Differentiation between vasculoprotective and uterotrophic effects of ligands with different binding affinities to estrogen receptors α and β

Abstract

Estrogen-based drug therapy in cardiovascular diseases has been difficult because it has not been possible to separate the wanted vasculoprotective effect from the unwanted effects of the hormone to the reproductive system. Here, we demonstrate that, after endothelial denudation of rat carotid artery, the mRNA of the classical estrogen receptor (ERα) is constitutively expressed at a low level whereas the expression of the novel ERβ mRNA increases >40-fold. Under in situ hybridization and immunohistochemistry, ERβ mRNA and protein colocalize with the smooth muscle cells in the media and neointima. Treatment of ovariectomized female rats with the isoflavone phytoestrogen genistein, which shows 20-fold higher binding affinity to ERβ than to ERα, or with 17β-estradiol, which does not differentiate between the two receptors, provides similar dose-dependent vasculoprotective effect in rat carotid injury model. In addition in concentrations <10 μM, both ligands are equally inhibitory to the replication and migration of smooth muscle cells in vitro. However, only treatment with 17β-estradiol, but not with genistein, is accompanied with a dose-dependent uterotrophic effect. The results suggest that preferential targeting to ERβ will provide vasculoprotective estrogen analogs devoid of effects to the reproductive system.

Vascular intimal dysplasia and remodeling are characteristic features of injury after percutaneous transluminal coronary angioplasty (1) and in chronic allograft rejection (2, 3). The initial response to injury is inflammatory and involves the attraction of lymphocytes, macrophages, and thrombocytes to the site of injury and the secretion of cytokines, eicosanoids, and growth factors (4). Under the influence of growth factors and cytokines, smooth muscle cells (SMCs) proliferate and migrate into the intima and contribute to intimal hyperplasia and vascular stenosis.

Estrogen has several protective effects on the vascular wall. Some of these are rapid, presumably direct membrane effects whereas others require transcriptional activation of genes (5, 6). The inhibitory effect of estrogen on the replication, migration, and extracellular matrix deposition by vascular SMC, the key events in vascular fibroproliferative dysplasias, is presumably a genomic effect mediated through a variety of mechanisms, including regulation of several growth factors and/or their receptors and possibly by a direct antiproliferative effect of estrogen on SMC (5, 7).

The development of vasculoprotective drug therapies based on the protective effect of estrogen has been difficult, as it has not been possible to differentiate the desired vasculoprotective effect of estrogen from its effects on the reproductive system. In this communication, we demonstrate that this is possible by preferential targeting to the novel estrogen receptor β (ERβ).

MATERIALS AND METHODS

Carotid Denudation Injury.

Carotid denudations were made to Wistar rats purchased from the Laboratory Animal Center (University of Helsinki). The details of the operation have been described (8). The rats were killed before the operation (day 0) and at 15 min and 3, 7, 14, and 30 days after the denudation injury.

Treatment of Ovariectomized Rats with 17β-Estradiol (17β-E₂) and Genistein.

Female rats were ovariectomized and placed on soy-bean free diet (Special Diet Services, Essex, U.K.) for 7 days, and both carotids were denuded. One group of animals received 17β-E₂ (Sigma), and the other group received genistein (kindly donated by William Helferich, Michigan State University, or purchased from Plantech, (York, U.K.) at decreasing doses from 2.5 mg/kg/day s.c. downwards whereas the third group received vehicle only and served as control. 17β-E₂ and genistein were dissolved in dimethylsulfoxide (Sigma). All animals were weighed daily, and both drugs were administered s.c., using the following doses: 2.5, 0.25, 0.025, and 0.0025 mg/kg in one s.c. injection per day. The animals were killed 7 days after injury, the uterus and left carotid were removed, the uterus was weighed, and both organs were processed for histology.

Histological Specimens, Stainings, and Quantitation of the Response.

