoptosis, and the expression of cell cycle proteins, growth factors, pro- and antiinflammatory cytokines, proteases, and protease inhibitors (1, 16, 17, 18, 19, 20, 21, 22, 23). Local expression and differential regulation of these genes influence cell proliferation, differentiation, migration, inflammatory response, angiogenesis, expression of adhesion molecules, extracellular matrix (ECM) turnover, apoptosis, etc., processes that are central to leiomyoma growth and regression (1, 2, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32). Microarray studies, including our own small scale array study, have also identified the expression profile of additional genes targeted by GnRHa in the murine gonadotrope tumor cell line L?T2, the human breast tumor cell line MCF-7, leiomyoma, and myometrium (10, 33, 34).
We designed the present study to further define the molecular environments of leiomyoma and matched myometrium during the early to midluteal phase of the menstrual cycle, which is characterized by elevated production of ovarian steroids, compared with tissues obtained from hormonally suppressed patients receiving GnRHa therapy. We also evaluated the direct action of GnRHa on global gene expression and their regulation in leiomyoma and myometrial cells isolated from the untreated tissue cohort. These approaches enabled us to identify the expression profiles of genes targeted by GnRHa. We validated the expression of 10 of these genes in these cohorts and concluded that local expression and activation of these genes may represent features differentiating leiomyoma and myometrial molecular environments during growth as well as GnRHa-induced regression.
Materials and Methods
Portions of leiomyoma and matched myometrium were collected from premenopausal women (n = 6) who were scheduled to undergo hysterectomy for indications related to symptomatic leiomyomas. Three of the patients received GnRHa therapy for 3 months before surgery. The untreated patients did not receive any medications (including hormonal therapy) in the previous 3 months before surgery, and based on endometrial histology and the patient’s last menstrual period, they were in early to midsecretory phase of their menstrual cycles (Fig. 1). To maintain a standard, all leiomyomas selected for this study were between 2–3 cm in diameter. After collection, the tissues were divided into several pieces and immediately snap-frozen and stored in liquid nitrogen for additional processing, fixed and paraffin embedded for histological evaluation and immunohistochemistry, or used for isolation of leiomyoma and myometrial smooth muscle cells and culturing (16, 18). The tissues were collected at the University of Florida-affiliated Shands Hospital with prior approval from the institutional review board.
FIG. 1. The chart describes the experimental protocol for the microarray analysis of leiomyoma and matched myometrial tissues from GnRHa-treated and untreated cohorts as well as LSMC and MSMC treated with GnRH, TGF-?, and TGF-? type II receptor antisense, data analysis, and verification using real-time PCR, Western blotting, and immunohistochemistry (see Ref.35 for the results of TGF-? and TGF-? type II receptor antisense cohorts).
Isolation and culture of leiomyoma and myometrial smooth muscle cells (LSMC and MSMC)
To determine the direct action of GnRHa on global gene expression in LSMC and MSMC, the cells were isolated from untreated group and cultured as previously described (16, 20) (Fig. 1). Before use in these experiments, the primary cultures of LSMC and MSMC were characterized using immunofluorescence microscopy and antibodies to -smooth muscle actin, desmin, and vimentin (16, 18). The cells were cultured in six-well plates at an approximate density of 106 cells/well in DMEM-supplemented medium containing 10% fetal bovine serum. After reaching visual confluence, often after 2–3 d, the cells were washed in serum-free medium and incubated for 24 h under serum-free, phenol red-free conditions (20). The cells were then treated with 0.1 μM GnRHa (leuprolide acetate, Sigma-Aldrich Corp., St. Louis, MO) for a period of 2, 6, or 12 h (16).
cDNA microarray and gene expression profiling
Total cellular RNA was isolated from the tissues and cells using TRIzol (Invitrogen Life Technologies, Inc., Carlsbad, CA). The isolated RNA was treated with deoxyribonuclease I (Roche, Indianapolis, IN) at 1 U/10 μg RNA, heat-inactivated, and subjected to additional purification using the RNeasy kit (Qiagen, Valencia, CA). The purified RNA was then subjected to amplification by RT using SuperScript Choice system (Invitrogen Life Technologies, Inc.), with final concentrations in 20 μl first strand reaction of 100 pmol HPLC-purified T7-(dT)24 primer (Genset Corp., La Jolla, CA), 8 μg RNA, 1x first strand buffer, 10 mM dithiothreitol, 500 μM of each deoxy-NTP, and 400 U Superscript II reverse transcriptase (T7 Megascript kit. Ambion, Austin, TX). The second strand cDNA synthesis was performed in a 150-μl reaction consisting of (final concentrations) 1x second strand reaction buffer, 200 μM of each deoxy-NTP, 10 U DNA ligase, 40 U DNA polymerase I, and 2 U ribonuclease H (Invitrogen Life Technologies, Inc.), and double-stranded cDNA was purified by phenol/chloroform extraction using phase lock gels (Eppendorf-5 Prime, Inc., Westbury, NY) and ethanol precipitation (10).
Five micrograms of purified cDNA were reverse transcribed using an Enzo BioArray high yield RNA transcript labeling kit (Affymetrix, Santa Clara, CA), and the product was purified in RNeasy spin columns (Qiagen) according to the manufacturer’s instructions. After an overnight ethanol precipitation, cRNA was resuspended in 15 μl diethyl pyrocarbonate-treated water (Ambion) and quantified using UV-visible spectrophotometer. After quantification of cRNA to reflect any carryover of unlabeled total RNA according to an equation provided by Affymetrix (adjusted cRNA yield = cRNA (μg) measured after in vitro transcription (starting total RNA) (fraction of cDNA reaction used in in vitro transcription), 20 μg cRNA were fragmented (0.5 μg/μl) according to Affymetrix instructions using the 5x fragmentation buffer containing 200 mM Tris acetate (pH 8.1), 500 mM potassium acetate, and 150 mM magnesium acetate (Sigma-Aldrich Corp.). Twenty micrograms of the adjusted fragmented cRNA were added to a 300 μl hybridization mixture containing (final concentrations) 0.1 mg/ml herring sperm DNA (Promega Corp., Madison, WI), 0.5 mg/ml acetylated BSA (Invitrogen Life Technologies, Inc.), and 2x 2-(N-morpholino)ethanesulfonic acid hybridization buffer (Sigma-Aldrich Corp.). Two hundred microliters of the mixture were hybridized to the human U95A Affymetrix GeneChip arrays, purchased at the same time from the same lot number and used within 2 wk of purchase to maintain a standard between this and the experiments described by Luo et al. (35). In addition, an aliquot of random samples were first hybridized to an Affymetrix Test 2 Array to determine sample quality. After meeting recommended criteria for use of the expression arrays, hybridization was performed for 16 h at 45 C, followed by washing, staining, signal amplification with biotinylated antistreptavidin antibody, and staining according to the manufacturer’s protocol.
Microarray data analysis
The Chips were scanned to obtain the raw hybridization values using Affymetrix Genepix 5000A scanner. Differences in fluorescence spot intensities representing the rate of hybridization between the 25-bp oligonucleotides and their mismatches were analyzed by multiple decision matrices to determine the presence or absence of gene expression and to derive an average difference score representing the relative level of gene expression. The fluorescence spot intensities, qualities, and local background were assessed automatically by Genepix software with manual supervision to detect any inaccuracies in automated spot detection. Background and noise corrections were made to account for nonspecific hybridization and minor variations in hybridization conditions. The net hybridization values for each array were subjected to global normalization as previously described (10). To identify changes in the pattern of gene expression, the average and SD of globally normalized values were calculated, followed by subtraction of the mean value from each observation and division by the SD. The mean transformed expression value of each gene in transformed dataset was set at 0, and the SD was set at 1 (10).
The transformed gene expression values were subjected to Affymetrix Analysis Suite version 5.0. Briefly, probe sets that were flagged as absent on all arrays using default settings were removed from the datasets. After application of this filtering, the dataset was reduced from 12,625 to 8,580 probe sets. The gene expression value of the remaining probe sets was then subjected to unsupervised and supervised learning, discrimination analysis, and cross-validation (36, 37, 38, 39, 40). After variation filtering, the coefficient of variation was calculated for each probe set across all chips, and the probe sets were ranked by the coefficient of variation of the observed single intensities. The expression values of the selected genes were then subjected to statistical analysis in an R programming environment that assesses multiple test corrections to identify statistically significant gene expression values (38, 39, 40). The gene expression values with a statistical significance of P 0.02 (by ANOVA and Tukey’s test) between leiomyoma and myometrium from GnRH-treated and untreated cohorts and P 0.005 between GnRHa-treated and untreated cells (control) were selected. The validity of gene sets identified at these P values in predicting treatment class was established using "leave-one-out" cross-validation where the data from one array was left out of the training set, and probe sets with differential hybridization signal intensities were identified from the remaining arrays (40). Hierarchical clustering and K-means analysis were performed and viewed with the algorithms in the software packages Cluster and TreeView (36).
Gene classification and ontology assessment
The selected differentially expressed and regulated genes in the above cohorts were subjected to functional annotation and visualization using Database for Annotation, Visualization, and Integrated Discovery (DAVID; www.david.niaid.nih.gov) software. The integrated GoCharts assigns genes to specific ontology functional categories based on selected classifications.
Quantitative real-time PCR
Real-time PCR was used for verification of 10 differentially expressed and regulated genes identified in leiomyoma and myometrium as well as LSMC and MSMC from untreated and GnRHa-treated cohorts. The selection of these genes was based not only on their expression values (up or down-regulation), but on classification and biological functions important to leiomyoma growth and regression, and on regulation by ovarian steroids, GnRHa and TGF-? (35), as indicated in the literature. They are IL-11, EGR3 (early growth response 3), TGIF (TGF-?-induced factor, TIEG (TGF-?-inducible early gene response), CITED2 [cAMP response element binding protein-binding protein (CBP)/p300-interacting transactivator with ED-rich tail], Nur77, p27, p57, Gas-1 (growth arrest-specific 1) and G protein-coupled receptor kinase 5 (GPRK), representing cytokines, transcription factors, cell cycle regulators, and signal transduction. Real-time PCR was carried out as previously described using TaqMan and ABI PRISM 7700 Sequence System and Sequence Detection System 1.6 software (Applied Biosystems, Foster City, CA) (16). Results were analyzed using the comparative method and after normalization of expression values to 18S rRNA expression according to the manufacturer’s guidelines (Applied Biosystems) as previously described (16).
Western blotting and immunohistochemical localization
For Western blotting, total protein was isolated from small portions of GnRHa-treated and untreated leiomyoma and myometrium as well as the GnRHa-treated and untreated cells and subjected to analysis as previously described (16, 17, 18). The blots were incubated with anti-TIEG antibody (provided by Dr. Thomas Spelsberg, Mayo Clinic, Rochester, MN) (41); TGIF, EGR3, Nur77, CITED2, p27, p57, and Gas1 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); and IL-11 antibodies (R&D Systems, Minneapolis, MN). The immunostained proteins were visualized using enhanced chemiluminescence reagents (Amersham Biosciences, Piscataway, NJ) as previously described (16).
Tissue sections prepared from formalin-fixed and paraffin-embedded leiomyoma and myometrium were subjected to microwave treatment and immunostained using antibodies to IL-11, TGIF, TIEG, EGR3, CITED2, Nur77, p27, p57, and Gas1 at 5 μg immunoglobulin G/ml for 2–3 h at room temperature. After additional standard processing, a chromogenic reaction was detected with 3,3'-diaminobenzidine tetrahydrochloride solution (16). In some instances the slides were counterstained with hematoxylin. Omission of primary antibodies and incubation of tissue sections with nonimmune mouse, rabbit, and goat IgG instead of primary antibodies at the same concentration served as controls (16, 17, 18).
Results
Gene expression profiles in leiomyoma and myometrium
Using global gene expression profiling, we characterized the gene expression profile of leiomyoma and matched myometrium and their transcriptional changes in response to hormonal transition induced by GnRHa therapy. The initial simultaneous assessment of the gene expression values in leiomyoma, myometrium, and their isolated smooth muscle cells from untreated as well as GnRHa- and TGF-?-treated (35) cohorts revealed uniform expression of transcripts for the housekeeping genes glyceraldehyde-3-phosphate dehydrogenase, ?-actin, and a large number of ribosomal proteins, indicating that the expression profile is consistent with established standards for gene expression analysis. After unsupervised and supervised learning, the gene expression values were subjected to statistical analysis in R programming and ANOVA with false discovery rate selected at P 0.02.
Using the above analysis, we identified a total of 153 genes, including 19 established sequence tags (EST), or 1.23% of the genes, and 122 genes including 21 EST, or 0.98% of the genes on the array, as differentially expressed in leiomyoma compared with matched myometrium from untreated and GnRHa-treated tissues, respectively. Hierarchical clustering and Tree-View analysis separated the genes in each cohort into distinctive clusters with sufficient variability, allowing division into their respective subgroups (Fig. 2). Of the 153 (excluding 19 EST) genes in untreated cohorts, the expression of 82 genes was up-regulated and that of 52 genes was down-regulated in leiomyoma compared with myometrium (Table 1). Of the 122 genes (excluding 21 EST) in leiomyoma and myometrium from patients who received GnRHa therapy, the expression of 34 genes was up-regulated and that of 67 genes was down-regulated in leiomyoma compared with myometrium, respectively (Table 2). Analysis of the variance-normalized mean (K-means) separated the differentially expressed and regulated genes in these cohorts into four distinctive clusters, with genes in clusters A and D displaying a tissue-specific response, whereas genes in clusters B and C showed a regulatory response to GnRHa therapy (Fig. 3). To further differentiate the molecular environment of leiomyoma from myometrium and their responses to GnRHa therapy, we compared the gene expression profiles in GnRHa-treated and corresponding untreated tissues. The results indicated that the expressions of 170 (excluding 26 EST) and 167 (excluding 31 EST) genes were targeted by GnRHa therapy in leiomyoma and myometrium compared with their respective untreated cohorts (Tables 3 and 4). Of these genes, 96 and 89 transcripts were down-regulated in leiomyoma and myometrium, respectively, by GnRHa therapy compared with their respective untreated tissues, and three transcripts were commonly found among the tissues in these cohorts, with different regulatory patterns of expression (compare Tables 3 and 4).
FIG. 2. Hierarchical clustering analysis of differentially expressed genes in leiomyoma and matched myometrium from untreated (f and m 315, 316, and 317) and GnRHa-treated (f and m 287, 312, and 314) groups identified after unsupervised and supervised analysis in an R programming environment and ANOVA with a false discovery rate of P 0.02 (38 39 40 ). Each column represents data from a single cohort, with shades of red and green indicating up- or down-regulation of a given gene according to the color scheme shown below. Genes represented by rows were clustered according to their similarities in expression patterns for each tissue and treatment. The dendrogram displaying similarity of gene expression among the cohorts is shown on top of the overview image, and relatedness of the arrays is denoted by distance to the node linking the arrays. The gene-tree shown at the left of the image corresponds to the degree of similarity (Pearson correlation) of the pattern of expression for genes across the experiments. The clustering divided the genes into five clusters, designated A–D, and their zoomed images are presented in A–D. Genes that appear more than once are represented by multiple clones on arrays.
TABLE 1. Categorical list of differentially expressed genes in leiomyoma and matched myometrium
TABLE 2. Categorical list of differentially expressed genes in leiomyoma and matched myometrium in response to GnRHa therapy
FIG. 3. K-means clustering of genes regulated in leiomyoma and matched myometrium during growth and under the influence of GnRH therapy. The gene expression values identified in these cohorts described in Fig. 1 were subjected to k-means clustering, grouping the genes into clusters based on similarity of expression in GnRHa-treated and untreated cohorts. The analysis grouped the genes into four clusters (A–D). The rows represent the genes, and columns represent the samples (f and m) from GnRHa-treated (287, 312, and 314) and untreated (315, 316, and 317) cohorts, with shades of red and green indicating up- or down-regulation of a given gene, which are clustered according to their similarities in expression patterns.
TABLE 3. Categorical list of differentially expressed genes in leiomyoma from GnRHa-treated vs. untreated group
TABLE 4. Categorical list of differentially expressed genes in myometrium from GnRHa-treated vs. untreated group
Since the present study was completed, a few microarray studies have reported the gene expression profiles of leiomyoma and myometrium (9, 10, 11, 12, 13, 14). We performed a comparative analysis using the genes identified at P 0.02 in the untreated leiomyoma and matched myometrium of our study and the list of genes reported in four of these studies, searching for a set of commonly expressed genes. The comparison identified two genes in our study in common with at least one of these studies. However, lowering the false discovery rate of our selection to P 0.05 enabled us to identify a larger number of genes (422, including 49 EST), of which 11 transcripts were found in common with other studies (Table 5).
TABLE 5. List of common gene found in present and other studies identified by their respective citations
Gene ontology assessment and division of differentially expressed genes into similar functional categories indicated that the products of a large percentage of these genes (40–67%), in leiomyoma and myometrium from both GnRHa-treated and untreated cohorts are involved in metabolic processes, catalytic activities, binding, signal transduction, transcriptional and translational activities, cell cycle regulation, cell and tissue structure, etc. (Tables 1–4 and 6). In addition, 15–23% of the genes were either functionally unclassified, or their roles in biological process are still unknown.
