Influence of Estradiol and Gestagens on Oxidative Stress in the Rat Uterus

Gynecologic Oncology 86, 250 –258 (2002) doi:10.1006/gyno.2002.6753

Influence of Estradiol and Gestagens on Oxidative Stress in the Rat Uterus M. A. Go´mez-Zubeldia,* ,1 S. Corrales,† J. Arbue´s,† A. G. Nogales,‡ and J. C. Milla´n† *Department of Nursing, †Department of Obstetrics and Gynaecology, School of Medicine, and ‡Department of Mathematics, University of Extremadura, Badajoz, Spain Received July 16, 2001

Objective. We studied the effect of ovariectomy, estradiol (E2), and E2 ⴙ medroxyprogesterone (MPA) on the Wistar rat uterus. Methods. We used 15 adult female rats. The study was divided into the following four stages: (a) extirpation of the upper half of the left hemi-uterus (basal state) and ovariectomy; (b) animals were maintained for 15 days without treatment, performance of a new laparotomy, and extirpation of the remaining left hemi-uterus (OVX state); (c) beginning of E2 replacement therapy (ERT) (8 ␮g/day) for 15 days, followed by extirpation of the upper half of the right hemi-uterus (ERT state); and (d) the administration of E2 was continued, and oral treatment with MPA was begun (20 ␮g/day) to last for a further 15 days. At the end of the combined hormone replacement therapy (HRT) the remaining right hemiuterus was extirpated (HRT state). At the end of each intervention, the plasma concentrations of E2 and PRG were measured. Results and Discussion. The ovariectomy significantly reduced the malonaldehyde (MDA) levels (P < 0.0008) and catalase activity (P < 0.0006). The ERT very significantly (P < 0.0033) raised the catalase and MDA levels; these significance levels were maintained after the Bonferroni method was applied (overall error 5%). The HRT reduced the levels of MDA and catalase, but not significantly after the Bonferroni test was applied. Conclusions. Uterine oxidative stress is increased by E2, resulting in a significant increase in MDA. This may be modulated in part by the catalase activity. Although it cannot be confirmed categorically, MPA seems to intervene by decreasing the said oxidative stress. © 2002 Elsevier Science (USA) Key Words: estradiol; progestagens; malonaldehyde; catalase; uterus.

INTRODUCTION Reactive oxygen species are generated endogenously under physiological and pathological conditions, as well as by exposure to exogenous changes [1]. The organism possesses defense mechanisms against these reactive oxygen species. Under physiological conditions, there is a delicate balance 1 To whom correspondence and reprint requests should be addressed at School of Medicine, University of Extremadura, Avenida Elvas s/n. 06071 Badajoz, Spain.

0090-8258/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

250

between the production of oxygen free radicals (OFR) and the organism’s antioxidant systems [2– 4]. An alteration in this balance will give rise to oxidative phenomena which may damage bio-molecules such as DNA, proteins, and lipids, for which reason OFR have been implicated in diverse human pathologies, as cancer and arteriosclerosis [5–9]. Recent studies [10 –13] have pointed to the involvement of ovarian steroid hormones, especially the estrogens, in the phenomena of oxidative stress. With respect to estrogens, studies have shown that their effect on OFR production may take place in contrary senses depending both on the structure of the estrogen and on the administered dose or the organ being studied [14 –16]. Thus, estradiol may have either an antioxidant or a prooxidant effect [14]. Its antioxidant effect (which would be responsible for its beneficial actions) seems to take place by means of proton donation from its phenolic group, which would inhibit the phenomenon of peroxidation of plasma lipoproteins, basically of the LDL [17–21]. The prooxidant effect generates OFR, fundamentally in such organs as the kidney, uterus, or breast, and seems to take place, among other mechanisms, by way of its own metabolism via the catecholestrogen pathway [22–24]. The now classic works of Liehr et al. [25] showed how the metabolism of estrogens and their hydroxylation by various isoenzymes of cytochrome P450 on carbons 2 and 4 convert them into catecholestrogens, which, through an oxidative metabolism which is reversible in nature, give rise to the formation of quinones and semiquinones capable of inducing a flow of free radicals in cytochrome P450. These radicals can lead to the possible formation of tumors via damage to DNA and proteins [15]. This mechanism seems to be regulated by certain antioxidant enzymes (superoxide dismutase (SOD) and catalase) and by the reduction of certain metal ions, such as Cu 2⫹ and Fe 3⫹ [2], which can trigger a cascade (Fenton’s reaction) leading to the production of more hydroxyl radicals. These in turn would increase on the one hand the hydroxylation of the adenine or guanine bases of DNA (recognized mutagenic events [27–29]) and on the other the damage to the unsaturated fatty acids of the membrane which would complete the vicious

EFFECT OF ESTRADIOL AND GESTAGENS ON UTERINE OXIDATIVE STRESS

FIG. 1.

