microRNA 21: response to hormonal therapies and regulatory function in leiomyoma, transformed leiomyoma and leiomyosarcoma cells.

Aberrant expression of microRNAs (miRNAs), including miR-21, and alteration of their target genes stability have been associated with cellular transformation and tumorigenesis. We investigated the expression, regulation and function of miR-21 in leiomyomas which develop from myometrial cellular transformation. The results indicated that miR-21 is over-expressed in leiomyomas with specific elevation during the secretory phase of the menstrual cycle and in women who received Depo-Provera and oral contraceptives, but reduced due to GnRHa therapy (P < 0.05). Bioinformatic analysis of microarray gene expression profiles previously obtained from the above cohorts, and myometrial smooth muscle cells (MSMC) and leiomyoma smooth muscle cells (LSMC) treated with GnRHa, transforming growth factor (TGF)-beta and TGF-beta receptor type II (TGF-betaRII) antisense oligomer, indicated that a number of miR-21-predicted target genes were co-expressed and differentially regulated in these cohorts. Gain- and loss-of-function of miR-21 in MSMC, LSMC, transformed LSMC and leiomyosarcoma cell line (SKLM-S1) resulted in differential expression of many genes, including some of the miR-21-predicted/validated target genes, PTEN, PDCD4 and E2F1, and TGF-betaRII, in a cell-specific manner. Gain-of miR-21 function in MSMC and LSMC reduced TGF-beta-induced expression of fibromodulin and TGF-beta-induced factor (P < 0.05), and moderately altered the rate of cell growth and caspase-3/7 activity in these cells. We concluded that miR-21 is aberrantly expressed and hormonally regulated in leiomyomas where, through functional interaction with ovarian steroids and the TGF-beta signaling pathway, either directly or indirectly regulates a number of genes whose products are critical in leiomyoma growth and regression as well as their potential cellular transformation.

Introduction

microRNAs (miRNAs) have emerged as a key regulator of gene expression stability (Engels and Hutvagner, 2006; Zeng, 2006; Fujita and Iba, 2008; Selbach ). Several thousand miRNAs have been identified and/or predicted and the expression of a large number of them has been profiled in various normal cells and tissues and functionally associated with normal cellular activities such as cell growth, differentiation and apoptosis (Engels and Hutvagner, 2006; Baek ; Connolly ; Fujita and Iba, 2008; Ku and McManus, 2008; Selbach ). Conversely, aberrant expression of some miRNAs has been associated with a number of disorders; more specifically, cancer, where they regulate gene expression involved in cellular transformation and tumorigenesis (Calin and Croce, 2006a, b; Engels and Hutvagner, 2006; Zeng, 2006; Cho, 2007). Furthermore, the genes encoding a number of miRNAs are located at chromosomal fragile sites and at regions of cytogenetic abnormalities associated with cancers (Calin and Croce, 2006b). Functionally, miRNAs through base-pairing with complementary sites at 3′-untranslated region's (UTR) of specific target genes cause their translational repression, and in some cases, mRNA degradation.

Expression profiling has also identified a few hundred miRNAs expressed in myometrium and leiomyomas with altered expression of a number of them, including miR-21 in leiomyomas (Wang ; Marsh ; Pan ). Altered expression of miR-21 has been reported in many tumors of various origins and functionally predicted to target the expression of several hundred genes whose products are involved in cellular proliferation, apoptosis, migration, transformation and tumorigenesis (Corsten ; Iorio ; Lui ; Meng ; Si ; Asangani ; Chan ; Dillhoff ; Frankel ; Gabriely ; Lu ). Among the experimentally validated genes targeted by miR-21 are phosphatase and tensin homolog (PTEN), programmed cell death 4 (PDCD4), transcription factor E2F1, tissue inhibitor of matrix metalloproteinase (TIMP3), Sprouty2 (SPRY2), tropomyosin 1 and maspin in a number of cell types (Meng ; Gabriely ; Lu ; Sayed ; Zhu ). The expression of miR-21 has also been reported to be regulated by ovarian steroids in MSMC and LSMC, and in the MCF-7 breast cancer cell line (Pan ; Wickramasinghe ).

In addition, leiomyomas when compared with myometrium are also characterized by differential expression of a large number of genes, with specific gene polymorphisms and non-random chromosomal abnormalities (Ligon and Morton, 2001; Luo ; Al Hendy and Salama, 2006; Denschlag ; Villanova ; Pan ). As such, altered expression of miRNAs, including miR-21 in leiomyoma, may have a direct regulatory function on the expression of specific target genes whose products play a key role in cellular transformation and their subsequent growth and/or apoptosis (Pan ; Peng ). To provide further support for regulatory function of miR-21 in leiomyoma, the present study was designed to (i) examine the expression and hormonal regulation of miR-21 in paired myometrium and leiomyomas and (ii) through gain- and loss-of function determine the expression of genes regulated by miR-21 in MSMC and transition into LSMC, transformed LSMC (t-LSMC) and the human leiomyosarcoma cell line SKLM. In addition, through bioinformatic data mining, we assessed the profile of miR-21-predicted target genes in these tissues, as well as in LSMC and MSMC treated with GnRHa, transforming growth factor (TGF)-β1 and TGF-β receptor type II (TGF-βRII) siRNA. We also examined the influence of gain-of function of miR-21 on the expression of genes downstream from TGF-β receptor signaling and cell growth and apoptosis in these cells.