Evaluation of vascular histological changes was made from midcarotid sections at day 0 and at 15 min and 3, 7, 14, and 30 days after the denudation injury. The carotids were removed “en block,” were fixed in 3% paraformaldehyde solution for 4 h, were transferred to saline, and were processed for paraffin embedding. Thin sections (2 μm thick) were stained with Mayer’s hematoxylin-eosin, and the number of cell nuclei in the adventitia, media and intima was quantitated by using 400× magnification. The result was expressed as the number of nuclei per vessel circumference. For uterus histology, 2-μm-thick sections were processed as above, and specimens from nontreated ovariectomized rats were compared blind to specimens from ovariectomized rats receiving 2.5 mg/kg/day of genistein or 17β-E₂ for 7 days (3–5 specimens of each type).

BrdUrd Labeling.

An aqueous solution of 5-bromo-2′-deoxyuridine and 5-fluoro-2′deoxyuridine (Zymed) was used for labeling of proliferating cells after denudation. For “pulse labeling,” a dose of 300 μl of labeling suspension was injected i.v. according to manufacturer’s instructions, the rats were killed exactly 3 h after the pulse, and the specimens were processed as described (8).

Probe Preparation.

The complementary RNA probes were synthesized according to manufacturer’s directions (Promega) in the presence of ³⁵S-UTP (Amersham Pharmacia) by using the following cDNA fragments as templates. For ERβ, a 400-bp EcoRI-AccI fragment (from the 5′ untranslated region, starting nucleotide sequence GAATTC, ending nucleotide CTACGT) of the rat ERβ cDNA subcloned to pBluescript II KS (+) (Stratagene) vector was linearized with EcoRI or AccI enzymes for the production of antisense and sense transcripts, respectively. For ERα, a 200-bp BstxI-EcoRI fragment (from the 3′ untranslated region, F-domain region, starting nucleotide sequence ATGGGA, ending nucleotide sequence TAGCAG) of the rat ERα cDNA subcloned to pBluescript II KS (+) vector was linearized with SacI or EcoRI enzymes before synthesis of antisense and sense probes, respectively. RNA probes transcripted from opposite strands of the same plasmid template were adjusted to the same specific radioactivity.

In Situ Hybridization.

Carotids were fixed in 3% paraformaldehyde solution for 4 h and were paraffin-embedded. To ensure that the different specimens were comparable, tissue specimens of different time points were placed on organosilanized microscopy slide were and hybridized in identical conditions, either for ERα or ERβ. After deparaffinization and rehydration, the sections were denatured in 0.2 M HCl, were heat-denatured in 2× standard saline citrate (SSC) at +70°C, and were treated with proteinase K (1 μg/ml). The sections then were postfixed with 4% paraformaldehyde, were acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine and were dehydrated and air-dried. The slides were hybridized with antisense or sense RNA probes overnight at 60°C, were washed in 4× SSC, were treated with RNase A (20 μg/ml), and were washed sequentially in SSC solutions with 1 mM DTT. Finally, the slides were rinsed in 0.1× SSC with 1 mM DTT, were dehydrated in graded ethanols, and were air-dried. The slides were dipped into the autoradiography emulsion (NBT-3, Eastman Kodak), were exposed for 7–14 days, and were developed, counterstained, dehydrated, and mounted with Permount. Control carotids were compared with denuded carotids removed at 15 min and 3, 7, 14, and 30 days after denudation.

Immunohistochemical Staining of ER Subtypes.

Frozen sections from male rat carotids at day 0 and at 15 min and 3, 7, and 14 days after denudation injury (five animals per each time point) and SMCs cultured on coverslips were stained for ERβ (rabbit antibody PAI-311 to N-terminal amino acid residues 55–70 of rat and mouse ERβ; Affinity BioReagents, Golden, CO) and ERα (mouse monoclonal antibody to recombinant ERα protein, Novocastra Laboratories, Balliol Business Park, Newcastle upon Tyne, U.K.; and Dako monoclonal antibody to a 67-kDa sequence in the A/B region, M7047, Dako A/S).

