Our purpose was to determine the effects of gender and exercise training on endothelial nitric oxide synthase (eNOS) and superoxide dismutase (SOD) protein content of porcine skeletal muscle arteries and to evaluate the role of 17β-estradiol (E2) in these effects. We measured eNOS and SOD content with immunoblots and immunohistochemistry in femoral and brachial arteries of trained and sedentary male and female pigs and measured estrogen receptor (ER) mRNA and α-ER and β-ER protein in aortas of male and female pigs. Results indicate that female arteries contain more eNOS than male arteries and that exercise training increases eNOS content independent of gender. Male and female pigs expressed similar levels of α-ER mRNA and protein and similar amounts β-ER protein in their arteries. E2 concentrations as measured by RIA were 180 ± 34 pg/ml in male sera and ∼5 pg/ml in female sera, and neither was changed by training. However, bioassay indicated that biologically active estrogen equivalent to only 35 ± 5 pg/ml was present in male sera. E2 in female pigs, whether measured by RIA or bioassay, was ∼24 pg/ml at peak estrous and 2 pg/ml on day 5 diestrus. The free fraction of E2 in sera did not explain the low measurements, relative to RIA, of E2. We conclude that 1) gender has significant influence on eNOS and SOD content of porcine skeletal muscle arteries; 2) the effects of gender and exercise training vary among arteries of different anatomic origin; 3) male sera contains compounds that cause RIA to overestimate circulating estrogenic activity; and 4) relative to human men, the male pig is not biologically estrogenized by high levels of E2 reported by RIA, whereas in female pigs E2 levels are lower than in the blood of human women.
- skeletal muscle blood flow
- vascular smooth muscle
- nitric oxide
- estradiol-free fraction
- biological activity
- endothelial nitric oxide synthase
men have a higher incidence of atherosclerosis and coronary heart disease than do premenopausal women of similar age (40). After menopause, the incidence of cardiovascular disease increases in women so that there is no longer a gender difference (30, 43). Present evidence suggests that female hormones mediate these beneficial effects, but which hormones are most important has not been established. These beneficial effects of the female gender are associated with modification of reactivity of arteries due to differences in vascular smooth muscle and endothelial and adventitial cells in the arteries (34). It has been proposed that the influence of estrogens on endothelial cell phenotype and function plays an important role in the protective effects of female gender on vascular disease (18, 40). Taddei et al. (53) proposed that protective female gender effects may abate with menopause because endothelium dysfunction is produced by endogenous estrogen deficiency. Consistent with the proposal of a role for estrogen in normal endothelial phenotype are recent reports that estrogen increases endothelial cell nitric oxide synthase (eNOS) gene expression (5, 24, 32), basal release of nitric oxide (NO) by vascular cells (18, 19), and endothelium-dependent vasodilation (5, 18). Also, the hypothesis that plasma estrogen can influence endothelial function is supported by recent reports of improved vascular function (39) and acetylcholine (Ach)-induced, endothelium-mediated vasodilation in male to female transsexuals treated with estrogen therapy (38).
In view of these results, we were intrigued by an observation that brachial arteries of female pigs were less responsive to endothelium-dependent relaxation effects of bradykinin (BK) and Ach than brachial arteries from male pigs (29). These responses were indeed endothelium-dependent, because removal of the endothelium abolished relaxation responses of the brachial arteries to both BK and Ach. Another interesting observation in these experiments was that inhibition of NO synthase (NOS) by treatment with arginine analogs (NG-nitro-l-arginine methyl ester) revealed that NOS contributes fractionally more to endothelium-dependent relaxation in brachial arteries from female pigs than in male brachial arteries. Finally, we reported that male pigs had greater plasma concentrations of 17β-estradiol (E2) than female pigs (29). These concentrations of E2 suggest the possibility that brachial arteries of male pigs exhibit greater endothelium-dependent relaxation than female brachial arteries because males have higher plasma concentrations of E2. However, two other sets of data argue against this postulate. First, male pigs were phenotypic boars and exhibited no evidence of being estrogenized, such as mammary development. Second, endothelium-dependent relaxations are greater in coronary arteries (2) and femoral arteries (29) of female pigs than of male pigs. The purpose of the study reported herein was to further investigate determinants of gender differences in endothelium-dependent relaxation of porcine femoral and brachial arteries and interactions of gender with exercise training effects on expression of eNOS in these arteries. Because SOD can increase bioavailability of NO and because previous studies indicate that exercise training modulates SOD protein levels in aorta and coronary arterioles of pigs (46, 47), we also examined effects of gender and exercise on SOD levels of femoral and brachial arteries.
We conducted five experiments. First, we measured eNOS and SOD protein content of femoral and brachial arteries isolated from male and female pigs by using immunoblots and immunohistochemistry. Our hypothesis was that endothelial cells of female arteries express more eNOS and/or SOD than male endothelial cells. Second, we measured estrogen receptor (ER) mRNA in aortic endothelial cells of male and female pigs and ER protein in aortas of male and female pigs to test the hypothesis that male arteries are less responsive to circulating E2 because they express fewer or no ER receptors. Third, we measured eNOS and SOD protein content in femoral and brachial arteries of trained male and female pigs to determine the role of altered eNOS and/or SOD expression in the training-induced changes. Fourth, we tested the hypothesis that only a small fraction of E2 detected by RIA in the blood of male pigs is free, unbound to serum-binding proteins. Finally, because we found that free fractions of E2 did not explain how the male pigs could show high levels of E2 in their blood without overt estrogenization of phenotype, we used a sensitive bioassay to measure biological estrogenic activity in male and female blood to test the hypothesis that E2 content, as measured by RIA, overestimates estrogenic activity in male pig blood. Our results appeared to confirm this hypothesis and suggested that not all of the E2 measured in male pig serum by RIA is biologically active.