Study design.

circle and give rise to the formation of lipid hydroperoxides (lipid peroxidation) [26]. Although the above reasons may explain the endometrial carcinogenic effect of estrogens when they are administered in isolation, it is not possible with current knowledge to understand the mechanism by which gestagens can turn this effect around and even diminish it [30], as there are only weak hints that seem to link this hormone with the phenomena of oxidative stress. In this sense, it has been related to the phenomenon of luteolysis [31, 32], implantation [33], macrophage response [34], sperm capacitation [35], etc. We therefore set out to study in the Wistar rat uterus the effect of the absence of steroid hormones (OVX) and the administration of estradiol (E2) alone or conjointly with medroxyprogesterone acetate (MPA) on the catalase antioxidant enzyme system and malonaldehyde (MDA) concentrations (an indicator of the lipid peroxidation phenomenon (LP)). MATERIALS AND METHODS

251

(a) After performing a laparotomy, we extirpated the upper half of the left uterine horn (basal state), proceeding subsequently to the OVX. (b) After the first partial hysterectomy and OVX, the animals were maintained for 15 days with no treatment, followed by a new laparotomy and extirpation of the remaining left uterine horn (OVX state). After closing the laparotomy, and before the anesthesia wore off, we shaved the animals’ lumbar region. (c) On the day following the second surgical procedure, we began the transdermal administration of 15 days of 17-␤estradiol (E2) (Oestraclin) at a dose of 8 ␮g/day. The application of each animal’s daily dose was performed at approximately the same time (10:00 AM). The calculation of the amount of gel to apply was made in accordance with the E2 concentration per gram of gel (2 mg E2/g gel). After the period of application of the E2 gel, we extirpated the upper half of the right hemi-uterus (ERT state). This was carried out after the application of the E2 dose corresponding to that day. (d) Finally, the animals continued with the E2 administration with the same doses and conditions, but associating with the E2, also on a daily basis, the administration of MPA at a dose of 20 ␮g/day orally. This administration was carried out using an oral-esophageal feeding tube. After the period of combined hormone replacement therapy (HRT), we extirpated the remaining right hemiuterus. At the end of each surgical procedure, we extracted 1 ml of blood from the tail vein. All the surgical procedures were performed by median line laparotomy with the animals under ip general anesthesia, with doses of ketamine, diazepan, and atropine given according to the animal’s weight. Nevertheless, given the mortality that was obtained in the second procedure, and which we shall give details of below, this dose was reduced by 1/3.

Animals

Methods of Analysis

We used 15 adult female Wistar rats, with an approximate weight of 250 g. The animals were exposed to constant cycles of light and darkness and fed a standard diet (Sander). The European Union norms on handling and treatment of laboratory animals were respected at all times.

The catalase activity was determined spectrophotometrically [36], with the results in micrograms per gram of uterine tissue. The MDA level was determined by high-performance liquid chromatography using the technique described by Esterbauer et al.[37]. The plasma E2 and PRG levels were determined by radioimmunoassay using the respective CIS Bio International kits (ORIS Group, Gif-Sur-Yvette-Cedex, France).

Experimental Design The design was aimed at observing in the same animal the effect of bilateral ovariectomy (OVX), E2 administration (ERT), and E2 ⫹ MPA administration (HRT) on the uterine variables catalase and MDA. The bicorn constitution of the rat uterus makes this a feasible design. The study was divided into four stages according to the protocol represented in Fig. 1. Uterine catalase and MDA and plasma E2 and progesterone (PRG) were determined at each stage.

Statistical Analysis The statistical analysis was performed using the Basic Statistics and Nonparametric Statistics modules of the statistical package STATISTICA, version 5 (1997). As basic descriptive measures of the different variables, we opted for the following: determining the value of the median (a central measure which, unlike the mean, is not affected by

´ MEZ-ZUBELDIA ET AL. GO

252

TABLE 1 Descriptive Analysis of the Variables in This Study Variable

Valid N

Median

Minimum

Maximum

Lower quartile

Upper quartile

Estradiol (basal) Estradiol (OVX) Estradiol (ERT) Estradiol (HRT) PRG (basal) PRG (OVX) PRG (ERT) PRG(HRT) Catalase (basal) Catalase (OVX) Catalase (ERT) Catalase (HRT) MDA (basal) MDA (OVX) MDA (ERT) MDA (HRT)

15 15 11 10 15 15 11 10 15 15 11 10 15 15 11 10

36.7 17.15 160.4 180.55 143.7 26.25 32.22 62.78 47.59 20.07 37.15 28.59 23.82 11.82 34.5 23.645

25.4 9.01 112.8 108.3 106.1 6.18 13.3 42.64 26.03 13.4 31.01 20.44 11.6 11.02 25.72 17.38

125.5 24.1 286.6 354.8 205.6 59.13 42.64 87.39 60.23 30.54 49.49 38.57 27.18 17.67 37.34 25.02

32.4 13.3 149 136.5 126.2 15.97 18.82 58.54 41.33 16.87 35.06 27.31 22.8 11.45 33.6 21.97