TABLE 6. Gene ontology assessment and division of the genes in leiomyoma (LYM) and myometrium (MYM) in untreated and in response to GnRH therapy as well as genes regulated in LSMC and MSMC in response to GnRH treatment into similar functional categories with the percentage of gene in each group presented
Time-dependent action of GnRHa on gene expression profile of LSMC and MSMC
To obtain a comprehensive picture of transcriptional changes induced by GnRHa’s direct action in leiomyoma and myometrium, LSMC and MSMC were treated with GnRHa (0.1 μM) for 2, 6, and 12 h, and their isolated RNA were subjected to microarray analysis. Based on the same data analysis criteria described above with a false discovery rate of P 0.005, we identified 281 genes, including 36 EST, or 2.2% of the genes on the array, displaying differential expression in response to a time-dependent GnRHa action compared with untreated controls in LSMC ad MSMC (Fig. 4). Hierarchical clustering analysis also separated these genes into different clusters in response to the time-dependent action of GnRHa in LSMC and MSMC, with expression patterns sufficiently different to cluster into their respective subgroups (Fig. 4). Analysis of the variance-normalized mean (K-means) separated the differentially expressed and regulated genes in these cohorts into four distinctive clusters, with genes in clusters A and D displaying a cell-specific response, whereas genes in clusters B and C showed regulatory behaviors to GnRHa’s time-dependent action (Fig. 5). Among these genes, the transcripts of 48 genes were identified as commonly expressed in LSMC and the original leiomyomas from the untreated cohort, although they many be differently regulated (Table 7).
FIG. 4. Hierarchical clustering analysis of 281 differentially expressed and regulated genes in LSMC (f) and MSMC (m) in response to GnRHa (0.01 μM) treatment for 2, 6, and 12 h or in untreated controls (C). The genes were identified after unsupervised and supervised analyses of the expression values, statistical analysis in an R programming environment, and ANOVA with a false discovery rate selected at P 0.005 (38 39 ). Each column represents data from a single treatment, with shades of red and green indicating up- or down-regulation of a given gene. Genes represented by rows were clustered according to their similarities in expression patterns for each cell type and treatment. The dendrogram displaying similarity of gene expression among the cohorts is shown on top of the overview image, with array relatedness denoted by distance to the node linking the arrays. The gene-tree shown at the left of the image corresponds to the degree of similarity (Pearson correlation) of the pattern of expression for genes across the experiments. The clustering divided the genes into four clusters, designated as A–D, and their zoomed images are presented in A–D. Genes that appear more than once are represented by multiple clones on arrays.
FIG. 5. K-means clustering of the gene expression values identified in response to time-dependent action of GnRHa in LSMC and MSMC described in Fig. 1. The analysis grouped the genes into six clusters (A–F) based on similarity of expression in response to time-dependent action of GnRHa. The rows represent the genes, and columns represent the samples (f and m) and treatments with GnRHa for 2, 6, and 12 h and untreated controls (C), with shades of red and green indicating up- or down-regulation of a given gene, which are clustered according to their similarities in expression patterns. Line graphs display the SD for each cluster in MSMC and LSMC in response to GnRHa time-dependent actions compared with untreated control (Ctrl).
TABLE 7. Common genes between GnRHa-treated leiomyoma and LSMC
Gene ontology and functional annotation of these genes into similar functional categories also indicated that in LSMC and MSMC, similar to their original tissues, the majority of the gene products are involved in cellular processes, catalytic activities, binding, signal transduction, transcriptional and translational activities, metabolism, cell cycle regulation, and cellular structure (Table 6). The time-dependent actions of GnRHa on the expression of a selective group of genes in these functional categories in LSMC and MSMC are shown in Fig. 6.
FIG. 6. The expression profile of a selected group of genes representing growth factors/cytokines/polypeptide hormones/receptors (first row), intracellular signal transduction pathways (second row), transcription factors (third row), cell cycle regulators (fourth row), and cell adhesion/ECM/cytoskeletons (fifth row) in response to the time-dependent action of GnRHa in LSMC and MSMC. Values on the y-axis represent an arbitrary unit derived from the mean gene expression value for each factor after supervised analysis, statistical analysis in an R programming environment, and ANOVA as described in Fig. 4, with gene expression values for the untreated controls (Ctrl) set at 1.
Verification of gene transcripts in leiomyoma, myometrium, LSMC, and MSMC
Among the differentially expressed and regulated genes identified in these tissues and cells, we selected 10 genes for verification using real-time PCR, Western blotting, and immunohistochemistry. The genes selected for validation were IL-11, CITED2, Nur77, EGR3, TGIF, TIEG, p27, p57, Gas-1, and GPRK5, representing cytokines, transcription factors, cell cycle regulators, and signal transduction. The pattern of expression of these genes in leiomyoma and myometrium from untreated and GnRHa-treated cohorts (Fig. 7) and in LSMC and MSMC treated with GnRHa for 2, 6, and 12 h (Fig. 8), as determined by real-time PCR, closely overlapped with their expression profiles identified by the microarray analysis.
FIG. 7. Comparative analysis of the expression profile of 10 genes identified as described in Fig. 1 as differentially expressed in response to GnRH therapy in leiomyoma and matched myometrium and in an untreated group by microarray and real-time PCR. Values on the y-axis represent an arbitrary unit derived from the mean expression value for each gene, with values for the untreated controls (Crtl) set at 1. Total RNA isolated from these tissues was used for both microarray analysis and real-time PCR validating the expression of IL-11, EGR3, CITED2, Nur77, TIEG, TGIF, p27, p57, Gas-1, and GPRK5. On the y-axis, untreated myometrium and leiomyoma are designated Unt-MM and Un-LM, respectively, and GnRH-treated myometrium and leiomyoma are designated GnRH-Trt MM and GnRH-Trt LM, respectively.
FIG. 8. Comparative analysis of the expression profile of 10 genes identified as described in Fig. 4 as differentially expressed and regulated in response to GnRHa time-dependent action in LSMC and MSMC by microarray and real-time PCR. Values on the y-axis represent an arbitrary unit derived from the mean expression value for each gene, and values on the x-axis represent the time course of GnRHa (0.1 μM) treatment (2, 6, and 12 h) with untreated control (Crtl) gene expression values set at 1. Total RNA isolated from these cells was used for both microarray analysis and real-time PCR to validate the expression of IL-11, EGR3, TIEG, TGIF, CITED2, Nur77, p27, p57, Gas-1, and GPRK5.
Western blotting also indicated that leiomyoma and myometrium as well as LSMC and MSMC locally produce IL-11, TGIF, TIEG, Nur77, EGR3, CITED2, p27, p57, and Gas-1 proteins (data are not shown). Immunohistochemically, IL-11, TGIF, TIEG, Nur77, EGR3, CITED2, p27, p57, and Gas-1 were localized in various cell types in leiomyoma and myometrium, including LSMC, MSMC, and vasculature and connective tissue fibroblasts (Fig. 9). We did not have access to antibody to GPRK5 and have not yet attempted to quantitate the levels of IL-11, TGIF, TIEG, Nur77, EGR3, CITED2, p27, p57, and Gas-1 production in leiomyoma and myometrium as well as in LSMC and MSMC in response to GnRHa treatment. However, these results provided additional support for the microarray and real-time PCR data, indicating that various cell types contribute to overall expression of these genes in leiomyoma and myometrium. In addition to these genes, we validated the expression of 15 more genes with real-time PCR including CTGF, Able-interactive (Abi2), fibromodulin, Runx1, and Runx2 (35, 42).
FIG. 9. Immunohistochemical localization of IL-11, TGIF, TIEG, Nur77, EGR3, p27, p57, and Gas-1 in leiomyoma and myometrium. Note the presence of immunoreactive IL-11, TGIF, TIEG, Nur77, EGR3, p27, p57, and Gas-1 in association with leiomyoma and myometrial smooth muscle cells, and cellular components of connective tissue and vasculature. Both nuclear (EGR3, Nur77, p27, and p57) and cytoplasmic (IL-11) staining was observed. Incubation of tissue sections with nonimmune mouse (A), rabbit (B), and goat (figure not shown) IgGs instead of primary antibodies during immunostaining served as controls (Ctrl), reduced the staining intensity. Magnifications: x150.
Discussion
In the present study we characterized gene expression fingerprints of leiomyoma and matched myometrium from the early to midsecretory phase of the menstrual cycle, a period associated with their rapid growth, their response to GnRHa therapy, and their response to direct action of GnRHa in isolated LSMC and MSMC prepared from untreated tissues. Combining global normalization and unsupervised assessment of the gene expression values derived from all the cohorts enabled us to sort potential candidate genes before their putative identification in each cohort. Transcripts of many of the genes on the array were found in leiomyoma and myometrium as well as in LSMC and MSMC. However, leiomyoma/LSMC were not distinguished as a single class from myometrium/MSMC based on single gene markers uniformly expressed only in leiomyoma and/or myometrium. This is not unique to leiomyoma/myometrium, because many large scale gene expression profiling studies have shown the existence of a significant degree of shared gene expression between various tumors and their normal tissue counterparts. However, supervised assessment and multiple test correction in a statistical R programming environment (37, 38, 39, 40) with a reduced false discovery rate, allowed us to identify a specific set of differentially expressed and regulated genes in descending order of significance in each cohort.
The analysis separated these genes into several clusters with sufficient differences, allowing their subdivision into respective subgroups in leiomyoma, myometrium, their isolated cells, as well as changes due to GnRHa therapy at the tissue and cellular levels. We identified 153 genes (excluding 19 EST) in these cultures, as differentially expressed in leiomyoma compared with myometrium, 82 of which were up-regulated and 52 of which were down-regulated in leiomyoma. GnRHa therapy affected the expression of 122 genes (excluding 21 EST), with 34 up-regulated and 67 down-regulated genes in leiomyoma compared with myometrium. However, the gene expression profiles from the untreated group differed substantially compared with those from GnRHa-treated leiomyoma/myometrium, pointing out a unique molecular environment that is targeted by GnRHa therapy. Analysis of the variance-normalized mean gene expression values divided these genes into four clusters, with two clusters showing treatment-specific, and the other clusters showing a tissue-specific, response to GnRHa therapy. Similar behavior was observed with gene clusters identified in LSMC and MSMC in response to GnRHa action in vitro. The significance of these findings relates to clinical observations indicating that GnRHa therapy affects both leiomyoma and myometrium, with nonmyoma tissue regressing more in response to therapy (7). Our gene expression profiling supports the clinical observations and points out that GnRHa therapy targets different genes in leiomyoma and myometrium despite their grouping in a similar functional category. Our recent microarray study with a small scale array of 1200 known genes (10) provides support for our current study; however, we are not aware of any other study using large scale gene expression profiling in leiomyoma and myometrium from women who received GnRHa therapy for additional comparison.
Since our study was completed, a few other microarray studies have reported the gene expression profiles of leiomyoma and myometrium from the proliferative and secretory phases of the menstrual cycle (9, 11, 12, 14). To broaden the scope of this study, we compared the genes identified in untreated leiomyoma and matched myometrium of our study with the datasets reported in four of these studies (9, 11, 12, 14). This comparison resulted in identification of only a few genes in common among these studies. Although intrinsic individual tissue variation may contribute to the differences among these studies, a standard of experimental process, utilization of different microarray platforms, use of tissues from different phases of the menstrual cycle, differences in leiomyoma size, and, most importantly, the method of data acquisition and analysis (9, 11, 12, 14) are among other key contributing factors accounting for different study results (38, 39, 40). To maintain a standard, we used leiomyoma of uniform size (2–3 cm in diameter) and matched myometrium, and the untreated cohorts were collected from the early to midsecretory phase of the menstrual cycle, a period associated with maximum growth of leiomyoma. However, lowering the false discovery rate of our study from P 0.02 to P 0.05 allowed identification of more transcripts and the appearance of additional common genes with other studies (9, 11, 12, 14). Considering the presence of a large number of probe sets on these arrays (i.e. 6,800–12,500), selection of genes based only on fold change (9) or higher statistical levels (11, 12, 13, 14) is no better than what one would expect by chance alone (38, 39, 40). Because we employed a similar data analysis, a larger number of genes was found in common with our previous microarray study, which used only a small scale array containing about 1,200 known genes (10). We recognize that exclusion of moderately regulated genes during microarray data analysis does not reflect a lack of functional importance, because a number of genes previously identified in leiomyoma and myometrium by conventional methods are not included among the differentially expressed genes in our study and other reports (1, 2, 9, 10, 11, 12, 13, 14). However, the expression of newly identified genes requires verification, and their regulation would allow linking their potential biological functions in leiomyoma growth and regression.
Using isolated LSMC and MSMC prepared from the untreated tissues allowed us to identify novel regulatory functions for GnRHa in leiomyoma and myometrium and to discover a wide range of genes whose expression has not previously been recognized to be the target of GnRHa direct action. Similar to their distinct clustering at tissue levels, the differentially expressed and regulated genes identified in LSMC and MSMC were also divided into clusters according to time-dependent response to GnRHa action. The genes in these clusters were either rapidly induced by GnRHa treatment or required prolong exposure, whereas others displayed biphasic patterns of temporal regulation in both treatment- and cell-specific fashions. Despite differences in their profiles, substantial similarity existed in the functional grouping of the genes affected by GnRHa therapy in leiomyoma/myometrium and GnRHa direct action on LSMC/MSMC (in vitro), with the expression of 48 genes commonly identified in tissues and cells. We speculate that the hypoestrogenic condition created by GnRHa therapy and contributions of other cell types to overall gene expression at the tissue level may account for the difference in profiles of gene expression between tissues and cell cultures. Gene functional annotation indicated that products of the majority of genes in these clusters are involved in transcriptional and signal transduction activities, cell cycle regulation, ECM turnover, cell-cell communication, transport, and enzyme regulatory activities.
Among the genes in these functional categories are several growth factors, cytokines, and chemokines, and polypeptide hormones, identified in leiomyoma, myometrium, and their isolated smooth muscle cells, and these were the target of GnRHa action in vivo and in vitro. Using several conventional methods, previous reports, including our own, have documented the expression of platelet-derived growth factor, epidermal growth factor, IGFs, vascular endothelial growth factor, fibroblast growth factor, TGF-?s, connective tissue growth factor, TNF-, IFN-, monocyte chemoattractant protein-1, and IL-8 as well as some of their receptors in leiomyoma and myometrium (1, 2, 16, 17, 20, 23, 26, 27). However, the expression of some of these and other genes in this category did not meet the selection criteria of our study, a common discrepancy often observed in microarray analysis, particularly in identifying moderately expressed and regulated genes (38, 39, 40). For example, the expressions of TGF-? isoforms, TGF-? receptors, and components of their signaling pathway that are well documented in leiomyoma and myometrium as well as in their isolated smooth muscle cells (17, 18, 19, 20, 43) are not consistently identified in microarray studies (9, 10, 11, 12, 13, 14), although in our current accompanying (35) and previous (10) studies we identified most of the members of the TGF-? system. However, microarray gene expression data may not always reflect the actual expression level of a given gene, as shown in some cases in our study compared with real-time PCR findings. Among the cytokines whose expression was identified and validated in our study is IL-11. IL-11 is recognized to play key regulatory functions in inflammation, angiogenesis, and tissue remodeling (44, 45, 46, 47, 48), events central to leiomyoma growth. IL-11 is a member of the IL-6 family and is produced by various cell types, including the uterus. Its overexpression is reported to cause subepithelial airway fibrosis, particularly through interaction with IL-13 and TGF-? (44, 45, 46, 47, 48, 49). We have provided evidence that IL-11, similar to TGF-? and IL-13, is overexpressed in leiomyoma compared with myometrium, and GnRHa therapy suppresses their expression in these tissues (17, 20, 43, 50). At the cellular level, unlike the expression of TGF-? and IL-13, GnRHa increases IL-11 expression in LSMC and MSMC within 2–6 h of treatment, which sharply declines to control levels after 12 h. Although the nature of differential regulation of IL-11 at the tissue and cellular levels requires detailed investigation, prolonged treatment with GnRHa, the contribution of other cell types, and the influence of other autocrine/paracrine regulators may account for the difference in IL-11 expression between in vivo and in vitro conditions.
Other differentially expressed and regulated genes identified in our study functionally belong to signal transduction pathways that are recruited and activated by various growth factors/cytokines/chemokines, polypeptide hormones, ECM, and adhesion molecules. However, only the expression of a few of these and other signal transduction pathways has been documented in leiomyoma and myometrium (1, 16, 19, 29), and little is known about their recruitment and activation in LSMC and MSMC. We have previously identified the expression of Smads, MAPK, and focal adhesion kinase in leiomyomas and myometrium and provided evidence for their regulation and activation by GnRHa in LSMC and MSMC (16, 18, 19). In this study we validated the expression of GPRK5, identified as one of the differentially expressed and regulated genes in leiomyoma and myometrium and demonstrated that GnRHa therapy and in vitro treatment of LSMC and MSMC with GnRHa inhibit GPRK5 expression. GPRKs, consisting of six members (GPRK1 to GPRK6), act as key regulators of signaling via GPRKs and are widely expressed in various tissues and cells (51, 52, 53, 54). Previous studies have demonstrated that pregnant and nonpregnant human myometrium as well as cultured myometrial cells express GPRK2, GPRK4, and GPRK5; however, GPRK3 and GPRK4, -?, and - were not detected in myometrium (52, 53). GPRK5 has been shown to serve as a substrate for protein kinase C (PKC), although PKC-mediated phosphorylation inhibits GPRK5 (21, 54). In addition, the extreme N terminus of GPRK5 contains a binding site for Ca2+/calmodulin, where, upon binding, it inhibits GPRK activity, a mechanism suggested to regulate GPRK activity (54). Because GnRH receptors are members of the GPCR family and recruit and activate the components of several signaling pathways, including PKC and Ca2+/calmodulin, their regulatory interaction with GPRKs may serve in mediating GnRH receptor downstream signaling in LSMC and MSMC.
