If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Evans Memorial Department of Medicine, and Whitaker Cardiovascular Institute, Room W507, Boston University Medical Center, 80 E. Concord St., Boston, MA 02118, USA
Evans Memorial Department of Medicine, and Whitaker Cardiovascular Institute, Room W507, Boston University Medical Center, 80 E. Concord St., Boston, MA 02118, USA
Evans Memorial Department of Medicine, and Whitaker Cardiovascular Institute, Room W507, Boston University Medical Center, 80 E. Concord St., Boston, MA 02118, USA
Risk for coronary artery disease is reduced by exposure to estrogens, although the mechanisms of protection are not fully defined. Recent observations have shown that physiologic concentrations of 17-beta-estradiol (E2) exhibit antioxidant activity in vitro, slowing the formation of atherogenic, oxidized low-density lipoprotein (LDL). Using concentrations physiologically relevant for premenopausal women, we compared the antioxidant potency of estrone (E1), E2, and estriol (E3) as measured by their ability to inhibit LDL oxidation. Plasma was incubated with 10 nmol/l estrogens for 4 h at 37°C, followed by LDL isolation and Cu2+-mediated oxidation in conjugated diene assays. Only E2 demonstrated antioxidant activity at these physiologic concentrations. Resistance to oxidation was not associated with sparing of endogenous α-tocopherol during plasma incubations. Incubation of plasma with radiolabeled estrogens yielded similar association of E1 and E2 with LDL which was 5–8-fold greater than the association of E3. Chromatographic analysis revealed the association of authenic E1 with LDL, while plasma-derived E2 esters were the major form of E2 associated with LDL which was resistant to oxidation. Thus, conjugation in plasma and association of E2 esters with LDL appear to be specific for E2 among these estrogens and render this LDL resistant to oxidation by Cu2+. This antioxidant activity may be another means whereby E2 protects against coronary artery disease in women.
]. Furthermore, exposure of postmenopausal women to estrogen treatment reduces the number of clinical events in this population, even in those with established CAD [
]. Although the mechanisms responsible for these observations are not fully understood, these studies suggest that estrogens play a significant role in reduction of established risk factors for CAD [
]. Early atherosclerotic lesions are characterized by an intimal accumulation of cholesteryl ester-laden `foam cells' formed presumably as a result of scavenger receptor-mediated internalization of LDL that has been oxidatively modified by cells of the vascular wall [
Antiatherogenic effect of probucol unrelated to its hypocholesterolemic effect: evidence that antioxidants in vivo can selectively inhibit low density lipoprotein degradation in macrophage-rich fatty streaks and slow the progression of atherosclerosis in the Watanabe heritable hyperlipidemic rabbit.
Studies demonstrating the antioxidant properties of estrogens support the hypothesis that these hormones protect against CAD by means other than lipid-lowering. Estrogens with a phenolic structure protect LDL from both cellular and Cu2+-mediated oxidation in vitro, and of the three principal estrogens in premenopausal women, 17-beta-estradiol (E2) is more potent than either estrone (E1) or estriol (E3) [
]. In addition, E2 has been shown to be more potent in protecting against LDL oxidation than α-tocopherol, the principal lipid-soluble antioxidant in LDL [
]. While concentrations of estrogens required to demonstrate antioxidant activity in these in vitro studies (≥1 μmol/l) greatly exceed levels observed even during pregnancy, recent reports have shown that chronic or acute hormonal replacement with physiologic concentrations of E2 in ovariectomized swine [
]is associated with increased LDL resistance to oxidation ex vivo. We have recently found that antioxidant protection with physiologic E2 concentrations against Cu2+-mediated LDL oxidation requires hydrophobic modification of E2 by plasma components [
]. In the present study, we compared plasma-mediated modification and association of E1, E2, and E3 with LDL and the subsequent Cu2+-mediated LDL oxidation in vitro and found that antioxidant protection of LDL by physiologic concentrations of estrogens is specific to E2 and is associated with the exclusive hydrophobic derivatization of E2 in plasma.