Materials and Methods

Tissues

Portions of matched leiomyoma and myometrium were collected from premenopausal women (n = 23) who were scheduled to undergo hysterectomy for indications related to symptomatic leiomyomas. The patients' age ranged from 27 to 49 years (median = 38). Of these patients, 13 were not taking any medications, including hormonal therapy for the previous 3 months prior to surgery; based on last menstrual period and endometrial histology, they were from the mid-proliferative (n = 7) and early-mid-secretory (n = 6) phases of the menstrual cycle. The hormonal therapies administered to women for medical management of their symptomatic leiomyomas included: GnRHa (n = 4), Depo-Provera (n = 3) and oral contraceptives (OCPs) (n = 3). All the leiomyomas used in this study were 3–5 cm in diameter and were collected at the University of Florida affiliated Shands Hospital with prior approval from the Institutional Review Board. Immediately after collection, the tissues were snapped frozen and kept in liquid nitrogen for further analysis, or used for isolation of smooth muscle cells.

Isolation and culture of leiomyoma and myometrial smooth muscle cells

Myometrial smooth muscle cells (MSMC) and leiomyoma smooth muscle cells (LSMC) were isolated from small portions of matched leiomyoma and myometrium from women who did not receive any hormonal therapies (Chegini ). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplement with 10% fetal bovine serum (FBS) and until reaching confluence with change of media every 2–3 days. Prior to use, the cells were characterized using antibodies to α-smooth muscle actin, desmin and vimentin. Spontaneously transformed LSMC (t-LSMC) originally established in our laboratory (Pan ) and a human leiomyosarcoma cell line, SKLM-S1 (purchased from ATCC), were also cultured in DMEM supplement with 10% FBS. All the supplies for isolation and culturing of the cells were purchased from Sigma-Aldrich (St Louis, MO, USA), Invitrogen (Carlsbad, CA, USA) and Fisher Scientific (Atlanta, GA, USA).

Transfection with miR-21 oligonucleotides

On the basis of the type of experiments, MSMC, LSMC, t-LSMC and SKLM-S1 were seeded at various densities in 96-, 24- and 6-well plates and in 10 cm Petri dishes and cultured until reaching 70% confluence. The cells were washed and incubated in antibiotic-free media for 24 h. Prior to large-scale transfection, the transfection condition was optimized using cells cultured in 24-well plates in Opti-MEM I serum-reduced medium, BLOCK-iT Fluorescent Oligo and Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Following optimization, the cells were transfected with 50 nM of 2′-O-methoxyethyl (2′-O-MOE) modified pre-miR-21, anti-miR-21 and their respective negative controls (Ambion) for 6 h. The media were removed and the cells were cultured with full growth medium for 24, 48 and 72 h. Total RNA was isolated from these cells and subjected to real-time PCR. On the basis of the result of miR-21 expression in transfected MSMC, LSMC, t-LSMC and SKLM-S1, all the experiments were carried out for 48 h unless otherwise stated.

Bioinformatic analysis

The computational algorithms TargetScan (http://www.targetscan.org/) and miRDB (http://mirdb.org) were utilized for selection of miR-21-predicted target genes. The search results were downloaded to a local database and filtered using Microsoft Access. In addition, a few genes experimentally validated as miR-21 targets in other cell types were identified through PubMed search and added to the database (Supplementary data, Table SI). Table SI consists of 265 genes utilized for data mining and bioinformatic analysis throughout this study.

Through data mining, we searched for the expression of the 265 genes in the microarray database previously collected from (i) paired leiomyoma and myometrium from Caucasians and African Americans and from women who received GnRHa therapy; (ii) LSMC and MSMC treated with GnRHa (0.1 µM) for 2, 6 and 12 h; (iii) LSMC and MSMC treated with TGF-β1 (2.5 ng/ml) for 2, 6 and 12 h; and (iv) LSMC and MSMC treated with TGF-βRII antisense oligomer (1 µM) for 24 h and then treated with TGF-β1 (2.5 ng/ml) for an additional 24 h (Luo , b; Pan ). The normalized and transformed expression values of 194 genes were commonly identified in these cohorts (all 265 were expressed in paired leiomyoma and myometrium from Caucasians and African Americans), which were sorted based on a 1.5-fold change cut-off (Supplementary data, Table SI) and analyzed by cluster and tree-view. The gene list was also subject to functional analysis using the Ingenuity Pathways Software (Ingenuity® Systems, Redwood City, CA, USA).

Real-time quantitative RT–PCR

Using real-time PCR, the relative expression of miR-21 was determined in paired leiomyoma and myometrium from the proliferative and secretory phases of the menstrual cycle, from women who received GnRHa and Depo-Provera or OCPs. Transfection efficiency of anti- and pre-miR-21 in LSMC, MSMC, t-LSMC and SKLM was also determined using real-time PCR. Total RNA was extracted from these tissues and cells using Trizol (Invitrogen), and RNA quantity and quality were determined using an ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) and an Agilent Bio-Analyzer, respectively. Reverse transcription (RT) was performed with the Taqman microRNA RT Kit (Applied Biosystems) using a miR-21 stem-loop primer (Ambion). Each reaction was performed on 10 ng or 1 µg of total RNA. Real-time PCR was carried out using Taqman gene expression master mix and Taqman miRNA and/or gene expression assays, respectively (Applied Biosystems). Reactions were incubated for 10 min at 95°C, followed by 40 cycles of 15 s at 95°C, and for 1 min at 60°C. Analysis of relative miRNA and mRNA expression was performed using the ΔΔCT method with U6 and 18s rRNA as endogenous controls, respectively, according to the manufacturer's guidelines.

