Ovarian Ablation Alone Promotes Aortic Intimal Hyperplasia and Accumulation of Fibroblast Growth Factor
Background—Estrogen-mediated cardiovascular protection is incompletely explained by its beneficial lipid-modifying effects. Previous studies interrogating direct vascular effects of estrogens have used models of either diet- or injury-induced atherosclerosis. The purpose of this study was to determine the influence of ovarian ablation alone on vascular remodeling. We hypothesized that estrogens are atheroprotective, independent of their influence on lipid metabolism, by directly influencing the production and effects of a prototypical atherogenic mediator, basic fibroblast growth factor (bFGF).
Methods and Results—Twenty-five female sheep were randomized to sham operation, ovariectomy, or ovariectomy plus 17β-estradiol replacement. Serum cholesterol and triglyceride levels were serially measured for 1 year and were similar among groups and in the normal range (30 to 60 mg/dL). At 6, 9, and 12 months, ovariectomy resulted in aortoiliac intimal hyperplasia compared with sham (P<0.01) and hormone replacement (P<0.01) groups. The neointima of ovariectomized animals was characterized immunohistochemically by increased vascular smooth muscle cells (VSMCs). Levels of bFGF protein were determined in adjacent aortic segments. Ovariectomized sheep had 2-fold more FGF than sham or ovariectomized sheep that received hormone replacement. In vitro, estradiol inhibited the mitogenic effect of bFGF on human aortic VSMCs.
Conclusions—Without dietary manipulation, ovarian ablation alone induces aortic intimal hyperplasia in the ewe. Estradiol abrogates this response independently of its effects on serum lipids. Hormone replacement decreases the accumulation of the atherogenic peptide bFGF in vivo and inhibits the mitogenic response of VSMCs to bFGF in vitro. These results suggest that estrogens may provide atheroprotection both by modulating local production and by attenuating the influence of bFGF on VSMC growth.
Premenopausal women and postmenopausal women on hormone replacement therapy are relatively protected against coronary disease.1 An abundance of epidemiological studies support the role of estrogens in mediating this protection.2 Although these hormones exert beneficial lipid-modifying effects, they do not alone appear to account for the reduction in cardiovascular deaths observed in estrogen users.1 Indeed, estrogens may have profound genomic and nongenomic effects on the vascular wall and may act at several levels in response to vascular injury.3 However, many of the studies investigating direct vascular effects of the hormone are premised on models of diet-induced atherosclerosis. As such, nonlipid effects of estrogens are confounded by hyperlipidemic animals.
Vascular injury results in the activation of cytokines and growth factors that may promote vascular smooth muscle cell (VSMC) proliferation, migration, and intimal hyperplasia.4 Basic fibroblast growth factor (bFGF) has been identified in human atheromatous lesions5 and is well known to be an important mediator of the response to vessel injury. Produced by several vascular cells, bFGF has well-established proliferative effects on endothelial cells and VSMCs.6 Accumulating evidence supports the coupling of estrogens to peptide growth factor signaling pathways.7 Estrogens may upregulate growth factor production, including bFGF, in endometrial8 and breast9 cancer cell lines. Conversely, hormone replacement therapy reduces the levels of bFGF in human postmenopausal endometrium.10 Yet, the majority of studies implicating steroid–peptide hormone cross talk are focused on tumor, bone, and reproductive system models. Currently, little information exists regarding the influence of estrogens on growth factors, in particular bFGF, in vascular tissue.
In this study, we sought to determine the influence of ovarian ablation alone on vascular remodeling. We demonstrate that surgical menopause results in aortic intimal hyperplasia characterized by increased VSMCs and accumulation of bFGF. Physiological estrogen replacement, independently of changes in serum lipids, prevents the development of intimal hyperplasia, intimal VSMC accumulation, and vessel production of bFGF. In addition, physiological estradiol inhibits the bFGF-induced mitogenic response of human VSMCs in culture. These results suggest that estrogen may provide atheroprotection both by modulating local production and by attenuating the influence of bFGF on VSMC growth.