Results indicate that conduit arteries of female pigs have a greater eNOS content than do the same arteries of male pigs and that exercise training increased eNOS protein content in both male and female arteries. ER receptor content of aortic tissues appear to be similar in samples from males and females. Interestingly, we observed that total estrogenic activity present in male pig serum was lower, and sometimes much lower, than the E2 content indicated by RIA, whereas E2-free fractions in male and female blood were similar in magnitude, although statistically significantly different.
Experiments were completed on adult male (n = 73) and female (n = 72) Yucatan miniature swine (Charles River) weighing 25–40 kg obtained from the breeder. Animals were procured in lots of 8 or 16 animals and familiarized with treadmill exercise over a 1- to 2-wk period of time. Treadmill performance tests were administered to each animal to evaluate exercise tolerance. Each lot of pigs was then randomly divided into two groups: exercise trained (Ex) or sedentary (Sed). Ex pigs underwent a progressive treadmill training program, which lasted 13–21 wk, that in our hands produces adaptations in miniature swine that are classically associated with the Ex state in all mammals (26–29). Sed pigs were restricted to their pens (6 × 12 ft.) for 13–21 wk. Pigs were procured at 9–10 mo of age and were 13–14 mo of age at the termination of experiments. All experimental procedures involving animals were approved by the University of Missouri Animal Care and Use Committee in accordance with the “Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research, and Training.”
Training program. Ex pigs trained on the treadmill 5 days/wk and were given positive reinforcement for exercise by feeding them after each training bout. The speed and duration of daily running were progressively increased over the first 8–12 wk at a rate dependent on the tolerance of each pig. A typical training session during week 12 consisted of the following 85-min workout: 1) a 5-min warm-up run at 2.5 miles/h (mph); 2) a 15-min sprint at speeds of 5–8 mph; 3) a 60-min endurance run at 4–5 mph; and 4) a 5-min cool-down run at 2 mph. Ranges of running speed are presented because the Ex program was customized to each pig's exercise ability. Treadmill performance tests were administered to the Sed and Ex pigs before initiation of training and at the completion of Sed or Ex periods.
Treadmill performance test. The performance test consisted of four stages of exercise (26–29). During stage 1, pigs ran at 3.1 mph, 0% grade, for 5 min. Pigs ran for 10 min at stage 2 (speed = 3.1 mph, grade = 10%) and then for 10 min at stage 3 (speed = 4.3 mph, grade = 10%). Finally, pigs ran at stage 4 (speed = 6 mph, grade = 10%) until the pigs could not maintain treadmill speed. Heart rates were recorded throughout the treadmill performance test along with total duration of exercise.
Efficacy of training. The effectiveness of the training program was determined by comparing the exercise tolerance (as reflected in the treadmill performance test), heart and body weights, and skeletal muscle oxidative capacity of Ex and Sed groups. At the time of death, samples were taken from the middle of the lateral and long head of triceps brachii and the deltoid muscles and stored at -70°C until processed. Citrate synthase activity was measured from whole muscle homogenate by using the spectrophotometric method of Srere (52).
Estrous cycles were measured by heat checking 10 female pigs on a daily basis with standard procedures using a male pig to determine when female pigs were receptive. Female pigs were exposed individually to a Yucatan boar daily and considered in heat when they responded to the presence of the boar (4). After estrous cycle duration in each of the 10 female pigs was established, they were instrumented with vascular access ports so that we could obtain blood samples at eight defined times during the estrous cycle (1, 41).
Implanting vascular access port. Pigs were sedated with an intramuscular injection of Telazol7 (5.0 mg/kg), Xylazine7 (2.2 mg/kg), and atropine (0.05 mg/kg), intubated, and placed under isoflurane anesthesia. With the use of aseptic techniques, the right jugular vein was catheterized and the tip of the catheter positioned cranial to the right atrium with the aid of a fluoroscope (1, 41). Catheters were routed subcutaneously to the left side of the neck where a subcutaneous pocket was made for the vascular access port. After implantation access ports and catheters were filled with a 50% dextrose/saline solution (2.8 ml) containing 200 U/ml of heparin and 1 mg/ml Vancomycin to maintain patency and prevent infection. Penicillin G procaine (600 IU) was given intramuscularly once per day for 3 days postsurgery.
Previous heat checking demonstrated an average 20-day estrous cycle for these pigs; therefore, blood samples were collected on days 17–20, 0–3, and 5 (day 20/day 0 = day of peak serum estradiol levels) to measure changing levels of estrogen. Ten milliliters of blood were collected, centrifuged, and stored (-80°C) before analysis.
Blood sampling regimen. In the male and female pigs not instrumented with vascular access ports, jugular vein blood samples were obtained via venipuncture and collected in vacutainers containing EDTA at the conclusion of training or cage confinement. Samples were collected after a 12-h fast. Plasma was separated by centrifugation (Beckman TJ-6R centrifuge, Palo Alto, CA) at 4°C for 15 min at 3,750 rpm. Plasma was stored at -70°C until assay.
