| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a Departments of Obstetrics and Gynecology and Cell Biology, Duke University Medical Center, Durham, North Carolina 27710
ABSTRACT
2-Methoxyestradiol (2-ME) is an estradiol metabolite with antiangiogenic and antitumor activity. It is formed by granulosa cell (GC) catechol-O-methyltransferase activity and is present in the normal follicle at high concentrations. In this unique microenvironment, it may regulate selected cell types via autocrine and/or paracrine action. To assess the possibility that 2-ME or estradiol might exert differential mitotic and/or apoptotic effects on endothelial cells and GCs, we compared their actions on primary cultures of hormone- and/or growth factor-stimulated porcine GCs (pGCs) as well as two types of endothelial cells, primary cultures of porcine endothelial cells (pECs), and a spontaneously transformed rabbit endothelial vascular cell (REVC) line. The 2-ME, but not estradiol, dose dependently suppressed tritiated thymidine (3H-T) incorporation into epidermal growth factor (EGF)-stimulated REVCs and EGF/insulin (INS)-stimulated pECs. In contrast, 2-ME did not attenuate incorporation in FSH/INS-stimulated pGCs. It reduced incorporation by approximately 50% in EGF/INS-stimulated pGCs, indicating that responsiveness to 2-ME in normal cells can be modulated by hormone and growth factor treatment. Estradiol was not antimitotic to pGCs. As indicated by 4',6-diamido-2-phenylindole hydrochloride nuclear staining, estradiol was nonapoptotic in either cell type, and 2-ME significantly increased apoptosis of REVCs, but not of pGCs. In a cell migration assay, REVC movement was attenuated by 2-ME, but not by estradiol. In summary, the results show that antimitotic as well as proapoptotic responses to 2-ME vary with cell type and, in the case of pGC antimitotic activity, with the regulatory microenvironment. Thus, they provide a rationale for autocrine and/or paracrine action of 2-ME at its site of production in vivo, and they strongly support the concept of 2-ME as a candidate ovarian angiogenesis inhibitor.
apoptosis, estradiol, follicle-stimulating hormone, granulosa cells, ovary
INTRODUCTION
Angiogenesis is the generation of new blood vessels by capillary growth sprouting from preexisting vessels [1, 2]. Angiogenesis and antiangiogenesis are prevalent during the cyclic events of the female reproductive cycle associated with follicular and luteal growth and apoptosis as well as endometrial proliferation and regression [3–6]. The molecular regulation of these processes is very poorly understood, and whether any aspects of regulation are unique to specific organs or tissues within the reproductive system is not known. Antiangiogenesis has been the subject of intense interest because of its implications for restricting tumor growth [2]. Several very exciting large-molecular-weight, antiangiogenic compounds have been identified as a result of this work, including angiostatin and endostatin [7, 8]. To our knowledge, no evidence exists that these or similarly acting endogenous, large-molecular-weight compounds are organ- or tissue-specific with respect to their production or action. AGM-1470 (O-chloracetylcarbamoyl fumagillol, also designated TNP-470), a synthetic compound with broad antiangiogenic action, was recently shown to inhibit implantation in the mouse [9], but it was not shown to inhibit luteal development in the mouse or the primate [9, 10].
Because the formation of new blood vessels plays an obligatory role in the successful growth and development of ovulatory follicles, a more complete understanding of the regulatory influences that promote as well as inhibit follicular angiogenesis is important. 2-Methoxyestradiol (2-ME) is a naturally occurring compound that is formed during metabolism of estradiol [11]. It is a major metabolite of estradiol in human serum; its concentration peaks at midcycle, like that of estradiol [12, 13]; and it has been reliably quantitated by gas chromatography-mass spectroscopy at maximal concentrations of 0.03 and 0.82 µM in human and mare follicular fluid, respectively [14, 15]. Evidence is growing that 2-ME is a potent inhibitor of angiogenesis and tumor growth, and that its ability to suppress tumorigenesis is due to its antiangiogenic effect on vascular endothelial cell proliferation [16]. Moreover, it has also been shown directly to kill cultured human leukemic cells [17] and is currently being evaluated in phase I clinical trials. The current study was designed primarily to evaluate whether 2-ME might have differential antimitotic and/or proapoptotic actions on selected, nontumorigenic cell types found at or near its site of synthesis (e.g., granulosa cells [GCs] and endothelial cells). A partial report of this work was presented previously in abstract form [18].