2. Materials and methods
2.1 Chemicals
Unlabeled estrogens were purchased from Steraloids (Wilton, NH) and [3H]E1 (specific activity, 76 Ci/mmol), [3H]E2 (specific activity, 110 Ci/mmol), and [3H]E3 (specific activity, 90 Ci/mmol) were from Dupont New England Nuclear (Boston, MA). Sephadex G-25 was obtained from Pharmacia Biotech (Uppsala, Sweden) and Chelex-100 resin was from Bio-Rad Laboratories (Hercules, CA). High-performance liquid chromatography (HPLC) grade methanol and hexane were from Aldrich (Milwaukee, WI) and Ecolite+ scintillation cocktail was obtained from ICN Biomedicals (Costa Mesa, CA). All other chemicals were purchased from Sigma (St. Louis, MO).
2.2 LDL isolation and incubation with estrogens
LDL exposure to estrogens was carried out as previously described [
]. Briefly, heparinized plasma was prepared from healthy, normolipidemic, fasted (14 h) males by centrifugation (1200×g) at 4°C for 15 min. Ethanolic solutions of unlabeled or radiolabeled E1, E2, or E3 or vehicle alone were dried under N2 and reconstituted with plasma to yield a final concentration of 10 nmol/l or 100 nmol/l in selected experiments. This plasma was incubated for 4 h at 37°C sealed under N2 gas. LDL was subsequently isolated by single vertical spin discontinuous density gradient ultracentrifugation [
], and size exclusion gel filtration was performed using Sephadex G-25 columns. Contaminating metal ions were removed with Chelex-100 resin followed by filtration through a 0.45 μm filter and protein determination using a modified procedure of Lowry [
]. LDL isolated in this manner contained no detectable lipid hydroperoxides (see below) and was used immediately for experiments. In selected experiments, plasma was was heated to 60°C for 35 min to inactivate acyltransferase activity prior to incubation with estrogens or coincubated with 1.4 mmol/l 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) and estrogens prior to determination of LDL resistance to oxidation and estrogen incorporation.
2.3 Oxidative modification of LDL
Resistance to oxidation was measured spectrophotometrically as accumulation of conjugated dienes at 37°C using 100 μg LDL protein per ml in phosphate buffered saline (PBS; 154 mmol/l NaCl, 10 mmol/l NaH2PO4, pH 7.4) and a final concentration of 3.3 μmol/l CuCl2. Duration of the lag phase was calculated as previously described [
2.4 LDL antioxidant and cholesteryl ester hydroperoxide content
Isolated LDL (250 μg protein) lipids were extracted with methanol–hexane (1:10, v/v) and centrifuged at 500×g for 10 min at 4°C. An aliquot of the hexane extract was dried under N2 and resuspended in 0.1 ml ethanol. LDL α-tocopherol was determined by HPLC with electrochemical detection as described [
2.5 Association and HPLC analysis of [3H]estrogens associated with LDL
Plasma was incubated with 10 or 100 nmol/l [3H]estrogens for 4 h as described above. The association of radiolabeled estrogens with isolated LDL was measured using scintillation spectroscopy (LKB 1214 Betarack) after removal of unbound 3H in LDL preparations by gel filtration with Sephadex G-25. An aliquot of LDL was precipitated with heparin–MnCl2 [
]. Electrophoresis of the apolipoprotein B-containing precipitate showed no detectable albumin and, <7% of the radiolabel in the LDL isolate was present in the albumin-containing supernatant. LDL (200 μg) was extracted with methanol and hexane as described above. Extracts were chromatographed using a 25 cm LC-18 column (Supelco) and a Hewlett-Packard Series 1050 HPLC system using absolute methanol as mobile phase at a flow rate of 1.0 ml/min with UV detection at 210 nm. Radiochromatograms of LDL-associated [3H]estrogens were generated by scintillation counting 0.3 ml fractions of eluate.
2.6 Data analysis
All data are reported as mean±S.E.M. The measures of LDL oxidative resistance and LDL estrogen content were compared among treatment groups by ANOVA and a post hoc Bonferroni t-test. Statistical significance was accepted if the null hypothesis was rejected with P<0.05.