49.38 19.2 237.3 206.5 176.6 42.86 35.6 74.24 51.57 23.38 41.01 35.41 24.75 13.56 35.32 24.7

extreme data); the lower (25%) and upper (75%) quartiles, whose difference is often used as a measure of the dispersion; and the maximum and minimum values. Given the asymmetry observed in the distributions and the nature of the variables under consideration, together with the small size of the sample, we considered it statistically more appropriate to use nonparametric methods. In particular we performed four simultaneous Friedman tests to compare each of the four variables of interest (E2, PRG, catalase, and MDA) for the four states (basal, OVX, ERT, and HRT). We also planned to compare each of the four variables in a state with the same variable in the previous state (making a total of 12 comparisons). In these comparisons, we used the Wilcoxon test for paired samples at a maximum total significance level of 5%, for which we used the well-known Bonferroni method. In each comparison, we eliminated those cases in which any of the variables were unavailable. All the tests were considered significant for P ⬍ 0.05 and highly significant for P ⬍ 0.001. All the P values provided are two-sided. RESULTS Table 1 lists the values of the median, lower and upper quartiles, and maxima and minima of the four variables in the four states of the study, together with the number of animals making up each of those states. All of the variables (E2, PRG, catalase, and MDA) were found to be highly significant in their Friedman test comparisons over the course of the four states of the study (basal, OVX, ERT, and HRT). The magnitude of the P values obtained for each variable, together with the application of the Bonferroni correction [39] to adjust the overall error to 5%, allows us to confirm that the simultaneous test of the four null hypotheses was highly significant (Table 2).

As has been noted, to look for the causes of the significance, we performed multiple comparisons for each variable. Specifically, we compared each variable in each state with the same variable in the previous state. This made a total of 12 comparisons to which we applied the Wilcoxon test for paired samples, followed by the Bonferroni method for multiple comparisons (in the said comparisons we recovered some of the data that we had earlier eliminated because of having lost some of the animals). Table 3 lists the results. One observes that the Wilcoxon test in general establishes highly significant differences in practically all of the comparisons. Exceptions were the plasma E2 levels (states ERT vs HRT) and plasma PRG (states OVX vs ERT). Both cases were to be expected. The Wilcoxon test comparisons for the catalase and MDA variables were in all cases significant. This a priori reflects the possible influence on these uterine parameters of OVX, E2, and the joint administration of MPA and E2. Nonetheless, the Bonferroni correction, which in our case set the significance level at P ⫽ 0.00416, led to the loss of significance in the differences between states ERT and HRT for all the variables (see the “Decision” column of Table 3). TABLE 2 P Values of the Friedman Tests (Before and After Bonferroni Correction)

Friedman test for Estradiol Progesterone Catalase MDA

Individual P value P P P P

⬍ ⬍ ⬍ ⬍

0.00001 0.00002 0.00000 0.00001

P values after Bonferroni correction P P P P

⬍ ⬍ ⬍ ⬍

0.00012 0.00024 0.00012 0.00012

253

EFFECT OF ESTRADIOL AND GESTAGENS ON UTERINE OXIDATIVE STRESS

TABLE 3 Planned Comparisons: Wilcoxon Matched Pairs ⴙ Bonferroni

Estradiol

PRG

Catalase

MDA

H0

N

Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Conclusions

15 11 10 15 11 10 15 11 10 15 11 10

Figure 2 shows the values of the median, lower and upper quartiles, and maxima and minima of the four variables studied in the four states of interest. We show this figure because we consider that this descriptive display of the results allows one to see how the four variables evolved over the course of the study and subsequently to understand some of the results. Thus, in Fig. 2 one observes that the MDA values that most conformed to the study’s intermediate range were obtained in the basal state, in which the plasma E2 and PRG concentrations must be regarded as physiological. It is noteworthy that the catalase activities in this state were the highest and most scattered of the study, coinciding in this with the PRG concentrations which were also the highest and had the greatest dispersion. As was to be expected, ovariectomy led to the E2 and PRG concentrations being significantly lower (P ⬍ 0.00065) than in the basal state. These lower hormone concentrations were accompanied by a very significant decline in catalase activity and MDA (P ⬍ 0.0065 and P ⬍ 0.0008, respectively), presenting in both cases the lowest values of the whole study. As can be observed, there was also a reduction in the dispersion of the data for the four variables (very sharp in the case of E2 and moderate in the cases of PRG, catalase, and MDA). The transdermal administration of E2 to these OVX animals led to a very significant increase (P ⬍ 0.0033) in plasma E2 concentrations. They reached values between 120 and 280 pg/ml. This major increase in plasma E2, unaccompanied by any increase in plasma PRG concentration, led in our animals to a very significant (P ⬍ 0.0033) increase in MDA, despite the also significant (P ⬍ 0.0033) increase in catalase activity. The association of MPA with ERT induced a marked, but not significant (P ⫽ 0.005065), increase in the PRG concentration. This increase was accompanied by a decline in MDA to values similar to those found in the basal state. Nonetheless, while the decline was significant (P ⫽ 0.005065) in the