Western blot analysis

Western blot analysis was carried out following gain- and loss-of function of miR-21 in LSMC, MSMC, t-LSMC and SKLM. Total protein was isolated from these cells using lysis buffer containing protease inhibitor, centrifuged and the supernatants were collected; their total protein content was determined using a conventional method as described previously (Luo ; Pan ). Equal amounts of sample proteins (20 µg) were subjected to SDS–PAGE, transferred onto polyvinylidene difluoride membranes, and following further processing, the blots were probed with antibodies generated against a select number of miR-21 target genes; including E2F1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), PTEN (Santa Cruz), PDCD4 (Santa Cruz and Abcam, Cambridge, MA, USA) and TGF-βRII (Santa Cruz Biotechnology and Abcam) at 1:200 and 1:500 (TGF-βRII) dilutions. The membranes were also probed with an antibody generated against β actin (Sigma-Aldrich) at a 1:10 000 dilution serving as a loading control. Blots were washed and exposed to corresponding horseradish peroxidase-conjugated IgG; immunostained proteins were visualized using enhanced chemiluminescence reagents (Amersham-Pharmacia Biotech, Piscataway, NJ, USA), and band densities were determined as described previously (Levens ; Luo , 2006).

Influence of miR-21 gain-of function on TGF-βRII downstream gene expression

MSMC, LSMC, t-LSMC and SKLM-S1 were cultured in 6-well plates and in 10 cm Petri dishes until reaching subconfluence. Following gain- and loss-of-miR-21 function as described above, the cells were serum-starved for 24 h and then treated with TGF-β1 (2.5 ng/ml) for 24 h. Total RNA and protein were isolated from these cells and subjected to real-time PCR and western blot analysis, respectively, to determine the expression of fibromodulin (FMOD) and TGF-β-induced factor (TGIF)—two known targets of TGF-β regulatory function (Luo , 2006; Levens ). Western blotting was carried out using antibodies generated against FMOD and TGIF at 1:1000 dilution (Santa Cruz Biochemical) and processed as described above.

Cell proliferation and caspase-3/7 production assay

MSMC, LSMC, t-LSMC and SKLM-S1 were cultured in 96-well plates at a density of 5000 cells/well for 24 h, and washed and subjected to gain- and loss-of function of miR-21 as described above. After 48 h of incubation, the cells were subjected to Cell Titer 96 Aqueous Non-Radioactive Cell Proliferation Assay (Promega, Madison, WI, USA) according to the manufacturer's protocol. The rate of absorbance was determined using a multiplate reader and the results were presented as fold change when compared with their respective untreated controls.

The cells with miR-21 gain- and loss-of functions were also treated with 100 µl of Caspase Glo 3/7 reagent according to the manufacturer's instructions (Promega). MSMC treated with 1 µM Staurosporine (Sigma) for 24 h was used as a positive control according to the Caspase Glo 3/7 Assay Technical Bulletin. The caspase-3/7 activities were determined using a multiplate reader and relative luminescence intensity was presented as fold change when compared with untreated controls.

Gene expression microarray

Total RNA isolated from LSMC, MSMC, t-LSMC and SKLM-S1 with miR-21 gain- and loss-of functions was subjected to large-scale gene expression profiling. Following amplification and second-strand cDNA synthesis, 5 µg of purified cDNA was reverse transcribed using a high-yield RNA transcript labeling kit (Ambion). The product was purified and 20 µg of cRNA (0.5 µg/µl) was fragmented, mixed with 300 µl of hybridization mixture and 200 µl of the mixture was hybridized to Human Ref-8 v3 Expression BeadChip (Illumina, Inc., San Diego, CA, USA), consisting of 24 526 oligonucleotide probe sets representing 18 630 transcripts. The beads were processed after meeting recommended criteria for use of the expression arrays according to the manufacturer's protocol.

The expression values were background-subtracted, globally normalized using BeadStudio version 1.5.1.3 (Illumina), and as recommended, the expression values with a differential score of ≥13 were independently removed from each cohort. Transformed gene expression values were subjected to supervised assessments and differentially expressed genes were selected based on P ≤ 0.05 [analysis of variance (ANOVA), Tukey test] and a 1.5-fold change cut-off (Pan ). Data presented in Fig. 5 were derived from the gene expression values selected based on two-way ANOVA, supervised assessments by removing gene expression values ≤50 and sorted based on a 1.5-fold cut-off in MSMC with miR-21 gain-of function divided to control. These values were subjected to cluster and tree-view analysis and functional analysis using Ingenuity Pathways Software (Ingenuity® Systems).

Statistical analysis

Whenever appropriate, the results were reported as mean ± SEM of at least three experiments performed using different MSMC and LSMC preparations from three different patients. Comparisons between two or among multiple groups were made using paired t-test and ANOVA followed by a Tukey's honest significant difference post hoc pairwise multiple comparison, respectively, and P < 0.05 was considered statistically significant.