MATERIALS AND METHODS
Cell Culture and Cell Growth Determinations
Porcine GCs (pGCs) were isolated from antral follicles (diameter, 1–4 mm) as previously described [19] with the following modifications. To better utilize GCs of the large aggregates composing the hemifollicular linings, the entire follicle cell harvest was layered on a 12.5% (v/v) Percoll/FD (Ham F-12/Dulbecco modified Eagle medium [DMEM], 1:1) solution for 2.5 min, during which time the large aggregates settled into the lower phase. Following removal of the supernatant cells and their application to the top of a separate 50% Percoll/FD solution, the aggregates were pelleted by centrifugation in a rapid on/off mode on reaching 1000 x g. They were then resuspended in 1.0 ml of FD, pipetted 5–10 times through a 1.0-ml plastic pipette tip, and incubated with 10 µl/ml of an enzyme mixture (0.28% collagenase, 0.2% hyaluronidase, and 0.125 mg/ml of DNase) for approximately 5 min at 37°C, followed by repeated pipetting. This procedure yielded individual GCs and small, three- to four-cell aggregates. In some cases, slightly longer incubation times and/or increased enzyme amounts were required. The GCs that had been applied to the 50% Percoll/FD were separated from red blood cells by centrifugation at 1000 x g for 30 min at 4°C. The band of cells at the media/50% Percoll interface (
2 ml) was combined with the enzymatically dispersed cells, diluted with FD to fill a 15-ml conical tube, and pelleted by centrifugation at 150 x g for 5 min. The resuspended cells were counted and cultured as previously described [20]. Viable GCs (2.5 x 105 cells) were plated in 24- or 48-well plates (24-well plates, Falcon, Becton-Dickinson, Lincoln Park, NJ; 48-well plates, Costar, Cambridge, MA) and cultured for 1 day in FD/10% fetal bovine serum (FBS) containing 1.0 µg/ml of insulin (INS). On Day 2, the medium was changed to FD/2.5% FBS plus INS and the steroidal and nonsteroidal test compounds (10.0 ng/ml of EGF [R&D Systems, Minneapolis, MN], 100 ng/ml of NIH ovine FSH 17 or NIH ovine FSH 19). After approximately 18 h, mitotic activity was assessed in triplicate cultures per treatment using methyl 3H-T (6.7 Ci/mmole; ICN, Costa Mesa, CA) [20]. The DNA was quantitated fluorometrically using Hoechst 33258 (Sigma, St. Louis, MO) [21]. Primary cultures of porcine aortic endothelial cells (kindly provided by Dr. Jeffrey L. Platt) were cultured as described above, except that they were maintained in high-glucose DMEM/serum [22]. A spontaneously transformed rabbit endothelial vascular cell (REVC) line [23] was obtained from the Lineberger Cancer Research Center (University of North Carolina at Chapel Hill, NC). These cells do not require INS but were otherwise maintained as the pGCs were. For analysis of 3H-T incorporation, the REVCs were seeded at 1.0 or 5.0 x 104 cells/well in a 96-well plate (Costar). Except where indicated, the chemical reagents were purchased from Sigma and the cell culture media solutions from Gibco BRL (Gaithersburg, MD).
Morphological Assessment and Quantification of Apoptosis
Cells (5.0 x 104) were grown in 24-well plates (as described above) on 12-mm borosilicate glass coverslips (thickness, 0.13–0.17 mm), the surfaces of which are optimal for fluorescent analysis (Deutsche Spiegelglas; Carolina Biological Supply Co., Burlington, NC). We used 4',6-diamido-2-phenylindole hydrochloride (DAPI) staining, as described by Aharoni et al. [24], to detect cells undergoing apoptosis, with the modification that the final wash step in PBS was extended to 30 min. Specimens were viewed under a fluorescent microscope (excitation, 370 nm; emission, 520 nm) using a 63x objective. Cells that contained highly fluorescent and condensed or blebbed nuclei were defined as apoptotic. At least six random fields of each treated culture were analyzed, scoring 200–300 total cells. Cultures were analyzed in duplicate.