P

␣ ⫽ 0.05/12

0.000656 0.0041666 0.003348 0.0041666 0.878483 0.0041666 0.000656 0.0041666 0.476912 0.0041666 0.005065 0.0041666 0.000656 0.0041666 0.003348 0.0041666 0.005065 0.0041666 0.000806 0.0041666 0.003348 0.0041666 0.005065 0.0041666 Estradiol: Basal ⫽ OVX ⫽ ERT ⫽ HRT Progesterone: Basal ⫽ OVX ⫽ ERT ⫽ HRT Catalase: Basal ⫽ OVX ⫽ ERT ⫽ HRT MDA: Basal ⫽ OVX ⫽ ERT ⫽ HRT

Decision Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT Basal ⫽ OVX OVX ⫽ ERT ERT ⫽ HRT

Wilcoxon test, it was not so after the Bonferroni correction. There was also an increase in catalase activity relative to the ERT state which, while significant in the Wilcoxon test, was no longer significant after the Bonferroni correction. DISCUSSION We believe that the method used with respect to the uterus is of great interest since it allows each animal to be its own control within the study. We believe that this can improve the reliability of the results since it avoids the problem of variability between individuals. This type of method is not, however, exempt from potential limitations in “in vivo” studies. In the present case, the most important potential limitation is the size of the sample, especially in the OVX state where the uterus is in its natural state of atrophy. This makes a more detailed study of the oxidation–antioxidation phenomenon impossible since it forces a selection of the analytical variables to be measured. In this sense, and in the light of the findings of an earlier work [38] in which we found correlations only between MDA, uterine catalase activity, and the plasma E2 concentration, we here decided to measure only these two uterine variables. Nonetheless, as we shall see below, it would have been interesting to measure the SOD activity. With respect to the animal model used, two of the workers with greatest impact in this field, Liehr and Yager, defend the rat (although not expressly the rat uterus) as a model with which to study carcinogenic processes induced by steroid hormones (estrogens and androgens). Also, the rat uterus is an organ that is affected by the action of steroid hormones (estrogens and gestagens), and while carcinogenic processes are not provoked as often as in the mouse, we consider that it indeed may serve as a model for studying the action of steroid hormones.

´ MEZ-ZUBELDIA ET AL. GO

254

FIG. 2.

Box and whisker plots for E2, PRG, catalase, and MDA.

We used the variable plasma concentration because we consider it to be more reliable than the administered dose since it obviates the possible interference effects that might exist in the absorption of the two hormones due to the route of administration.

There were two reasons for using 17-␤-estradiol in our study: (a) because it is the natural hormone, it is the best indicator of the effect of this hormone in both physiological and pathological situations; and (b) because there is currently

EFFECT OF ESTRADIOL AND GESTAGENS ON UTERINE OXIDATIVE STRESS

a manifest tendency in our professional milieu to use this natural hormone in both ERT and HRT. With respect to the gestagen used, we chose MPA as it is the preparation that enjoys the widest acceptance in our milieu. We are currently, however, performing trials in which we are administering oral doses of natural progesterone which we consider, like 17-␤-estradiol, to be more physiological. One observes that the final number of animals that terminated the study was reduced from 15 to 10. This reduction was due to the high intraoperative mortality during the second surgical procedure (4 animals) due, we believe, to processes of respiratory insufficiency provoked by the dose of anesthetic that was administered. We reduced the dose by a third, and no further intraoperative deaths occurred. The fifth animal died while under HRT treatment, and we do not know the cause of death despite having performed a necropsy. In view of the above, we eliminated from each study those cases in which the values of any of the variables were unavailable. The 15 original animals were thus reduced to 10. This number makes it advisable to regard the P values obtained and, in general, the conclusions of our study as a first relative approximation. A possible criticism that can be directed at our study is that the plasma E2 concentration (except in the OVX state) presents a marked dispersion of the values determined for the different specimens studied. This dispersion was also important in the case of the plasma PRG. In this latter case, however, it was most notable in the basal state, when our influence would have been minimal. We consider that this dispersion of the values of the E2 and PRG concentrations may be explained on the basis of two facts: (a) In the basal state, we did not take into account which point of the cycle each animal was at. While this a priori seems to be a drawback, we consider that it is not lacking in advantages since it allows us to observe what occurs physiologically in the female rat relative to uterine oxidative stress when the moment in the cycle is random, i.e., with very varied concentrations of the two hormones. Observe that this dispersion in the E2 and PRG values at the basal time might explain the dispersion found in the values of MDA and uterine catalase, also in the basal state. (b) The method of E2 administration (transdermically) is perhaps not the most suitable for studying lipid peroxidation in the rat uterus nor for the working methods that we applied, since, despite the care we took in calculating the daily doses, this method is probably subject to major individual variations in the dermal absorption capacity. It is possible that, taking into account that what we wanted to observe was the action of 17-␤-estradiol rather than any other estrogenic compound, the application of E2 implants would have been a better way of applying the hormone. Nonetheless, we consider that neither is this method free of drawbacks, such as, among others, the progressive alteration over time of the daily dose administered as a consequence of the also progressive decrease in the implant’s exchange surface. The dispersion is less notable in the case of the oral administration of MPA. In this case, although there was also a