Air-dried serial sections (4–6 μm) or cultured SMCs were fixed in acetone at −20°C for 20 or 10 min, respectively, and were stored at −20°C until used. Before staining, the slides were refixed with chloroform and were air-dried. After incubation with 1.5% normal goat serum (Vector Laboratories, Burlingame, CA), frozen sections were incubated with PAI-311 antibody (1:200, in Tris-buffered saline, 0.1% BSA, and 2% goat serum) at +4°C for 12 h, were washed with Tris-saline, were incubated with biotinylated goat anti-rabbit antibody at room temperature for 30 min, and were washed again; avidin-peroxidase complex (Vectastain Elite ABC Kit, Vector Laboratories) was added for 30 min at room temperature. The color reaction was developed by 3-amino-9-ethylcarbazole (Sigma) and 0.1% hydrogen peroxide. The specimens were counterstained with hematoxylin, and coverslips were mounted (Aquamount, BDH). The intensity of immunohistochemical staining was scored as percent of stained cells of all cells.

In case of monoclonal primary antibody, the slides were incubated with 1.5% normal horse serum (Vector Laboratories) and thereafter with a mouse monoclonal antibody (1:50, diluted in Tris-buffered saline, 0.1 BSA, and 2% horse serum) at +4°C for 12 h. The subsequent steps were performed as above, except that biotinylated horse anti-mouse antibody was used as secondary antibody.

Specificity Controls of Immunostaining.

To establish the specificity of the antibodies to rat ER subtypes, their reactivity to rat uterus and bladder specimens was tested, the former known to express primarily ERα and the latter, ERβ (9). Additional controls for anti-ERβ antibody staining were performed by using 10- to 20-fold molar excess of the neutralizing synthetic peptide to PAI-311 (Affinity BioReagents) and omitting primary antibody. These control stainings showed no immunoreactivity.

In Vitro Cultures.

Adult rat aorta SMCs at the 10–13th passage were used for the in vitro studies. The cells were seeded on 96-well plates (5,000 cells/well) in DMEM (Euroclone, Pero, Italy) supplemented with 5% fetal calf serum on day −2. The next day the growth medium was changed to serum-free and phenol red-free DMEM (Euroclone) containing 0.5% BSA. Platelet-derived growth factor-B (PDGF-B) (20 ng/ml, Upstate Biotechnology, Lake Placid, NY) and serially diluted genistein and 17β-E₂ were added together with ³H-TdR (Amersham Pharmacia) at day 0, and incubation continued for 24 h. Incorporated radioactivity was measured by 10% trichloroacetic acid precipitation and was counted by using a liquid scintillation counter (RackBeta, Wallac, Turku, Finland).

The migration of SMCs was quantitated by using a Transwell culture chamber (Costar) in which the upper and lower culture chamber are separated by a polycarbonate filter with 8-μm pores. The chambers were first coated with collagen (Rat Tail Collagen, Type 1, Upstate Biotechnology) at a concentration of 20 μg/ml for 24 h at +4°C. Primary adult rat aortic SMC (50 μl, 1 million cells/ml) were seeded in the upper chamber. PDGF-B (60 ng/ml) (Upstate Biotechnology) was added to the lower chamber in DMEM supplemented with 0.5% FCS and 0.1% BSA. Genistein or 17β-E₂ were added to the upper chamber, and, after 24 h, the filters were removed, were fixed with methanol, and were stained with Mayers hematoxylin and eosin. Migrated cells on the lower side of the filter were quantitated by counting specified cross-sectional fields with a light microscope using 400× magnification.

For immunohistochemical stainings, SMCs (13th passage) were grown until 70% confluency on glass slides in DMEM supplemented with 10% FCS. Cells then were cultured in serum-free media for 24 h and were stimulated with 10% FCS or were left unstimulated for another 24 h.

Statistical Analysis.

The linear regression analysis (10, 11) was applied to evaluate the significances of changes in the number of intimal nuclei and labeled nuclei, as well as the weight of the uterus between the genistein and estrogen treated animals. Here, statistical difference between regression coefficients (slopes) of linear plots have been calculated. P values <0.05 were regarded as statistically significant.

RESULTS

Expression and Localization of ERα and ERβ mRNA After Carotid Denudation Trauma in Male Rats.