Results

This study consisted of three integrated parts aiming to analyze the expression, regulation and functional significance of miR-21 in myometrium and leiomyoma, their isolated smooth muscle cells, spontaneously transformed LSMC as well as a leiomyosarcoma cell line.

miR-21 expression is altered in leiomyomas

Microarray profiling identified the expression of a few hundred miRNAs with altered expression of several of them, including miR-21, in leiomyomas, LSMC, t-LSMC and SKLM-S1 when compared with myometrium and MSMC (Wang ; Marsh ; Pan ). As part of continuation of this work and the first aim of the present study, we further assessed the expression and regulation of miR-21 in paired leiomyoma and myometrium during the menstrual cycle and in women who received hormonal therapies for medical management of their symptomatic leiomyomas. As shown in Fig. 1, miR-21 is expressed at elevated levels in leiomyomas when compared with myometrium from both proliferative and secretory phases of the menstrual cycle (P < 0.05), confirming the above studies. In addition, the relative expression of miR-21 in leiomyomas, but not myometrium, was higher in women who received Depo-Provera and OCPs and lower in women who received GnRHa therapy (Fig. 1, P < 0.05). The pattern of miR-21 expression levels determined by real-time PCR in these cohorts correlated with the pattern obtained from microarray profiling, with the exception of a moderate increase in miR-21 expression in leiomyomas from the OCP group (data not shown).

Figure 1

Relative expression of miR-21 in paired leiomyoma (LYO) and myometrium (MYO) from the proliferative (n = 6) and secretory (n = 6) phases of the menstrual cycle, and in women who received GnRHa therapy (GnRHa; n = 3), Depo-Provera (n = 4) and those taking OCPs (n = 3).

Lines indicate a significant difference between LYO and corresponding MYO in each cohort. Different letters indicate a significant difference among treatments (P < 0.05; paired t-test and ANOVA)

miR-21-predicted target genes are expressed and regulated by GnRHa therapy

As a continued part of the first aim, we focused on the expression profile of genes predicted/validated as miR-21 targets (Supplementary data, Table SI) in myometrium and leiomyoma. Utilizing the list of 265 genes predicted and/or validated as miR-21 targets (Supplementary data, Table SI) through data mining of microarray data sets previously obtained in: (i) paired leiomyomas and myometrium from Caucasians and African Americans and women who received GnRHa therapy; (ii), LSMC and MSMC treated with GnRHa (0.1 µM) or with TGF-β1 (2.5 ng/ml) for 2, 6 and 12 h; and (iii) LSMC and MSMC treated with TGF-βRII siRNA (1 µM) for 24 h and then treated with TGF-β1 (2.5 ng/ml) for an additional 24 h, we identified the expression of 194 of these genes commonly expressed among the cohorts with cluster and tree-view analyses shown in the Supplementary data, Figs S1–3. The number of genes differentially expressed (≥1.5-fold change cut-off) in leiomyomas when compared with myometrium from Caucasians and African Americans (Supplementary data, Fig. S1A), in women who received GnRHa therapy when compared with the untreated group (Supplementary data, Fig. S1B), and in MSMC and LSMC treated with GnRHa (Supplementary data, Fig. S3A) is presented as part of the Supplementary Table SI. Noticeable among the differentially expressed genes are PDCD4, PTEN, SPRY1, E2F1, STAT3 and TGF-βRII, suggesting the possible regulation of some of the miR-21 target genes by GnRHa. Pathway analysis identified and mapped 21 functional networks among these genes, with 15 networks containing a majority of the genes, including networks 1, 2 and 3 with network 1 having most of the TGF-β-related genes (network 1 and merged networks 1, 2 and 3 are shown in two figures in Supplementary Table SI). The biological functions of the genes in these networks are shown in a bar graph (Supplementary data, Table SI), which include regulation of cell growth and proliferation, cell death, connective tissue development and function, cell assembly, transcriptional regulation, and cancer.

Experimental validation of miR-21 target genes

As a second part of the aim of this study, we focused on experimental validation of a few miR-21 target genes in leiomyoma and myometrial cells. To accomplish this, we isolated MSMC and LSMC (primary cultures through third passage), and t-LSMC and SKLM-S1, and transfected the cells with pre- and anti-miR-21 (gain-and loss-of functions, respectively) for 24, 48 and 72 h. Prior to transfection, the basal expression of miR-21 was determined in each cell type and, as indicated, MSMC expressed a lower level of miR-21 when compared with LSMC, t-LSMC and SKLM (Fig. 2A). As expected, gain-of-miR-21 function significantly increased, whereas loss-of function decreased the expression of miR-21, when compared with scrambled-oligo-transfected cells after 48 h (P < 0.05; Fig. 2B) and 72 h (data not shown) of incubation. Since PTEN, PDCD4 and E2F1 have been experimentally validated as miR-21 targets in several cell types, we next assessed whether gain- or loss-of function of miR-21 altered their expression and the expression of TGF-βRII in MSMC, LSMC, t-LSMC and SKLM-S1. These cells expressed PTEN, PDCD4, E2F1 and TGF-βRII mRNAs and proteins at varying levels (Figs 3 and 4) and gain- and loss-of-miR-21 function differentially altered their expression in a cell-specific manner when compared with scrambled-oligo-transfected controls (Figs 3 and 4; P < 0.05). However, the level of E2F1 production in MSMC and LSMC, and PDCD4 production in MSMC, LSMC, t-LSMC and SKLM-S1 were very low with weak band intensity (data not shown). The degree of miR-21 functional regulation of PTEN, PDCD4, E2F1 and TGF-βRII expression in these cells is possibly due in part to differences in the baseline expression of the genes, or cell-type-dependent action of miR-21; i.e. more effective inhibition of TGF-βRII in SKLM-S1 and MSMC, when compared with LSMC with least inhibition in t-LSMC.