Tissue sampling. After completion of Ex or Sed confinement and ∼24 h after the last exercise bout, pigs were sedated with ketamine (35 mg/kg; Fort Dodge) and Rompun (2.25 mg/kg; Bayer), anesthetized with thiopental (10 mg/kg; Abbott Labs), and euthanized by removal of the heart in full compliance with American Veterinary Medical Association Panel on Euthanasia guidelines. Segments of femoral and brachial artery of ∼2- to 3-mm outer diameter and 3–5 mm in length were carefully removed and trimmed of fat and connective tissue, frozen in liquid N2, and stored at -80°C until processed. Artery samples were taken from the same locations of male and female Sed and Ex pigs. In addition, a 15-cm segment of thoracic aorta was isolated, starting ∼3cm distal to the end of the arch of the aorta, and placed in Krebs bicarbonate buffer (4°C), previously aerated with 95% O2-5% CO2 gas mixture. Aortic endothelial cell lysates were collected as previously described (47).
RT-PCR using aortic endothelial cell samples. First-strand cDNA synthesis was initiated by using 1 μg of total aortic endothelial cell RNA and an oligo(dT)12–18 primer (Super-Script preamplification system, GIBCO-BRL) in a total volume of 20 μl.A5-μl aliquot of cDNA product was used in PCR in a total reaction volume of 50 μl, containing (in mM) 50 KCL, 20 Tris · HCl (pH 8.4), 4 MgCl2, 0.2 dNTP, and 0.2 forward and reverse primers, as well as 2.5 U Taq DNA polymerase (Promega). ER-α primers were based on a reported sequence for porcine ER-α (6) and were designed to amplify a 325-bp product. The primers were sense, 5′-CATGTTGCTGGCTACATCATCTCG-3′; anti-sense, 5′-CACCACGTTCTTGCACTTCATG-3′. PCR was initiated with a 5-min denaturation step (94°C) and a 5-min annealing step (63°C) followed by cycles of 72°C (elongation, 2 min), 94°C (1 min), and 63°C (1 min). The final step was a 10-min, 72°C elongation. PCR products were separated by electrophoresis (1.5% agarose gels) and visualized under ultraviolet light via ethidium bromide staining. Semiquantitative analysis was performed by scanning densitometry of photographic images (Polaroid) of ethidium bromide-stained gels. Preliminary experiments confirmed that, under these conditions, 30 cycles of PCR produced results that were in the linear range for product accumulation. The relationship between volume of the PCR preparation loaded on the agarose gel and the densitometric signal was determined, and an optimal loading volume of 3 μl was found for the conditions outlined above. In subsequent analyses, 3 μl of PCR preparation was loaded for all samples. To validate the densitometric method, we also performed semiquantitative analyses by radiolabeling experiments with 32P-labeled deoxycytidine triphosphate ([α-32P]dCTP) in PCR (data not shown) as we have detailed previously (46). The two methods yielded similar qualitative and quantitative results for group differences. Only the densitometric analysis data are illustrated in this manuscript.
Immunoblots. Aortic endothelial cell samples were prepared for immunoblots as previously described (47). Artery (brachial and femoral) samples were homogenized in protein solubilization buffer consisting of 50 mM Tris · HCl, pH 7.4, 6 M Urea, and 2% SDS by using a ground glass homogenizer. After a 2-h incubation at 45°C and centrifugation (10 000 g, 1 min), protein concentration of the supernatants was determined by using the bicinchoninic acid assay (Pierce). Samples were processed for immunoblot analysis as described previously (47). For gender, artery location, and exercise comparisons, equal numbers of samples from the treatment groups were loaded in the same gel for comparison.
Immunohistochemistry. Samples of aorta were dissected and immersed in 10% formalin for a minimum of 24 h. Rings 3 mm in length were processed routinely to paraffin embedment. Five-micron sections were cut with an automated microtome (Microm), floated onto positively charged slides (Fischer), and deparaffinized. The slides were steamed in citrate buffer at pH 6.0 (Dako target retrieval solution S1699) for 30 min to achieve antigen retrieval and then cooled for 30 min. The slides were stained manually with sequential Tris buffer and water wash steps performed after each protocol step. Sections were incubated with avidin biotin two-step blocking solution (Vector SP-2001) to inhibit background staining and in 3% hydrogen peroxide to inhibit endogenous peroxidase. Nonserum protein block (Dako X909) was applied to inhibit nonspecific protein binding. The primary antibodies utilized were mouse monoclonal antibody for eNOS (BD Transduction Laboratories) (21) and rabbit polyclonal antibodies for SOD1, SOD2 (Stressgen Biotechnology) (46), ER-α, and ER-β (Santa Cruz Biotechnology). All primary antibodies were diluted 1:800 and incubated with the tissue sections overnight at 4°C. After appropriate washing steps were completed, the sections were incubated with biotinylated anti-mouse or rabbit link secondary antibody in PBS containing 15 mM sodium azide and peroxidase-labeled streptavidin (Dako LSAB+ kit, peroxidase, K0690). Diaminobenzidine (Dako) applied for 5 min allowed visualization of primary antibody staining. Sections were counterstained with Mayer's hematoxylin stain for 1 min, dehydrated, and coverslipped. For negative controls, histological sections were prepared as described above, but incubation in primary antibody was excluded from the protocol. Sections were examined and photographed by using an Olympus BX40 photomicroscope.