The intensity of expression and localization of ERα and ERβ mRNA were investigated by in situ hybridization from paraformaldehyde-fixed paraffin-embedded specimens (Fig. 1). Three control nondenuded carotids were compared with denuded carotids removed at 15 min and at 3, 7, 14, and 28 days after denudation, three at each time point.

Figure 1.

Expression of ERα (A, B, E, and F) and ERβ (C, D, G, and H) mRNA in male rat carotid arteries 15 min (A, B, C, and D) and 7 days (E, F, G, and H) post-denudation. Bright-field images are shown in the left panels, and the corresponding dark field images are shown in the right panels. Lumen (Lu) is facing up. Note the strongly enhanced expression of ERβ in the media (Med) and particularly in the vascular intima (Int). (×400.)

ERα and ERβ mRNAs were expressed at a low level in the vascular tunica media, and the level of expression of ERα mRNA remained unaltered throughout the experiment (Table 1 and Fig. 1). Three-fold up-regulation of ERβ mRNA first was observed 3 days after denudation in the media, and the level of expression in the media increased to 8-fold on day 7, after which it declined. Even more prominent changes in the expression levels were observed in the hyperplastic intima/neointima. Although the ERα expression in the intima increased only slightly, the level of ERβ expression increased nearly 40-fold on day 7 (Table 1 and Fig. 1), after which it declined but remained elevated even after 28 days post-injury.

Table 1.

Expression of ERα and ERβ mRNA in denuded male rat carotid at different time points after injury

Time after injury	ERα		ERβ
	Neointima	Media	Neointima	Media
	Number of grains per 400 μm2178
Control	n.a.	3.43  ±  1.21	n.a.	7.13  ±  3.19
15 min	n.a.	3.75  ±  0.8	n.a.	8.25  ±  1.98
3 days	n.a.	5.75  ±  1.38	n.a.	28.5  ±  3.93
7 days	11.0  ±  1.87	5.25  ±  0.48	303.0  ±  19.8	79.0  ±  7.72
14 days	5.5  ±  1.04	4.75  ±  1.11	68.0  ±  5.42	42.0  ±  8.87
28 days	9.81  ±  2.73	6.73  ±  1.45	87.1  ±  6.31	57.3  ±  7.11

n.a., not applicable (no neointima present). 

¹⁷⁸Values are expressed as means from three rats per time point and three countings of each specimen ± SEM. 

In situ hybridization demonstrated, furthermore, that the enhanced expression of ERβ mRNA nearly completely colocalized during the vascular proliferative response with the SMCs in the media and in the neointima (Fig. 1). More than 95% of the cells in the neointima were identified as α-actin-positive SMCs with some leukocyte common antigen positive inflammatory cells on the luminal side, as repopulation of endothelium in this model does not occur until after day 14 (data not shown).

Expression and Localization of ERβ Protein in the Denuded Carotid.

Male rat carotids were denuded as above, the animals were killed at the same time points post-injury, and the mid-carotid sections from five animals per time point were investigated. Nondenuded carotids served as controls. At 0 h and 15 min post-denudation, only occasional cells in the media (<0.5%) were stained with PAI-311 antibody, and no positive cells were observed in the adventitia. At 7 and 14 days post-denudation, virtually all cells in the intima, 40–65% of the cells in the media—>90% of which were α-actin-positive SMCs (8)—but only occasional cells in the adventitia (<2%) were stained with this antibody. The staining was entirely abolished by preincubation of the antibody with the peptide used for immunization and by omitting the primary antibody (Fig. 2). Staining of parallel specimens with Novocastra monoclonal antibody to human ERα showed no reaction, although controls of uterine tissue were positive.

Figure 2.

Immunostaining of ERβ protein in carotid tissue 15 min post-denudation (A), 7 days post-denudation (B), 14 days post-denudation (C), and 7 days post-denudation after blocking with neutralizing peptide (D). Arrows indicate internal and external elastic laminae. (×400.)

Dose Responses to 17β-E₂ and Genistein on Post-Denudation Carotid Trauma and on Uterine Weight in Ovariectomized Female Rats.