255

dispersion of the values that were obtained, it was considerably less intense than that found in the basal state and was roughly similar in magnitude to that found in the states OVX and ERT, when we were not administering MPA. Nevertheless, despite being aware of the aforementioned limitations of the study, we consider that there exist some tendencies in our results that we believe may have a certain interest. Thus, the MDA and catalase activity results found in the OVX and ERT states corroborate the findings of other workers [14, 15, 22, 24] that 17-␤-estradiol in certain organs (the uterus in our case) behaves as a prooxidant molecule that induces an increase in oxidative stress. One of the possible pathways for this increased oxidative stress is probably the molecule’s metabolism itself (transformation into catecholestrogens and subsequently into quinones and semiquinones) which leads to an increase in OFR production. If these are not neutralized, they will induce an increase in the phenomenon of lipid peroxidation (raised MDA concentrations) and a possible action on the bases of the DNA leading to its mutation [26]. Indeed, the relationship between the increase in free radicals (hydroxyl basically) and some tumoral processes has been amply described: Mobley et al. [40], using 8-oxo-2⬘-deoxyguanosine as a genotoxic marker of oxidative damage, found that both 2-hydroxyestradiol and 4-hydroxyestradiol increased oxidative stress, but that damage resulted only at concentrations above physiological levels (⬎10 ␮M), and the presence of exogenous antioxidants such as glutathione, SOD, and catalase reduced to a great extent the DNA damage induced by these high concentrations of 2-hydroxyestradiol. In this sense the study corroborated the findings of other workers [41, 42]. Liehr and Ricci [43] found that a high concentration of 4-hydroxyestradiol in human breast tissue induces the formation of the hydroxyl radical, while other workers [44, 45] found an increase of the concentrations of the bases hydroxyguanine and hydroxyadenine in breast cancers with respect to controls, thereby demonstrating a major activity of the hydroxyl radical. Likewise, there has also been found in breast cancers an increase in products of the decomposition of lipid peroxides, such as the case of MDA [46]. Liehr et al. [47], working with human uterine myometrium, implicated the catecholestrogens, especially 4-OH-E2, in the process of initiation and promotion of uterine myoma. Similarly, the also highly significant increase found at that stage of the study (ERT vs OVX) in catalase activity may indicate that this lipid peroxidation phenomenon is modulated by, among other factors, the activity of the enzyme. Indeed, in an earlier study [38] also carried out on the rat uterus, we found that there is a highly significant linear correlation between the parameters uterine MDA, plasma E2, and uterine catalase activity. This correlation was positive between MDA and plasma E2 and negative between MDA and catalase activity. The divergent results found by different workers [10, 38, 48 –50] for the activity of this enzyme when they administer

256

´ MEZ-ZUBELDIA ET AL. GO

estrogenic compounds to certain rodents and in different tissues suggest that this enzyme’s activity may be subject to possible enzymatic induction phenomena. The duration of treatment [51] the E2 concentrations reached [38], or the degree of prior activity of other enzymes [51], among other factors, may also be factors that explain the differences observed in the levels of catalase activity. Hence our results are in agreement with the evidence [52, 53] that implicates estrogens in the initiation of tumors in such tissues as the kidney, breast, or endometrium by means of the redox cycle of estrogen metabolism and the subsequent production of hydroxyl radicals. As was remarked above, after the application of the Bonferroni correction, we could not confirm that there were any significant differences between HRT and ERT for any of the four variables of the study. From a descriptive point of view, however, we observed that the association of MPA with ERT (which augments the plasma PRG concentrations to values in the range 40 –90 ng/ml) seems to modify both the response of catalase activity to the isolated estrogen stimulus and the lipid peroxidation phenomenon as measured by the uterine MDA concentration. In this sense, we consider that this nonsignificant increase in plasma PRG seems to induce, at levels of significance (P ⬍ 0.005065), a decline in uterine MDA to values similar to those found at the basal state. This decline occurs despite the decrease, also significant, in catalase activity (corrected P value of 0.06078). We therefore consider that it is not totally irrational to have the impression that the association of MPA with E2 may modify the oxidative response of the uterus to E2 administration. In this sense, the not very ample literature on the role of gestagens in oxidative stress phenomena also seems to support this idea, implicating both PRG in OFR production mechanisms and vice versa. With respect to the regulatory action of the progestagens in oxidation–reduction phenomena, Tranquilli et al. [54], working with different HRT regimes, found that the progestagen MPA induced a significant decline in peroxide production and concluded that the decline observed in LP during HRT was not due to the estrogens alone but also depended on the action of the gestagens. Laloraya et al. [33], working with the rat uterus, found at the moment of implantation (days 5 and 8) high SOD levels and low oxygen radical levels in PRG-treated animals. In contrast, the administration of E2 and PRG provoked an increase in the generation of oxygen radicals and a fall in SOD. This indicates that both estrogens and PRG regulate the generation of superoxide radicals by means of the SOD activity. Recent studies, such as that of Bekesi et al. [34], seem to show that the hormones E2, PRG, and testosterone induce in granulocytes a decrease in the production of superoxide radicals and that this inhibition may depend on the PRG concentration that is reached. In this sense, Sugino et al. [55], working with PRG concentrations of 10, 50, and 100 ng/ml, found that