Figure 2

Relative level of miR-21 endogenous expression in MSMC, LSMC, t-LSMC and SKLM-S1 (A) and following gain- and loss- of function of miR-21 (B).

Cells were cultured and at 70% confluence, were transfected with 2′-O-MOE miR-21 inhibitor (anti-miR-21), pre-miR-21 or their respective scrambled control (Ctrl) oligonucleotides (50 nM) for 48 h. Total RNA was isolated from these cells and the level of miR-21 expression was determined using real-time PCR. The expression value in MSMC (A) and controls for each cohort independently (B) was set at 1. Different letters or number of asterisks indicate a significant difference among cell types and treatments, respectively, when compared with their respective controls (Ctrl) with P < 0.05 considered significant.

Figure 3

Western blot analysis of PTEN, E2F1 and TGF-βRII production in MSMC, LSMC, t-LSMC and SKLM-S1 with gain- and loss-of-function of miR-21.

Total cellular protein was isolated from cells transfected with 2′-O-MOE anti-miR-21 (A), pre-miR-21 (P) or their respective scrambled oligomer controls (AC and PC) for 48 h and subjected to immunoblot analysis. Note considerable changes in the level of production of PTEN in MSMC and LSMC; E2F1 in t-LSMC and SKLM-S1; and TGF-βRII in MSMC, LSMC and SKLM-S1 (see densitometric analysis in Fig. 4) when compared with controls.

Figure 4

The level of (mean ± SEM) expression of E2F1 (A and B), PTEN (C and D), TGF-βRII (E and F) and PDCD4 (G) mRNA (A, C, E and G) and protein (B, D, and F) determined by real-time PCR and western blot analysis, respectively, in MSMC, LSMC, t-LSMC and SKLM with gain- and loss-of-function of miR-21 as described in Fig. 3. Note considerable alterations in the level of gene expression in pre-mir-21 versus anti-mir-21 treated cells when compared with scrambled oligomer-treated (Ctrl) cells. The basal expression of E2F1, PTEN and PDCD4 is shown in (H). The expression values in MSMC (H) and in controls for other cohorts were independently set at 1 for each cell type. Different letters or number of asterisks indicate a significant difference among cell types (A–G) or genes analyzed (H), respectively, when compared with their respective controls (Ctrl) with P < 0.05 considered significant.

miR-21 overall regulation of gene expression in MSMC and LSMC

To further assess the overall regulatory function of miR-21 on MSMC, LSMC, t-LSMC and SKLM gene expression, total RNA was isolated from these cells (with gain- and loss-of function of miR-21) and subjected to gene microarray analysis. Based on P ≤ 0.05 (ANOVA) and a 1.5-fold change cut-off, the expression of 117, 807, 705 and 998 genes was altered (directly and/or indirectly) as a result of miR-21 gain-of function in MSMC, LSMC, t-LSMC and SKLM-S1, respectively, when compared with scrambled-oligo-transfected controls (Supplementary data, Table SIIA–D). Of these genes, the expression of 83, 609, 434 and 702 genes was down-regulated and of 34, 198, 271 and 296 genes was up-regulated in MSMC, LSMC, t-LSMC and SKLM, respectively (Supplementary data, Table SIIA–D). Loss-of function of miR-21 also altered the expression of a number of genes, some displaying an inverse pattern of expression when compared with cells with gain-of miR-21 function (Supplemental data, Table SII). Cluster/tree-view analysis of 406 differentially regulated genes in MSMC with gain-of miR-21 function versus control is shown in Fig. 5 with the enlarged region showing several genes, including TGF-βRII (TGFBR2), whose expression was altered in MSMC, LSMC and SKLM-S1, and to a lesser extent in t-LSMC (Fig. 3).

Figure 5

Cluster and tree-view analyses of 406 up- and down-regulated genes selected based on two-way ANOVA with a ≥1.5-fold change cut-off in MSMC with gain-of miR-21 function/control in MSMC, LSMC, t-LSMC and SKLM-S1.

The cells were transfected with 2′-O-MOE anti-miR-21(Anti), pre-miR-21 (Pre-mir) or their respective scrambled oligomers (Ctrl) for 48 h. Enlarged area shows the heat map and the list of genes, including TGFBR2 (TGF-βRII), whose expression was altered in these cohorts. The lists of up- or down-regulated genes, the Ingenuity pathways analysis and the biological function of the genes in MSMC, LSMC, t-LSMC and SKLM-S1 are presented in Supplementary Table SIIA–D, respectively.