3. Results
We have recently reported that plasma components are required for antioxidant protection of LDL by physiologic concentrations of E2 [
]. In order to compare the antioxidant potency of the structurally related estrogens E1 and E3 to that reported for E2 at physiologically relevant concentrations, we incubated plasma with 10 nmol/l E1, E2, and E3 prior to LDL isolation and measured LDL resistance to oxidation in conjugated diene assays. Consistent with our previous observations, plasma incubated with 10 nmol/l E2 for 4 h at 37°C yielded LDL that was significantly more resistant to Cu2+-mediated oxidation as lag time to oxidation of E2-exposed LDL increased by 20% above vehicle-treated LDL (Fig. 1; P<0.05 vs. control, n=4). However, LDL resistance to oxidation was not significantly altered by exposure to either E1 or E3 in plasma (lag time 115±3 and 117±4 min, respectively) and was comparable to that of control LDL (lag time 115±4 min). Moreover, heat inactivation of plasma (60°C for 35 min) or coincubation with 1.4 mmol/l DTNB inhibited the antioxidant protection afforded by E2 described above, actually decreasing the lag time by 22% and 20%, respectively, compared to control incubations (data not shown).
Fig. 1Plasma was incubated with 10 nmol/l E1, E2, E3, or vehicle for 4 h at 37°C. LDL was isolated by ultracentrifugation, Chelex treated, and low molecular weight contaminants were removed by gel filtration with Sephadex G-25 columns. Resistance to oxidation was measured in conjugated diene assays using 100 μg/ml LDL protein and 3.3 μmol/l copper. Data are representative of 4–6 experiments. For clarity, only every fourth data point is shown. Lag time to oxidation is significantly increased in LDL isolated from plasma incubations with E2. P<0.05 by ANOVA.
The increased LDL resistance to oxidation afforded by E2 suggests some stable alteration in the nature of LDL rendering increased antioxidant protection. Two important determinants of LDL susceptibility to oxidation in vitro are the content of lipid-soluble antioxidants, principally α-tocopherol [
Content of antioxidants, preformed lipid hydroperoxides, and cholesterol as predictors of the susceptibility of human LDL to metal ion-dependent and -independent oxidation.
]. Thus, we investigated the LDL content of α-tocopherol and lipid hydroperoxides after exposure of plasma to 10 nmol/l estrogens to determine if differences in LDL resistance to oxidation resulted from selective changes in endogenous antioxidants or lipid hydroperoxides. LDL isolated from plasma incubated with vehicle for 4 h had an α-tocopherol content of 11.9±0.9 nmol/mg protein, consistent with previous reports [
]. There was no difference in the LDL α-tocopherol content with E1, E2, or E3 treatment (11.9±0.9, 11.5±0.7 and 12.1±0.9 nmol/mg protein, respectively, n=4). Similarly, LDL samples were free of detectable cholesteryl ester hydroperoxides, thus precluding prevention of LDL lipid hydroperoxide formation as a mechanism of antioxidant activity specific for E2.
To determine if the selective antioxidant protection of LDL by E2 was related to the degree of association with LDL compared to that of other estrogens, we incubated plasma for 4 h at 37°C with [3H]estrogens and measured the radiolabel associated with LDL after isolation and gel filtration. At both 10 and 100 nmol/l estrogen concentrations, E1 and E2 demonstrated comparable levels of association with LDL that were equivalent to approximately 1–2% of the total estrogen added, while E3, the most polar of these estrogens, showed 5–8-fold less association with LDL than either of the more non-polar estrogens (Table 1). For all three estrogens the association with LDL in plasma was dose-related (Table 1).
Table 1Plasma-mediated association of estrogens with low density lipoprotein
Plasma treatment
LDL estrogen content (pmol/mg protein)
10 nmol/l
100 nmol/l
Vehicle
ND
ND
Estrone
0.081±0.004
0.872±0.062
Estradiol
0.067±0.002
0.861±0.136
Estriol
0.018±0.008*
0.105±0.004*
Plasma was incubated with the indicated concentrations of radiolabelled estrogens and LDL isolated and gel filtered as described in Section 2. Estrogen incorporation was determined by scintillation counting.