This study used 25 skeletally mature, 8-year-old Warhill ewes housed at the College of Veterinary Medicine and Biomedical Sciences at the Colorado State University (CSU). The sheep were raised on the range in flocks and were maintained on a natural grass and hay diet. The study design was approved by the Animal Care and Use committee of CSU. The sheep were randomized to 3 groups: sham (n=6), ovariectomy (OVx, n=9), and ovariectomy plus estradiol implants (OVxE, n=10). The initial procedures were performed under general endotracheal anesthesia maintained with isoflurane and 100% oxygen. Ovariectomy or sham operations were performed with a ventral midline approach with standard techniques. A 5-cm capsule was inserted subcutaneously near the iliac crest. Each capsule was prepared from Silastic tubing, 0.335×0.465 cm OD, and filled with pure crystalline 17β-estradiol in the hormone replacement animals and with saline in the sham and ovariectomy alone animals. When able to ambulate, ewes remained in the barn for 1 week and then were transported to their flocks. The general medical management and nursing care were provided by the staff of the CSU Veterinary Teaching Hospital. During the observation period, serial profiles of triglycerides, cholesterol, and luteinizing hormone were obtained at 0-, 6-, 9-, and 12-month intervals.
At 6 (n=6), 9 (n=13), or 12 (n=6) months, the ewes were anesthetized and given 5000 U of heparin sodium IV. Euthanasia was performed according to the guidelines set forth by the American Veterinary Medical Association Panel on Euthanasia, with 10 mL of pentobarbital sodium (400 mg/mL). The aortoiliac segment was removed and flushed with PBS. One-centimeter sections were cut and immediately frozen with liquid nitrogen and stored at −70°C. The adjacent bifurcation was perfusion-fixed with 10% neutral buffered formalin for 20 minutes at room temperature and then immersed in the formalin fixative for an additional 24 hours at 4°C. The specimens were removed and cut into 1-cm segments at the bifurcation to prepare for histological processing. After paraffin embedment, 4-μm sections were placed on slides for Masson’s trichrome staining.
To standardize morphometric analysis, the lumen, intima, and media were defined microscopically and their dimensions were calculated by use of a computer-linked digitizer (NIH Image 1.59b4 application program). Cross-sectional areas of intima and media were not used in our measurements because of the large caliber of the sheep aorta. At the magnification necessary to view the whole circumference of the vessel (×1), the internal elastic lamina (IEL) could not be reliably assessed. At ×10 magnification, all layers of the vessel wall at a given segment, including the IEL, could be identified and analyzed. As such, intimal hyperplasia was expressed quantitatively by comparing the ratio of the intima to the sum of the intima and media at 6 different points. Vessel bifurcations can be problematic because of hemodynamic forces that may promote intimal changes. In particular, the aortoiliac bifurcation results in flow-related changes in the posterior wall.11 Therefore, measurements were taken from the anterior and medial segments of the vessel wall to standardize our technique and to minimize the effect of tensile forces on our observations. Triplicate measurements were performed separately by 2 blinded investigators to avoid observational bias.
To characterize the effect of ovarian ablation with or without hormone replacement on vascular smooth muscle, aortic segments were stained for muscle-specific actin. Paraffin-embedded sections were deparaffinized with xylene and PBS washes. After proteolysis with pronase (Sigma), blocking with normal rabbit serum (10%), and PBS washes, segments were incubated with muscle-specific actin primary polyclonal antibody (Enzo Diagnostics) for 10 minutes. After PBS washings, secondary incubation was performed with the peroxidase-antiperoxidase technique.
Frozen aortic segments were homogenized on ice in a buffer containing 50 mmol/L Tris-HCl, 2 mmol/L EGTA, 1 mmol/L benzamidine, and 1 mmol/L PMSF (Sigma). After centrifugation at 4°C and 2000 rpm for 10 minutes, the supernatant was removed and frozen. Aliquots (10 μL) were used to measure protein concentration via a bicinchoninic acid assay (Pierce). For Western blot analysis, tissue homogenates were thawed and mixed 1:1 with a Tris-SDS buffer (Bio-Rad). Electrophoresis was performed on linear gradient SDS-polyacrylamide gels (Bio-Rad). After transfer to 0.45-mm polyvinylidene difluoride membrane (Millipore), membranes were blocked in 5% nonfat milk at 4°C overnight. Primary incubation with rabbit anti–bFGF-2 1/1000 (Santa Cruz Biotechnology) was performed at room temperature for 1 hour. After sequential washing in 0.1% Tween-20 in PBS, membranes were incubated in horseradish peroxidase–linked secondary antibodies (R&D Systems) for 45 minutes and detected with the enhanced chemiluminescence system (Amersham).