Functional pathway analysis mapped several networks among the genes targeted by gain-of function of miR-21 in MSMC, LSMC, t-LSMC and SKLM-S1 (see figures embedded into Supplementary Table SIIA–D, respectively; and network summary in Supplementary Table SII). A majority of the focus genes function as transcriptional, translational and signal transduction mediators, and in cell cycle regulation, ECM turnover, cell–cell communication and metabolic activities etc. (see bar graphs embedded into Supplementary Table SIIA–D). The list of 15 genes, including TGFβRII, whose expression was commonly altered as a result of miR-21 gain-of function in MSMC, LSMC, t-LSMC and SKLM-s1, is presented in Supplementary data, Table SIIE. Additionally, of the 265 genes predicted/validated as miR-21 targets (Supplementary data, Table SI), the expression of 9, 10, 31 and 10 genes, including TGF-βRII, was altered in MSMC, LSMC, t-LSMC and SKLM-S1, respectively, as a result of miR-21 gain-of function (Supplemental data, Table SIIF).

Regulation of miR-21-predicted target genes by TGF-β and influence of gain-of function of miR-21

Since TGF-βRII appeared to be a target of miR-21 regulatory function (Figs 3 and 4), as a third aim of this study, we assessed whether the expression of genes downstream from TGF-β receptor signaling is altered as a result of miR-21 gain-of function. As illustrated in Fig. 6A–C, 2′-O-MOE modified pre-miR-21-transfected MSMC and LSMC (P) treated with TGF-β1 (P-TGF) produced significantly lower levels of FMOD and transforming growth interacting factor (TGIF) when compared with cells transfected with negative control oligo (PC) and PC treated with TGF-β1 (PC-TGF) (Fig. 6; P < 0.05). We extended this aim and assessed the expression pattern of the genes predicted/validated as miR-21 targets (Supplementary data, Table SI) in microarray profiles previously obtained from LSMC and MSMC treated with TGF-β (2.5 ng/ml) for 2, 6 and 12 h, and cells treated with TGF-βRII antisense siRNA for 24 h followed by TGF-β treatment. The data mining indicated that 35 genes were differentially expressed and altered by at least 1.5-fold in MSMC and LSMC treated with TGF-β (Supplementary data, Fig. S2A) and TGF-βRII siRNA (Supplementary data, Fig. S2B).

Figure 6

Gain-of function of miR-21 alters TGF-βRII and TGF-β-regulated genes.

MSMC and LSMC were transfected with 2′-O-MOE pre-mir-21 (P) or scrambled oligomer controls (PC) for 48 h, and then treated with TGF-β1 (2.5 ng/ml) for 24 h (P-TGF and PC-TGF). Total protein isolated from these cells was subjected to western blot analysis of FMOD and TGIF (A) and their band densities (mean ± SEM) are reported in (B and C), respectively. β-Actin was used as a loading control. The control (bars representing PC) band intensity values are independently set at 1 for each protein. Different letters indicate a significant difference among treatments when compared with their respective controls (Ctrl), with P < 0.05 considered significant.

miR-21 gain- and loss-of functions on cell growth and caspase activity

Aberrant expression of miR-21 has been associated with alteration of cell growth and apoptosis in various cell types. To follow up on the functional significance of miR-21, we determined the influence of miR-21 gain- and loss-of functions on cell growth and apoptosis of MSMC, LSMC, t-LSMC and SKLM-S1. The results indicated that miR-21 gain- and loss-of functions have a moderate effect on the rate of cell proliferation and caspase-3/7 activity, which occurred in a cell-type-specific manner (Fig. 7, P < 0.05).

Figure 7

Cell growth and apoptosis in response to miR-21 gain- and loss-of functions.

MSMC, LSMC, t-LSMC and SKLM were cultured and transfected with anti-miR21, pre-miR-21 or their respective scrambled oligomer (Ctrl) for 48 h and the rate of growth (A) and caspase-3/7 activity (B) were determined. Bar graphs (mean ± SEM) show that gain-of function of miR-21 (pre-mir-21) in LSMC and SKLM moderately increased the rate of cell growth and reduced caspase-3/7 activities, when compared with loss-of-function of miR-21 (anti-mir21) and Ctrl, respectively. The control values are set at 1 independently for each group. Different letters indicate a significant difference among cell types and treatments, respectively, when compared with their respective controls (Ctrl) with P < 0.05 considered significant.

Discussion

Accumulated evidence has associated the aberrant expression of many miRNAs, including miR-21, with various pathological disorders, more specifically cancers, and with the regulation of the expression of specific genes involved in various aspects of tumorigenesis, i.e. cellular transformation, proliferation, apoptosis and migration (Si ; Asangani ; Chan ; Connolly ; Dillhoff ; Fujita ; Gabriely ; Sayed ; Talotta ; Zhu ). In the present study, we confirmed that miR-21 is aberrantly expressed in leiomyomas when compared with myometrium (Wang ; Marsh ; Pan ). We further demonstrated a specific increase in expression of miR-21 in leiomyomas, but not in myometrium, during the secretory phase of the menstrual cycle and in women who received Depo-Provera and OCPs, with lower expression in women who received GnRHa therapy. It was previously observed that miR-21 is differentially regulated by ovarian steroids in MSMC and LSMC, and this included an increase in expression of miR-21 following medroxyprogesterone acetate treatment (Pan ). Although the in vitro condition does not reflect the complex in vivo milieu, these results imply that miR-21 is hormonally regulated in leiomyomas, potentially influencing the expression of genes targeted by miR-21.