*P<0.05 versus estrone and estradiol incubation by ANOVA (n=4).
Using scintillation spectroscopy, we next investigated the chemical nature of plasma-mediated association of estrogens with LDL. Of the three estrogens employed in this study, E1 is the most non-polar. When LDL isolated from plasma treated with 10 nmol/l [3H]E1 was extracted with methanol–hexane, 0.065±0.007 pmol E1 per mg LDL was recovered in the methanol phase, while 0.020±0.003 pmol E1 per mg LDL was found in the hexane phase (76% and 24%, respectively), closely approximating the partitioning of authenic E1 when incubated in PBS alone and extracted as above (Fig. 2). However, incubation of plasma with E2 resulted in the conversion of E2 into a form that was more hydrophobic as evidenced by its increased solubility in hexane compared to authenic E2 in PBS (Fig. 2). This derivatized E2 was the principal form associated with LDL. Approximately 76% of the radiolabel that associated with LDL in E2–plasma incubations was found in the hexane phase, nearly 12-fold the amount of authentic E2 that partitions into hexane (Fig. 2). Although some hydrophobic modification of E3 in plasma was evident by its enhanced solubility in hexane compared to authentic E3 (Fig. 2), E3 did not appreciably associate with LDL in plasma compared to the more non-polar estrogens.
Fig. 2Purified [3H]estrogens in (A) PBS and (B) LDL isolated from plasma incubated with [3H]estrogens at 10 nmol/l were extracted with methanol–hexane as described in Section 2. Partitioning of radiolabel was measured by scintillation spectroscopy. The total association in the combined methanol and hexane extracts reflects the observations of Table 1. Values in (A) are the mean of two separate determinations which differed by <3%; in (B), n=4.
To characterize further the nature of the estrogen species which associated with LDL in plasma, aliquots of the extracted LDL lipids were analyzed by HPLC. The radiochromatograms in Fig. 3 demonstrate that E2 is specific in its modification to hydrophobic forms that associate with LDL. E1 associates with LDL with equal or greater avidity compared to E2 (Table 1) and appears only as the authentic, parent compound in both the methanol and hexane phases with a single elution time of 3.6 min (Fig. 3). However, E2 is primarily found in LDL as hydrophobic derivatives, with only 25–30% of the radiolabel eluting as authenic E2. This association and derivatization of E2 is qualitatively reproducible with three or four major hydrophobic peaks separated by HPLC in each of four samples examined, while amounts of each peak vary among individuals.
Fig. 3Methanol and hexane fractions from Fig. 2 were analyzed by HPLC using absolute methanol as mobile phase at a flow rate of 1 ml/min. Fractions were collected every 20 s and radiochromatograms generated by scintillation counting. One chromatogram representative of three is shown. The distribution of radiolabelled E1 associated with LDL reflects normal partitioning of E1 in a methanol–hexane extraction, whereas the majority of radiolabeled E2 is shifted into the hexane phase where only 6% of E2 is normally partitioned. Only small quantities of E3 associate with LDL in plasma and appears equally distributed as authenic E3 and a derivatized hydrophobic form.
To confirm that association of E2 with LDL was an enzymatic process, we subjected samples to heat inactivation (60°C for 35 min) or co-incubation with 1.4 mmol/l DTNB to inhibit acyltransferase activity in plasma. Heat inactivation or DTNB treatment inhibited the association of E2 with LDL by 65 and 72%, respectively (mean of two observations, varying by <7%). This reduction in E2 association with LDL was primarily limited to hydrophobic forms of E2. Hexane extracts of LDL from these experiments accounted for approximately 82% of the reduction in E2 association. Thus, hydrophobic derivitization of estrogens and association with LDL is specific to E2, and this process is prevented by either inhibition or inactivation of acyltransferase activity in plasma.