Nuclear translocation of activated signaling molecules results in phosphorylation and activation of transcription factors, major elements in the signaling networks that regulate specific gene expression. In our previous (10) and current studies, we identified several genes in this category in leiomyoma and myometrium whose expression was targeted by GnRHa in LSMC and MSMC. Many of these transcription factors are involved in ovarian steroid-, polypeptide hormone-, cytokine-, growth factor-, and ECM receptor-mediated actions, by regulating the promoter of their target genes in various normal and cancer cells. However, little is known about the expression and regulation of these and other transcription factors in leiomyoma and myometrium. For this reason we placed a greater emphasis on the expression of genes in this category and verified the expression of Nur77, CITED2, EGR3, TIEG, and TGIF in leiomyoma and myometrium and their temporal regulation by GnRHa in LSMC and MSMC.
Nur77 (also known as NR4A1, TR3, NGFI-B, and NAK-1) is a member of the orphan nuclear receptor superfamily originally identified as an immediate-early gene in serum-treated fibroblasts (55, 56, 57, 58, 59, 60, 61, 62). It is also identified as nerve growth factor-inducible gene, which is constitutively expressed in various tissues and is strongly induced by several stimuli, resulting in the regulation of gene expression related to inflammation, angiogenesis, apoptosis, and steroidogenesis, including steroid 21- and 17-hydroxylases and 20-hydroxysteroid dehydrogenase in the hypothalamic-pituitary-adrenal axis (55, 56, 57, 58, 59, 60, 61, 62). In the anterior pituitary, Nur77 is reported to mediate the stimulatory effect of CRH and the negative feedback regulation of proopiomelanocortin transcription by glucocorticoids as well as GnRH-induced GnRH receptor expression (56, 59). LH-induced Nur77 is also reported to regulate cytochrome P450 expression in granulosa and Leydig cells (59, 60, 61). More importantly, overexpression of Nur77 is implicated as an important regulator of apoptosis in different cells. In response to apoptotic stimuli, Nur77 translocation from nucleus to mitochondria results in cytochrome c release and apoptosis of LNCaP human prostate cancer cells (63, 64, 65). Nur77 is expressed at relatively similar levels in myometrium and leiomyoma and is significantly increased in response to GnRHa therapy. GnRHa also resulted in a rapid induction of Nur77 in MSMC and LSMC, subsequently declining to control levels or in LSMC to below control levels. Interestingly, GnRH is reported to regulate Nur77 expression in T3-1 and L?T2 gonadotrope cell lines through protein kinase A pathway and GnRH receptor promoter via a mechanism involving steroidogenic factor-1 with Nur77 acting as a negative regulator of this response (59). In a recent study, activation of the MAPK pathway involving Raf-1, MAPK kinase 2, and extracellular signal-regulated kinase 2 was reported to regulate Nur77 activation, resulting in nonapoptotic program cell death (64). We have shown that GnRH receptor signaling through MAPK and transcriptional activation of c-Fos and c-Jun regulate the expression of several specific genes in LSMC and MSMC (16). This suggests that GnRH-mediated action through this pathway may regulate Nur77 expression, thus influencing the outcome of cellular growth arrest and/or apoptosis in leiomyoma.
Recently, a new family of transcriptional coregulators, the CITED family, was discovered that interact with the first cysteine-histidine-rich region of CBP/p300 (66, 67). The CITED family contains four members and appears to act as key transcriptional modulators in embryogenesis, inflammation, and stress responses (66) by affecting the transcriptional activity of many transcription factors ranging from activating protein-2, estrogen receptor, and hypoxia-inducible factor 1 and LIM (68). We identified CITED2 among the differentially expressed and regulated genes in leiomyoma, myometrium, and their isolated cells and in response to GnRHa treatment in vivo and in vitro. Unlike GnRHa therapy, which increased CITED2 expression in leiomyoma and myometrium, GnRHa had a biphasic effect on CITED2 expression in MSMC while inhibiting expression in LSMC. Although in vitro culture conditions may directly influence the expression of regulatory molecules that either interact with or regulate CITED2 expression, the molecular mechanism resulting in differential expression of CITED2 in vivo and in vitro by GnRHa requires additional investigation. Interestingly, the expression of several growth factors, cytokines, and hypoxia-inducible factor 1 are the target of estrogen receptor and progesterone receptor regulatory actions, and CITED2, acting as a repressor of their expression, may serve as an important mediator of processes that regulate inflammatory response, angiogenesis, and tissue remodeling in leiomyoma. Additionally, CBP/p300, which serve as promiscuous coactivators for an increasing number of transcription factors, resulting in proliferation, differentiation, and apoptosis in response to diverse biological factors, including estrogen receptor- and progesterone receptor-dependent transcriptional activity, is specifically recruited by Nur77 acting as dimers after protein kinase A activation (55, 69, 70).
In our previous microarray study, we reported that EGR1, a prototype of a family of zinc finger transcription factors that includes EGR2, EGR3, EGR4, and NGFI-B (71, 72), is differentially expressed in leiomyoma and myometrium (10). In this study we provide evidence for the expression of EGR3 and differential regulation in response to GnRHa therapy in leiomyoma and myometrium as well as in LSMC and MSMC in vitro. A recent report demonstrated that EGR1 expression is elevated in leiomyoma compared with corresponding myometrium in women who received GnRHa therapy (73), supporting our previous microarray data (10). EGR expression is rapidly and transiently induced by a large number of growth factors, cytokines, polypeptide hormones, and injurious stimuli, and the kinetics of their expression are essentially identical to those of the c-fos protooncogene (71, 72, 74). In addition, induction of EGR1 occurs primarily at the level of transcription and is mediated in part through MAPKs, including extracellular signal-regulated kinase, c-Jun N-terminal kinase, and p38 pathways (71, 72). We have demonstrated that GnRHa, through the activation of MAPK, regulates the expression of c-Fos and c-Jun as well as fibronectin, collagen, and plasminogen activator inhibitor-1 expression (16). In human fibrosarcoma and glioblastoma cells, EGR directly influences the expression of fibronectin, TGF-?1, and plasminogen activator inhibitor-1 and may regulate the expression of platelet-derived growth factor, tissue factor, and membrane type 1 matrix metalloproteinase (72, 75). Estrogen also induces EGR3 in various cancer cells, whereas progesterone inhibits EGR3 in Schwann cells (74, 76). Constitutive transgenic expression of EGR3 has been shown to increase thymocyte apoptosis, possibly through potent activation of Fas ligand expression (77). Given the roles of ovarian steroids and a large number of growth factors, cytokines, and polypeptide hormones in leiomyoma growth and suppression by GnRHa, their differential influence on EGR1 and EGR3 expression may represent a mechanism resulting in a balance between the rate of cell proliferation and apoptosis as well as tissue turnover, affecting leiomyoma growth and regression.
We also provide the first evidence of the expression and regulation of TIEG and TGIF, novel three-zinc finger Kruppel-like transcriptional repressors, and key regulators of TGF-? receptor signaling (41, 78, 79, 80, 81) by GnRHa in leiomyoma, myometrium, LSMC, and MSMC. TIEG regulates TGF-? receptor signaling through a negative feedback mechanism by repressing the inhibitory Smad7 (41). In addition, TGIF, through direct binding to DNA or interaction with TGF-?-activated Smads, represses TGF-?-responsive gene expression (80, 81). Because GnRHa suppresses TGF-?s and TGF-? receptors while enhancing Smad7 expression in leiomyoma and myometrium and in LSMC and MSMC, differential regulation of TIEG and TGIF may serve as an additional downstream mechanism altering TGF-? actions in leiomyoma (35).
The expression, activation, and direct interaction of these and other transcription factors with DNA result in regulation of the expression of various genes whose products influence cell growth, inflammation, angiogenesis, apoptosis, and tissue turnover. In our previous (10, 16) and present study we identified several differentially expressed and regulated genes in leiomyoma, myometrium, LSMC, and MSMC whose promoters are the targets of these transcription factors. Among these genes are members of cell cycle regulatory proteins that play a central role in leiomyoma growth and regression (1, 2, 31), including p27, p57, and Gas-1. We identified p27, p57, and Gas-1 as differentially expressed and regulated in leiomyoma and myometrium as well as in LSMC and MSMC and in response to GnRHa treatment. Although p27, p57, and Gas-1 function as major regulators of cell cycle progression, several studies have also shown Cip/Kip proteins to function as transcriptional cofactors by regulating the activity of nuclear factor-B, signal transducer and activator of transcription-3, Myc, Rb, C/EBP, CBP/p300, E2F, and activating protein-1 (82). A recent report suggests that p21, p27, and p57 are involved in the regulation of apoptosis (83), and their differential regulation in leiomyoma and myometrium is consistent with GnRHa induction of an apoptosis-related gene in LSMC and MSMC (1, 2, 22, 31). However, our results are the first to document the expression of Gas-1 in leiomyoma and myometrium and its regulation in LSMC and MSMC in response to the time-dependent action of GnRHa. GnRHa has been demonstrated to alter cell cycle progression and programmed cell death in several cell types, including leiomyoma smooth muscle cells (1, 22, 31), and our results provide additional support for the involvement of specific cell cycle- and apoptosis-related genes in leiomyoma growth and regression.
Leiomyoma growth and GnRHa therapy resulting in leiomyoma regression also involve ECM turnover. In our previous (10) and present studies and recent studies by other groups (9, 11, 12, 13, 14), several genes in this category were identified displaying differential expression in leiomyoma and myometrium and were targeted by GnRH therapy (1, 16, 24, 45, 42, 84). These include the expression of several collagens; a small leucine-rich repeat family of proteoglycans, decorin, biglycan, osteomodulin, fibromodulin, versican, and osteoadherin/osteoglycin, fibronectin, desmin, and vimentin; several member of proteases, such as matrix metalloproteinases, and their inhibitors; a disintegrin-like and metalloproteinase protein, etc. We have also reported that GnRHa regulates the expression of fibronectin, collagen type I, plasminogen activator inhibitor-I, matrix metalloproteinases, and matrix metalloproteinase inhibitors (1, 16, 24), as well as decorin, versican, desmin, and vimentin (our unpublished observations) in leiomyoma and myometrium, involving the activation of MAPK in LSMC and MSMC (16). Because ECM turnover is a key regulator of the outcome of tissue fibrosis, and many cytokines, chemokines, growth factors, and polypeptide hormones, through specific intracellular signal transduction and activation of transcription factors, influence the expression of ECM and proteases, additional investigation is underway to elucidate their regulatory interactions affecting processes that may influence leiomyoma growth and regression.
In summary, in the present study we provide a comprehensive assessment of the gene expression profile of leiomyoma and matched myometrium during the early to midluteal phase of the menstrual cycle, a period characterized by elevated production of ovarian steroids and maximal leiomyoma growth, compared with that in tissues obtained from hormonally suppressed patients receiving GnRHa therapy and in response to the direct action of GnRHa on LSMC and MSMC. We identified several common and tissue-specific gene clusters in these cohorts, suggesting their coregulation by the same factors and/or mechanism(s) in the same cluster. We validated the expression of several genes whose products are important in signal transduction, transcription, cell cycle regulation, apoptosis, and ECM turnover, events critical to the development, growth, and regression of leiomyoma. Based on these and our previous observations, we propose that the product of these specific genes, by regulating local inflammatory and apoptotic processes leading to elaboration of profibrotic cytokine production, such as TGF-?, is central to the establishment and progression of leiomyoma growth and regression.
Acknowledgments
We thank Dr. Thomas Spelsberg, (Mayo Clinic, Rochester, MN) for the gift of the antibody to TIEG, and Drs. Henry Baker (Department of Microbiology and Molecular Genetics), Mick Popp, and Li Liu (Interdisciplinary Center for Biotechnology Research) at University of Florida for their helpful advice concerning microarray data analysis.