Human VSMCs were isolated from segments of thoracic aorta harvested from 2 premenopausal female transplant donors as previously described.12 Cells were trypsinized and plated in gelatin-coated 96-well microtiter plates at a density of 3000 cells/well with “complete medium” (phenol red–free DMEM, penicillin G 100 U/mL, streptomycin sulfate 100 μg/mL, amphotericin 250 ng/mL, glutamine 200 mM, and 5% FCS and pooled human cord serum). After 8 hours, the medium was changed to serum-free DMEM with glutamine. Serum-free conditions were maintained for 48 hours to allow for synchronized growth arrest. The medium was then changed to DMEM with 5% FCS with 10 ng/mL bFGF (R&D) with and without 17β-estradiol (Sigma) or 10 μmol/L tamoxifen (Sigma). After 24 hours, rates of proliferation were determined with the CellTiter 96 assay (Promega). This technique is equivalent to tritiated thymidine incorporation and direct cell counting in determining viable cell numbers.12 After the addition of 20 μL of methyl-tetrazolium salt, plates were incubated at 37°C for 90 minutes. Absorbance was then recorded at 490 nm with a microtiter plate reader. Results, reported as optical densities, represent experiments done in quadruplicate with both cell lines at first and second passage.
Data are presented as mean±SEM. ANOVA with Bonferroni-Dunn post hoc analysis was used to analyze differences between individual means. Statistical significance was accepted within 95% confidence limits.
Estradiol levels in sheep are lower than those in premenopausal and postmenopausal women. Basal levels of ovine estrogen are 2 to 3 pg/mL, with a luteinizing peak of 8 to 10 pg/mL.13 The estradiol implants maintain serum levels at 4 to 6 pg/mL.13 Available estradiol assays may not accurately detect levels <20 pg/mL. Because luteinizing hormone levels are a reliable measure of ovarian function,14 they were used to evaluate for the completeness of ovarian ablation as well as the adequacy of estrogen replacement. In the ovariectomized sheep, luteinizing hormone increased from 0.15±0.07 ng/mL before treatment to 5.36±1.08 ng/mL at 6 months, 5.86±0.31 ng/mL at 9 months, and 6.48±0.23 ng/mL at 12 months (P<0.01 at each time point), thus confirming ovarian ablation. In the hormone replacement group, luteinizing hormone levels remained similar to the pretreatment level (0.51±0.07 ng/mL) at 6 (0.11±0.11 ng/mL), 9 (0.35±0.21 ng/mL), and 12 months (0.52±0.33 ng/mL). As expected, because of negative pituitary feedback, estrogen replacement inhibited luteinizing hormone release. When the 2 groups were compared, luteinizing hormone was increased in sheep without replacement (P<0.01 versus OVxE) at 6, 9, and 12 months.
Estrogen Has No Effect on Serum Lipids
Serial measurements of blood cholesterol and triglyceride were obtained in the ovariectomized animals with and without hormone replacement and did not change from baseline values. In addition, lipid levels were within normal accepted ranges and comparable to sham groups (30 to 60 mg/dL). To verify that other lipid parameters were not affected, we obtained blood samples in the 9-month animals for more specific parameters of lipid metabolism (Table⇓). There were no differences in LDL, HDL, and lipoprotein (a) levels among sham, ovariectomy, and ovariectomy plus estrogen replacement groups.