Measurement of E2and biologically active E2
Serum E2 has traditionally been measured by RIA, but because of reports of extremely high E2 concentrations in male pig serum, we chose to use a variety of methods to determine not only total E2 present in serum but also the free, biologically active fraction of E2 and the effective concentration that was biologically active in bioassay.
RIA. We measured E2 in serum and plasma by RIA according to the method of Kirby et al. (23), validated for porcine serum by Laughlin et al. (29). Some samples were brought to pH 9 by adding NaHCO3 (60 mM) to inhibit solvolysis prior to ether extraction. Aliquots of 50–200 μl of serum or plasma were extracted twice with 3 ml of methyl tertiary butyl ether. Radiolabeled estradiol and the estradiol antibody were obtained from ICN Pharmaceuticals (Costa Mesa, CA). The intra-assay and interassay coefficients of variation were 4.9 and 13.4%, respectively. Sensitivity of the assay was 0.5 pg/ml, and recoveries from a spectrum of differing amounts of E2 were ∼88%. Parallelism between the standard curve and different volumes of a porcine serum quality control standard was not different (slopes = -1.99 ± 0.07 for standard curve and -2.22 ± 0.34 for porcine serum). Measurements of E2 in a test set of comparisons from eight female pigs, from which both plasma and serum were obtained at the same time, indicated statistically similar values for E2. E2 in serum was 4.4 ± 1.2 pg/ml and in plasma was 5.0 ± 1.0 pg/ml.
LH-20 chromatography-RIA. To ensure that high E2 concentrations measured by RIA were not due to contaminating estrone, which can cross react in E2 RIA, we first separated extracted serum or plasma on Sephadex LH-20 (Amersham Pharmacia, Piscataway, NJ), as described by Nagel et al. (36), by using 85% benzene:15% methanol as the elution solvent. Fractions collected from the column were dried before being assayed by RIA.
Ultrafiltration dialysis. The free fraction of E2 was measured in serum or plasma after addition of tritiated E2 by centrifugal ultrafiltration dialysis, according to the method of Montano et al. (35). The percentage of free hormone was calculated from relative amounts of labeled hormone present in the ultrafiltrate and in the total, nondialyzed samples. Measurements of the free fraction of E2 in a test set of comparisons from 18 animals from which both plasma and serum were obtained at the same time gave identical results. The free fraction in serum was 3.59 ± 0.10% and in plasma was 3.64 ± 0.17% (means ± SE; n = 18).
Bioassay of estrogenic activity. Serum was extracted with methyl-tertiary butyl ether as for RIA, and an aliquot of the same extract was also tested for estrogenic activity by bioassay using estrogen-responsive MCF-7 human breast cancer cells, with modifications to Grady et al. (16). Plating of the cells, preparation of dose-response series by serial dilution, feeding of the cells, and parts of the final DNA assay were all performed by robotics with a Tomtec Quadra 96. Cells were plated on day 0 in estrogen-free medium in 96-well plates. On day 1, the medium was replaced with fresh estrogen-free medium. Serum extracts were dried and resuspended in estrogen-free tissue culture medium, and then serially diluted to yield a series of serum-equivalent concentrations. Cells were fed with this medium (180 μl/well) for 4 days with daily medium changes on days 3–6. On day 7, cell growth was determined by DNA content/well, using the diphenylamine assay of Nataragan et al. (37), and absorbance was read on a Bio-Tek PowerWave-x 96-well plate reader. Serum estrogenic activity was determined against a standard dose-response curve of purified E2 (Sigma Chemical). High-performance liquid chromatography (HPLC) and bioassay separation of estrogenic activities was performed as described previously (16).
Data analysis. All values are means ± SE. Between-group differences were assessed by using repeated-measures ANOVA or Student's t-tests where appropriate. Differences with P < 0.05 were considered significant. Where between-group variances were not homogeneous, the data were log-transformed before ANOVA and post hoc tests; in each case, variance was homogeneous after the log transform.
Gender Effects on Endothelial Proteins in Sed Animals
Experimental design. To determine whether endothelial cells of female arteries express more eNOS and/or SOD than male endothelial cells, we measured eNOS and SOD protein content of femoral and brachial arteries isolated from male and female pigs. Then, because previous results indicated that Ex resulted in increased endothelium-dependent relaxation in female, but not male, brachial arteries (29), we also measured eNOS and SOD protein content in femoral and brachial arteries of Ex male and female pigs to determine the role of altered eNOS and SOD expression in the training-induced changes. In both of these experiments, protein content was assessed with immunoblots and immunohistochemistry. Results indicated that endothelial cells in arteries from female pigs appeared to have greater content of eNOS and SOD protein relative to arteries of male pigs just as do arteries of other female mammals described above. Of interest, anatomical location also appears to influence expression of these proteins.
Gender and anatomical location effects on arterial eNOS, SOD-1, SOD-2 proteins. A single 130-kDa band was evident in immunoblots for eNOS. As illustrated in Fig. 1, the eNOS content of brachial and femoral arteries from female pigs was 1.9- and 2.4-times greater than in the respective arteries from male pigs (P = 0.006 and 0.004, respectively). Aortic endothelial cells from female pigs also had significantly higher eNOS protein levels compared with males (5.9-times, P = 0.01; data not shown). Greater eNOS content of endothelial cells of female arteries was also apparent in immunohistochemistry results (Fig. 2).