References
Chegini N 2000 Implication of growth factor and cytokine networks in leiomyomas. In: Hill J, ed. Cytokines in human reproduction. New York: Wiley & Sons; 133–162
Maruo T, Ohara N, Wang J, Matsuo H 2004 Sex steroidal regulation of uterine leiomyoma growth and apoptosis. Hum Reprod Update 10:207–220
Takeuchi H, Kobori H, Kikuchi I, Sato Y, Mitsuhashi N 2000 A prospective randomized study comparing endocrinological and clinical effects of two types of GnRH agonists in cases of uterine leiomyomas or endometriosis. J Obstet Gynecol Res 26:325–331
Steinauer J, Pritts EA, Jackson R, Jacoby AF 2004 Systematic review of mifepristone for the treatment of uterine leiomyomata. Obstet Gynecol 103:1331–1336
Palomba S, Russo T, Orio Jr F, Tauchmanova L, Zupi E, Panici PL, Nappi C, Colao A, Lombardi G, Zullo F 2002 Effectiveness of combined GnRH analogue plus raloxifene administration in the treatment of uterine leiomyomas: a prospective, randomized, single-blind, placebo-controlled clinical trial. Hum Reprod 17:3213–3219
DeManno D, Elger W, Garg R, Lee R, Schneider B, Hess-Stumpp H, Schubert G, Chwalisz K 2003 Asoprisnil (J867): a selective progesterone receptor modulator for gynecological therapy. Steroids 68:1019–1032
Carr BR, Marshburn PB, Weatherall PT, Bradshaw KD, Breslau NA, Byrd W, Roark M, Steinkampf MP 1993 An evaluation of the effect of gonadotropin-releasing hormone analogs and medroxyprogesterone acetate on uterine leiomyomata volume by magnetic resonance imaging: a prospective, randomized, double blind, placebo-controlled, crossover trial. J Clin Endocrinol Metab 76:1217–1223
Ligon AH, Morton CC 2001 Leiomyomata: heritability and cytogenetic studies. Hum Reprod Update 7:8–14
Tsibris JCM, Segars J, Coppola D, Mane S, Wilbanks GD, O’Brien WF, Spellacy WN 2002 Insights from gene arrays on the development and growth regulation of uterine leiomyomata. Fertil Steril 78:114–121
Chegini N, Verala J, Luo X, Xu J, Williams RS 2003 Gene expression profile of leiomyoma and myometrium and the effect of gonadotropin releasing hormone analogue therapy. J Soc Gynecol Invest 10:161–171
Wang H, Mahadevappa M, Yamamoto K, Wen Y, Chen B, Warrington JA, Polan ML 2003 Distinctive proliferative phase differences in gene expression in human myometrium and leiomyomata. Fertil Steril 80:266–276
Weston G, Trajstman AC, Gargett CE, Manuelpillai U, Vollenhoven BJ, Rogers PA 2003 Fibroids display an anti-angiogenic gene expression profile when compared with adjacent myometrium. Mol Hum Reprod 9:541–549
Ahn WS, Kim KW, Bae SM, Yoon JH, Lee JM, Namkoong SE, Kim JH, Kim CK, Lee YJ, Kim YW 2003 Targeted cellular process profiling approach for uterine leiomyoma using cDNA microarray, proteomics and gene ontology analysis. Int J Exp Pathol 84:267–279
Quade BJ, Wang TY, Sornberger K, Dal Cin P, Mutter GL, Morton CC 2004 Molecular pathogenesis of uterine smooth muscle tumors from transcriptional profiling. Genes Chromosomes Cancer 40:97–108
Chegini N, Rong H, Dou Q, Kipersztok S, Williams RS 1996 Gonadotropin releasing hormone (GnRH) and GnRH receptor gene expression in human myometrial and leiomyomta and the direct action of GnRH analogs on myometrial smooth muscle cells interaction with ovarian steroids in vitro. J Clin Endocrinol Metab 81:3215–3221
Ding L, Xu J, Luo L, Chegini N 2004 Gonadotropin releasing hormone and transforming growth factor ? activate MAPK/ERK and differentially regulate fibronectin, type I collagen, and PAI-1 expression in leiomyoma and myometrial smooth muscle cells. J Clin Endocrinol Metab 89:5549–5557
Chegini N, Luo X, Ding L, Ripley D 2003 The expression of Smads and transforming growth factor ? receptors in leiomyoma and myometrium and the effect of gonadotropin releasing hormone analogue therapy. Mol Cell Endocrinol 209:9–16
Xu J, Luo X, Chegini N 2003 Differential expression, regulation, and induction of Smads, transforming growth factor-? signal transduction pathway in leiomyoma, and myometrial smooth muscle cells and alteration by gonadotropin-releasing hormone analog. J Clin Endocrinol Metab 88:1350–1361
Chegini N, Kornberg L 2003 Gonadotropin releasing hormone analogue therapy alters signal transduction pathways involving mitogen-activated protein and focal adhesion kinases in leiomyoma. J Soc Gynecol Invest 10:21–26
Chegini N, Ma C, Tang XM, Williams RS 2002 Effects of GnRH analogues, ’add-back’ steroid therapy, antiestrogen and antiprogestins on leiomyoma and myometrial smooth muscle cell growth and transforming growth factor-? expression. Mol Hum Reprod 8:1071–1078
Klausen C, Chang JP, Habibi HR 2002 Multiplicity of gonadotropin-releasing hormone signaling: a comparative perspective. Prog Brain Res 141:111–128
Mizutani T, Sugihara A, Nakamuro K, Terada N 1998 Suppression of cell proliferation and induction of apoptosis in uterine leiomyoma by gonadotropin-releasing hormone agonist (leuprolide acetate). J Clin Endocrinol Metab 83:1253–1255
Wu X, Blanck A, Olovsson M, Moller B, Lindblom B 2001 Expression of basic fibroblast growth factor (bFGF), FGF receptor 1 and FGF receptor 2 in uterine leiomyomas and myometrium during the menstrual cycle, after menopause and GnRHa treatment. Acta Obstet Gynecol Scand 80:497–504
Dou Q, Tarnuzzer RW, Williams RS, Schultz GS, Chegini N 1997 Differential expression of matrix metalloproteinases and their tissue inhibitors in leiomyomata: a mechanism for gonadotrophin releasing hormone agonist-induced tumour regression. Mol Hum Reprod 3:1005–1014
Chegini N, Tang XM, Ma C 1999 Regulation of transforming growth factor-?1 expression by granulocyte macrophage-colony-stimulating factor in leiomyoma and myometrial smooth muscle cells. J Clin Endocrinol Metab 84:4138–4143
Senturk LM, Sozen I, Gutierrez L, Arici A 2001 Interleukin 8 production and interleukin 8 receptor expression in human myometrium and leiomyoma. Am J Obstet Gynecol 184:559–566
Sozen I, Olive DL, Arici A 1998 Expression and hormonal regulation of monocyte chemotactic protein-1 in myometrium and leiomyomata. Fertil Steril 69:1095–1102
Gustavsson I, Englund K, Faxen M, Sjoblom P, Lindblom B, Blanck A 2000 Tissue differences but limited sex steroid responsiveness of c-fos and c-jun in human fibroids and myometrium. Mol Hum Reprod 6:55–59
Orii A, Masutani H, Nikaido T, Zhai YL, Kato K, Kariya M, Konishi I, Yodoi J, Fujii S 2002 Altered post-translational modification of redox factor 1 protein in human uterine smooth muscle tumors. J Clin Endocrinol Metab 87:3754–3759
Fukuhara K, Kariya M, Kita M, Shime H, Kanamori T, Kosaka C, Orii A, Fujita J, Fujii S 2002 Secreted frizzled related protein 1 is overexpressed in uterine leiomyomas, associated with a high estrogenic environment and unrelated to proliferative activity. J Clin Endocrinol Metab 87:1729–1736
Zhai YL, Nikaido T, Shiozawa T, Orii A, Fujii S 1999 Expression of cyclins and cyclin-dependent kinases in smooth muscle tumors of the uterus. Int J Cancer 84:244–250
Ma C, Chegini N 1999 Matrix metalloproteinases and tissue inhibitor of MMPs (TIMPs): expression in human myometrial smooth muscle cell and regulation by transforming growth factor b1. Mol Hum Reprod 5:950–954
Kakar SS, Winters SJ, Zacharias W, Miller DM, Flynn S 2003 Identification of distinct gene expression profiles associated with treatment of L?T2 cells with gonadotropin-releasing hormone agonist using microarray analysis. Gene 308:67–77
Wurmbach E, Yuen T, Ebersole BJ, Sealfon SC 2001 Gonadotropin-releasing hormone receptor-coupled gene network organization. J Biol Chem 276:47195–47201
Luo X, Ding L, Xu J, Chegini N 2005 Gene expression profiling of leiomyoma and myometrial smooth muscle cells in response to transforming growth factor-?. Endocrinology 146:1097–1118
Eisen MB, Spellman PT, Brown PO, Botstein D 1998 Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95:14863–14868
Tusher VG, Tibshirani R, Chu G 2001 Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98:5116–5121
Pavlidis P 2003 Using ANOVA for gene selection from microarray studies of the nervous system. Methods 31:282–289
Peterson LE 2003 Partitioning large-sample microarray-based gene expression using principal components analysis. Comput Methods Programs Biomed 70:107–119
Butte A 2002 The use and analysis of microarray data. Nat Rev Drug Discov 1:951–960
Johnsen SA, Subramaniam M, Janknecht R, Spelsberg TC 2002 TGF-? inducible early gene enhances TGF-?/Smad-dependent transcriptional responses. Oncogene 21:5783–5790
Levens E, Luo X, Ding L, Williams RS, Chegini N 2004 Differential expression of fibromodulin and Abl-interactor 2 in leiomyoma and myometrium and regulation by gonadotropin releasing hormone analogue (GnRHa) therapy. Fertil Steril 82(Suppl 2):S88
Dou Q, Zhao Y, Tarnuzzer RW, Rong H, Williams RS, Schultz GS, Chegini N 1996 Suppression of transforming growth factor-? (TGF-?) and TGF-? receptor messenger ribonucleic acid and protein expression in leiomyomata in women receiving gonadotropin-releasing hormone agonist therapy. J Clin Endocrinol Metab 81:3222–3230
Leng SX, Elias JA 1997 Interleukin-11. Int J Biochem Cell Biol 29:1059–1062
Tang W, Geba GP, Zheng T, Ray P, Homer RJ, Kuhn III C, Flavell RA, Elias JA 1996 Targeted expression of IL-11 in the murine airway causes lymphocytic inflammation, bronchial remodeling, and airways obstruction. J Clin Invest 98:2845–2853
Zhu Z, Lee CG, Zheng T, Chupp G, Wang J, Homer RJ, Noble PW, Hamid Q, Elias JA 2001 Airway inflammation and remodeling in asthma. Lessons from interleukin 11 and interleukin 13 transgenic mice. Am J Respir Crit Care Med 164:S67–S70
Zimmerman MA, Selzman CH, Reznikov LL, Raeburn CD, Barsness K, McIntyre Jr RC, Hamiel CR, Harken AH 2002 Interleukin-11 attenuates human vascular smooth muscle cell proliferation. Am J Physiol 283:H175–H180
Bamba S, Andoh A, Yasui H, Makino J, Kim S, Fujiyama Y 2003 Regulation of IL-11 expression in intestinal myofibroblasts: role of c-Jun AP-1- and MAPK-dependent pathways. Am J Physiol 285:G529–G538
Karpovich N, Chobotova K, Carver J, Heath JK, Barlow DH, Mardon HJ 2003 Expression and function of interleukin-11 and its receptor in the human endometrium. Mol Hum Reprod 9:75–80
Ding L, Luo X Chegini N 2004 The expression of IL-13 and IL-15 in leiomyoma and myometrium and their influence on TGF-? and proteases expression in leiomyoma and myometrial smooth muscle cells and SKLM, leiomyosarcoma cell line. J Soc Gynecol Invest 11:319A
Mak JC, Chuang TT, Harris CA, Barnes PJ 2002 Increased expression of G protein-coupled receptor kinases in cystic fibrosis lung. Eur J Pharmacol 436:165–172
Simon V, Mhaouty-Kodja S, Legrand C, Cohen-Tannoudji J 2001 Concomitant increase of G protein-coupled receptor kinase activity and uncoupling of ?-adrenergic receptors in rat myometrium at parturition. Endocrinology 142:1899–1905
Simon V, Robin MT, Legrand C, Cohen-Tannoudji J 2003 Endogenous G protein-coupled receptor kinase 6 triggers homologous ?-adrenergic receptor desensitization in primary uterine smooth muscle cells. Endocrinology 144:3058–3066
Krasel C, Dammeier S, Winstel R, Brockmann J, Mischak H, Lohse MJ 2001 Phosphorylation of GRK2 by protein kinase C abolishes its inhibition by Calmodulin. J Biol Chem 276:1911–1915
Maira M, Martens C, Batsché E, Gauthier Y, Drouin J 2003 Dimer-specific potentiation of NGFI-B (Nur77) transcriptional activity by the protein kinase A pathway and AF-1-dependent coactivator recruitment. Mol Cell Biol 23:763–776
Drouin, J, Maira MH, Philips A 1998 Novel mechanism of action for Nur77 and antagonism by glucocorticoids: a convergent mechanism for CRH activation and glucocorticoid repression of POMC gene transcription. J. Steroid Biochem Mol Biol 65:59–63
Fernandez P, Brunel MF, Jimenez MA, Saez JM, Cereghini S, Zakin MM 2000 Nuclear receptors Nor1 and NGFI-B/Nur77 play similar, albeit distinct, roles in the hypothalamo-pituitary-adrenal axis. Endocrinology 141:2392–2400
Gelman L, Zhou G, Fajas L, Raspe E, Fruchart JC, Auwerx J 1999 p300 Interacts with the N- and C-terminal part of PPAR2 in a ligand-independent and -dependent manner, respectively. J Biol Chem 274:7681–7688
Sadie H, Styger G, Hapgood J 2003 Expression of the mouse gonadotropin-releasing hormone receptor gene in T3–1 gonadotrope cells is stimulated by cyclic 3',5'-adenosine monophosphate and protein kinase A, and is modulated by Steroidogenic factor-1 and Nur77. Endocrinology 144:1958–1971
Wilson TE, Mouw AR, Weaver CA, Milbrandt J, Parker KL 1993 The orphan nuclear receptor NGFI-B regulates expression of the gene encoding steroid 21-hydroxylase. Mol Cell Biol 13:861–868
Song KH, Park JI, Lee MO, Soh J, Lee K, Choi HS 2001 LH induces orphan nuclear receptor Nur77 gene expression in testicular Leydig cells. Endocrinology 142:5116–123
Zhang, P, Mellon SH 1997 Multiple orphan nuclear receptors converge to regulate rat P450c17 gene transcription: novel mechanisms for orphan nuclear receptor action. Mol Endocrinol 11:891–904
Rajpal A, Cho YA, Yelent B, Koza-Taylor PH, Li D, Chen E, Whang M, Kang C, Turi TG, Winoto A 2003 Transcriptional activation of known and novel apoptotic pathways by Nur77 orphan steroid receptor. EMBO J 22:6526–6536
Castro-Obregon S, Rao RV, del Rio G, Chen SF, Poksay KS, Rabizadeh S, Vesce S, Zhang XK, Swanson RA, Bredesen DE 2004 Alternative, nonapoptotic programmed cell death: mediation by arrestin 2, ERK2, and Nur77. J Biol Chem 279:17543–17553
Li H, Kolluri SK, Gu J, Dawson MI, Cao X, Hobbs PD, Lin B, Chen GLJ, Lin F, Xie Z, Fontana JA, Reed JC, Zhang X 2000 Cytochrome c release and apoptosis induced by mitochondrial targeting of orphan receptor TR3 (nur77, NGFI-B). Science 289:1159–1164
Tien ES, Davis JW, Vanden Heuvel JP 2004 Identification of the CREB-binding protein/p300-interacting protein CITED2 as a peroxisome proliferator-activated receptor coregulator. J Biol Chem 279:24053–24063
Kranc KR, Bamforth SD, Braganca J, Norbury C, van Lohuizen M, Bhattacharya S 2003 Transcriptional coactivator Cited2 induces Bmi1 and Mel18 and controls fibroblast proliferation via Ink4a/ARF. Mol Cell Biol 23:7658–7666
Yin Z, Haynie J, Yang X, Han B, Kiatchoosakun S, Restivo J, Yuan S, Prabhakar NR, Herrup K, Conlon RA, Hoit BD, Watanabe M, Yang YC 2002 The essential role of Cited2, a negative regulator for HIF-1, in heart development and neurulation. Proc Natl Acad Sci USA 99:10488–10493
Kranc K, Shikama R, Lyon NJ, La Thangue NB 1997 The p300/CBP family: integrating signals with transcription factors and chromatin. Trends Cell Biol 7:230–236
Puri PL, Avantaggiati ML, Balsano C, Sang N, Graessmann A, Giordano A, Levrero M 1997 p300 is required for MyoD-dependent cell cycle arrest and muscle-specific gene transcription. EMBO J 16:369–383
Hjoberg J, Le L, Imrich A, Subramaniam V, Mathew SI, Vallone J, Haley KJ, Green FHY, Shore SA, Silverman ES 2004 Induction of early growth-response factor 1 by platelet-derived growth factor in human airway smooth muscle. Am J Physiol 286:L817–L825
Thiel G, Cibelli G 2002 Regulation of life and death by the zinc finger transcription factor Egr-1. J Cell Physiol 193:287–292
Shozu M, Murakami K, Segawa T, Kasai T, Ishikawa H, Shinohara K, Okada M, Inoue M 2004 Decreased expression of early growth response-1 and its role in uterine leiomyoma growth. Cancer Res 64:4677–4684
Inoue A, Omoto Y, Yamaguchi Y, Kiyama R, Hayashi SI 2004 Transcription factor EGR3 is involved in the estrogen-signaling pathway in breast cancer cells. J Mol Endocrinol 32:649–661
Liu C, Yao J, de Belle I, Huang RP, Adamson E, Mercola D 1999 The transcription factor EGR-1 suppresses transformation of human fibrosarcoma HT1080 cells by coordinated induction of transforming growth factor-?1, fibronectin, and plasminogen activator inhibitor-1. J Biol Chem 274:4400–4411
Mercier G, Turque N, Schumacher M 2001 Early activation of transcription factor expression in Schwann cells by progesterone. Mol Brain Res 97:137–148
Xi H, Kersh GJ 2004 Sustained early growth response gene 3 expression inhibits the survival of CD4/CD8 double-positive thymocytes. J Immunol 173:340–348
Cook T, Urrutia R 2000 TIEG proteins join the Smads as TGF-?-regulated transcription factors that control pancreatic cell growth. Am J Physiol 278:G513–G521
Ribeiro A, Bronk SF, Roberts PJ, Urrutia R, Gores GJ 1999 The transforming growth factor ?1-inducible transcription factor TIEG1, mediates apoptosis through oxidative stress. Hepatology 30:1490–1497
Chen F, Ogawa K, Nagarajan RP, Zhang M, Kuang C, Chen Y 2003 Regulation of TG-interacting factor by transforming growth factor-?. Biochem J 371:257–263
Melhuish TA, Gallo CM, Wotton D 2001 TGIF2 interacts with histone deacetylase 1 and represses transcription. J Biol Chem 276:32109–32114
Coqueret O 2003 New roles for p21 and p27 cell-cycle inhibitors: a function for each cell compartment? Trends Cell Biol 13:65–70
Blagosklonny MV 2003 Apoptosis, proliferation, differentiation: in search of the order. Semin Cancer Biol 13:97–105
Stewart EA, Friedman AJ, Peck K, Nowak RA 1994 Relative overexpression of collagen type I and collagen type III messenger ribonucleic acids by uterine leiomyomas during the proliferative phase of the menstrual cycle. J Clin Endocrinol Metab 79:900–906(Xiaoping Luo, Li Ding, Ji)
Address all correspondence and requests for reprints to: Dr. Nasser Chegini, Department of Obstetrics and Gynecology, University of Florida, Box 100294, Gainesville, Florida 32610. E-mail: cheginin@obgyn.ufl.edu.
Abstract
Gene microarray was used to characterize the molecular environment of leiomyoma and matched myometrium during growth and in response to GnRH analog (GnRHa) therapy as well as GnRHa direct action on primary cultures of leiomyoma and myometrial smooth muscle cells (LSMC and MSMC). Unsupervised and supervised analysis of gene expression values and statistical analysis in R programming with a false discovery rate of P 0.02 resulted in identification of 153 and 122 differentially expressed genes in leiomyoma and myometrium in untreated and GnRHa-treated cohorts, respectively. The expression of 170 and 164 genes was affected by GnRHa therapy in these tissues compared with their respective untreated group. GnRHa (0.1 μM), in a time-dependent manner (2, 6, and 12 h), targeted the expression of 281 genes (P 0.005) in LSMC and MSMC, 48 of which genes were found in common with GnRHa-treated tissues. Functional annotations assigned these genes as key regulators of processes involving transcription, translational, signal transduction, structural activities, and apoptosis. We validated the expression of IL-11, early growth response 3, TGF-?-induced factor, TGF-?-inducible early gene response, CITED2 (cAMP response element binding protein-binding protein/p300-interacting transactivator with ED-rich tail), Nur77, growth arrest-specific 1, p27, p57, and G protein-coupled receptor kinase 5, representing cytokine, common transcription factors, cell cycle regulators, and signal transduction, at tissue levels and in LSMC and MSMC in response to GnRHa time-dependent action using real-time PCR, Western blotting, and immunohistochemistry. In conclusion, using different, complementary approaches, we characterized leiomyoma and myometrium molecular fingerprints and identified several previously unrecognized genes as targets of GnRHa action, implying that local expression and activation of these genes may represent features differentiating leiomyoma and myometrial environments during growth and GnRHa-induced regression.