Estrogen Inhibits Ovariectomy-Induced Intimal Hyperplasia
The typical histological morphology of intimal proliferation after ovariectomy is shown in Figure 1⇓ (left) and is compared with the normal intima in estradiol-treated sheep (right). Ratios of the intima to the sum of the intima and media from several points in a given tissue section were compared among the groups (Figure 2A⇓). Cumulatively, ovariectomy resulted in a 2.5-fold increase in intimal hyperplasia compared with sham and estrogen replacement groups (P<0.01). There was no difference between the sham and hormone replacement groups (P=0.85). Because neointimal lesions evolve with time, subgroup analysis was performed for the 6-, 9-, and 12-month sheep (Figure 2B⇓). At 6 months, ovariectomy resulted in intimal thickening (0.12±0.02) compared with the sham (0.04±0.01) and hormone replacement (0.06±0.01) groups; at 9 months, ovariectomy resulted in intimal thickening (0.14±0.01) compared with sham (0.06±0.01) and hormone replacement (0.06±0.01) groups; and at 12 months, ovariectomy resulted in intimal thickening (0.14±0.02) compared with sham (0.05±0.01) and hormone replacement (0.04±0.01) groups. The intima-to-media ratios of sham and replacement groups were not different at 6 (P=0.38), 9 (P=0.64), and 12 (P=0.50) months. To verify that the intimal ratio adequately assessed intimal thickening, we also compared medial ratios among the animals (media/intima+media). There were no differences between groups (P=0.3), suggesting that the intimal ratios represented true changes in intimal size.
Although a wide array of vascular cells are important in atherogenesis, we were interested in the relationship between estrogen and vascular smooth muscle. Immunohistochemical staining for smooth muscle–specific actin was performed to characterize the intima in the respective groups. In the sham and hormone replacement animals, the VSMCs were confined to the medial layer (Figure 3A⇓ and 3B⇓). In the ovariectomized animals, staining of aortic segments identified VSMCs penetrating across the IEL into the enlarged intimal layer (Figure 3C⇓).
Estrogen Inhibits Aortic bFGF Production
To explain the intimal hyperplasia with observed estrogen deprivation, we examined the ability of estrogens to regulate production of the prototypical atherogenic peptide, bFGF. Aortas from sheep 9 months after ovariectomy were analyzed for expression of bFGF protein. Western blots, with 17-kDa human recombinant bFGF as a standard, demonstrated an increased expression of bFGF in ovariectomized animals compared with hormone replacement and sham animals (Figure 4⇓). By delivering a known quantity of bFGF standard to the gels, we quantified levels of aortic bFGF, normalized to total protein concentrations, by computer-assisted densitometry. Ovariectomized sheep had 2-fold more bFGF (1.19±0.15 pg/μg total protein) than sham or ovariectomized sheep on hormone replacement (0.61±0.09 and 0.58±0.06 pg/μg total protein, P<0.01).
Estradiol Inhibits bFGF-Induced Human VSMC Proliferation
Human aortic VSMCs from 2 premenopausal female human transplant donors were grown in culture to determine the effect of estrogens on growth factor–regulated VSMC proliferation (Figure 5⇓). Estrogen treatment had no effect on growth of unstimulated VSMCs (data not shown). bFGF is a known smooth muscle mitogen and induced VSMC proliferation (0.18±0.01) compared with control (0.11±0.01). With 17β-estradiol treatment, bFGF-induced growth was attenuated in a dose-dependent fashion. Growth inhibition was seen at physiological doses, beginning as low as 10 pg/mL (0.12±0.003 versus bFGF alone, P<0.01). To further support the notion that this growth inhibition was secondary to an estradiol effect, we performed parallel assays with concurrent administration of tamoxifen, an estrogen antagonist. When given alone, tamoxifen (10 μmol/L) had no effect on growth of unstimulated VSMCs. When given with bFGF, tamoxifen resulted in a nonsignificant decrease in VSMC proliferation (0.18±0.02). Simultaneous delivery of bFGF (10 ng/mL), estradiol (10 pg/mL), and tamoxifen (10 μmol/L) reversed the suppressive effects of estrogen on bFGF-induced VSMC growth (0.16±0.01 versus 0.12±0.01, P<0.01).