Cell Migration Assay
The REVC migration was measured using a modified Boyden Chamber assay, as described by Clyman et al. [25]. Confluent REVCs were trypsinized, pelleted by centrifugation, and resuspended in DMEM/2.5% FBS to a final concentration of 2 x 106 cells/ml. Aliquots (50 µl) of cells (1 x 105) were mixed with an equal volume of DMEM/2.5% FBS containing the test compound, and these 100-µl cell treatment aliquots were applied to the filter insert wells (polycarbonated filter wells; pore size, 8 µm; Costar). Each filter insert was then placed in the culture well containing 500 µl of DMEM/2.5% FBS. After 5 h of incubation at 37°C, the migrated cells on the lower surface of the filter inserts were fixed and stained. Treatment cell counts were compared by counting all cells in a given microscopic field and a constant number of fields for all treatment groups within a given experiment. The total cell count for the control groups was 700–800.
Statistics
Based on the comparisons required, data were analyzed by linear regression, Student t-test (two-tailed), or analysis of variance followed by Tukey test. Because the computer program used to portray the results did not allow graphical presentation of SEM estimates, the SEM (expressed as a percentage of the sample mean) was calculated and averaged for the treatment groups and is provided as additional information in the figure legends to provide a sense of within-treatment and among-treatment variability.
RESULTS
Growth Inhibition by 2-ME
We routinely utilized hormone- or growth factor-stimulated cells, because they provided a broader range of mitotic activity over which to evaluate the postulated antimitotic and/or proapoptotic effects. The 2-ME caused a highly significant (P < 0.001), dose-dependent attenuation of 3H-T incorporation into epidermal growth factor (EGF)-stimulated cells of the REVC line, whereas estradiol, at the same concentrations, was without effect (Fig. 1A). Relative to EGF alone, 2-ME first significantly attenuated incorporation at 0.5 µM (P < 0.001). Moreover, 2-ME did not alter DNA levels, and the 3H-T inhibition pattern was unchanged, when incorporation was normalized to culture DNA (Fig. 1), thus indicating the decrease in 3H-T incorporation was not due to significant cell loss secondary to detachment. The differing responses to 2-ME and estradiol were reproduced in primary cultures of EGF/INS-stimulated porcine endothelial cells (Fig. 2), and the inhibition by 2-ME was, again, highly significantly dose dependent (P < 0.001).
|
|
In contrast, 2-ME had no effect on 3H-T incorporation by the FSH/INS-stimulated pGC (105.0% ± 4.2% [mean ± SEM] of control; pooled data from 0.1–2.0 µM) (Fig. 3A). Relative to the mitogenic response achieved with FSH/INS, EGF/INS-stimulated cells responded suboptimally (57.6% of incorporation by FSH/INS-stimulated controls; 5440 ± 327 vs. 9450 ± 337 cpm/culture; n = 3; P < 0.001). In this group, 0.1–2.0 µM 2-ME attenuated incorporation to 52.1% of the EGF/INS-stimulated control value (P < 0.05) (Fig. 3B). In turn, this level of incorporation was highly significantly less that that found in FSH/INS-stimulated cells at each dose of 2-ME (P < 0.01), averaging 28.7% that of FSH/INS/2-ME-treated cells. The response to 0.5 µM 2-ME in the EGF/INS-treated group (Fig. 3B) was an exception to the trend in this study; it was not observed in the other two replicates of this experiment. Comparison of 2-ME action among the various cell types indicates that attenuation by 2-ME in EGF/INS-stimulated pGCs is intermediate between that of FSH/INS-stimulated pGCs and EGF-stimulated endothelial cell models in the presence or absence of INS (Figs. 1–3). As in the endothelial cells, estradiol (0.1–2.0 µM) did not have significant inhibitory effects on pGCs (34.6% ± 5.9% and 2.0% ± 6.5% greater than control in two replicate experiments).