the 100 ng/ml concentration is the only one that inhibited superoxide radical production significantly. These works may therefore corroborate our impression that we discussed above that the dose we administered in the present work was not sufficient to produce differences that might have endowed the tendencies observed in the catalase activity and uterine MDA with statistical significance relative to the previous state. They are also coherent with recent clinical findings [56 –58] that there is an increased risk of endometrial adenocarcinoma when estrogens are administered chronically at low doses (0.3 mg/day), and even despite the addition of a gestagen if this is administered at low doses or cyclically [59]. Finally, it has already been noted that the relationship between PRG and OFR is not limited to the influence of the hormone, but that there also seems to be a relationship between OFR production and the mechanisms of regulation of PRG secretion. Thus Sawada and Carlson [32] found that LH regulates PRG secretion by the corpus luteum via the secretion of superoxide radicals (low levels raise PRG secretion and high levels lower it). This mechanism is also regulated by the SOD and catalase enzymes. These findings have been supported by the results of other workers [31, 33]. We should thus state that, under the working conditions of the present study (with the MPA dose that was administered and the small number of animals), we were unable to demonstrate categorically that the association of MPA produced significant modifications in the levels of plasma PRG or uterine MDA and catalase relative to the isolated administration of estrogen alone. It should be noted, however, that applying the Bonferroni correction guarantees an overall significance level of less than 0.05 (but not exactly equal to 0.05) and that this implies a relative loss of power and makes the acceptance of the null hypothesis even more questionable. Nevertheless we consider, despite our results not being statistically conclusive, that the gestagens do seem to play a role in oxidation–reduction phenomena. CONCLUSIONS In this work we have corroborated the findings that estrogens induce an increase in uterine oxidative stress which gives rise to an increase in the phenomenon of lipid peroxidation (higher MDA levels). This increased lipid peroxidation is partially modulated by the catalase activity. The results also support the view that the gestagens (in our case MPA) seem, although we cannot confirm it categorically, to intervene by diminishing this oxidative stress. This influence may be mediated via a possible increase in catalase activity or through other mechanisms that we did not measure, such as the SOD activity or by a decrease in superoxide radical production. We therefore consider that, in the light of our results and those of other workers, it would be advisable to perform both in vivo and in vitro studies to shed light on the role of gestagens

EFFECT OF ESTRADIOL AND GESTAGENS ON UTERINE OXIDATIVE STRESS

in the phenomenon of lipid peroxidation. This would further our knowledge of the mechanisms involved in the preventive action of gestagens on estrogen-dependent endometrial pathologies.

22. 23.

REFERENCES 1. Halliwell B, Gutteridge JM. Free radicals in biology and medicine. New York: Oxford Univ. Press, 1985:279 –313.

24.

2. Hu JJ, Dubin N, Kurland D, Ma BL, Roush GC. The effects of hydrogen peroxide on DNA repair activities. Mutat Res 1995;336:193–201.

25.

3. Stahl W, Sies H. Antioxidant defense: Vitamins E and C and carotenoids. Diabetes 1997;46(2):S14 – 8.

26.

4. Johnson IT. Antioxidants and anticarcinogens. Eur J Cancer 1998;7(2): S55–S62. 5. Dreher D, Junod AF. Role of oxygen free radicals in cancer development. Eur J Cancer 1996;32A(1):30 – 8.

27.

6. Han X, Liehr J. Microsome-mediated 8-hydroxylation of guanine bases of DNA by steroid estrogens: correlation of DNA damage by free radicals with metabolic activation to quinones. Carcinogenesis 1995;16(10):2571– 4. 7. Vendrell FL, Ten J, Cano A, Tarin JJ. Protective effect of antioxidants in vitro and in vivo against oxidative damage to DNA and/or chromosomes. Prog Obstet Gynecol 1998;41:471– 8. 8. Asada Y, Kumura S, Ohishi N, Kasugai M, Suganuma N, et al. Effect of ovariectomy on serum lipids peroxide levels in women. J Clin Biochem Nutr 1990;8:247–52. 9. Holvoet P. Endothelial dysfunction, oxidation of low-density lipoprotein, and cardiovascular disease. Ther Apher 1999;3(4):287–93. ¨ llinger K. Impact of oestradiol and proges10. Dabrosin CH, Hammar M, O terone on antioxidant activity in normal human breast epithelial cells in culture. Free Rad Res 1998;28:241–9.

28.

29.

30.

11. Roy D, Liehr JG. Estrogen, DNA damage and mutations. Mutat Res 1999;8(424, 1/2):107–15.

31.

12. Barth A, Landmann G, Liepold K, Zapf H, Muller D, et al. Influence of oestrogens on formation of reactive oxygen species in liver microsomes of differently aged male Wistar rats. Exp Toxicol Pathol 1999;51(4/5): 282– 8.

32.

13. Huh K, Shin US, Choi JW, Lee SI. Effect of sex hormones on lipid peroxidation in rat liver. Arch Pharm Res 1994;17(2):109 –14.

33.