4. Discussion
The data presented here demonstrate that, of the three principal estrogens in premenopausal women, antioxidant protection of LDL against Cu2+-mediated oxidation is specific for E2 at physiologically relevant concentrations. This antioxidant activity is associated with the accumulation of hydrophobic derivatives of E2 in LDL. Although E1 and E2 showed comparable association with LDL, E1 demonstrated neither protection of LDL against Cu2+-mediated oxidation nor evidence of hydrophobic conjugation, which we have recently shown to be required for antioxidant protection by physiologic concentrations of E2 [
]. It is likely that E3, the most hydrophilic of these estrogens, did not afford antioxidant activity due to its poor association with lipoproteins. These observations suggest that the specific combination of esterification and association of E2 with LDL in plasma is associated with alterations of the lipoprotein particle which render it more resistant to oxidation by Cu2+.
We have recently shown that plasma incubations with E2 resulted in dose-dependent formation and association with LDL of hydrophobic conjugates with characteristics of fatty acid esters of E2 with the hydroxyl group of C-17 as the preferential site for fatty acid conjugation [
]. Furthermore, modification of E2 was required for antioxidant activity as heat inactivation of plasma, or incubation with DTNB which blocks thiols in the active site of lecithin–cholesterol acyltransferase (LCAT) inhibited the formation and association of E2 esters with LDL by 65–90% and completely blocked subsequent antioxidant protection of LDL. The observations in the present report are in concert with the structural differences among the three estrogens used here and support our hypothesis that esterification and association of estrogens with LDL is required for antioxidant protection at physiologic concentrations. In particular, E2 incubated in plasma was the only estrogen converted to a hydrophobic form which substantially associated with LDL and enhanced its resistance to oxidation. Although E1 and E2 associated with LDL to a similar degree, C-17 of E1 is a ketone and is resistant to esterification (Fig. 4). Thus, all of the radiolabel recovered from incubation of plasma with [3H]E1 eluted as authenic E1, while the majority of [3H]E2 was modified to the hydrophobic ester (Fig. 3). Finally, although E3 is similar to E2 in that C-17 bears a hydroxyl group and demonstrates some degree of modification during exposure to plasma (Fig. 2, Fig. 4), it poorly associates with lipoproteins, including HDL [
], which we propose mediates the hydrophobic modification of estrogens through the activity of LCAT.
Fig. 4While all three estrogens have at C-3 a phenolic hydroxyl group important in the free radical scavenging behavior of supraphysiologic concentrations of estrogens, our data indicate that esterification at C-17 is required for antioxidant protection of LDL at physiologic concentrations of these hormones (see [24]and Fig. 1). While E2 and E3 are esterified in plasma, E1 is a ketone at carbon 17 which is resistant to esterification and shows no antioxidant protection of LDL.
]. Incubation of isolated lipoproteins with physiologic concentrations of estrogens in the absence of plasma components led to E1 and E2 concentrations in LDL that were approximately 5–10 times those of E3. The authors showed that E2 association with LDL was linear and reached equilibrium by 10 h. In contrast, when incubated with HDL, E2 appeared to become more lipophilic over time and remained in non-equilibrium even after 96 h [
Formation of pregnenolone- and dehydroepiandrosterone-fatty acid esters by lecithin–cholesterol acyltransferase in human plasma high density lipoproteins.
], the authors suggest that the enhanced association of E2 with HDL involved esterification of E2 by HDL-associated LCAT, as conjugation was inhibited by DTNB and the steroid esters were similar to those found in HDL3 [
]. Under these conditions using isolated LDL in saline buffer, there was no such modification or accumulation of E2 esters. This is in agreement with our previous report showing that modification and association of E2 with LDL requires plasma where there is significant interaction among lipoproteins and/or transport components (e.g. cholesteryl ester transfer protein) that may shuttle modified steroid hormones between lipoproteins [
]and plasma (Table 1), esterification and antioxidant activity appear specific for E2 as analysis of LDL extracts showed that E1 associated with LDL in an unaltered form and provided no antioxidant activity (Fig. 1, Fig. 3).