Through data mining, a number of miR-21-predicted/validated target genes, including PDCD4, PTEN, SPRY1, SPRY2, E2F1 and TGFβRII were identified as differentially expressed in leiomyomas when compared with myometrium from African Americans and Caucasians, and from patients who received GnRHa therapy. The co-expression and ethnic and hormonal regulation of miR-21 and predicted target genes in leiomyomas imply their potential functional interactions. As such, using isolated MSMC and LSMC as well as t-LSMC and SKLM-S1 with gain- and loss-of function of miR-21 revealed that a number of genes, including some of miR-21-predicted targets are potentially regulated by miR-21. However, there was a considerable difference in the number and types of genes targeted by miR-21 gain-of function among these cells, which could be due to variation in the level of their endogenous miR-21 expression, or the degree of cellular transformation (benign versus cancerous as in LSMC versus t-LSMC and SKLM-S1). Since altered expression of miR-21 has been associated with cellular transformation and tumorigenesis (Cho, 2007; Connolly ), and miR-21 differentially regulates specific genes in a cell-type-based manner, miR-21 could influence cellular events necessary for transformation of myometrial cells into leiomyomas and further into leiomyosarcoma. Since TGF-β plays a central role in pathogenesis of leiomyoma and miR-21 appeared to regulate TGF-βRII expression, we extended our investigation and through data mining we identified that a number of miR-21-predicted/validated target genes differentially expressed in MSMC and LSMC treated with TGF-β and TGF-βRII antisense oligomer. Our preliminary results have also indicate that TGF-β regulates the expression of miR-21 in MSMC and LSMC; these results, in combination with the finding of our study, suggest a direct and/or indirect regulatory interaction between TGF-β and miR-21 in leiomyoma cells.

The biological significance of regulatory interactions between TGF-β and miR-21 and their target genes, PTEN, E2F1, PDCD4 and TGF-βRII, may lie in their functions in various cellular activities critical to the outcome of tumorigenesis and tissue fibrosis, including in leiomyomas (Chegini, 2005; Luo ; Luo ). However, miR-21 differentially regulated the expression of PDCD4, PTEN, E2F1 and TGF-βRII in MSMC, LSMC, t-LSMC and SKLM, which could be either due to variations in basal endogenous expression of these genes, or a cell-type-dependent action of miR-21. As such, a feedback regulatory mechanism between miR-21 and these genes, specifically TGF-β, may result in balancing their different functions in a cell- and tissue-dependent context (i.e. inflammatory response, cell growth regulation and tissue turnover), leading to either cellular transformation and tumorigenesis, tissue fibrosis or tumor regression (Fig. 8). Although the TGF-β system is well recognized as a key regulator of establishment and progression of tissue fibrosis, up-regulation of miR-21 has also been observed in several fibrotic disorders over-expressing TGF-β, including cardiac hypertrophy, vascular walls balloon injury and neointimal lesion (for review, see Krichevsky and Gabriely, 2009). However, controversies exist regarding miR-21 functional regulation of cellular events, and gene expression changes leading to tissue fibrosis. Inhibition of miR-21 reduced hypertrophy of cardiomyocytes and neointima formation, but resulted in cardiac cellular hypertrophy; conversely, over-expression of miR-21 in cardiomyocytes altered the expression of sets of genes and resulted in the development of cellular hypertrophy (Thum ; Krichevsky and Gabriely, 2009).

Figure 8

Schematic diagram describing the potential regulatory interactions between ovarian steroids, GnRHa, TGF-β and miR-21.

miR-21 and TGF-β are differentially regulated by ovarian steroids and GnRHa (Chegini ; Pan ; Luo and Chegini, 2008). TGF-β increases the expression of miR-21 which suppresses the expression of TGF-βRII, resulting in reduced TGF-β receptor-mediated signaling. Regression of TGF-βRII by miR-21 either directly or indirectly alters TGF-β-induced miR-21 through a feedback mechanism, leading to differential expression of PTEN, E2F1 and PDCD4 and other miR-21 target genes. As such, altered expression of miR-21, TGF-β, TGF-βRII, PTEN, E2F1 and PDCD4 could then alter the balance between the rate of cell growth and apoptosis, inflammatory reaction and extracellular matrix accumulation, resulting in leiomyoma growth and regression, or mediate myometrial and leiomyoma cellular transformation. Using Ingenuity pathway analysis provided a functional network among TGF-βRII-, PTEN-, E2F1- and PDCD4-mediated actions, which include interactions with other gene products and signaling pathways (i.e. Smad, AKT, JUK, NFkβ, cbp/p300, VEGF, 14-3-3 and proteasome) that regulate the above cellular processes.

