Role of sex steroids in modulating tumor necrosis factor alpha induced changes in vascular function and blood pressure

Introduction

Preeclampsia is defined as new onset hypertension with proteinuria during pregnancy. Blood pressure typically increases during the third trimester and remits after delivery of the placenta. Increased arterial pressure during preeclampsia is accompanied by elevated circulating inflammatory cytokines such as TNF-α, IL-1 and IL-6 [1;2]. Previously, our laboratory demonstrated that a two-fold increase in circulating TNF-α in a normal pregnant rat is sufficient to cause hypertension and increased renal vascular resistance [3;4]. Importantly, this renal hemodynamic and blood pressure response occurs only in pregnant, but not virgin rats suggesting that placental factors or the hormonal environment of pregnancy may be required for cytokine mediated hypertension during pregnancy in the rat. Currently, the physiological mechanisms that facilitate enhanced blood pressure sensitivity to TNF-α during pregnancy are unknown. One possible mechanism may be related to the increased circulating levels of sex steroids. The role of sex steroids in promoting changes in vascular function and blood pressure regulation remains a controversial area of investigation. The physiological response to estrogenic hormones varies between animal models, vascular beds, and with concentration. While a large body of evidence points to a vascular protective and anti-hypertensive role for ovarian hormones, others suggest a potential pro-inflammatory role for both estrogens and progestans. For example, a recent report shows that 17β-estradiol facilitates TNF-α induced leukocyte adhesion to endothelial cells by increasing adhesion molecule expression [5]. Whether the increase in blood pressure caused by TNF-α in pregnant rats requires elevated estrogen and progesterone is not clear. In order to examine this, we asked whether administration of sex steroids (estrogen and/or progesterone) to mimic levels observed during pregnancy would be sufficient to promote TNF-α induced hypertension and vascular dysfunction in ovariectomized (OVX) rats.

Methods

Results

We have previously reported that infusion of TNF-α in late pregnant rats causes an increase in blood pressure in pregnant but not virgin rats [3;4;11]. The results showed a 27 mmHg increase in blood pressure caused by a two-fold increase in serum TNF-α levels [4]. In the present study, OVX rats were subjected to a 5 day infusion of TNF-α that increased serum levels of TNF-α from 4.5 ± 0.3 pg/ml to 52.5 ± 18.5 pg/ml. The increased TNF-α did not result in elevated MAP (126± 4 vs. 125± 5 mmHg) (Figure 1). Serum levels of estrogen in OVX and OVX rats with TNF-α were 62±12 and 49±16 pg/ml, respectively. Serum progesterone levels were 37±6 and 41±6 ng/ml, in OVX rats or OVX rats with TNF-α.

We next asked whether the presence of increased estrogen is necessary to mediate TNF-α induced hypertension. 17β–estradiol was administered to OVX rats for a period of seventeen days to mimic levels observed during pregnancy in the rat. TNF-α or vehicle was administered for five days beginning on day 12. Plasma estrogen levels were significantly elevated in 17β-estradiol OVX rats treated with either vehicle (89 ± 16 pg/ml) or TNF-α (146 ± 10 pg/ml). In rats administered 17β-estradiol, progesterone levels were not different from control (20±7 and 56±10 ng/ml compared to OVX and OVX + TNF-α, respectively) OVX rats infused with TNF-α and treated with 17β-estradiol had elevated serum TNF-α levels (49 ± 9 pg/ml compared to vehicle 7.2 ± 2 pg/ml) but this did not significantly increase mean arterial pressure (122 ± 7 mmHg vs. 111 ± 4 mmHg) (Figure 1).

In a similar study OVX rats were administered progesterone over the seventeen days at levels to mimic pregnancy. Progesterone administration increased serum levels in OVX rats (75±9 ng/ml) and in OVX rats infused with TNF-α (99±11 ng/ml). 17β-estradiol levels were not different from OVX (44±8 pg/ml) or OVX with TNF-α (53±10 pg/ml). TNF-α infused animals had higher serum TNF-α levels compared to vehicle treated controls (30.8 ± 7.1 pg/ml vs. 1.8 ± 0.2 pg/ml); however, MAP was not significantly altered (129 ± 4 mmHg OVX + progesterone vs. 130 ± 4 mmHg OVX + progesterone + TNF-α) (Figure 1).