Introduction
LEIOMYOMAS ARE BENIGN uterine tumors originating from the transformation of myometrial smooth muscle cells and/or connective tissue fibroblasts during the reproductive years. Despite recent progress in our understanding of leiomyoma’s molecular environment, the identity of a molecules(s) responsible for initiating the transformation of these cells into leiomyoma remains unknown. However, ovarian steroids are essential for leiomyoma growth, and creating a hypoestrogenic condition with GnRH analog (GnRHa) therapy is often used for their medical management (1, 2, 3, 4, 5, 6, 7). Hypoestrogenic conditions created by GnRHa therapy affect both leiomyoma and myometrium; however, clinical observations indicate a difference in their responses to changes in the hormonal milieu (7). In addition to GnRHa therapy, clinical and preclinical assessments of selective estrogen and progesterone receptor modulators, either alone or in combination with GnRHa, have shown efficacy in leiomyoma regression (4, 5, 6).
With respect to the leiomyoma molecular environment, several genome-wide allele-typing studies have evaluated the association between genomic instability and the pathogenesis of leiomyoma (for review, see Ref.8). These studies have led to the identification of several candidate genes; however, in the majority of cases evidence of genomic instability is either lacking or inconsistent, implying that the existence of unrecognized pathways leads to the development of leiomyoma. Additional studies have provided support for various autocrine/paracrine regulators in the pathogenesis of leiomyoma, including local estrogen production, growth factors, cytokines, chemokines, and their receptors, whose expression are regulated by ovarian steroids (1, 2). These studies in many instances demonstrated altered expression of these factors and/or their receptors in leiomyoma compared with normal myometrium. In recent years, cDNA microarray has been used as a high throughput method to identify a large number of differentially expressed and regulated genes in various tissues and cells. Using this approach, several recent studies, including our own, have also assisted in fingerprinting the gene expression profile of leiomyoma and myometrium during the menstrual cycle (9, 10, 11, 12, 13, 14). However, only the expression of a few of these newly identified genes has been validated, and their regulation and correlation with pathogenesis of leiomyoma remain to be investigated.
With respect to GnRHa therapeutic action, it is traditionally believed to act primarily at the level of the pituitary-gonadal axis, and by suppressing ovarian steroid production, it causes leiomyoma regression. However, the identification of GnRH and GnRH receptor expression in several peripheral tissues, including the uterus, has indicated an autocrine/paracrine role for GnRH and additional sites of action for GnRHa therapy (15, 16, 17, 18, 19, 20). We provided support for this concept by demonstrating the expression of GnRH as well as GnRH I and II receptors mRNA in leiomyoma and myometrium and their isolated smooth muscle cells (15, 16). Several in vitro studies, including our own, have also demonstrated GnRHa direct action on various cell types derived from peripheral tissues resulting in alteration of cell growth, apoptosis, and the expression of cell cycle proteins, growth factors, pro- and antiinflammatory cytokines, proteases, and protease inhibitors (1, 16, 17, 18, 19, 20, 21, 22, 23). Local expression and differential regulation of these genes influence cell proliferation, differentiation, migration, inflammatory response, angiogenesis, expression of adhesion molecules, extracellular matrix (ECM) turnover, apoptosis, etc., processes that are central to leiomyoma growth and regression (1, 2, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32). Microarray studies, including our own small scale array study, have also identified the expression profile of additional genes targeted by GnRHa in the murine gonadotrope tumor cell line L?T2, the human breast tumor cell line MCF-7, leiomyoma, and myometrium (10, 33, 34).
We designed the present study to further define the molecular environments of leiomyoma and matched myometrium during the early to midluteal phase of the menstrual cycle, which is characterized by elevated production of ovarian steroids, compared with tissues obtained from hormonally suppressed patients receiving GnRHa therapy. We also evaluated the direct action of GnRHa on global gene expression and their regulation in leiomyoma and myometrial cells isolated from the untreated tissue cohort. These approaches enabled us to identify the expression profiles of genes targeted by GnRHa. We validated the expression of 10 of these genes in these cohorts and concluded that local expression and activation of these genes may represent features differentiating leiomyoma and myometrial molecular environments during growth as well as GnRHa-induced regression.
Materials and Methods
Portions of leiomyoma and matched myometrium were collected from premenopausal women (n = 6) who were scheduled to undergo hysterectomy for indications related to symptomatic leiomyomas. Three of the patients received GnRHa therapy for 3 months before surgery. The untreated patients did not receive any medications (including hormonal therapy) in the previous 3 months before surgery, and based on endometrial histology and the patient’s last menstrual period, they were in early to midsecretory phase of their menstrual cycles (Fig. 1). To maintain a standard, all leiomyomas selected for this study were between 2–3 cm in diameter. After collection, the tissues were divided into several pieces and immediately snap-frozen and stored in liquid nitrogen for additional processing, fixed and paraffin embedded for histological evaluation and immunohistochemistry, or used for isolation of leiomyoma and myometrial smooth muscle cells and culturing (16, 18). The tissues were collected at the University of Florida-affiliated Shands Hospital with prior approval from the institutional review board.
FIG. 1. The chart describes the experimental protocol for the microarray analysis of leiomyoma and matched myometrial tissues from GnRHa-treated and untreated cohorts as well as LSMC and MSMC treated with GnRH, TGF-?, and TGF-? type II receptor antisense, data analysis, and verification using real-time PCR, Western blotting, and immunohistochemistry (see Ref.35 for the results of TGF-? and TGF-? type II receptor antisense cohorts).
Isolation and culture of leiomyoma and myometrial smooth muscle cells (LSMC and MSMC)
To determine the direct action of GnRHa on global gene expression in LSMC and MSMC, the cells were isolated from untreated group and cultured as previously described (16, 20) (Fig. 1). Before use in these experiments, the primary cultures of LSMC and MSMC were characterized using immunofluorescence microscopy and antibodies to -smooth muscle actin, desmin, and vimentin (16, 18). The cells were cultured in six-well plates at an approximate density of 106 cells/well in DMEM-supplemented medium containing 10% fetal bovine serum. After reaching visual confluence, often after 2–3 d, the cells were washed in serum-free medium and incubated for 24 h under serum-free, phenol red-free conditions (20). The cells were then treated with 0.1 μM GnRHa (leuprolide acetate, Sigma-Aldrich Corp., St. Louis, MO) for a period of 2, 6, or 12 h (16).
cDNA microarray and gene expression profiling
Total cellular RNA was isolated from the tissues and cells using TRIzol (Invitrogen Life Technologies, Inc., Carlsbad, CA). The isolated RNA was treated with deoxyribonuclease I (Roche, Indianapolis, IN) at 1 U/10 μg RNA, heat-inactivated, and subjected to additional purification using the RNeasy kit (Qiagen, Valencia, CA). The purified RNA was then subjected to amplification by RT using SuperScript Choice system (Invitrogen Life Technologies, Inc.), with final concentrations in 20 μl first strand reaction of 100 pmol HPLC-purified T7-(dT)24 primer (Genset Corp., La Jolla, CA), 8 μg RNA, 1x first strand buffer, 10 mM dithiothreitol, 500 μM of each deoxy-NTP, and 400 U Superscript II reverse transcriptase (T7 Megascript kit. Ambion, Austin, TX). The second strand cDNA synthesis was performed in a 150-μl reaction consisting of (final concentrations) 1x second strand reaction buffer, 200 μM of each deoxy-NTP, 10 U DNA ligase, 40 U DNA polymerase I, and 2 U ribonuclease H (Invitrogen Life Technologies, Inc.), and double-stranded cDNA was purified by phenol/chloroform extraction using phase lock gels (Eppendorf-5 Prime, Inc., Westbury, NY) and ethanol precipitation (10).
Five micrograms of purified cDNA were reverse transcribed using an Enzo BioArray high yield RNA transcript labeling kit (Affymetrix, Santa Clara, CA), and the product was purified in RNeasy spin columns (Qiagen) according to the manufacturer’s instructions. After an overnight ethanol precipitation, cRNA was resuspended in 15 μl diethyl pyrocarbonate-treated water (Ambion) and quantified using UV-visible spectrophotometer. After quantification of cRNA to reflect any carryover of unlabeled total RNA according to an equation provided by Affymetrix (adjusted cRNA yield = cRNA (μg) measured after in vitro transcription (starting total RNA) (fraction of cDNA reaction used in in vitro transcription), 20 μg cRNA were fragmented (0.5 μg/μl) according to Affymetrix instructions using the 5x fragmentation buffer containing 200 mM Tris acetate (pH 8.1), 500 mM potassium acetate, and 150 mM magnesium acetate (Sigma-Aldrich Corp.). Twenty micrograms of the adjusted fragmented cRNA were added to a 300 μl hybridization mixture containing (final concentrations) 0.1 mg/ml herring sperm DNA (Promega Corp., Madison, WI), 0.5 mg/ml acetylated BSA (Invitrogen Life Technologies, Inc.), and 2x 2-(N-morpholino)ethanesulfonic acid hybridization buffer (Sigma-Aldrich Corp.). Two hundred microliters of the mixture were hybridized to the human U95A Affymetrix GeneChip arrays, purchased at the same time from the same lot number and used within 2 wk of purchase to maintain a standard between this and the experiments described by Luo et al. (35). In addition, an aliquot of random samples were first hybridized to an Affymetrix Test 2 Array to determine sample quality. After meeting recommended criteria for use of the expression arrays, hybridization was performed for 16 h at 45 C, followed by washing, staining, signal amplification with biotinylated antistreptavidin antibody, and staining according to the manufacturer’s protocol.
Microarray data analysis
The Chips were scanned to obtain the raw hybridization values using Affymetrix Genepix 5000A scanner. Differences in fluorescence spot intensities representing the rate of hybridization between the 25-bp oligonucleotides and their mismatches were analyzed by multiple decision matrices to determine the presence or absence of gene expression and to derive an average difference score representing the relative level of gene expression. The fluorescence spot intensities, qualities, and local background were assessed automatically by Genepix software with manual supervision to detect any inaccuracies in automated spot detection. Background and noise corrections were made to account for nonspecific hybridization and minor variations in hybridization conditions. The net hybridization values for each array were subjected to global normalization as previously described (10). To identify changes in the pattern of gene expression, the average and SD of globally normalized values were calculated, followed by subtraction of the mean value from each observation and division by the SD. The mean transformed expression value of each gene in transformed dataset was set at 0, and the SD was set at 1 (10).
The transformed gene expression values were subjected to Affymetrix Analysis Suite version 5.0. Briefly, probe sets that were flagged as absent on all arrays using default settings were removed from the datasets. After application of this filtering, the dataset was reduced from 12,625 to 8,580 probe sets. The gene expression value of the remaining probe sets was then subjected to unsupervised and supervised learning, discrimination analysis, and cross-validation (36, 37, 38, 39, 40). After variation filtering, the coefficient of variation was calculated for each probe set across all chips, and the probe sets were ranked by the coefficient of variation of the observed single intensities. The expression values of the selected genes were then subjected to statistical analysis in an R programming environment that assesses multiple test corrections to identify statistically significant gene expression values (38, 39, 40). The gene expression values with a statistical significance of P 0.02 (by ANOVA and Tukey’s test) between leiomyoma and myometrium from GnRH-treated and untreated cohorts and P 0.005 between GnRHa-treated and untreated cells (control) were selected. The validity of gene sets identified at these P values in predicting treatment class was established using "leave-one-out" cross-validation where the data from one array was left out of the training set, and probe sets with differential hybridization signal intensities were identified from the remaining arrays (40). Hierarchical clustering and K-means analysis were performed and viewed with the algorithms in the software packages Cluster and TreeView (36).
Gene classification and ontology assessment
The selected differentially expressed and regulated genes in the above cohorts were subjected to functional annotation and visualization using Database for Annotation, Visualization, and Integrated Discovery (DAVID; www.david.niaid.nih.gov) software. The integrated GoCharts assigns genes to specific ontology functional categories based on selected classifications.
Quantitative real-time PCR
Real-time PCR was used for verification of 10 differentially expressed and regulated genes identified in leiomyoma and myometrium as well as LSMC and MSMC from untreated and GnRHa-treated cohorts. The selection of these genes was based not only on their expression values (up or down-regulation), but on classification and biological functions important to leiomyoma growth and regression, and on regulation by ovarian steroids, GnRHa and TGF-? (35), as indicated in the literature. They are IL-11, EGR3 (early growth response 3), TGIF (TGF-?-induced factor, TIEG (TGF-?-inducible early gene response), CITED2 [cAMP response element binding protein-binding protein (CBP)/p300-interacting transactivator with ED-rich tail], Nur77, p27, p57, Gas-1 (growth arrest-specific 1) and G protein-coupled receptor kinase 5 (GPRK), representing cytokines, transcription factors, cell cycle regulators, and signal transduction. Real-time PCR was carried out as previously described using TaqMan and ABI PRISM 7700 Sequence System and Sequence Detection System 1.6 software (Applied Biosystems, Foster City, CA) (16). Results were analyzed using the comparative method and after normalization of expression values to 18S rRNA expression according to the manufacturer’s guidelines (Applied Biosystems) as previously described (16).
Western blotting and immunohistochemical localization
For Western blotting, total protein was isolated from small portions of GnRHa-treated and untreated leiomyoma and myometrium as well as the GnRHa-treated and untreated cells and subjected to analysis as previously described (16, 17, 18). The blots were incubated with anti-TIEG antibody (provided by Dr. Thomas Spelsberg, Mayo Clinic, Rochester, MN) (41); TGIF, EGR3, Nur77, CITED2, p27, p57, and Gas1 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); and IL-11 antibodies (R&D Systems, Minneapolis, MN). The immunostained proteins were visualized using enhanced chemiluminescence reagents (Amersham Biosciences, Piscataway, NJ) as previously described (16).
Tissue sections prepared from formalin-fixed and paraffin-embedded leiomyoma and myometrium were subjected to microwave treatment and immunostained using antibodies to IL-11, TGIF, TIEG, EGR3, CITED2, Nur77, p27, p57, and Gas1 at 5 μg immunoglobulin G/ml for 2–3 h at room temperature. After additional standard processing, a chromogenic reaction was detected with 3,3'-diaminobenzidine tetrahydrochloride solution (16). In some instances the slides were counterstained with hematoxylin. Omission of primary antibodies and incubation of tissue sections with nonimmune mouse, rabbit, and goat IgG instead of primary antibodies at the same concentration served as controls (16, 17, 18).
Results
Gene expression profiles in leiomyoma and myometrium
Using global gene expression profiling, we characterized the gene expression profile of leiomyoma and matched myometrium and their transcriptional changes in response to hormonal transition induced by GnRHa therapy. The initial simultaneous assessment of the gene expression values in leiomyoma, myometrium, and their isolated smooth muscle cells from untreated as well as GnRHa- and TGF-?-treated (35) cohorts revealed uniform expression of transcripts for the housekeeping genes glyceraldehyde-3-phosphate dehydrogenase, ?-actin, and a large number of ribosomal proteins, indicating that the expression profile is consistent with established standards for gene expression analysis. After unsupervised and supervised learning, the gene expression values were subjected to statistical analysis in R programming and ANOVA with false discovery rate selected at P 0.02.
Using the above analysis, we identified a total of 153 genes, including 19 established sequence tags (EST), or 1.23% of the genes, and 122 genes including 21 EST, or 0.98% of the genes on the array, as differentially expressed in leiomyoma compared with matched myometrium from untreated and GnRHa-treated tissues, respectively. Hierarchical clustering and Tree-View analysis separated the genes in each cohort into distinctive clusters with sufficient variability, allowing division into their respective subgroups (Fig. 2). Of the 153 (excluding 19 EST) genes in untreated cohorts, the expression of 82 genes was up-regulated and that of 52 genes was down-regulated in leiomyoma compared with myometrium (Table 1). Of the 122 genes (excluding 21 EST) in leiomyoma and myometrium from patients who received GnRHa therapy, the expression of 34 genes was up-regulated and that of 67 genes was down-regulated in leiomyoma compared with myometrium, respectively (Table 2). Analysis of the variance-normalized mean (K-means) separated the differentially expressed and regulated genes in these cohorts into four distinctive clusters, with genes in clusters A and D displaying a tissue-specific response, whereas genes in clusters B and C showed a regulatory response to GnRHa therapy (Fig. 3). To further differentiate the molecular environment of leiomyoma from myometrium and their responses to GnRHa therapy, we compared the gene expression profiles in GnRHa-treated and corresponding untreated tissues. The results indicated that the expressions of 170 (excluding 26 EST) and 167 (excluding 31 EST) genes were targeted by GnRHa therapy in leiomyoma and myometrium compared with their respective untreated cohorts (Tables 3 and 4). Of these genes, 96 and 89 transcripts were down-regulated in leiomyoma and myometrium, respectively, by GnRHa therapy compared with their respective untreated tissues, and three transcripts were commonly found among the tissues in these cohorts, with different regulatory patterns of expression (compare Tables 3 and 4).
FIG. 2. Hierarchical clustering analysis of differentially expressed genes in leiomyoma and matched myometrium from untreated (f and m 315, 316, and 317) and GnRHa-treated (f and m 287, 312, and 314) groups identified after unsupervised and supervised analysis in an R programming environment and ANOVA with a false discovery rate of P 0.02 (38 39 40 ). Each column represents data from a single cohort, with shades of red and green indicating up- or down-regulation of a given gene according to the color scheme shown below. Genes represented by rows were clustered according to their similarities in expression patterns for each tissue and treatment. The dendrogram displaying similarity of gene expression among the cohorts is shown on top of the overview image, and relatedness of the arrays is denoted by distance to the node linking the arrays. The gene-tree shown at the left of the image corresponds to the degree of similarity (Pearson correlation) of the pattern of expression for genes across the experiments. The clustering divided the genes into five clusters, designated A–D, and their zoomed images are presented in A–D. Genes that appear more than once are represented by multiple clones on arrays.