In this study, we demonstrate that in the ewe, ovarian ablation alone, without dietary manipulation, immunological insult, or mechanical injury, resulted in intimal hyperplasia, which is abrogated by estrogen replacement therapy. Several investigators have reported inhibition of atherosclerosis by estrogen without changes in serum lipoprotein levels.15 16 However, the cholesterol-fed animals in these studies all exhibited supranormal lipid levels. In rabbit transplant and balloon-injury models, estrogen therapy had no effect on intimal hyperplasia when cholesterol levels were normalized to human levels.17 Conversely, after carotid artery balloon injury, without cholesterol supplementation, estrogens inhibited intimal hyperplasia at physiological doses.18 Although estrogen replacement has been associated with decreasing levels of total cholesterol and increasing levels of HDL cholesterol and triglycerides in healthy postmenopausal women,1 we did not observe any changes in the lipid profiles of these sheep. Acknowledging that the correlation between sheep and human lipid levels in regard to cardiovascular disease is obscure, the observed intimal effects occurred within a normolipemic milieu, suggesting that these effects are independent of hormonal influence on lipid metabolism.
Although estrogen may affect the biological activity of several important vascular cells involved in atherogenesis, including monocytes, platelets, macrophages, lymphocytes, and fibroblasts,3 the present study focuses on the influence of the hormone on vascular smooth muscle. Ovariectomy resulted in intimal hyperplasia characterized by an increase in intimal VSMCs. bFGF is 1 prototypical atherogenic peptide that is an important mediator of VSMC proliferation and intimal hyperplasia.6 After ovarian ablation in the ewe, bFGF is upregulated in the hyperplastic aortic wall. Hormone replacement attenuates this accumulation of bFGF (Figure 4⇑). Although the increased levels of bFGF in ovariectomized sheep may simply be a marker for the amount of neointima present, it may alternatively explain, in part, the observation of increased smooth muscle in their neointima (Figure 3⇑). As such, the lack of intimal VSMCs in hormone replacement animals may have resulted from the estrogen-associated decrease in production of vessel bFGF. Heretofore, few reports have suggested a relationship between growth factors and estrogen in vascular tissue. In rabbit aortic allografts, estradiol may attenuate the vessel accumulation of insulin-like growth factor (IGF).19 Yet, in rat aortic allografts, although estradiol attenuated both early (3 days) and late (21 days) intimal, medial, and adventitial IGF protein expression, the hormone had no consistent suppression of platelet-derived growth factor (PDGF) or bFGF expression.20 At 21 days, however, estradiol did suppress medial expression of bFGF, suggesting that hormonal influence over bFGF production may indeed exist, albeit with a time course different from that of IGF.
In addition to downregulating bFGF production, estrogen might also alter the response of a target cell, specifically VSMCs, to available bFGF. To test this hypothesis and extend our observations to a more clinically relevant situation, we examined the influence of bFGF and physiological estrogen on human premenopausal aortic VSMC proliferation (Figure 5⇑). In vitro, the ability of estradiol to inhibit bFGF-induced proliferation in human female arterial VSMCs has not been reported previously. Investigators have reported estradiol inhibition of VSMC proliferation in response to bFGF in other cell culture conditions, including A10 cells, a fetal rat aortic cell line,21 concurrent administration with epidermal growth factor in human aortic VSMCs,22 and postmenopausal saphenous vein VSMCs.23 Conversely, estrogens may promote VSMC proliferation, both independently and synergistically with PDGF, in uterine and aortic VSMCs.24 Although these latter observations arose from supraphysiological doses of estradiol in VSMCs isolated from guinea pigs, estrogens may have different effects with different growth factors, ie, they may inhibit VSMC proliferation induced by bFGF but not by PDGF.
Morphometrically, the intimal thickening 6 months after ovariectomy appears to be similar to that observed at 12 months (Figure 2B⇑), suggesting that a relatively short interval of estrogen depletion can effect profound changes in the vessel wall. These observations are apparently related to ovarian ablation and not a function of aging, because sham-operated animals did not develop intimal hyperplasia throughout the experimental period. The ovariectomized sheep did not demonstrate some of the histopathological findings associated with advanced atherosclerosis, limiting broad conclusions correlating our observations in sheep with those of human atherosclerosis. An additional important point concerning our model is that immunoassays for bFGF were performed on homogenates of whole aortic segments. As such, we are unable to conclude which particular cell may be most affected by hormone treatment. Estrogen receptors have been identified on several bFGF producers, including mononuclear cells, endothelial cells, and VSMCs.25 Estrogens regulate monocyte release of interleukin-1 and tumor necrosis factor, endothelial release of prostacyclin and nitric oxide, and VSMC production of procollagen I and III.3 Thus, hormone replacement may act on several cellular sources to decrease bFGF production in the vessel wall.