A single 17-kDa band was evident in immunoblots for SOD-1. SOD-1 content was greater in female femoral arteries (2.41-times male; P = 0.005), but male and female brachial arteries had similar amounts of SOD-1 (P = 0.434; Fig. 1). A single 22-kDa band was evident in immunoblots for SOD-2. Female pigs had significantly greater SOD-2 content than males in both brachial (1.48-times, P = 0.013) and femoral (1.57-times, P = 0.008) arteries (Fig. 1). Figure 3 presents immunohistochemical results demonstrating that the endothelium and smooth muscle exhibited immunoreactivity for SOD-1 in the arteries of both males and females. Immunohistochemical staining for SOD-2 revealed punctuate staining (consistent with SOD-2 being most concentrated in mitochondria) in the endothelium and smooth muscle in arteries of both males and female pigs (data not shown).
Figure 4 illustrates the relative levels of eNOS, SOD-1, and SOD-2 proteins in brachial vs. femoral arteries within a given gender. There are only two significant differences in protein content between brachial and femoral artery: SOD-1 was significantly lower in femoral compared with brachial arteries in male pigs (0.61-times brachial levels, P = 0.001), and SOD-2 was significantly higher in femoral compared with brachial arteries in female pigs (1.26-times brachial levels, P = 0.048).
Gender effects on ER-α mRNA levels in aortic endothelial cells. Figure 5 illustrates the data from RT-PCR experiments designed to evaluate relative levels of ER-α mRNA in aortic endothelial cells from male and female pigs. A single PCR product of 325 bp was evident. This is consistent with the predicted product size based on the primers designed. Identification of this PCR product was confirmed as ER-α by direct sequencing after band excision from an agarose gel. The gel-purified band shared 100% homology with the pig ER-α gene (6) over 300 sequenced bases. Negative control experiments leaving either reverse transcriptase enzyme or cDNA out of the reactions yielded no detectable PCR products (data not shown). There was no significant effect of gender on α-ER mRNA levels (female 0.84-times male levels, P = 0.189) under conditions optimized for product formation and densitometric analysis (Fig. 5).
Gender effects on ER protein levels in aorta. Representative panels of immunohistochemistry for α-ER and β-ER protein of abdominal aorta are presented in Fig. 6. These results show similar staining (immunoreactivity) in both endothelial and smooth muscle cells for α-ER protein (Fig. 6, A and C) and for β-ER protein in endothelial cells (Fig. 6, B and D) in aorta from both females and males.
Exercise Training Effects
Exercise training produced the expected adaptations in exercise endurance, heart weight, and skeletal muscle oxidative capacity in pigs of both genders. Thus in both male and female pigs, average heart weight and heart weight-to-body weight ratio was greater in Ex than in Sed pigs resulting from greater heart weights in Ex animals (Table 1). The heart weight-to-body weight ratio of Sed male pigs was greater than that for Sed female pigs. Treadmill performance tests revealed that endurance times (total run time) were significantly longer in Ex pigs (Table 1). Citrate synthase activity of the long and lateral heads of the triceps brachii muscle and the deltoid muscle of Ex pigs was 20–40% greater than values in Sed male and female pigs, confirming the shift in skeletal muscle oxidative capacity that characterizes effective exercise training. Finally, as shown in the last two columns of Table 1, exercise training did not alter E2 levels in the blood of male or female pigs.
Figure 7 illustrates the effects of gender and exercise training on eNOS, SOD-1, and SOD-2 protein levels in both brachial and femoral arteries. Exercise training increased eNOS protein levels in both the brachial and femoral arteries of male (1.80- and 2.08-times Sed males, P = 0.021 and 0.030, respectively) and female (1.67- and 1.84-times Sed females, P = 0.013 and 0.001, respectively) pigs. Exercise training did not significantly affect SOD-1 protein levels in brachial or femoral arteries of either males or females (Fig. 7). Levels of SOD-2 protein were not significantly affected by exercise training in either artery type in males nor in brachial arteries of female pigs. However, femoral arteries of Ex female pigs had significantly higher levels of SOD-2 protein than their Sed counterparts (Ex pigs 1.39-times Sed levels, P = 0.0002; Fig. 7).
Circulating Estradiol and Estrogenic Activity in Male and Female Pigs
Experimental rationale. Results of the experiments above indicated that endothelial cells in arteries from female pigs appeared to have greater content of eNOS and SOD protein relative to arteries of male pigs just as do arteries of other female mammals described above. These results raised the question: Why does endothelial phenotype of male pigs not appear to reflect the high levels of E2 in their serum? To test the hypothesis that male arteries do not reflect the high levels of circulating E2 because they express fewer or no ER, we measured ER mRNA in aortic endothelial cells and ER protein in aortas of male and female pigs. We reasoned that the level of ER expression in the male must be one-tenth or less that of female arteries to explain our results. However, our results revealed indistinguishable levels of ER mRNA (RT-PCR) and levels of both α-ER and β-ER protein by immunocytochemistry in arteries of male and female pigs.