Female adult rats were ovariectomized on day −7 and were placed on phytoestrogen-deficient diet, and carotid denudation was performed on day 0. Ten denuded carotids of rats receiving only vehicle (DMSO 200 μl/kg/day) were compared with denuded carotids of rats receiving 17β-estradiol (17β-E₂) or genistein at escalating doses of 0.0025, 0.025, 0.25, and 2.5 mg/kg/day, three to five carotids at each dose level.

Both 17β-E₂ and genistein had a dose-dependent effect on the nuclear number in the intima (Fig. 3) but no measurable effect on the number of nuclei in the media (data not shown). Genistein was equally efficacious compared with 17β-E₂ in its inhibitory effect in the intimal nuclear number (r² = 0.946 vs. 0.746, respectively; P = 0.5). In addition, both ligands reduced dose-dependently and to the same extent as the replication rate of SMC in the intima, as quantitated after pulse labeling by the number of BrdUrd-incorporating cells in histological sections (r² = 0.993 and 0.630, respectively; P = 0.3).

Figure 3.

Dose responses of 0.0025–2.5 mg/kg/day of 17β-E₂ (open circles) and genistein (closed circles) on the nuclear number in the vascular intima 7 days after denudation injury of carotid arteries in ovariectomized rats. The figures indicate the number of nuclei per vessel circumference, means of three to five determinations per point ± SEM. Statistical significance between regression coefficients (slopes) is nonsignificant (P = 0.5).

In the dose area used, only 17β-E₂ displayed a dose-dependent stimulatory effect on uterine weight whereas genistein had no effect (r² = 0.954 vs. 0.096, respectively; P = 0.0003) (Fig. 4). Moreover, at the dose of 2.5 mg/kg/day, only 17β-E₂ but not genistein induced typical estrogen-related morphological changes in the uterus. After 7 days of 17β-E₂ treatment of ovariectomized animals on phytoestrogen-deficient diet, uterine histology was consistent with estrogen stimulation: i.e., hyperplastic luminal epithelium, thickening of stromal layer, and some stromal inflammatory reaction. These changes did not occur in the uteri after the 7-day genistein treatment or in control animals (Fig. 5). In genistein-treated animals, the luminal epithelium was slightly taller and had retained its single-layer columnar structure, compared with the thin atrophic epithelium in the control animals (Fig. 5).

Figure 4.

Dose responses of 0.0025–2.5 mg/kg/day of 17β-E₂ (open circles) and genistein (closed circles) on the uterine weight 7 days after denudation injury of carotid arteries in ovarictomized rats. Values indicate means of three to five determinations per point ± SEM. Statistical significance between regression coefficients (slopes) is significant (P = 0.0003).

Figure 5.

Effects of 2.5 mg/kg/day 17β-E₂ and genistein on uterine histology after 7-day treatment of ovariectomized rats. Note hyperplastic luminal epithelium in the uterus of rats receiving 17β-E₂ (B) that was not observed in the uteri of nontreated rats (A) or in rats receiving genistein (C). Hematoxylin-eosin. (×400.)

Dose-Responses to 17β-E₂ and Genistein on Vascular SMC Proliferation and Migration in Vitro.

Finally, the antiproliferative and antimigratory effects of genistein vs. 17β-E₂ were confirmed in vitro with cultured vascular SMC at the 10–15th passage. For proliferation, primary adult rat aortic SMC, >90% of which expressed SMC α-actin, were plated in culture on day −2 of the experiment, were serum-starved for 2 days, and were stimulated on day 0 with 20 μg/ml PDGF-B, the strongest known growth stimulating factor for SMC, or were left unstimulated. The estrogenic compounds were added to the cultures on day 0. The cultures, made in triplicate, were pulsed with ³H-TdR for 24 h and were harvested on day 1. The presence of ER proteins in the cultured SMCs was confirmed by immunohistochemical staining. Of serum-starved and FCS-stimulated cells, 80% expressed ERβ while only 1% of serum-starved cells and none of the FCS-stimulated cells were ERα-positive.