|
Limited specificity studies were performed because of the detailed analysis already completed by Fotsis et al. [16] showing 2-ME to be the first steroid found to have high antiangiogenic activity by itself. Estriol and pregnanediol were evaluated over a dose range of 1.0–20.0 µM and did not have significant inhibitory effects on REVCs. At 20.0 µM, incorporation by estriol- or pregnanediol-treated cells was 81.2% ± 31.5% and 82.9% ± 20.5% of control, respectively. At 10.0 and 20.0 µM, 2-methoxyestriol attenuated 3H-T incorporation to 69.9% ± 12.2% and 38.7% ± 16.0% of control, respectively, indicating activity of the 2-methoxy configuration in a different parent compound, albeit at extremely high concentrations.
Induction of Apoptosis
Nuclear condensation and fragmentation are characteristic markers of apoptosis. To evaluate the possible apoptotic effect of 2-ME, we used the nuclear stain DAPI. Based on a time-course study, 2.0 µM 2-ME produced a plateau in apoptosis over 6–18 h (data not shown). We chose 16–18 h as the time interval for all comparative determinations, because 15 and 20 h were the time points used in studies of cultured rat GCs and bovine pulmonary artery endothelial cells, respectively [24, 26]. Figure 4 shows that apoptotic cells could be identified by their highly condensed or fragmented nuclei, in contrast to those of normal cells, which showed much weaker, homogeneous staining. In EGF-stimulated REVCs, 0.1, 0.5, 1.0, and 2.0 µM 2-ME increased the percentage of apoptotic cells to the same level (range, 9.4% at 0.1 µM to 10.6% at 2.0 µM), suggesting that under these conditions, the proapoptotic response to 2-ME may be threshold in nature. These points were pooled for presentation in Figure 5A and averaged 5.8-fold greater than the control values (P < 0.01). At the same concentrations, 2-ME did not increase apoptosis in EGF/INS-stimulated pGCs (data not shown) or in FSH/INS-stimulated pGCs (range of pooled data, 1.2% at 0.1 µM to 0.9% at 2.0 µM) (Fig. 5B). In a comparison of cell types, 2-ME increased the average percentage of apoptotic cells in REVCs (9.3%) over that in pGCs (1.4%) by 7.2-fold (P < 0.001) (Fig. 5). Estradiol was not proapoptotic in either cell system (data not shown).
|
|
Attenuation of Endothelial Cell Migration
Endothelial cell migration is essential to promote angiogenesis. In addition to the demonstrated growth-inhibitory and apoptotic effects of 2-ME, we evaluated the possibility that it could also block migration of REVCs. At 1.0 as well as 2.0 µM, 2-ME highly significantly reduced migration relative to the untreated control (P < 0.001). At 2.0 µM, 2-ME attenuated migration to approximately 40% (P < 0.001) of the control value, whereas this dose of estradiol was inactive (Fig. 6).
|
DISCUSSION
During follicular development, capillary networks enter the thecal compartment from the stroma, grow, and just before ovulation, migrate to approximate the basal lamina separating the thecal from the granulosal compartment. At the time of ovulation, this process likely is regulated primarily by vascular endothelial growth factor (VEGF) under the control of gonadotropins. VEGF mRNA and/or protein expression have been demonstrated using the preovulatory ovary or isolated GCs in the rat and primate, respectively [27–30]. To our knowledge, no comparable information exists concerning the regulation of antiangiogenesis in the ovary in terms of reports specifically linking candidate effectors to the local inhibition of endothelial cell mitosis or migration. Because the ovarian follicle is a site of 2-ME production, we hypothesized that it would minimally affect GC mitotic activity or survival in culture but have antimitotic and/or proapoptotic effects on endothelial cells under similar conditions in vitro. This would contribute evidence for possible differential cellular and antiangiogenic responses in the ovarian follicle and a paradigm unique to the normal ovary in vivo.