The importance of estrogens in mediating cardiovascular protection in premenopausal women and postmenopausal women on hormone replacement therapy is strongly supported epidemiologically. This protection is not adequately explained by favorable changes in lipid metabolism. Within the response-to-injury paradigm of atherogenesis, estrogens may intervene at several levels of vascular remodeling. Our present study demonstrates that ovarian ablation alone may promote intimal hyperplasia, which is abrogated with hormone replacement. Because we used physiological concentrations of estrogens in these studies of range-fed sheep, our observations may have greater relevance to human atherosclerosis than many previous cell culture or animal studies that used supraphysiological concentrations of estrogen in hyperlipidemic animals. Furthermore, we expand the relationship between estrogens and vascular peptides by demonstrating that estrogen may modulate atherogenesis by altering vessel accumulation and target cell responses of the prototypical atherogenic peptide bFGF.
This work was supported by a CSU Veterinary Medicine Research Council grant, National Institutes of Health grants GM-49222 and GM-08315, and the University of Colorado Health Sciences Center Dean’s Academic Enrichment Fund. We thank Lauren Kaufman, Denise Parker, Kelley Fantle, and Drs Renee Dewell, Kyung Joo, Yvonne Hodges, and Brian Shames for their support and help.
- Received March 6, 1998.
- Revision received June 8, 1998.
- Accepted June 11, 1998.
- Copyright © 1998 by American Heart Association
Grady D, Rubin SM, Petitti DB, Fox CS, Black D, Ettinger B, Ernster VL, Cummings SR. Hormone therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med. 1992;117:1016–1037.
Hughes SE, Crossman D, Hall PA. Expression of basic and acidic fibroblast growth factors and their receptors in normal and atherosclerotic human arteries. Cardiovasc Res. 1993;27:1214–1219.
Lindner V, Lappi DA, Baird A, Majack RA, Reidy MA. Role of basic fibroblast growth factor in vascular lesion formation. Circ Res. 1991;68:106–113.
Ignar-Trowbridge DM, Pimentel M, Teng CT, Korach KS, McLachlan JA. Cross talk between peptide growth factor and estrogen receptor signaling systems. Environ Health. 1995;103:35–38.
Turizillo AM, Nett TM. Effects of estradiol on concentrations of gonadotropin releasing hormone receptor messenger ribonucleic acid following removal of progesterone. Endocrine. 1995;3:765–768.
Karsch FJ, Legan SJ, Ryan KD, Foster DL. Importance of estradiol and progesterone in regulating LH secretion and estrous behavior during the sheep estrous cycle. Biol Reprod. 1980;23:404–413.
Adams MR, Kaplan JR, Manuck SB, Koritnik DR, Parks JS, Wolfe MS, Clarkson TB. Inhibition of coronary artery atherosclerosis by 17β-estradiol in ovariectomized monkeys: lack of an effect of added progesterone. Arteriosclerosis. 1990;10:1051–1057.
Sulistiyani, Adelman SJ, Chandrasekaran A, Jayo J, St. Clair RW. Effect of 17α-dihydroequilin sulfate, a conjugated equine estrogen, and ethynylestradiol on atherosclerosis in cholesterol-fed rabbits. Arterioscler Thromb Vasc Biol. 1995;15:837–846.
Oparil S, Levine RL, Chen SJ, Durand J, Chen YF. Sexually dimorphic response of the balloon-injured rat carotid artery to hormone treatment. Circulation. 1997;95:1301–1307.
Lou H, Zhao Y, Delafontaine P, Kodama T, Katz N, Ramwell PW, Foegh ML. Estrogen effects on insulin-like growth factor-I (IGF-I)–induced cell proliferation and IGF-I expression in native and allograft vessels. Circulation. 1997;96:927–933.