TABLE 1. Categorical list of differentially expressed genes in leiomyoma and matched myometrium
TABLE 2. Categorical list of differentially expressed genes in leiomyoma and matched myometrium in response to GnRHa therapy
FIG. 3. K-means clustering of genes regulated in leiomyoma and matched myometrium during growth and under the influence of GnRH therapy. The gene expression values identified in these cohorts described in Fig. 1 were subjected to k-means clustering, grouping the genes into clusters based on similarity of expression in GnRHa-treated and untreated cohorts. The analysis grouped the genes into four clusters (A–D). The rows represent the genes, and columns represent the samples (f and m) from GnRHa-treated (287, 312, and 314) and untreated (315, 316, and 317) cohorts, with shades of red and green indicating up- or down-regulation of a given gene, which are clustered according to their similarities in expression patterns.
TABLE 3. Categorical list of differentially expressed genes in leiomyoma from GnRHa-treated vs. untreated group
TABLE 4. Categorical list of differentially expressed genes in myometrium from GnRHa-treated vs. untreated group
Since the present study was completed, a few microarray studies have reported the gene expression profiles of leiomyoma and myometrium (9, 10, 11, 12, 13, 14). We performed a comparative analysis using the genes identified at P 0.02 in the untreated leiomyoma and matched myometrium of our study and the list of genes reported in four of these studies, searching for a set of commonly expressed genes. The comparison identified two genes in our study in common with at least one of these studies. However, lowering the false discovery rate of our selection to P 0.05 enabled us to identify a larger number of genes (422, including 49 EST), of which 11 transcripts were found in common with other studies (Table 5).
TABLE 5. List of common gene found in present and other studies identified by their respective citations
Gene ontology assessment and division of differentially expressed genes into similar functional categories indicated that the products of a large percentage of these genes (40–67%), in leiomyoma and myometrium from both GnRHa-treated and untreated cohorts are involved in metabolic processes, catalytic activities, binding, signal transduction, transcriptional and translational activities, cell cycle regulation, cell and tissue structure, etc. (Tables 1–4 and 6). In addition, 15–23% of the genes were either functionally unclassified, or their roles in biological process are still unknown.
TABLE 6. Gene ontology assessment and division of the genes in leiomyoma (LYM) and myometrium (MYM) in untreated and in response to GnRH therapy as well as genes regulated in LSMC and MSMC in response to GnRH treatment into similar functional categories with the percentage of gene in each group presented
Time-dependent action of GnRHa on gene expression profile of LSMC and MSMC
To obtain a comprehensive picture of transcriptional changes induced by GnRHa’s direct action in leiomyoma and myometrium, LSMC and MSMC were treated with GnRHa (0.1 μM) for 2, 6, and 12 h, and their isolated RNA were subjected to microarray analysis. Based on the same data analysis criteria described above with a false discovery rate of P 0.005, we identified 281 genes, including 36 EST, or 2.2% of the genes on the array, displaying differential expression in response to a time-dependent GnRHa action compared with untreated controls in LSMC ad MSMC (Fig. 4). Hierarchical clustering analysis also separated these genes into different clusters in response to the time-dependent action of GnRHa in LSMC and MSMC, with expression patterns sufficiently different to cluster into their respective subgroups (Fig. 4). Analysis of the variance-normalized mean (K-means) separated the differentially expressed and regulated genes in these cohorts into four distinctive clusters, with genes in clusters A and D displaying a cell-specific response, whereas genes in clusters B and C showed regulatory behaviors to GnRHa’s time-dependent action (Fig. 5). Among these genes, the transcripts of 48 genes were identified as commonly expressed in LSMC and the original leiomyomas from the untreated cohort, although they many be differently regulated (Table 7).
FIG. 4. Hierarchical clustering analysis of 281 differentially expressed and regulated genes in LSMC (f) and MSMC (m) in response to GnRHa (0.01 μM) treatment for 2, 6, and 12 h or in untreated controls (C). The genes were identified after unsupervised and supervised analyses of the expression values, statistical analysis in an R programming environment, and ANOVA with a false discovery rate selected at P 0.005 (38 39 ). Each column represents data from a single treatment, with shades of red and green indicating up- or down-regulation of a given gene. Genes represented by rows were clustered according to their similarities in expression patterns for each cell type and treatment. The dendrogram displaying similarity of gene expression among the cohorts is shown on top of the overview image, with array relatedness denoted by distance to the node linking the arrays. The gene-tree shown at the left of the image corresponds to the degree of similarity (Pearson correlation) of the pattern of expression for genes across the experiments. The clustering divided the genes into four clusters, designated as A–D, and their zoomed images are presented in A–D. Genes that appear more than once are represented by multiple clones on arrays.
FIG. 5. K-means clustering of the gene expression values identified in response to time-dependent action of GnRHa in LSMC and MSMC described in Fig. 1. The analysis grouped the genes into six clusters (A–F) based on similarity of expression in response to time-dependent action of GnRHa. The rows represent the genes, and columns represent the samples (f and m) and treatments with GnRHa for 2, 6, and 12 h and untreated controls (C), with shades of red and green indicating up- or down-regulation of a given gene, which are clustered according to their similarities in expression patterns. Line graphs display the SD for each cluster in MSMC and LSMC in response to GnRHa time-dependent actions compared with untreated control (Ctrl).
TABLE 7. Common genes between GnRHa-treated leiomyoma and LSMC
Gene ontology and functional annotation of these genes into similar functional categories also indicated that in LSMC and MSMC, similar to their original tissues, the majority of the gene products are involved in cellular processes, catalytic activities, binding, signal transduction, transcriptional and translational activities, metabolism, cell cycle regulation, and cellular structure (Table 6). The time-dependent actions of GnRHa on the expression of a selective group of genes in these functional categories in LSMC and MSMC are shown in Fig. 6.
FIG. 6. The expression profile of a selected group of genes representing growth factors/cytokines/polypeptide hormones/receptors (first row), intracellular signal transduction pathways (second row), transcription factors (third row), cell cycle regulators (fourth row), and cell adhesion/ECM/cytoskeletons (fifth row) in response to the time-dependent action of GnRHa in LSMC and MSMC. Values on the y-axis represent an arbitrary unit derived from the mean gene expression value for each factor after supervised analysis, statistical analysis in an R programming environment, and ANOVA as described in Fig. 4, with gene expression values for the untreated controls (Ctrl) set at 1.
Verification of gene transcripts in leiomyoma, myometrium, LSMC, and MSMC
Among the differentially expressed and regulated genes identified in these tissues and cells, we selected 10 genes for verification using real-time PCR, Western blotting, and immunohistochemistry. The genes selected for validation were IL-11, CITED2, Nur77, EGR3, TGIF, TIEG, p27, p57, Gas-1, and GPRK5, representing cytokines, transcription factors, cell cycle regulators, and signal transduction. The pattern of expression of these genes in leiomyoma and myometrium from untreated and GnRHa-treated cohorts (Fig. 7) and in LSMC and MSMC treated with GnRHa for 2, 6, and 12 h (Fig. 8), as determined by real-time PCR, closely overlapped with their expression profiles identified by the microarray analysis.
FIG. 7. Comparative analysis of the expression profile of 10 genes identified as described in Fig. 1 as differentially expressed in response to GnRH therapy in leiomyoma and matched myometrium and in an untreated group by microarray and real-time PCR. Values on the y-axis represent an arbitrary unit derived from the mean expression value for each gene, with values for the untreated controls (Crtl) set at 1. Total RNA isolated from these tissues was used for both microarray analysis and real-time PCR validating the expression of IL-11, EGR3, CITED2, Nur77, TIEG, TGIF, p27, p57, Gas-1, and GPRK5. On the y-axis, untreated myometrium and leiomyoma are designated Unt-MM and Un-LM, respectively, and GnRH-treated myometrium and leiomyoma are designated GnRH-Trt MM and GnRH-Trt LM, respectively.
FIG. 8. Comparative analysis of the expression profile of 10 genes identified as described in Fig. 4 as differentially expressed and regulated in response to GnRHa time-dependent action in LSMC and MSMC by microarray and real-time PCR. Values on the y-axis represent an arbitrary unit derived from the mean expression value for each gene, and values on the x-axis represent the time course of GnRHa (0.1 μM) treatment (2, 6, and 12 h) with untreated control (Crtl) gene expression values set at 1. Total RNA isolated from these cells was used for both microarray analysis and real-time PCR to validate the expression of IL-11, EGR3, TIEG, TGIF, CITED2, Nur77, p27, p57, Gas-1, and GPRK5.
Western blotting also indicated that leiomyoma and myometrium as well as LSMC and MSMC locally produce IL-11, TGIF, TIEG, Nur77, EGR3, CITED2, p27, p57, and Gas-1 proteins (data are not shown). Immunohistochemically, IL-11, TGIF, TIEG, Nur77, EGR3, CITED2, p27, p57, and Gas-1 were localized in various cell types in leiomyoma and myometrium, including LSMC, MSMC, and vasculature and connective tissue fibroblasts (Fig. 9). We did not have access to antibody to GPRK5 and have not yet attempted to quantitate the levels of IL-11, TGIF, TIEG, Nur77, EGR3, CITED2, p27, p57, and Gas-1 production in leiomyoma and myometrium as well as in LSMC and MSMC in response to GnRHa treatment. However, these results provided additional support for the microarray and real-time PCR data, indicating that various cell types contribute to overall expression of these genes in leiomyoma and myometrium. In addition to these genes, we validated the expression of 15 more genes with real-time PCR including CTGF, Able-interactive (Abi2), fibromodulin, Runx1, and Runx2 (35, 42).
FIG. 9. Immunohistochemical localization of IL-11, TGIF, TIEG, Nur77, EGR3, p27, p57, and Gas-1 in leiomyoma and myometrium. Note the presence of immunoreactive IL-11, TGIF, TIEG, Nur77, EGR3, p27, p57, and Gas-1 in association with leiomyoma and myometrial smooth muscle cells, and cellular components of connective tissue and vasculature. Both nuclear (EGR3, Nur77, p27, and p57) and cytoplasmic (IL-11) staining was observed. Incubation of tissue sections with nonimmune mouse (A), rabbit (B), and goat (figure not shown) IgGs instead of primary antibodies during immunostaining served as controls (Ctrl), reduced the staining intensity. Magnifications: x150.
Discussion
In the present study we characterized gene expression fingerprints of leiomyoma and matched myometrium from the early to midsecretory phase of the menstrual cycle, a period associated with their rapid growth, their response to GnRHa therapy, and their response to direct action of GnRHa in isolated LSMC and MSMC prepared from untreated tissues. Combining global normalization and unsupervised assessment of the gene expression values derived from all the cohorts enabled us to sort potential candidate genes before their putative identification in each cohort. Transcripts of many of the genes on the array were found in leiomyoma and myometrium as well as in LSMC and MSMC. However, leiomyoma/LSMC were not distinguished as a single class from myometrium/MSMC based on single gene markers uniformly expressed only in leiomyoma and/or myometrium. This is not unique to leiomyoma/myometrium, because many large scale gene expression profiling studies have shown the existence of a significant degree of shared gene expression between various tumors and their normal tissue counterparts. However, supervised assessment and multiple test correction in a statistical R programming environment (37, 38, 39, 40) with a reduced false discovery rate, allowed us to identify a specific set of differentially expressed and regulated genes in descending order of significance in each cohort.
The analysis separated these genes into several clusters with sufficient differences, allowing their subdivision into respective subgroups in leiomyoma, myometrium, their isolated cells, as well as changes due to GnRHa therapy at the tissue and cellular levels. We identified 153 genes (excluding 19 EST) in these cultures, as differentially expressed in leiomyoma compared with myometrium, 82 of which were up-regulated and 52 of which were down-regulated in leiomyoma. GnRHa therapy affected the expression of 122 genes (excluding 21 EST), with 34 up-regulated and 67 down-regulated genes in leiomyoma compared with myometrium. However, the gene expression profiles from the untreated group differed substantially compared with those from GnRHa-treated leiomyoma/myometrium, pointing out a unique molecular environment that is targeted by GnRHa therapy. Analysis of the variance-normalized mean gene expression values divided these genes into four clusters, with two clusters showing treatment-specific, and the other clusters showing a tissue-specific, response to GnRHa therapy. Similar behavior was observed with gene clusters identified in LSMC and MSMC in response to GnRHa action in vitro. The significance of these findings relates to clinical observations indicating that GnRHa therapy affects both leiomyoma and myometrium, with nonmyoma tissue regressing more in response to therapy (7). Our gene expression profiling supports the clinical observations and points out that GnRHa therapy targets different genes in leiomyoma and myometrium despite their grouping in a similar functional category. Our recent microarray study with a small scale array of 1200 known genes (10) provides support for our current study; however, we are not aware of any other study using large scale gene expression profiling in leiomyoma and myometrium from women who received GnRHa therapy for additional comparison.
Since our study was completed, a few other microarray studies have reported the gene expression profiles of leiomyoma and myometrium from the proliferative and secretory phases of the menstrual cycle (9, 11, 12, 14). To broaden the scope of this study, we compared the genes identified in untreated leiomyoma and matched myometrium of our study with the datasets reported in four of these studies (9, 11, 12, 14). This comparison resulted in identification of only a few genes in common among these studies. Although intrinsic individual tissue variation may contribute to the differences among these studies, a standard of experimental process, utilization of different microarray platforms, use of tissues from different phases of the menstrual cycle, differences in leiomyoma size, and, most importantly, the method of data acquisition and analysis (9, 11, 12, 14) are among other key contributing factors accounting for different study results (38, 39, 40). To maintain a standard, we used leiomyoma of uniform size (2–3 cm in diameter) and matched myometrium, and the untreated cohorts were collected from the early to midsecretory phase of the menstrual cycle, a period associated with maximum growth of leiomyoma. However, lowering the false discovery rate of our study from P 0.02 to P 0.05 allowed identification of more transcripts and the appearance of additional common genes with other studies (9, 11, 12, 14). Considering the presence of a large number of probe sets on these arrays (i.e. 6,800–12,500), selection of genes based only on fold change (9) or higher statistical levels (11, 12, 13, 14) is no better than what one would expect by chance alone (38, 39, 40). Because we employed a similar data analysis, a larger number of genes was found in common with our previous microarray study, which used only a small scale array containing about 1,200 known genes (10). We recognize that exclusion of moderately regulated genes during microarray data analysis does not reflect a lack of functional importance, because a number of genes previously identified in leiomyoma and myometrium by conventional methods are not included among the differentially expressed genes in our study and other reports (1, 2, 9, 10, 11, 12, 13, 14). However, the expression of newly identified genes requires verification, and their regulation would allow linking their potential biological functions in leiomyoma growth and regression.
Using isolated LSMC and MSMC prepared from the untreated tissues allowed us to identify novel regulatory functions for GnRHa in leiomyoma and myometrium and to discover a wide range of genes whose expression has not previously been recognized to be the target of GnRHa direct action. Similar to their distinct clustering at tissue levels, the differentially expressed and regulated genes identified in LSMC and MSMC were also divided into clusters according to time-dependent response to GnRHa action. The genes in these clusters were either rapidly induced by GnRHa treatment or required prolong exposure, whereas others displayed biphasic patterns of temporal regulation in both treatment- and cell-specific fashions. Despite differences in their profiles, substantial similarity existed in the functional grouping of the genes affected by GnRHa therapy in leiomyoma/myometrium and GnRHa direct action on LSMC/MSMC (in vitro), with the expression of 48 genes commonly identified in tissues and cells. We speculate that the hypoestrogenic condition created by GnRHa therapy and contributions of other cell types to overall gene expression at the tissue level may account for the difference in profiles of gene expression between tissues and cell cultures. Gene functional annotation indicated that products of the majority of genes in these clusters are involved in transcriptional and signal transduction activities, cell cycle regulation, ECM turnover, cell-cell communication, transport, and enzyme regulatory activities.
Among the genes in these functional categories are several growth factors, cytokines, and chemokines, and polypeptide hormones, identified in leiomyoma, myometrium, and their isolated smooth muscle cells, and these were the target of GnRHa action in vivo and in vitro. Using several conventional methods, previous reports, including our own, have documented the expression of platelet-derived growth factor, epidermal growth factor, IGFs, vascular endothelial growth factor, fibroblast growth factor, TGF-?s, connective tissue growth factor, TNF-, IFN-, monocyte chemoattractant protein-1, and IL-8 as well as some of their receptors in leiomyoma and myometrium (1, 2, 16, 17, 20, 23, 26, 27). However, the expression of some of these and other genes in this category did not meet the selection criteria of our study, a common discrepancy often observed in microarray analysis, particularly in identifying moderately expressed and regulated genes (38, 39, 40). For example, the expressions of TGF-? isoforms, TGF-? receptors, and components of their signaling pathway that are well documented in leiomyoma and myometrium as well as in their isolated smooth muscle cells (17, 18, 19, 20, 43) are not consistently identified in microarray studies (9, 10, 11, 12, 13, 14), although in our current accompanying (35) and previous (10) studies we identified most of the members of the TGF-? system. However, microarray gene expression data may not always reflect the actual expression level of a given gene, as shown in some cases in our study compared with real-time PCR findings. Among the cytokines whose expression was identified and validated in our study is IL-11. IL-11 is recognized to play key regulatory functions in inflammation, angiogenesis, and tissue remodeling (44, 45, 46, 47, 48), events central to leiomyoma growth. IL-11 is a member of the IL-6 family and is produced by various cell types, including the uterus. Its overexpression is reported to cause subepithelial airway fibrosis, particularly through interaction with IL-13 and TGF-? (44, 45, 46, 47, 48, 49). We have provided evidence that IL-11, similar to TGF-? and IL-13, is overexpressed in leiomyoma compared with myometrium, and GnRHa therapy suppresses their expression in these tissues (17, 20, 43, 50). At the cellular level, unlike the expression of TGF-? and IL-13, GnRHa increases IL-11 expression in LSMC and MSMC within 2–6 h of treatment, which sharply declines to control levels after 12 h. Although the nature of differential regulation of IL-11 at the tissue and cellular levels requires detailed investigation, prolonged treatment with GnRHa, the contribution of other cell types, and the influence of other autocrine/paracrine regulators may account for the difference in IL-11 expression between in vivo and in vitro conditions.