One representative experiment of three with similar results is shown in the upper panel of Fig. 6. At 3–30 nM concentration area, both 17β-E₂ and genistein seem to induce an additional 5–10% increase in the ³H-TdR incorporation rate (=DNA synthesis) of SMCs, followed by inhibition of proliferation. Visual inspection of the ³H-TdR incorporation rates in a semilogarithmic plot suggests that they are composed of two slopes. Within 3 nM to 2 μM area, the inhibition of DNA synthesis was weak, ≈10–20% of maximal incorporation, whereas, within 2–50 μM area, it was more pronounced (Fig. 6).

Figure 6.

(Upper) Effect of 17β-E₂ and genistein on proliferation of adult rat vascular SMCs to 20 ng/ml PDGF-B as quantitated by ³H-TdR incorporation. Shown is one representative experiment of three. Triplicate determinations are ±SD. (Lower) The effect of 17β-E2 and genistein on the migration of rat vascular SMCs. Shown is one representative experiment of two. The cells were plated on the upper chambers with 17β-E2 or genistein and were stimulated with 60 ng/ml of PDGF-BB in the lower chamber. The effect of the estrogenic compounds on migration rate (number of cells in the lower surface of the filter) was quantitated after 24 h with a light microscope using 400× magnification.

Alternatively, primary rat aortic SMCs at the 10–15th passage were plated on collagen-coated Transwell migration chambers. Migration was induced by 60 ng/ml PDGF-B in the lower chamber, and the estrogenic compounds were added to the upper chamber. The experiments were repeated once with a similar result. Both 17β-E₂ and genistein showed a dose-dependent inhibitory effect on SMC migration at the dose range from 0.003 μM to 50 μM. There was no difference in the rate of inhibition of SMC replication by 17β-E₂ and genistein in either one of these two experiments.

DISCUSSION

The vasculoprotective effect of estrogen was first shown in population studies in humans, where estrogen replacement therapy demonstrated a protective effect on atherosclerotic vascular disease in postmenopausal women, as later confirmed in ovariectomized monkeys (12–14). The vasculoprotective effect has since been documented more in detail in animal models and in vitro. Estrogen has been found to inhibit the intimal thickening after mechanical carotid balloon injury in rabbits, rats, and mice (15–17), as well as the immunologically induced vascular fibroproliferative dysplasia in rabbit aorta and heart allografts (18, 19). In vitro, it has been demonstrated that estrogen inhibits migration and replication of vascular SMCs (20–23). These observations are consistent with the findings in reporter gene assays that functional ERs are expressed in the vascular SMCs of bovine, rat, guinea pig, and human origin (24–28).

Both the “classical” estrogen receptor ERα and the novel ERβ have been detected in arteries and in cultured vascular SMC of cynomolgus monkeys as well as in rat arteries (29, 30). The expression of ERα has been reported to increase in rabbit cardiac allografts (31) whereas up-regulation of ERβ expression, but not of ERα, was detected in rat carotid artery after denudation injury (30). The finding on the vasculoprotective effect of estrogen in ERα-deficient mice (32), further suggests that ERα may not be solely responsible for the vasculoprotective effect of estrogen. In this study, we have investigated the specific role of ERβ in the prevention of vascular intimal hyperplasia after denudation injury of rat carotid artery.

Using a well established model of vascular injury previously validated in our laboratory (8, 33, 34), we have demonstrated that ERβ mRNA and protein are strongly up-regulated in the vascular wall as a consequence of injury whereas ERα remains expressed at a constitutive, low background level only. In situ hybridization and immunostainings demonstrate that ERβ mRNA and protein colocalize in the vascular SMCs in the media and neointima.