We then tested the hypothesis that only a small fraction of E2 detected by RIA in the blood of male pigs is free (unbound to serum-binding proteins). In previous studies, E2 content of serum of male and female pigs was measured with RIA, which measures total E2 in the serum (including free E2 and E2 bound to glycoproteins and serum albumin in the blood) (2, 29, 34). It is widely accepted that the biologically active form of E2 is the free, non-protein-bound fraction (54). If the free fraction for E2 in male pig serum were much lower than the free fraction in female serum, then our results would be explained. Free fraction of E2, determined by centrifugal ultrafiltration dialysis, indicated that the free fraction of E2 in males was only slightly lower than in females, 3.5 vs. 3.7%, respectively. Finally, we used a sensitive bioassay to test the hypothesis that total estrogenic activity present in male pig serum was lower than the E2 content indicated by RIA. In a subset of animals, we measured estrogenic activity and E2 content in the same blood samples of male and female pigs. E2 content was measured with standard RIA and with RIA after running serum on LH-20 columns to evaluate any influence of circulating estrone in our samples that could have been released by solvolysis.
Biologically active estrogenic activity in the blood was determined with a bioassay of estrogenic activity in a subset of male and female sera. Bioassay results indicated that total estrogenic activity in the blood of our male pigs (35 pg E2/ml serum) was similar to human male concentrations of E2 reported in the literature (10–50 pg/ml) (54) in stark contrast to published values in the hundreds to 1,500 pg E2/ml as detected by RIA in male pig sera. Finally, we also measured serum E2 concentrations (both total by RIA and biologically active by bioassay) throughout the estrous cycle in 10 female pigs to establish the peak concentration of E2 in Yucatan females just before ovulation for comparison to the values of E2 in males and human females.
Free fraction of E2 in serum. Determinations of percent free estradiol in pig blood revealed that there were significant differences (P = 0.021) between free estradiol levels in male and female pigs (Table 2). However, the free fraction of E2 in male pigs would have had to be one-tenth or less of the level in female pigs to explain the fact that the males were not apparently estrogenized by the high total E2 levels in their serum measured with RIA. Therefore, the differences in total E2 do not appear to be counteracted by a much lower free fraction in males. Interestingly, exercise produced significant increases in free fraction in males but significant decreases in females (Table 2).
Biologically active estrogen by bioassay in male and female blood. In blood samples from a subset of pigs used in the exercise experiments described above, total estrogenic activity was measured by bioassay in the same extracts used for RIA to compare biologically active estrogen to the RIA value for E2 measured in the serum. We used HPLC/bioassay to fractionate the serum and measured estrogenic activity in each fraction with a tissue culture bioassay. Figure 8 presents results for two male pigs that did not exhibit significant levels of estrone. As shown in Fig. 8 (pig 1,284) the level of estrogenic activity, corresponding to where E2 ran in the HPLC, was only 42 pg E2 equivalents/ml (7.5-min sample), whereas RIA indicated that E2 was 100 pg/ml. Also, in a sample chosen for its low measured E2 to provide a low background, almost no estrogenic activity of any other kind was observed in the serum (Fig. 8, bottom, pig 1,343), indicating that most or all of the biologically significant estrogenic activity, at least in these two serum samples, migrated with E2.
Analysis of group data reveals that, in female pigs, estrogen determination as E2 by RIA and by RIA after LH-20 was indistinguishable from estrogenic activity measured by bioactivity (Table 3). In contrast, in male pigs, all groups were statistically different, with estrogenic activity by bioassay less than half of the level of E2 determined by RIA and the value by RIA after LH-20 chromatography was intermediate (Table 3). The highest E2 we observed in a male serum sample, as measured by RIA, was 645 pg/ml. In this sample, the total estrogenic activity was equivalent to only 66 pg E2/ml (data not shown). So consistently in these samples, the biologically active estrogenic activity in male blood was lower and sometimes much lower than the E2 level indicated by RIA. In contrast to this pattern, in the unstaged females, biologically active estrogen content was identical to the E2 measured by RIA (Table 3).
Estradiol in female blood across full estrous cycles. The changes in serum estradiol observed in the 10 female pigs are presented as a function of days relative to the day of peak estradiol levels in Fig. 9. These results reveal that the peak value for these cycling female pigs was ∼24 pg/ml, whereas diestrous values were 1–5 pg/ml as determined by RIA. Table 4 presents results demonstrating that E2 in the serum of these 10 pigs was similar at peak and during diestrous when measured by RIA and by bioassay of estrogen equivalents in the circulation.
The purpose of this study was to test the hypothesis that female gender and exercise training increase eNOS and SOD-1 protein content of porcine skeletal muscle arteries and to evaluate the role of serum E2 in these effects. This hypothesis was stimulated by previous observations that gender and exercise training have important effects on endothelium-dependent relaxation in these arteries. Results provide six important findings. First, femoral and brachial arteries as well as aortic endothelial cells isolated from female pigs have significantly greater eNOS content than arteries from male pigs, consistent with gender-related effects in arteries of other mammals (5, 18, 19, 24, 32). Second, although female femoral arteries contained greater SOD-1 content than the femoral artery of male pigs, there was no gender difference between SOD-1 content in brachial arteries. Third, SOD-2 content was greater in brachial and femoral arteries of female pigs. Fourth, biologically active estrogen concentration in male pigs is similar to E2 values reported for human males, whereas female pigs exhibit estradiol concentrations during diestrous (2 pg/ml) and estrous (24 pg/ml) that are less than those of women (22, 54). Fifth, there was no effect of gender on α-ER mRNA levels in aortic endothelial cells or on α-ER or β-ER protein in smooth muscle or endothelial cells in aortas of male and female pigs. Sixth, although gender has important influences on endothelium-dependent dilator responses and eNOS protein content of femoral and brachial arteries, gender did not seem to interact with the effects of exercise training, because training increased eNOS content in both male and female arteries. As discussed in Effects of Gender on Expression of eNOS, SOD-1, and SOD-1 in Skeletal Muscle Arteries, our results combined with present literature indicate that mechanisms of gender differences in arteries perfusing skeletal muscle and of exercise training-induced adaptations in these arteries are not uniform across agonists, species, or anatomic origin of the artery.