Finally, we tested whether the presence of both progesterone and estrogen at levels to mimic pregnancy were required to cause hypertension during the infusion of TNF-α. OVX rats administered both 17β-estradiol and progesterone had significantly increased serum levels (99±8 pg/ml and 78±4 ng/ml, respectively). Similarly, TNF-α infused OVX rats receiving both 17β-estradiol and progesterone had significantly increased serum levels (114±11 pg/ml and 108±13 ng/ml, respectively). Although serum TNF-α levels were increased following infusion (3.0±0.04 pg/ml vs. 48.5 ± 19.4 pg/ml), again blood pressure was not increased by TNF-α (118 ± 3 mmHg vs.123 ± 4 mmHg) (Figure 1).

In addition to examining the interaction of sex steroids with TNF-α on blood pressure, we also determined whether there were any changes in vascular function. Increased TNF-α has been shown in numerous studies to alter endothelial dependent relaxation as well as vascular contraction. Concentration dependent relaxation to acetylcholine was not different between any of the experimental groups (Figure 2) suggesting that the five day infusion of TNF-α in OVX rats did not cause endothelial dysfunction. Similarly, relaxation responses to sodium nitroprusside were not impaired with infusion of TNF-α and sex steroid administration did not affect this response (Figure 2).

Because our earlier data showed that TNF-α infusion increased vascular contraction in pregnant but not virgin rats, we examined whether vascular contractility was altered by TNF-α or sex steroids. Our data show that there is a modest but significant increase in the contractile response to maximal concentrations of phenylephrine in the OVX rats infused with TNF-α when compared to OVX rats (Figure 3). No other differences in vascular function were detected among the other experimental groups.

Although there was no change in endothelial function after the five day infusion of TNF-α, it is possible that TNF-α infusion could increase cell adhesion molecule expression as an early marker of endothelial changes that could progress to vascular damage. Therefore, we tested whether soluble ICAM was increased in control and TNF-α infused rats and whether the presence of either estrogen or progesterone would alter levels of soluble ICAM. OVX rats infused with TNF-α had significantly increased soluble ICAM levels compared to control OVX rats (0.9±0.2 ng/ml, n=26 vs. 10391±536 ng/ml, n=9). Treatment with TNF-α and progesterone (10144±1549 ng/ml, n=4), estrogen (9474±1229 ng/ml, n=6), or both (11394±1267 ng/ml, n=8) did not significantly alter levels of soluble ICAM in this study when compared to OVX rats infused with TNF-α alone.

In a previous study we demonstrated that TNF-α induced hypertension in pregnant rats is due to a TNF-α mediated increase in endothelin-1. Therefore, we measured the expression of renal preproendothelin using quantitative real time PCR. The results show that there is no difference in preproendothelin expression among any of the experimental groups (Figure 4).

Discussion

Preeclampsia affects 5-8% of all pregnancies in the United States and is the most common cause for maternal and fetal mortality. While it is generally accepted that preeclampsia results from improper remodeling of the uterine spiral arteries leading to a reduction in uterine perfusion pressure, the mechanisms that account for this are not well understood. As a result, to date, the only effective treatment for preeclampsia is delivery of the fetus after which hypertension and proteinuria remit.

We have recently begun to uncover some of the pathophysiological mechanisms that are involved in preeclampsia using the reduced uterine perfusion pressure model in the rat (RUPP). In addition to elucidating a role for the renin angiotensin system [12] and endothelin [11;13] in pregnancy induced hypertension, we and others have recently demonstrated an important role for inflammatory cytokines [1-4;14-18]. During preeclampsia, the placenta has increased immune infiltrates [19] and levels of circulating TNF-α are elevated greater than two-fold. Similar to humans, the RUPP model also has a two-fold increase in TNF-α [1;20-22] suggesting it as a putative mechanism promoting hypertension during pregnancy. Furthermore, we have shown that if TNF-α is infused at a rate to mimic the increases reported during preeclampsia, pregnant rats, but not virgins, become hypertensive and have increased vascular contractility [3;4].