To investigate the specific role of ERβ vs. ERα in vasculoprotection, ERβ-specific ligands would be needed. Unfortunately, compounds that bind specifically to ERβ and do not interact with ERα are not available at present. Several steroidal and nonsteroidal compounds have been tested for the binding to the two ERs, and the compound showing the highest binding affinity to ERβ compared with its affinity to ERα is the isoflavone genistein. Genistein binds to ERβ almost as well as 17β-E₂ (RBAs 87 and 100, respectively) whereas its affinity to ERα is considerably lower than that of 17β-E₂ (RBAs 4 and 100, respectively) (35). With these two different ligands with ≈20-fold affinity difference for ERβ vs. ERα, we have clearly differentiated the vasculoprotective vs. uterotrophic effect. Within the dose range tested, both ligands demonstrated a similar dose-dependent vasculoprotective effect in vivo and a dose-dependent inhibitory effect on SMC replication and migration in response to PDGF in vitro, but only 17β-E₂, but not genistein, demonstrated a dose-dependent uterotrophic effect in vivo.

It also may be argued that genistein does not function here as an estrogen but exerts its effect via inhibitory action on protein tyrosine kinases. In numerous in vitro studies with vascular SMC, genistein has been shown to block several tyrosine kinase-dependent events that are critical in the regulation of SMC proliferation and contractility. These include inhibition of prostacyclin production, angiotensin II-induced Ras activation, and fibroblast growth factor 2 expression (36–38), contractions induced by phorbol ester, endothelin, angiotensin, and serotonin (39), serotonin-evoked increase in intracellular Ca²⁺ (40), and PDGF-induced SMC proliferation (41). Tyrosine kinase inhibition, however, requires rather high concentrations of genistein, >10 μM, whereas already a 100 nM concentration is sufficient to exert significant ER-mediated effects (35, 42). We believe that, in our study and in used doses, tyrosine kinase inhibition is not likely to play any role. An inhibitory effect on SMC replication and migration was observed already in concentrations equal to or less than 10 μM in vitro. It is also highly unlikely that genistein concentrations >10 μM were achieved in our in vivo experiment. In female rats receiving dietary genistein (750 μg/g of diet), which results in the daily dose of ≈20 mg/kg of body weight, the concentration in serum was only 2.2 μM (43).

Genistein is considered to be a weak estrogen, and the concentrations or doses of genistein required to produce an effect in the most commonly used target cells or organs, such as MCF-7 cells in vitro or rodent uterus in vivo, are 1,000- or 10,000-fold higher than those of 17β-E₂ (42, 44). However, in the present study, using ovariectomized female rats, genistein was as effective as 17β-E₂ in protecting the arterial wall against post-denudation intimal dysplasia. This apparent “discrepancy” may be explained by different expression levels of ER subtypes in different tissues and particularly after vascular trauma. The main estrogen receptor subtype in the uterus is ERα (9), where genistein has only low affinity. This is in clear contrast to the vascular wall after endothelial trauma, where ERβ mRNA and protein were both up-regulated as consequence of injury and colocalized with the SMCs of the media and neointima during the replicative and migratory bursts. In contrast, the expression of ERα remained essentially unchanged and at a low level as already reported by Lindner et al. in their mRNA studies (30). The >40-fold up-regulation of ERβ expression in the SMCs, compared with unchanged expression level of ERα in the vessel wall after injury, may thus allow preferential targeting of genistein to proliferating SMCs. This is further supported by the equal effects of the two ligands on the cultured SMCs, the majority of which express ERβ but not ERα. The only other tissue in which genistein has been reported to be active at the same concentration range as 17β-E₂ is male mouse urethroprostatic complex (45), rodent prostate also being a tissue in which ERβ is the predominant ER subtype (9). These findings are entirely compatible with our interpretation.

In conclusion, the present study provides further evidence that estrogens are potent vasculoprotective agents and that they regulate the proliferation of SMC in the vascular wall. Our results suggest that ERβ plays a pivotal role in vasculoprotection after intimal injury and may be the main receptor to mediate the effects of estrogen in the vascular wall. Furthermore, our findings indicate that ligands that bind preferentially to ERβ may be good candidates in the search for vasculoselective estrogens.

ABBREVIATIONS

ER

estrogen receptor

17β-E₂

17-estradiol

SMC

(vascular) smooth muscle cell

PDGF-B

platelet-derived growth factor-B