miR-21 has also been shown to regulate cellular dedifferentiation, and in cell lines with low or undetectable expression of miR-21, their differentiation resulted in increased miR-21 expression (Krichevsky and Gabriely, 2009). As such, the cell-specific function of miR-21 on PTEN, E2F1, PDCD4 and TGF-βRII expression in MSMC, LSMC, t-LSMC and SKLM-S1, representing various stages of progression from normal to benign, transformed and sarcoma, and differentiated versus undifferentiated status, may reflect the diverse actions of miR-21, including crosstalk with other miRNAs and target gene products. PDCD4 acts as a tumor suppressor and decreases benign and malignant tumor progression, whereas E2F1 acts as an oncogene or a tumor suppressor regulating cell cycle progression and apoptosis, depending on cellular context. PTEN antagonizes the action of phosphatidylinositol-triphosphate kinase (PI3K)-signaling pathways and Akt-signaling cascades, which increase cellular survival and proliferation. Previous studies have demonstrated that TGF-β down-regulates the expression of PTEN in various cell lines, whereas it increases PDCD4 expression, mediating apoptosis through caspase activation (Zhang ). Functional interactions between miR-21 and TGF-β in vascular smooth muscle cells, where TGF-β, acting in part through miR-21 regulation and Smad signaling, down-regulated the expression of PDCD4 (Davis ). MSMC and LSMC with gain-of function of miR-21 treated with TGF-β also expressed lower levels of FMOD and TGIF, two genes known to be regulated by TGF-β (Levens ; Luo ), comparable to TGF-βRII antisense-treated cells. Such cell- and gene-target-specific actions of miR-21 may be necessary since the products of many of the genes targeted by miR-21 also regulate the expression of other genes, i.e. TGF-β regulation of PTEN, E2F1 and PDCD4 (Hu ; Hjelmeland ; Garcia-Alvarez ; Bu ; Chow ; Davis ; Lacerte ; Petrocca ; Selbach ; Talotta ). Consistent with the level of gene expression in MSMC, LSMC, t-LSMC and SKLM-S1 with gain- and loss-of miR-21 functions, proteomic analysis has also indicated that a single miRNA repressed the production of many proteins; however, the level of repression was relatively low, and miRNA actions occurred either directly or indirectly through protein synthesis from thousands of genes (Baek ; Selbach ).

The consequence of miR-21 functional regulation of TGF-βRII, E2F1, PTEN and PDCD4 in leiomyoma cells could also involve cell survival, while limiting their cellular transformation and tumorigenesis when compared with t-LSMC and SKLM-S1. Aberrant expression of miR-21 has been associated with alteration of cell proliferation, migration, invasion and malignant transformation by targeting the expression of select genes whose products are integral to these processes (Meng ; Si ; Connolly ; Lu ; Petrocca ; Zhu ). Pathway analysis indicated that many of the genes targeted by miR-21 in MSMC, LSMC, t-LSMC and SKLM-S1 are functionally involved in regulating the above processes, as well as tissue development, connective tissue development/function, cell assembly, cell cycle, cell death and metabolism. However, gain- or loss-of function of miR-21 in MSMC, LSMC, t-LSMC and SKLM-S1 had a limited effect on the rate of cell proliferation, while moderately affecting caspase activities. In several cancer cell lines, inhibition of miR-21 decreased cell proliferation, migration and invasion, and reduced PDCD4 production, while increasing PTEN and matrix metalloproteinases 2 and 9 (Corsten ; Meng ; Asangani ; Frankel ; Lu ; Zhu ); we expected a similar trend at least in t-LSMC and SKLM-S1 with miR-21 gain-of function. The reason for differences in miR-21 action on the rate of cell proliferation/caspase activities could be due to cell type specificity and/or the extent of alteration of genes functionally regulating cell cycle progression and apoptosis. Additionally, it has been reported that NFIB, a target of miR-21 regulatory function, binds to the miR-21 promoter and negatively regulates its expression—a double-negative feedback regulation considered to sustain miR-21 expression (Fujita ). As such, miR-21 functional regulation of TGF-βRII, E2F1, PTEN and PDCD4 in leiomyoma cells could serve in mediating cell survival, while limiting cellular transformation and tumorigenesis when compared with t-LSMC and SKLM.

In summary, we provided further evidence that miR-21 is over-expressed in leiomyomas when compared with myometrium in a menstrual cycle-dependent manner, with altered expression in women who received GnRHa, Depo-Provera and OCPs. More specifically, we identified a number of genes as potential targets of miR-21 via direct and/or indirect actions, including PDCD4, E2F1, PTEN and TGF-βRII in MSMC, LSMC, t-LSMS and SKLM-S1, with potential regulation by GnRHa and TGF-β in these cells. Functionally, miR-21 had a moderate influence on the rate of proliferation and apoptosis of these cells, but down-regulated the expression of TGF-β downstream target genes. On the basis of our previous and present results, we propose a feedback regulatory interaction between miR-21, TGF-β and ovarian steroids, necessary to balance their different functions in a cell- and tissue-dependent context, influencing events such as inflammatory response, cell growth regulation and tissue turnover, leading either to cellular transformation and tumorigenesis, or tissue fibrosis (Fig. 8). Although the goal of this study was to better understand the regulatory function of miR-21 in leiomyomas, future investigation is required to address the direct interaction of miR-21 with TGF-βRII 3′-UTR and with other miR-21 target genes, as well as the biological significance of other aberrantly expressed miRNAs in leiomyoma and transformation into leiomyosarcoma.

Supplementary data

Supplementary data are available at http://molehr.oxfordjournals.org/.

Funding

This work was supported by grants RO1HD37432 and RO1HD58664 (N.C.) and RO3HD58779 (X.L.) from the National Institutes of Health.