The data presented here are in agreement with those of Larner et al. [
], who showed that authenic estrogens added to plasma poorly associate with LDL (<1% of total) whereas E2 esters preferentially associate with LDL due to their extreme hydrophobicity (43% of radiolabelled E2-17-stearate preferentially associated with LDL in human serum). Our preliminary data indicate that E2 associates to a greater degree with HDL than LDL in plasma (0.95±0.04 vs. 0.64±0.03 pmol E2 per mg protein, n=3) and is hydrophobically modified in subsequently isolated HDL. This esterification of E2 may maintain a concentration gradient for authenic E2 as the fatty acid esters are most likely to associate with the non-polar cholesterol-ester rich core of the particle sequestered from the pool of unassociated E2. However, we are skeptical that E2 or its hydrophobic derivatives that associate with lipoproteins directly provide antioxidant protection by free radical scavenging. There are three lines of evidence that lead to this conclusion. First, E1 associates with LDL to a slightly greater extent than E2 yet provides no significant antioxidant activity against Cu2+-mediated oxidation of LDL despite previous in vitro studies that have demonstrated similar degrees of free radical scavenging activity [
]. Secondly, our observations with LDL derived from plasma incubations with E2 demonstrated that while oxidation of LDL by Cu2+ was inhibited, no protection was afforded from oxidation by 2,2′-azobis(2-amidinopropane) dihydrochloride, a source of aqueous peroxyl radicals [
]. Finally, the amount of estrogen associated with LDL is quite small compared with what one would expect to provide free radical scavenging activity in the lipoprotein particle. Incubations of plasma with 100 nmol/l E2 indicate that, on average, <1% of LDL particles containes a hydrophobic form of E2 (Table 1). This is in agreement with both the experimental measurements of the association of estrogens and lipoproteins and the calculated values of hormone transport in plasma of premenopausal women [
]. The nature of this antioxidant mechanism remains unknown and requires further investigation.
To our knowledge, there have been no studies demonstrating a protective effect of E1 or E3 against atherosclerosis, while E2 is protective in both primates and rabbits [
]. In premenopausal women, E2 is the predominant estrogen with follicular concentrations exceeding 1 nmol/l while E1 and E3 concentrations remain <0.5 nmol/l in the non-pregnant female [
Transport of steriod hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma.
]. However, after menopause the ovaries cease to secrete bioactive estrogens and the majority of circulating estrogens are secreted as E1 from peripheral aromatization of adrenal androstenedione. We have shown that E1 is not esterified in plasma and is unable to provide antioxidant protection of LDL in spite of similar degrees of association with LDL in a plasma milieu. E3 is most likely ineffective as an antioxidant in LDL due its low plasma levels and poor affinity for the lipoproteins.
Thus, antioxidant protection is specific for E2 among the three principle estrogens in premenopausal women. This protection is associated with the exclusive formation and association of E2-fatty acid esters with LDL, a plasma-mediated phenomenon which is associated with increased resistance of LDL to Cu2+-mediated oxidation. This specific antioxidant behavior may be another means by which E2 protects premenopausal women from CAD.
Acknowledgements
This work was supported by a grants from the Council for Tobacco Research-USA, Inc. (4073), the American Heart Association, and the National Institutes of Health (HL52936). J.A.V. is an Established Investigator of the American Heart Association and J.F.K. is the recipient of a Clinical Investigator Development Award (HL03195) from the NIH.
Antiatherogenic effect of probucol unrelated to its hypocholesterolemic effect: evidence that antioxidants in vivo can selectively inhibit low density lipoprotein degradation in macrophage-rich fatty streaks and slow the progression of atherosclerosis in the Watanabe heritable hyperlipidemic rabbit.
Content of antioxidants, preformed lipid hydroperoxides, and cholesterol as predictors of the susceptibility of human LDL to metal ion-dependent and -independent oxidation.
Formation of pregnenolone- and dehydroepiandrosterone-fatty acid esters by lecithin–cholesterol acyltransferase in human plasma high density lipoproteins.
Transport of steriod hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma.
00020-3&id=gr1.gif){kind=link}
00020-3&id=gr2.jpg){kind=link}
00020-3&id=gr3.gif){kind=link}
00020-3&id=gr4.gif){kind=link}