Other differentially expressed and regulated genes identified in our study functionally belong to signal transduction pathways that are recruited and activated by various growth factors/cytokines/chemokines, polypeptide hormones, ECM, and adhesion molecules. However, only the expression of a few of these and other signal transduction pathways has been documented in leiomyoma and myometrium (1, 16, 19, 29), and little is known about their recruitment and activation in LSMC and MSMC. We have previously identified the expression of Smads, MAPK, and focal adhesion kinase in leiomyomas and myometrium and provided evidence for their regulation and activation by GnRHa in LSMC and MSMC (16, 18, 19). In this study we validated the expression of GPRK5, identified as one of the differentially expressed and regulated genes in leiomyoma and myometrium and demonstrated that GnRHa therapy and in vitro treatment of LSMC and MSMC with GnRHa inhibit GPRK5 expression. GPRKs, consisting of six members (GPRK1 to GPRK6), act as key regulators of signaling via GPRKs and are widely expressed in various tissues and cells (51, 52, 53, 54). Previous studies have demonstrated that pregnant and nonpregnant human myometrium as well as cultured myometrial cells express GPRK2, GPRK4, and GPRK5; however, GPRK3 and GPRK4, -?, and - were not detected in myometrium (52, 53). GPRK5 has been shown to serve as a substrate for protein kinase C (PKC), although PKC-mediated phosphorylation inhibits GPRK5 (21, 54). In addition, the extreme N terminus of GPRK5 contains a binding site for Ca2+/calmodulin, where, upon binding, it inhibits GPRK activity, a mechanism suggested to regulate GPRK activity (54). Because GnRH receptors are members of the GPCR family and recruit and activate the components of several signaling pathways, including PKC and Ca2+/calmodulin, their regulatory interaction with GPRKs may serve in mediating GnRH receptor downstream signaling in LSMC and MSMC.
Nuclear translocation of activated signaling molecules results in phosphorylation and activation of transcription factors, major elements in the signaling networks that regulate specific gene expression. In our previous (10) and current studies, we identified several genes in this category in leiomyoma and myometrium whose expression was targeted by GnRHa in LSMC and MSMC. Many of these transcription factors are involved in ovarian steroid-, polypeptide hormone-, cytokine-, growth factor-, and ECM receptor-mediated actions, by regulating the promoter of their target genes in various normal and cancer cells. However, little is known about the expression and regulation of these and other transcription factors in leiomyoma and myometrium. For this reason we placed a greater emphasis on the expression of genes in this category and verified the expression of Nur77, CITED2, EGR3, TIEG, and TGIF in leiomyoma and myometrium and their temporal regulation by GnRHa in LSMC and MSMC.
Nur77 (also known as NR4A1, TR3, NGFI-B, and NAK-1) is a member of the orphan nuclear receptor superfamily originally identified as an immediate-early gene in serum-treated fibroblasts (55, 56, 57, 58, 59, 60, 61, 62). It is also identified as nerve growth factor-inducible gene, which is constitutively expressed in various tissues and is strongly induced by several stimuli, resulting in the regulation of gene expression related to inflammation, angiogenesis, apoptosis, and steroidogenesis, including steroid 21- and 17-hydroxylases and 20-hydroxysteroid dehydrogenase in the hypothalamic-pituitary-adrenal axis (55, 56, 57, 58, 59, 60, 61, 62). In the anterior pituitary, Nur77 is reported to mediate the stimulatory effect of CRH and the negative feedback regulation of proopiomelanocortin transcription by glucocorticoids as well as GnRH-induced GnRH receptor expression (56, 59). LH-induced Nur77 is also reported to regulate cytochrome P450 expression in granulosa and Leydig cells (59, 60, 61). More importantly, overexpression of Nur77 is implicated as an important regulator of apoptosis in different cells. In response to apoptotic stimuli, Nur77 translocation from nucleus to mitochondria results in cytochrome c release and apoptosis of LNCaP human prostate cancer cells (63, 64, 65). Nur77 is expressed at relatively similar levels in myometrium and leiomyoma and is significantly increased in response to GnRHa therapy. GnRHa also resulted in a rapid induction of Nur77 in MSMC and LSMC, subsequently declining to control levels or in LSMC to below control levels. Interestingly, GnRH is reported to regulate Nur77 expression in T3-1 and L?T2 gonadotrope cell lines through protein kinase A pathway and GnRH receptor promoter via a mechanism involving steroidogenic factor-1 with Nur77 acting as a negative regulator of this response (59). In a recent study, activation of the MAPK pathway involving Raf-1, MAPK kinase 2, and extracellular signal-regulated kinase 2 was reported to regulate Nur77 activation, resulting in nonapoptotic program cell death (64). We have shown that GnRH receptor signaling through MAPK and transcriptional activation of c-Fos and c-Jun regulate the expression of several specific genes in LSMC and MSMC (16). This suggests that GnRH-mediated action through this pathway may regulate Nur77 expression, thus influencing the outcome of cellular growth arrest and/or apoptosis in leiomyoma.
Recently, a new family of transcriptional coregulators, the CITED family, was discovered that interact with the first cysteine-histidine-rich region of CBP/p300 (66, 67). The CITED family contains four members and appears to act as key transcriptional modulators in embryogenesis, inflammation, and stress responses (66) by affecting the transcriptional activity of many transcription factors ranging from activating protein-2, estrogen receptor, and hypoxia-inducible factor 1 and LIM (68). We identified CITED2 among the differentially expressed and regulated genes in leiomyoma, myometrium, and their isolated cells and in response to GnRHa treatment in vivo and in vitro. Unlike GnRHa therapy, which increased CITED2 expression in leiomyoma and myometrium, GnRHa had a biphasic effect on CITED2 expression in MSMC while inhibiting expression in LSMC. Although in vitro culture conditions may directly influence the expression of regulatory molecules that either interact with or regulate CITED2 expression, the molecular mechanism resulting in differential expression of CITED2 in vivo and in vitro by GnRHa requires additional investigation. Interestingly, the expression of several growth factors, cytokines, and hypoxia-inducible factor 1 are the target of estrogen receptor and progesterone receptor regulatory actions, and CITED2, acting as a repressor of their expression, may serve as an important mediator of processes that regulate inflammatory response, angiogenesis, and tissue remodeling in leiomyoma. Additionally, CBP/p300, which serve as promiscuous coactivators for an increasing number of transcription factors, resulting in proliferation, differentiation, and apoptosis in response to diverse biological factors, including estrogen receptor- and progesterone receptor-dependent transcriptional activity, is specifically recruited by Nur77 acting as dimers after protein kinase A activation (55, 69, 70).
In our previous microarray study, we reported that EGR1, a prototype of a family of zinc finger transcription factors that includes EGR2, EGR3, EGR4, and NGFI-B (71, 72), is differentially expressed in leiomyoma and myometrium (10). In this study we provide evidence for the expression of EGR3 and differential regulation in response to GnRHa therapy in leiomyoma and myometrium as well as in LSMC and MSMC in vitro. A recent report demonstrated that EGR1 expression is elevated in leiomyoma compared with corresponding myometrium in women who received GnRHa therapy (73), supporting our previous microarray data (10). EGR expression is rapidly and transiently induced by a large number of growth factors, cytokines, polypeptide hormones, and injurious stimuli, and the kinetics of their expression are essentially identical to those of the c-fos protooncogene (71, 72, 74). In addition, induction of EGR1 occurs primarily at the level of transcription and is mediated in part through MAPKs, including extracellular signal-regulated kinase, c-Jun N-terminal kinase, and p38 pathways (71, 72). We have demonstrated that GnRHa, through the activation of MAPK, regulates the expression of c-Fos and c-Jun as well as fibronectin, collagen, and plasminogen activator inhibitor-1 expression (16). In human fibrosarcoma and glioblastoma cells, EGR directly influences the expression of fibronectin, TGF-?1, and plasminogen activator inhibitor-1 and may regulate the expression of platelet-derived growth factor, tissue factor, and membrane type 1 matrix metalloproteinase (72, 75). Estrogen also induces EGR3 in various cancer cells, whereas progesterone inhibits EGR3 in Schwann cells (74, 76). Constitutive transgenic expression of EGR3 has been shown to increase thymocyte apoptosis, possibly through potent activation of Fas ligand expression (77). Given the roles of ovarian steroids and a large number of growth factors, cytokines, and polypeptide hormones in leiomyoma growth and suppression by GnRHa, their differential influence on EGR1 and EGR3 expression may represent a mechanism resulting in a balance between the rate of cell proliferation and apoptosis as well as tissue turnover, affecting leiomyoma growth and regression.
We also provide the first evidence of the expression and regulation of TIEG and TGIF, novel three-zinc finger Kruppel-like transcriptional repressors, and key regulators of TGF-? receptor signaling (41, 78, 79, 80, 81) by GnRHa in leiomyoma, myometrium, LSMC, and MSMC. TIEG regulates TGF-? receptor signaling through a negative feedback mechanism by repressing the inhibitory Smad7 (41). In addition, TGIF, through direct binding to DNA or interaction with TGF-?-activated Smads, represses TGF-?-responsive gene expression (80, 81). Because GnRHa suppresses TGF-?s and TGF-? receptors while enhancing Smad7 expression in leiomyoma and myometrium and in LSMC and MSMC, differential regulation of TIEG and TGIF may serve as an additional downstream mechanism altering TGF-? actions in leiomyoma (35).
The expression, activation, and direct interaction of these and other transcription factors with DNA result in regulation of the expression of various genes whose products influence cell growth, inflammation, angiogenesis, apoptosis, and tissue turnover. In our previous (10, 16) and present study we identified several differentially expressed and regulated genes in leiomyoma, myometrium, LSMC, and MSMC whose promoters are the targets of these transcription factors. Among these genes are members of cell cycle regulatory proteins that play a central role in leiomyoma growth and regression (1, 2, 31), including p27, p57, and Gas-1. We identified p27, p57, and Gas-1 as differentially expressed and regulated in leiomyoma and myometrium as well as in LSMC and MSMC and in response to GnRHa treatment. Although p27, p57, and Gas-1 function as major regulators of cell cycle progression, several studies have also shown Cip/Kip proteins to function as transcriptional cofactors by regulating the activity of nuclear factor-B, signal transducer and activator of transcription-3, Myc, Rb, C/EBP, CBP/p300, E2F, and activating protein-1 (82). A recent report suggests that p21, p27, and p57 are involved in the regulation of apoptosis (83), and their differential regulation in leiomyoma and myometrium is consistent with GnRHa induction of an apoptosis-related gene in LSMC and MSMC (1, 2, 22, 31). However, our results are the first to document the expression of Gas-1 in leiomyoma and myometrium and its regulation in LSMC and MSMC in response to the time-dependent action of GnRHa. GnRHa has been demonstrated to alter cell cycle progression and programmed cell death in several cell types, including leiomyoma smooth muscle cells (1, 22, 31), and our results provide additional support for the involvement of specific cell cycle- and apoptosis-related genes in leiomyoma growth and regression.
Leiomyoma growth and GnRHa therapy resulting in leiomyoma regression also involve ECM turnover. In our previous (10) and present studies and recent studies by other groups (9, 11, 12, 13, 14), several genes in this category were identified displaying differential expression in leiomyoma and myometrium and were targeted by GnRH therapy (1, 16, 24, 45, 42, 84). These include the expression of several collagens; a small leucine-rich repeat family of proteoglycans, decorin, biglycan, osteomodulin, fibromodulin, versican, and osteoadherin/osteoglycin, fibronectin, desmin, and vimentin; several member of proteases, such as matrix metalloproteinases, and their inhibitors; a disintegrin-like and metalloproteinase protein, etc. We have also reported that GnRHa regulates the expression of fibronectin, collagen type I, plasminogen activator inhibitor-I, matrix metalloproteinases, and matrix metalloproteinase inhibitors (1, 16, 24), as well as decorin, versican, desmin, and vimentin (our unpublished observations) in leiomyoma and myometrium, involving the activation of MAPK in LSMC and MSMC (16). Because ECM turnover is a key regulator of the outcome of tissue fibrosis, and many cytokines, chemokines, growth factors, and polypeptide hormones, through specific intracellular signal transduction and activation of transcription factors, influence the expression of ECM and proteases, additional investigation is underway to elucidate their regulatory interactions affecting processes that may influence leiomyoma growth and regression.
In summary, in the present study we provide a comprehensive assessment of the gene expression profile of leiomyoma and matched myometrium during the early to midluteal phase of the menstrual cycle, a period characterized by elevated production of ovarian steroids and maximal leiomyoma growth, compared with that in tissues obtained from hormonally suppressed patients receiving GnRHa therapy and in response to the direct action of GnRHa on LSMC and MSMC. We identified several common and tissue-specific gene clusters in these cohorts, suggesting their coregulation by the same factors and/or mechanism(s) in the same cluster. We validated the expression of several genes whose products are important in signal transduction, transcription, cell cycle regulation, apoptosis, and ECM turnover, events critical to the development, growth, and regression of leiomyoma. Based on these and our previous observations, we propose that the product of these specific genes, by regulating local inflammatory and apoptotic processes leading to elaboration of profibrotic cytokine production, such as TGF-?, is central to the establishment and progression of leiomyoma growth and regression.
Acknowledgments
We thank Dr. Thomas Spelsberg, (Mayo Clinic, Rochester, MN) for the gift of the antibody to TIEG, and Drs. Henry Baker (Department of Microbiology and Molecular Genetics), Mick Popp, and Li Liu (Interdisciplinary Center for Biotechnology Research) at University of Florida for their helpful advice concerning microarray data analysis.