Effects of Gender on Expression of eNOS, SOD-1, and SOD-1 in Skeletal Muscle Arteries
Barber and Miller (2) reported that endothelium-dependent relaxations of coronary arteries to BK and UK-14304 (α2-adrenergic analog) were greater and/or shifted leftward in arterial rings from female pigs compared with males. Similarly we found that femoral arteries of female pigs exhibited greater endothelium-mediated relaxation (BK- and Ach-induced) than did femoral arteries from male pigs (29). Thus there is substantial evidence that conduit arteries of female pigs exhibit greater endothelium-dependent responses than do males. However, in contrast to these observations, we found that male brachial arteries exhibited greater BK-induced relaxation and greater Ach-induced relaxation than did brachial arteries from female pigs (29). Blockade of the cyclooxygenase pathway and NOS pathways revealed that cyclooxygenase seems to be relatively unimportant in porcine brachial arteries, and the NOS pathway contributes more to endothelium-dependent relaxation in female brachial arteries than in males (29). Because present results (Figs. 1 and 2) indicate that female femoral arteries contain more than twice as much eNOS protein as male femoral arteries, it is reasonable to propose that greater release of NO contributes to the greater relaxation seen in female arteries (29). On the other hand, the greater endothelium-dependent relaxation responses in male brachial arteries cannot be explained by differences in eNOS content since female brachial arteries also had higher eNOS content than male brachial arteries (Fig. 1). These observations suggest that gender has different effects on vasomotor reactivity of porcine femoral and brachial arteries.
We are intrigued by the fact that the effects of gender on vasomotor reactivity (29) and on eNOS and SOD expression (Figs. 1 and 2) of femoral and brachial arteries do not appear to fully correspond. Ours is not the only observation of such differing effects of gender among peripheral arteries. For example, Vagnoni et al. (55) reported that estradiol stimulated an increased expression of eNOS protein in the sheep uterine artery but not in systemic arteries. Thus our results provide further evidence supporting the conclusion of Barber and Miller (2) that the mechanisms of gender differences in endothelial function differ by agonist, species, and anatomic origin of the artery.
Role of Estrogen in the Porcine Gender Differences
Barber and Miller (2) reported that the gender differences in porcine coronary endothelium-dependent relaxation may not be related to estrogen levels because male pigs had higher plasma concentrations of estrogen than female pigs, perhaps due to metabolism of testosterone by aromatase in the adipose tissue. It has been know for some time that boars appear to have high concentrations of estrogens in their blood. Indeed, domestic boars as well as stallions appear to be unique among mammals in having high blood and urinary concentration of estrogens (7–9, 51, 57). Paradoxically, these levels of estrogens in the boars are many times higher compared with either preovulatory females of the same species (20) or the males of other species (44, 45). The interstitial compartment of the testis is the site of expression of steroidogenic enzymes in the boar (11, 48) and therefore the major cellular site of both androgen and estrogen biosynthesis. The dynamic nature of the secretory patterns of free androgens (10), estrogens (14), and conjugated steroids (49) from the boar testis, especially the predominance of Δ5-C19-steroid production (3) and other conjugated estrogens (56) has been well documented. Our results (Table 1) initially seemed to confirm these observations since RIA data indicated plasma concentrations of E2 were greater in male pigs than in female pigs and these levels of plasma E2 are in the range of values reported in the literature (2, 44, 45, 60).
Nevertheless, as the experiments described herein proceeded, we used several assays of E2 to allow evaluation of the possibility that E2-RIA, as used by us and others, provides high values of circulating estrogens (≥1,000 pg/ml) in male pigs because of a lack of specificity of the estrogen antibody used in our E2-RIA. For example, in male pig blood, the antibody may react with estrone, or ether extraction of samples may cause solvolysis of estrogen conjugates, or the antibody may crossreact with a closely related Δ5-C19-steroid in the blood. We obtained similar values for E2 with RIA, LH-20-RIA, and bioassay in female samples (Table 3). In contrast, our results indicate that the levels of biologically active estrogen (estradiol and other estrogens) in male blood were much lower than indicated by RIA measures of E2 (Table 3). This activity, expressed as estradiol equivalents, was similar to the total amount of E2 (10–50 pg/ml) reported to circulate in human males (54) and in the range of the peak levels (estrous) of both E2 and total estrogenic activity measured in female blood (Table 4). These results indicate that one reason male pigs show no evidence of being estrogenized, despite the apparently elevated levels of E2, is that boars do not have excessive amounts of biologically active E2 in their blood. Thus present results indicate that male pigs appear to have E2 levels equal to or greater than female pigs and that endothelial cells from male and female pigs express similar amounts of α-ER mRNA and protein. These results suggest that male pigs are likely as sensitive to E2 as are female pigs and that other (non-estrogen) sex hormones and/or other gender-related processes play a role in establishing the greater eNOS content in arteries of female pigs.