Based on these earlier studies, we asked whether ovarian sex steroids are required to cause hypertension in TNF-α infused female rats. The major new finding of this study is that ovarian hormones (17β-estradiol and/or progesterone) at levels to mimic pregnancy are not sufficient in combination with TNF-α to cause hypertension. These data do not completely exclude the possibility that the combination of ovarian hormones and TNF-α induce hypertension, but rather suggest at minimum, that some other factor, perhaps placental in origin, is required to cause the hypertension. This may explain why in pregnant rats, infusion of TNF-α results in increased pressure while infusion of TNF-α in the present study does not. One possible candidate promoting TNF-α induced hypertension in pregnant rats is endothelin. Previously published work from our laboratory shows that TNF-α induced hypertension in pregnant rats is prevented by blockade of the endothelin system [11;13]. However, in the present study there was no difference in renal expression of preproendothelin despite increased serum TNF-α concentrations. These data further support the likelihood that factors in addition to TNF-α and sex steroids are required to cause hypertension during pregnancy.

In addition to blood pressure, we examined vascular responsiveness in TNF-α infused OVX rats. Our data shows no change in relaxation responses to acetylcholine or sodium nitroprusside and only a moderate increase in the contractile response to phenylephrine. The latter finding varies slightly from our earlier report showing that only pregnant rats infused with TNF-α have increased contractile responses [23]. One possible explanation for this modest discrepancy is that the circulating levels of TNF-α achieved in the present study are higher than the earlier experiments (33.9±4.6 pg/ml vs. 13.5±1.8 pg/ml). However, we believe this is unlikely since more recent experiments from our laboratory demonstrated that serum TNF-α levels as high as 73.6±11 pg/ml still caused hypertension in pregnant rats [24]. In all of these studies, the levels of TNF-α achieved are substantially lower than what is expected for an acute inflammatory reaction [25-27]. Perhaps a more likely possibility is that the experimental set up to examine vascular function between experiments differs slightly. For example, the current study examined carotid vascular rings instead of aortic strips and tested a higher maximal concentration of phenylephrine (10^-4 M versus 10^-5 M). This increased phenylephrine mediated contraction was only evident at the highest concentrations. Although we recognize that conduit vessels may not play a major role in blood pressure regulation, it should be noted that there is a direct correlation between carotid artery hemodynamics and blood pressure in humans [28].

The role of ovarian hormones in the development of cardiovascular disease has been somewhat controversial. A large majority of studies show that estrogens have vascular protective effects and contribute to lower blood pressures in pre-menopausal females. However, there is accumulating evidence to suggest that estrogens also have pro-inflammatory properties. For example, increased serum levels of the acute phase reactant C-reactive protein have been reported by numerous studies in women taking estrogen [29]. Proinflammatory effects of ovarian hormones are also supported by in vitro studies [5]. Experiments in human umbilical vein endothelial cells show that estrogen and progesterone promote increased leukocyte binding to TNF stimulated cells by increasing the expression of cell adhesion molecules including E-selectin, intercellular adhesion molecule type 1 (ICAM-1) and vascular cell adhesion molecule (VCAM-1). Our results show that TNF-α increases serum levels of soluble ICAM; however, serum levels were not affected by estrogen and/or progesterone.

The current study has some potential limitations. For example, in order to be consistent with our earlier work, pressure was measured on the final day of the protocol rather than monitored temporally. Therefore, the possibility for undetected changes in pressure over the course of the study exists. However, this is unlikely since both TNF-α and sex steroid levels were increased at the time of sacrifice when pressure was measured. A second limitation is the variability in estrogen and progesterone levels achieved in the OVX rats. The reason for the difficulty in consistently regulating these levels was not apparent; however, hormone levels may have been affected by the infusion of TNF-α. Despite this difficulty, it is important to note that the levels of estrogen and progesterone achieved are consistent with those reported during pregnancy in the rat [6]. Therefore this experimental design adequately tests the hypothesis.

In conclusion, the data from this study suggest that sex steroid levels mimicking levels observed during pregnancy, by themselves, are not sufficient to play a role in TNF-α induced hypertension. Therefore it is likely that some other factor or factors, possibly produced by the placenta during pregnancy, is required to cause increases in blood pressure or is necessary to interact with sex steroids to cause hypertension. Future experiments will be important to determine these factors.