14. Markides C, Roy D, Liehr J. Concentration dependence of prooxidant and antioxidant properties of catecholestrogens. Arch Biochem Biophys 1998; 360(1):105–12.

34.

15. Nathan L, Chaudhuri G. Antioxidant and prooxidant actions of estrogens: potential physiological and clinical implications. Semin Reprod Endocrinol 1998;16(4):309 –14.

35.

16. Schwenke DC. Aging, menopause, and free radicals. Semin Reprod Endocrinol 1998;16(4):281–308. 17. Ro¨ mer W, Oettel M, Droescher P, Schwarz S. Novel “scavestrogens” and their radical scavenging, and total antioxidative activities: dehydroderivatives of 17 estradiol and 17 estadiol. Steroids 1997;62:304 –10.

36. 37.

18. Ruiz-Larrea MB, Leal AM, Martin C, Martinez R, Lacort M. Antioxidant action of estrogen in rat hepatocytes. J Physiol Biochem 1997;53(2):225– 30.

38.

19. Sugioka K, Shimosegawa Y, Nakano M. Estrogens as natural antioxidant of membrane phospholipid peroxidation. FEBS Lett 1987;210:37–9. 20. Huber LA, Sheffler E, Poll T, Ziegler R, Dresel HA. 17-beta-estradiol inhibits LDL oxidation and cholesteryl ester formation in cultured macrophages. Free Rad Res Comun 1990;8:167–73. 21. Sack MN, Rader DJ, Cannon RO, III. Oestrogen and inhibition of oxida-

39. 40.

257

tion of low-density lipoproteins in postmenopausal women. Lancet 1994; 343:269 –70. Yager JD, Liehr JG. Molecular mechanims of estrogen carcinogenesis. Pharmacol Toxicol 1996;36:203–32. Kuhn-Velten WN, Pippirs U. Novel connections between NADPH-induced lipid peroxidation and cytochrome P450 inactivation, and antioxidant and enzyme prospective properties of estradiol in gonadal membranes. Free Rad Res 1997;26:125–33. Wyllie S, Liehr J. Release of iron from ferritin storage by redox cycling of stilbene and steroid estrogen metabolites: a mechanism of induction of free radical damage by estrogen. Arch Biochem Biophys 1997;346(2):180 – 6. Liehr J, et al. Cytocrome P450-mediated redox cycling of estrogens. J Biol Chem 1986;261(36):16865–70. Wang MY, Liehr JG. Identification of fatty acid hydroperoxide cofactors in the cytochrome P450-mediated oxidation of estrogens to quinone metabolites: role and balance of lipid peroxides during estrogen-induced carcinogenesis. J Biol Chem 1994;269:284 –91. Kamiya H, Miura H, Murata-Kamiya N, Ishikawa H, Sakaguchi T, Inoue H, et al. 8-Hydroxyadenine (7,8 dihydro-8-oxoadenine) induces misincorporation in in vitro DNA synthesis and mutations in NIH 3T3 cells. Nucleic Acids Res 1995;23(15):2893–9. Kuchino Y, Mori F, Kasai H, Inoue H, Iwai S, Miura K, Ohtsuka E, Nishimura S. Misreading of DNA templates containing 8-hydroxyadeoxyguanosine at the modified base and at adjacent residues. Nature 1987;7– 13;327(6117):77–9. Cheng KL, Cahill DS, Kasai H, Nishimura S, Loeb LA. 8-Hydroxyguanine, an abundant form of oxidative DNA damage causes G–T and A–C substitutions. J Biol Chem 1992;5;267(1):166 –72. Piegsa K, Calder A, Davis JA, McKay-Hart D, Wells M, Bryden F. Endometrial status in post-menopausal women on long-term continuous combined hormone replacement therapy (Kliofem). A comparative study of endometrial biopsy, outpatient hysteroscopy and transvaginal ultrasound. Eur J Obstet Gynecol Reprod Biol 1997;72(2):175– 80. Sawada M, Carlson JC. Studies on the mechanism controllong generation of superoxide radical in luteinized rat ovaries during regression. Endocrinology 1994;135(4):1645–50. Sawada M, Carlson JC. Intracellular progesterone secretion by the superoxide radical in the rat ha´ beas luteum. Endocrinology 1996;137(5): 1580 – 4. Laloraya M, Jain S, Thomas M, Kopergaonkar S, Pradeep Kumar G. Estrogen surge: a regulatory switch for superoxide radical generation at implantation. Biochem Mol Biol Int 1996;39(5):993– 40. Bekesi G, Kakucs R, Varbiro S, Racz K, Sprintz D, Feher J, Szekacs B. In vitro effects of different steroid hormones on superoxide anion production of human neutrophil granulocytes. Steroids 2000;65(12):889 –94. de Lamirande E, Harakat A, Gagnon C. Human sperm capacitation induced by biological fluids and progesterone, but not by NADH or NADPH, is associated with the production of superoxide anion. J Androl 1998;19(2):215–25. Aebi H. Catalase in vitro. Methods Enzymol 1984;105:121– 6. Estenbauer H, Lang J, Zadrarec S, Slater TF. Detection of malonaldehyde by high-performance liquid chromatography. Methods Enzymol 1984; 105:319 –28. Go´ mez-Zubeldia MA, Hinchado G, Arbue´ s JJ, Nogales AG, Millan JC. Influence of estradiol on oxidative stress in the castrated rat uterus. Gynecol Oncol 2001;80:227–32. Fleiss. The design and analysis of clinical experiments, New York: Wiley, 1986. Mobley JA, Bhat AS, Bmeggemeier RW. Measurement of oxidative DNA damage by catechol estrogens and analogues in vitro. Chem Res Toxicol 1999;Mar;12(3):270 –7.