Although exogenous 2-ME has antiangiogenic activity in isolated cell systems and attenuates solid-tumor growth by this mechanism [16, 26], as mentioned above, its most physiologically relevant action likely occurs at its site of production. In this locale, concentrations as high as 0.82 µM (247 ng/ml) have been reported in mare follicular fluid [15]. The GC is a principal cellular site of production; Spicer et al. [11] demonstrated that pGCs were able to convert 2-hydroxyestradiol to 2-ME. In partial contrast to the results reported here, those authors also showed that estradiol, as well as 2-ME at doses (1.0–4.0 µg/ml [3.3–13.2 µM]) appreciably beyond its reported maximal follicular fluid concentration, had pronounced antimitotic effects on EGF/INS-stimulated or unstimulated pGCs [31]. The most probable explanations for the difference involve issues related to the experimental conditions, principally concentrations of the steroids tested, timing of the experimental test period, media composition, and cell density. Fotsis et al. [16] reported that 2-ME attenuated the growth of cultured nonendothelial cell types, including bovine GCs, if they were subconfluent and not contact-inhibited. Our studies did not show a density-dependent response in FSH/INS-stimulated cells (data not shown). They did show, however, that pGCs optimally stimulated by the appropriate hormonal combination were least affected by the antimitotic and proapoptotic capabilities of 2-ME. The cellular mechanisms contributing to the treatment-dependent sensitivity of cells to 2-ME is worthy of further investigation and could be clinically relevant. In this regard, 2-ME induced apoptosis in phytohemagglutinin-stimulated human lymphocytes, but not in normal lymphocytes [17].
The dose range over which 2-ME is active indicates that it is unlikely to be acting via the estrogen receptor (ER), which is consistent with the observation that 2-ME has a less than 1000-fold binding affinity for the ER relative to that of estradiol [32]. It is possible that 2-ME could activate putative, nonclassical ERs, the actions of which have been hypothesized to explain the effects of methoxychlor and the catechol estrogen 2-hydroxyestradiol [33, 34], or it may act via yet-undiscovered mechanisms. Moreover, these observations invite the question of whether cell culture systems that do not involve classical ER action and require high effector concentrations in vitro are relevant models of follicle cell responses in vivo. The available evidence, however, indicates the models are appropriate. The maximal average concentrations of estradiol in human and porcine follicular fluid from healthy follicles are approximately 1700 ng/ml (6.0 µM) and 225 ng/ml (0.8 µM), respectively [35, 36]. Therefore, the concentration in human follicular fluid is approximately 3500-fold that in serum at midcycle (4500-fold for the pig at estrus) and in great excess of that required for maximal ER activation. Such extremely high concentrations also serve as substrate for the formation of high concentrations of estradiol metabolites, such as 2-ME. Furthermore, 2-ME is lipophilic and may concentrate in cell membranes or cellular compartments at even higher concentrations [26]. For example, 3H-2-ME is concentrated intracellularly by cultured cells to levels approximately 10-fold those in media [17]. Thus, the highly specialized microenvironment of the ovarian follicle in vivo provides the physiological situation wherein 2-ME may act independently of classical ER mechanisms to modulate angiogenesis as well as other cellular processes in any of the cell types within the follicle, including the oocyte. In the follicle interior, 2-ME likely is not a predominant regulator of cellular activity related to meiosis, mitosis, or apoptosis, but it is interesting to speculate that it could participate in the attenuation of GC mitosis and/or promotion of apoptosis under conditions of FSH and growth factor insufficiency. This action might be confined to a narrow time window, because inappropriate FSH stimulation, in turn, would be expected to limit 2-ME production. Exterior to the follicular basal lamina, 2-ME may limit inappropriate angiogenesis (i.e., hyperplasia of endothelial cells and/or premature migration of endothelial cells into the follicular antrum just before breakdown of the basal lamina at ovulation). Its action may be direct, as suggested by our results, and/or indirect, as suggested by the study of Banerjeei et al. [37], which showed that 2-ME decreased VEGF protein expression in estrogen-induced pituitary tumors.