Our male pig serum E2 results are even more interesting in light of the fact that estradiol as well as testosterone is required for full reproductive development in the male pig, whereas only estradiol is required in the female (15, 31). Whether human males also require estrogen for development of a normal phenotype is not known. However, it is intriguing that development of the male phenotype in pigs requires both estradiol and testosterone and the biologically active levels of estrogen are similar in males and females.
It is also interesting that the paradox presented by the male blood E2 content, which was high when measured by RIA but consistently lower when measured by bioassay of actual estrogenic activity, was not seen in the females. In the female pig, the two measures, RIA and bioassay, corresponded closely (Tables 3 and 4). Thus it appears that most or all of the estrogenic activity present in female serum was accounted for by the circulating E2 as measured by RIA. Dietary estrogens did not appear to contribute to circulation of biologically significant levels of phytoestrogens in the male or female pigs even though the feeds used in these studies were soy based and contained detectable estrogenic activity (data not shown). Clearly, our results indicate that in male pig blood, RIA-measured E2 values did not reflect estrogenic activity, whereas, in female pig blood, there was little contribution to estrogenic activity by any other circulating estrogens (including any potential dietary estrogens).
There have been previous observations where biologically active estrogen levels are less than estimates of total estrogen in the blood of mammals. For example, high levels of total E2 measured by RIA in the rodent fetus are counterbalanced by very high levels of the estradiol-binding protein in alphafetofetal serum. This leads to a free (unbound) bioavailable level of E2 that is biologically low, in the range of the male or diestrous female (35, 58). To determine whether differences in binding protein in the blood contributed to our findings, we examined the free fraction of E2 in our male and female blood samples. Results reveal that the free fraction of E2 in male pigs was very similar to that of the female pigs (Table 2).
There are at least two potential explanations of the observation that biologically active estrogen in male pig blood, as measured by a standard in vitro bioassay, was lower than E2 measured by RIA and the apparent absence of physiological estrogenic effects in male pigs. First, it is possible that a molecule is present in the male pig blood that crossreacts in the RIA for estradiol but exhibits little estrogenic activity. Secondly, it is possible that male pig serum contains an inhibitor of estrogenic activity that reduces the effective estrogenic activity of the E2 present in the serum, perhaps acting as a natural antiestrogen. Further work is required to determine an explanation for why biologically active estrogen in male pig blood, as measured by a standard in vitro bioassay, is lower than E2 measured by RIA, and the apparent absence of physiological estrogenic effects in male pigs.
Effects of Exercise Training
Exercise training has been reported to result in enhanced endothelium-dependent relaxation in rat aorta (12, 13), in rat skeletal muscle arteries (25), in brachial arteries of pigs (33), in peripheral arteries of humans (17), in dog aorta and coronary arteries (50, 59), and porcine coronary arteries (28). Present results (Fig. 7) indicate that exercise training increased eNOS expression by similar amounts in both male and female femoral and brachial arteries. The results provide no evidence for an interaction of gender with the effects of exercise training on eNOS expression. Exercise training did not alter content of SOD-1 protein in femoral or brachial arteries in either gender.
The fact that exercise training increased eNOS content of femoral and brachial arteries in both male and female pigs is interesting since previous results indicated that exercise training did not alter net endothelium-dependent relaxation of male femoral or brachial arteries or of female femoral arteries but increased endothelium-dependent relaxation of female brachial arteries (29). One purpose of this study was to determine the role of changes in eNOS expression in these events. It is not clear at this time why exercise training increases eNOS expression in both femoral and brachial arteries of female and male pigs, but endothelium-dependent relaxation is only enhanced by exercise training in the brachial artery of female pigs.
Exercise training did not alter estradiol levels in either male or female pigs. Training did produce slight changes in the binding of E2 in serum, because it increased free fraction in males but lowered it in females. Because the source of E2 binding proteins is believed to be the liver, these results suggest that training may have gender-specific effects on production of serum proteins by the liver.
The results of this study indicate that conduit arteries of female pigs have greater eNOS content than do the same arteries of male pigs. Although male pigs have high concentrations of RIA-measured E2 (estrogen-like molecules) in their blood, the levels of biologically active estrogen are similar to those in male humans. These results combined with the observation that there is no gender effect on α-ER mRNA or α-ER or β-ER protein levels suggest that eNOS expression in endothelial cells of arteries may be modulated by other female hormones instead of estrogen.
These data stimulate us to propose that, on the basis of circulating estrogenic activity, the male pig is not hyperestrogenized relative to the human male. Furthermore, our results indicate that female pigs may be hypoestrogenized relative to the human since E2 levels at peak estrous are substantially lower than values of human females at peak estrous. However, it is important to emphasize that these levels of estradiol or bioassayable estrogenic activity in serum of male pigs are not sufficiently high to interfere with spermatogenesis and male reproduction and that the levels of estradiol and estrogenic activity we measured in the female are not so low as to fail to generate the cyclic changes to the uterus and to sustain pregnancy, lactation, and female reproduction. Therefore, both genders in the pig may represent appropriate models of cardiovascular function and gender differences relative to humans.
The authors thank Pam Thorne, Denise Holliman, Tammy Strawn, and Lisa Code for technical contributions to this work.
This work was supported by National Institute of Health Grants HL-36088 and HL-52490 (to M. H. Laughlin) and RR-13223 (to M. Sturek), Heart and Stroke Foundation of Ontario Grant NA-4604 (to J. W. E. Rush), and the University of Missouri Food for the 21st Century VMFC-0018 (to W. V. Welshons).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2003 the American Physiological Society