258

´ MEZ-ZUBELDIA ET AL. GO

41. Nutter LM, Wu YY, Ngo EO, Sierra EE, Gutierrez PL, Abul-Hajj YJ. An o-quinone form of estrogen produces free radicals in human breast cancer cells: correlation with DNA damage. Toxicology 1994;Jan–Feb;7(1): 23– 8. 42. Wyllie S, Liehr JG. Release of iron from ferritin storage by redox cycling of stilbene and steroid estrogen metabolites: a mechanism of induction of free radical damage by estrogen. Arch Biochem Biophys 1997;346(2): 180 – 6. 43. Liehr JG, Ricci MJ. 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci USA 1996;93:3294 – 6. 44. Malins DC, Haimanot R. Major alterations in the nucleotide structure of DNA in cancer of the female breast. Cancer Res 1991;51:5430 –2. 45. Malins DC, Holmes EH, Polissar NL, Gunselman SJ. The etiology of breast cancer. Cancer 1993;71:3036 – 43. 46. Wang MY, Dhingra K, Hittelman WN, Liehr JG, de Andrade M, Li D. Lipid peroxidation induced putative malondialdehyde-DNA adducts in human breast tissues. Cancer Epidemiol Prev 1996;5:705–10. 47. Liehr JG, Ricci MJ, Jefcoate CR, Hannigan Ev, et al. 4-Hydroxylation of estradiol by human uterine myometrium and myoma microsomes: implications for the mechanism of uterine tumorogenesis. Proc Natl Acad Sci USA 1995;92:9220 – 4. 48. Sreelathakumari KT, Menon VP, Leelamma S. Lipid peroxide metabolism in oophorectomised rats. Ind J Med Res 1993;Dec;98:305– 8. 49. Pajovic SB, Saicic ZS, Spasic MB, Petrovic VM, Martinovic JV. Effects of progesterone and estradiol benzoate on glutathione dependent antioxidant enzyme activities in the brain of female rats. Gen Physiol Biophys 1999;Mar;18(1):35– 44. 50. Singh D, Pandey RS. Changes in catalase activity and hydrogen peroxide level in rat ovary during estrous cycle and induction of catalase in rat ovary by estradiol-17 beta. Ind J Exp Biol 1998;Apr;36(4):421–3.

51. McCormick ML, Oberley TD, Elwell JH, Oberley LW, Sun Y, Li JJ. Superoxide dismutase and catalase levels during estrogen-induced renal tumorigenesis, in renal tumors and their autonomous variants in the Syrian hamster. Carcinogenesis 1991;Jun;12(6):977– 83. 52. Sarabia SF, Liehr JG. Induction of monoamine oxidase B by 17-estradiol in the hamster kidney preceding carcinogenesis. Arch Biochem Biophys 1998;355,2:249 –52. 53. Yoshie Y, Ohshima H. Synergistic induction of DNA strand breakage by cathechol– estrogen and nitric oxide: implications for hormonal carcinogenesis. Free Rad Biol Med 1998;24(2):341– 8. 54. Tranquilli Al, Mazzanti L, Cugini AM, Cester N, Garzetti GG, Romanini C. Transdermal estradiol and medroxyprogesterone acetate in hormone replacement therapy are both antioxidants. Gynecol Endocrinol 1995;9(2): 137– 41. 55. Sugino N, Shimamura K, Tamura H, Ono M, Nakamura Y, Ogino K, Kato H. Progesterone inhibits superoxide radical production by mononuclear phagocytes in pseudopregnant rats. Endocrinology 1996;137(2):749 –54. 56. Gruber DM, Wagner G, Kurz C, Sator MO, Huber JC. Endometrial cancer after combined hormone replacement therapy. Maturitas 1999;Mar;15; 31(3):237. 57. Cushing KL, Weiss NS, Voigt LF, McKnight B, Beresford Sam Hutchinson F. Risk of endometrial cancer in relation to use of low-dose, unopposed estrogens. Obstet Gynecol 1998;91(1):35–9. 58. Weiderpass E, Baron JA, Adami HO, Magnusson C, Lindgren A, Bergstrom R, Correia N, Persson I. Low potency estrogen and risk of endometrial cancer: a case-control study. Lancet 1999;May;29;353(9167):1824– 8. 59. Weiderpass E, Adami HO, Baron JA, Magnusson C, Bergstrom R, Lindgren A, Correia N, Persson I. Risk of endometrial cancer following estrogen replacement with and without progestins. J. Natl Cancer Inst 1999;Jul;7;91(13):1131–7.