The demonstrated intracellular mechanisms that may link 2-ME action with participatory regulation of mitosis and/or apoptosis include its interaction with microtubules and/or modulation of selected enzymatic activities, such as stress-activated protein kinase, nitric oxide synthase, and superoxide dismutase (SOD) [17, 26, 35, 38–40]. The 2-ME binds to the colchicine site of tubulin with a median effective concentration of 33 µM and inhibits polymerization at substoichiometric concentrations under suboptimal reaction conditions [38]. This places 2-ME in the class of drugs such as paclitaxel, etoposide, vinblastine, and vincristine, which cause cell-cycle arrest by perturbing normal microtubule dynamics and inhibiting spindle assembly [39]. The demonstrated interaction of 2-ME with microtubules is also one of the mechanisms responsible for its proapoptotic activity and attenuation of endothelial cell migration [26]. The mechanism involving CuZnSOD (SOD 1) as a target of 2-ME action is most intriguing [17]. Inhibition of SOD 1, possibly by preferential 2-ME binding, causes accumulation of cellular O2-, which leads to free radical-mediated damage to mitochondrial membranes, release of mitochondrial cytochrome C, and apoptosis of human leukemic cells. In vivo, SOD 1 has important actions in the ovary. Gene knockout of SOD 1, but not of SOD 2, contributes to markedly suppressed fertility, and pharmacological inhibition of SOD 1 significantly attenuates ovulation and blood-follicle barrier function [41, 42]. However, the extent to which these functional effects can be linked directly to changes in specific ovarian cell types remains to be determined.
In summary, the results show the antimitotic as well as the proapoptotic responses to 2-ME vary with cell type and, in the case of pGC antimitotic activity, with the regulatory microenvironment. Thus, they provide a rationale for autocrine/paracrine action of 2-ME at its site of production in vivo and support strongly the concept of 2-ME as a candidate ovarian angiogenesis inhibitor.
ACKNOWLEDGMENTS
Ovine FSH preparations (NIDDK-oFSH-19-SIAFP AFP-4117A; NIADDK oFSH-17 AFP 6446C) were provided by the National Institute of Diabetes and Digestive and Kidney Diseases' National Hormone and Pituitary Program and by Dr. A.F. Parlow, National Institute of Child Health and Human Development, and the U.S. Department of Agriculture.
FOOTNOTES
First decision: 5 January 2001.
1 Supported in part by grants HD 11827 and T32 HD 07135 from the NICHD, DK 07012 from the NIDDK, the Andrew W. Mellon Foundation, and ACS-IRG-INMI58, an Institutional Research Grant Award to the Duke University Comprehensive Cancer Center. 
2 Correspondence: David W. Schomberg, Box 3323, Duke University
Medical Center, Durham, NC 27710. FAX: 919 681 6494; david.schomberg{at}duke.edu
Accepted: April 5, 2001.
Received: November 27, 2000.
REFERENCES
-dihydrotestosterone, gonadotropins and catecholamines. Mol Cell Endocrinol 1989; 64:119-126[CrossRef][Medline] and non-ERß mechanism. Endocrinology 1999; 140:3526-3533This article has been cited by other articles:
|
|
 |
|
  S. Salih, X. Xu, T. D. Veenstra, A. J. Duleba, H. Fouad, M. Nagamani, and A. Al-Hendy Lower Levels of Urinary 2-Hydroxyestrogens in Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., August 1, 2007; 92(8): 3285 - 3291. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Eichenlaub-Ritter, U. Winterscheidt, E. Vogt, Y. Shen, H.-R. Tinneberg, and R. Sorensen 2-Methoxyestradiol Induces Spindle Aberrations, Chromosome Congression Failure, and Nondisjunction in Mouse Oocytes Biol Reprod, May 1, 2007; 76(5): 784 - 793. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R Gruemmer, L Klein-Hitpass, and J Neulen Regulation of gene expression in endothelial cells: the role of human follicular fluid J. Mol. Endocrinol., February 1, 2005; 34(1): 37 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. L. Lattanzi, C. B. Santos, M. D. Mudry, and J. L. Baranao Exposure of Bovine Oocytes to the Endogenous Metabolite 2-Methoxyestradiol During In Vitro Maturation Inhibits Early Embryonic Development Biol Reprod, December 1, 2003; 69(6): 1793 - 1800. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