References
Chegini N 2000 Implication of growth factor and cytokine networks in leiomyomas. In: Hill J, ed. Cytokines in human reproduction. New York: Wiley & Sons; 133–162
Maruo T, Ohara N, Wang J, Matsuo H 2004 Sex steroidal regulation of uterine leiomyoma growth and apoptosis. Hum Reprod Update 10:207–220
Takeuchi H, Kobori H, Kikuchi I, Sato Y, Mitsuhashi N 2000 A prospective randomized study comparing endocrinological and clinical effects of two types of GnRH agonists in cases of uterine leiomyomas or endometriosis. J Obstet Gynecol Res 26:325–331
Steinauer J, Pritts EA, Jackson R, Jacoby AF 2004 Systematic review of mifepristone for the treatment of uterine leiomyomata. Obstet Gynecol 103:1331–1336
Palomba S, Russo T, Orio Jr F, Tauchmanova L, Zupi E, Panici PL, Nappi C, Colao A, Lombardi G, Zullo F 2002 Effectiveness of combined GnRH analogue plus raloxifene administration in the treatment of uterine leiomyomas: a prospective, randomized, single-blind, placebo-controlled clinical trial. Hum Reprod 17:3213–3219
DeManno D, Elger W, Garg R, Lee R, Schneider B, Hess-Stumpp H, Schubert G, Chwalisz K 2003 Asoprisnil (J867): a selective progesterone receptor modulator for gynecological therapy. Steroids 68:1019–1032
Carr BR, Marshburn PB, Weatherall PT, Bradshaw KD, Breslau NA, Byrd W, Roark M, Steinkampf MP 1993 An evaluation of the effect of gonadotropin-releasing hormone analogs and medroxyprogesterone acetate on uterine leiomyomata volume by magnetic resonance imaging: a prospective, randomized, double blind, placebo-controlled, crossover trial. J Clin Endocrinol Metab 76:1217–1223
Ligon AH, Morton CC 2001 Leiomyomata: heritability and cytogenetic studies. Hum Reprod Update 7:8–14
Tsibris JCM, Segars J, Coppola D, Mane S, Wilbanks GD, O’Brien WF, Spellacy WN 2002 Insights from gene arrays on the development and growth regulation of uterine leiomyomata. Fertil Steril 78:114–121
Chegini N, Verala J, Luo X, Xu J, Williams RS 2003 Gene expression profile of leiomyoma and myometrium and the effect of gonadotropin releasing hormone analogue therapy. J Soc Gynecol Invest 10:161–171
Wang H, Mahadevappa M, Yamamoto K, Wen Y, Chen B, Warrington JA, Polan ML 2003 Distinctive proliferative phase differences in gene expression in human myometrium and leiomyomata. Fertil Steril 80:266–276
Weston G, Trajstman AC, Gargett CE, Manuelpillai U, Vollenhoven BJ, Rogers PA 2003 Fibroids display an anti-angiogenic gene expression profile when compared with adjacent myometrium. Mol Hum Reprod 9:541–549
Ahn WS, Kim KW, Bae SM, Yoon JH, Lee JM, Namkoong SE, Kim JH, Kim CK, Lee YJ, Kim YW 2003 Targeted cellular process profiling approach for uterine leiomyoma using cDNA microarray, proteomics and gene ontology analysis. Int J Exp Pathol 84:267–279
Quade BJ, Wang TY, Sornberger K, Dal Cin P, Mutter GL, Morton CC 2004 Molecular pathogenesis of uterine smooth muscle tumors from transcriptional profiling. Genes Chromosomes Cancer 40:97–108
Chegini N, Rong H, Dou Q, Kipersztok S, Williams RS 1996 Gonadotropin releasing hormone (GnRH) and GnRH receptor gene expression in human myometrial and leiomyomta and the direct action of GnRH analogs on myometrial smooth muscle cells interaction with ovarian steroids in vitro. J Clin Endocrinol Metab 81:3215–3221
Ding L, Xu J, Luo L, Chegini N 2004 Gonadotropin releasing hormone and transforming growth factor ? activate MAPK/ERK and differentially regulate fibronectin, type I collagen, and PAI-1 expression in leiomyoma and myometrial smooth muscle cells. J Clin Endocrinol Metab 89:5549–5557
Chegini N, Luo X, Ding L, Ripley D 2003 The expression of Smads and transforming growth factor ? receptors in leiomyoma and myometrium and the effect of gonadotropin releasing hormone analogue therapy. Mol Cell Endocrinol 209:9–16
Xu J, Luo X, Chegini N 2003 Differential expression, regulation, and induction of Smads, transforming growth factor-? signal transduction pathway in leiomyoma, and myometrial smooth muscle cells and alteration by gonadotropin-releasing hormone analog. J Clin Endocrinol Metab 88:1350–1361
Chegini N, Kornberg L 2003 Gonadotropin releasing hormone analogue therapy alters signal transduction pathways involving mitogen-activated protein and focal adhesion kinases in leiomyoma. J Soc Gynecol Invest 10:21–26
Chegini N, Ma C, Tang XM, Williams RS 2002 Effects of GnRH analogues, ’add-back’ steroid therapy, antiestrogen and antiprogestins on leiomyoma and myometrial smooth muscle cell growth and transforming growth factor-? expression. Mol Hum Reprod 8:1071–1078
Klausen C, Chang JP, Habibi HR 2002 Multiplicity of gonadotropin-releasing hormone signaling: a comparative perspective. Prog Brain Res 141:111–128
Mizutani T, Sugihara A, Nakamuro K, Terada N 1998 Suppression of cell proliferation and induction of apoptosis in uterine leiomyoma by gonadotropin-releasing hormone agonist (leuprolide acetate). J Clin Endocrinol Metab 83:1253–1255
Wu X, Blanck A, Olovsson M, Moller B, Lindblom B 2001 Expression of basic fibroblast growth factor (bFGF), FGF receptor 1 and FGF receptor 2 in uterine leiomyomas and myometrium during the menstrual cycle, after menopause and GnRHa treatment. Acta Obstet Gynecol Scand 80:497–504
Dou Q, Tarnuzzer RW, Williams RS, Schultz GS, Chegini N 1997 Differential expression of matrix metalloproteinases and their tissue inhibitors in leiomyomata: a mechanism for gonadotrophin releasing hormone agonist-induced tumour regression. Mol Hum Reprod 3:1005–1014
Chegini N, Tang XM, Ma C 1999 Regulation of transforming growth factor-?1 expression by granulocyte macrophage-colony-stimulating factor in leiomyoma and myometrial smooth muscle cells. J Clin Endocrinol Metab 84:4138–4143
Senturk LM, Sozen I, Gutierrez L, Arici A 2001 Interleukin 8 production and interleukin 8 receptor expression in human myometrium and leiomyoma. Am J Obstet Gynecol 184:559–566
Sozen I, Olive DL, Arici A 1998 Expression and hormonal regulation of monocyte chemotactic protein-1 in myometrium and leiomyomata. Fertil Steril 69:1095–1102
Gustavsson I, Englund K, Faxen M, Sjoblom P, Lindblom B, Blanck A 2000 Tissue differences but limited sex steroid responsiveness of c-fos and c-jun in human fibroids and myometrium. Mol Hum Reprod 6:55–59
Orii A, Masutani H, Nikaido T, Zhai YL, Kato K, Kariya M, Konishi I, Yodoi J, Fujii S 2002 Altered post-translational modification of redox factor 1 protein in human uterine smooth muscle tumors. J Clin Endocrinol Metab 87:3754–3759
Fukuhara K, Kariya M, Kita M, Shime H, Kanamori T, Kosaka C, Orii A, Fujita J, Fujii S 2002 Secreted frizzled related protein 1 is overexpressed in uterine leiomyomas, associated with a high estrogenic environment and unrelated to proliferative activity. J Clin Endocrinol Metab 87:1729–1736
Zhai YL, Nikaido T, Shiozawa T, Orii A, Fujii S 1999 Expression of cyclins and cyclin-dependent kinases in smooth muscle tumors of the uterus. Int J Cancer 84:244–250
Ma C, Chegini N 1999 Matrix metalloproteinases and tissue inhibitor of MMPs (TIMPs): expression in human myometrial smooth muscle cell and regulation by transforming growth factor b1. Mol Hum Reprod 5:950–954
Kakar SS, Winters SJ, Zacharias W, Miller DM, Flynn S 2003 Identification of distinct gene expression profiles associated with treatment of L?T2 cells with gonadotropin-releasing hormone agonist using microarray analysis. Gene 308:67–77
Wurmbach E, Yuen T, Ebersole BJ, Sealfon SC 2001 Gonadotropin-releasing hormone receptor-coupled gene network organization. J Biol Chem 276:47195–47201
Luo X, Ding L, Xu J, Chegini N 2005 Gene expression profiling of leiomyoma and myometrial smooth muscle cells in response to transforming growth factor-?. Endocrinology 146:1097–1118
Eisen MB, Spellman PT, Brown PO, Botstein D 1998 Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95:14863–14868
Tusher VG, Tibshirani R, Chu G 2001 Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98:5116–5121
Pavlidis P 2003 Using ANOVA for gene selection from microarray studies of the nervous system. Methods 31:282–289
Peterson LE 2003 Partitioning large-sample microarray-based gene expression using principal components analysis. Comput Methods Programs Biomed 70:107–119
Butte A 2002 The use and analysis of microarray data. Nat Rev Drug Discov 1:951–960
Johnsen SA, Subramaniam M, Janknecht R, Spelsberg TC 2002 TGF-? inducible early gene enhances TGF-?/Smad-dependent transcriptional responses. Oncogene 21:5783–5790
Levens E, Luo X, Ding L, Williams RS, Chegini N 2004 Differential expression of fibromodulin and Abl-interactor 2 in leiomyoma and myometrium and regulation by gonadotropin releasing hormone analogue (GnRHa) therapy. Fertil Steril 82(Suppl 2):S88
Dou Q, Zhao Y, Tarnuzzer RW, Rong H, Williams RS, Schultz GS, Chegini N 1996 Suppression of transforming growth factor-? (TGF-?) and TGF-? receptor messenger ribonucleic acid and protein expression in leiomyomata in women receiving gonadotropin-releasing hormone agonist therapy. J Clin Endocrinol Metab 81:3222–3230
Leng SX, Elias JA 1997 Interleukin-11. Int J Biochem Cell Biol 29:1059–1062
Tang W, Geba GP, Zheng T, Ray P, Homer RJ, Kuhn III C, Flavell RA, Elias JA 1996 Targeted expression of IL-11 in the murine airway causes lymphocytic inflammation, bronchial remodeling, and airways obstruction. J Clin Invest 98:2845–2853
Zhu Z, Lee CG, Zheng T, Chupp G, Wang J, Homer RJ, Noble PW, Hamid Q, Elias JA 2001 Airway inflammation and remodeling in asthma. Lessons from interleukin 11 and interleukin 13 transgenic mice. Am J Respir Crit Care Med 164:S67–S70
Zimmerman MA, Selzman CH, Reznikov LL, Raeburn CD, Barsness K, McIntyre Jr RC, Hamiel CR, Harken AH 2002 Interleukin-11 attenuates human vascular smooth muscle cell proliferation. Am J Physiol 283:H175–H180
Bamba S, Andoh A, Yasui H, Makino J, Kim S, Fujiyama Y 2003 Regulation of IL-11 expression in intestinal myofibroblasts: role of c-Jun AP-1- and MAPK-dependent pathways. Am J Physiol 285:G529–G538
Karpovich N, Chobotova K, Carver J, Heath JK, Barlow DH, Mardon HJ 2003 Expression and function of interleukin-11 and its receptor in the human endometrium. Mol Hum Reprod 9:75–80
Ding L, Luo X Chegini N 2004 The expression of IL-13 and IL-15 in leiomyoma and myometrium and their influence on TGF-? and proteases expression in leiomyoma and myometrial smooth muscle cells and SKLM, leiomyosarcoma cell line. J Soc Gynecol Invest 11:319A
Mak JC, Chuang TT, Harris CA, Barnes PJ 2002 Increased expression of G protein-coupled receptor kinases in cystic fibrosis lung. Eur J Pharmacol 436:165–172
Simon V, Mhaouty-Kodja S, Legrand C, Cohen-Tannoudji J 2001 Concomitant increase of G protein-coupled receptor kinase activity and uncoupling of ?-adrenergic receptors in rat myometrium at parturition. Endocrinology 142:1899–1905
Simon V, Robin MT, Legrand C, Cohen-Tannoudji J 2003 Endogenous G protein-coupled receptor kinase 6 triggers homologous ?-adrenergic receptor desensitization in primary uterine smooth muscle cells. Endocrinology 144:3058–3066
Krasel C, Dammeier S, Winstel R, Brockmann J, Mischak H, Lohse MJ 2001 Phosphorylation of GRK2 by protein kinase C abolishes its inhibition by Calmodulin. J Biol Chem 276:1911–1915
Maira M, Martens C, Batsché E, Gauthier Y, Drouin J 2003 Dimer-specific potentiation of NGFI-B (Nur77) transcriptional activity by the protein kinase A pathway and AF-1-dependent coactivator recruitment. Mol Cell Biol 23:763–776
Drouin, J, Maira MH, Philips A 1998 Novel mechanism of action for Nur77 and antagonism by glucocorticoids: a convergent mechanism for CRH activation and glucocorticoid repression of POMC gene transcription. J. Steroid Biochem Mol Biol 65:59–63
Fernandez P, Brunel MF, Jimenez MA, Saez JM, Cereghini S, Zakin MM 2000 Nuclear receptors Nor1 and NGFI-B/Nur77 play similar, albeit distinct, roles in the hypothalamo-pituitary-adrenal axis. Endocrinology 141:2392–2400
Gelman L, Zhou G, Fajas L, Raspe E, Fruchart JC, Auwerx J 1999 p300 Interacts with the N- and C-terminal part of PPAR2 in a ligand-independent and -dependent manner, respectively. J Biol Chem 274:7681–7688
Sadie H, Styger G, Hapgood J 2003 Expression of the mouse gonadotropin-releasing hormone receptor gene in T3–1 gonadotrope cells is stimulated by cyclic 3',5'-adenosine monophosphate and protein kinase A, and is modulated by Steroidogenic factor-1 and Nur77. Endocrinology 144:1958–1971
Wilson TE, Mouw AR, Weaver CA, Milbrandt J, Parker KL 1993 The orphan nuclear receptor NGFI-B regulates expression of the gene encoding steroid 21-hydroxylase. Mol Cell Biol 13:861–868
Song KH, Park JI, Lee MO, Soh J, Lee K, Choi HS 2001 LH induces orphan nuclear receptor Nur77 gene expression in testicular Leydig cells. Endocrinology 142:5116–123
Zhang, P, Mellon SH 1997 Multiple orphan nuclear receptors converge to regulate rat P450c17 gene transcription: novel mechanisms for orphan nuclear receptor action. Mol Endocrinol 11:891–904
Rajpal A, Cho YA, Yelent B, Koza-Taylor PH, Li D, Chen E, Whang M, Kang C, Turi TG, Winoto A 2003 Transcriptional activation of known and novel apoptotic pathways by Nur77 orphan steroid receptor. EMBO J 22:6526–6536
Castro-Obregon S, Rao RV, del Rio G, Chen SF, Poksay KS, Rabizadeh S, Vesce S, Zhang XK, Swanson RA, Bredesen DE 2004 Alternative, nonapoptotic programmed cell death: mediation by arrestin 2, ERK2, and Nur77. J Biol Chem 279:17543–17553
Li H, Kolluri SK, Gu J, Dawson MI, Cao X, Hobbs PD, Lin B, Chen GLJ, Lin F, Xie Z, Fontana JA, Reed JC, Zhang X 2000 Cytochrome c release and apoptosis induced by mitochondrial targeting of orphan receptor TR3 (nur77, NGFI-B). Science 289:1159–1164
Tien ES, Davis JW, Vanden Heuvel JP 2004 Identification of the CREB-binding protein/p300-interacting protein CITED2 as a peroxisome proliferator-activated receptor coregulator. J Biol Chem 279:24053–24063
Kranc KR, Bamforth SD, Braganca J, Norbury C, van Lohuizen M, Bhattacharya S 2003 Transcriptional coactivator Cited2 induces Bmi1 and Mel18 and controls fibroblast proliferation via Ink4a/ARF. Mol Cell Biol 23:7658–7666
Yin Z, Haynie J, Yang X, Han B, Kiatchoosakun S, Restivo J, Yuan S, Prabhakar NR, Herrup K, Conlon RA, Hoit BD, Watanabe M, Yang YC 2002 The essential role of Cited2, a negative regulator for HIF-1, in heart development and neurulation. Proc Natl Acad Sci USA 99:10488–10493
Kranc K, Shikama R, Lyon NJ, La Thangue NB 1997 The p300/CBP family: integrating signals with transcription factors and chromatin. Trends Cell Biol 7:230–236
Puri PL, Avantaggiati ML, Balsano C, Sang N, Graessmann A, Giordano A, Levrero M 1997 p300 is required for MyoD-dependent cell cycle arrest and muscle-specific gene transcription. EMBO J 16:369–383
Hjoberg J, Le L, Imrich A, Subramaniam V, Mathew SI, Vallone J, Haley KJ, Green FHY, Shore SA, Silverman ES 2004 Induction of early growth-response factor 1 by platelet-derived growth factor in human airway smooth muscle. Am J Physiol 286:L817–L825
Thiel G, Cibelli G 2002 Regulation of life and death by the zinc finger transcription factor Egr-1. J Cell Physiol 193:287–292
Shozu M, Murakami K, Segawa T, Kasai T, Ishikawa H, Shinohara K, Okada M, Inoue M 2004 Decreased expression of early growth response-1 and its role in uterine leiomyoma growth. Cancer Res 64:4677–4684
Inoue A, Omoto Y, Yamaguchi Y, Kiyama R, Hayashi SI 2004 Transcription factor EGR3 is involved in the estrogen-signaling pathway in breast cancer cells. J Mol Endocrinol 32:649–661
Liu C, Yao J, de Belle I, Huang RP, Adamson E, Mercola D 1999 The transcription factor EGR-1 suppresses transformation of human fibrosarcoma HT1080 cells by coordinated induction of transforming growth factor-?1, fibronectin, and plasminogen activator inhibitor-1. J Biol Chem 274:4400–4411
Mercier G, Turque N, Schumacher M 2001 Early activation of transcription factor expression in Schwann cells by progesterone. Mol Brain Res 97:137–148
Xi H, Kersh GJ 2004 Sustained early growth response gene 3 expression inhibits the survival of CD4/CD8 double-positive thymocytes. J Immunol 173:340–348
Cook T, Urrutia R 2000 TIEG proteins join the Smads as TGF-?-regulated transcription factors that control pancreatic cell growth. Am J Physiol 278:G513–G521
Ribeiro A, Bronk SF, Roberts PJ, Urrutia R, Gores GJ 1999 The transforming growth factor ?1-inducible transcription factor TIEG1, mediates apoptosis through oxidative stress. Hepatology 30:1490–1497
Chen F, Ogawa K, Nagarajan RP, Zhang M, Kuang C, Chen Y 2003 Regulation of TG-interacting factor by transforming growth factor-?. Biochem J 371:257–263
Melhuish TA, Gallo CM, Wotton D 2001 TGIF2 interacts with histone deacetylase 1 and represses transcription. J Biol Chem 276:32109–32114
Coqueret O 2003 New roles for p21 and p27 cell-cycle inhibitors: a function for each cell compartment? Trends Cell Biol 13:65–70
Blagosklonny MV 2003 Apoptosis, proliferation, differentiation: in search of the order. Semin Cancer Biol 13:97–105
Stewart EA, Friedman AJ, Peck K, Nowak RA 1994 Relative overexpression of collagen type I and collagen type III messenger ribonucleic acids by uterine leiomyomas during the proliferative phase of the menstrual cycle. J Clin Endocrinol Metab 79:900–906(Xiaoping Luo, Li Ding, Ji)