Tumor necrosis factor–α–induced hypertension in pregnant rats results in decreased renal neuronal nitric oxide synthase expression

Abstract

Background:

Preeclampsia is associated with increases in plasma levels of tumor necrosis factor–α (TNF-α), a cytokine known to contribute to endothelial dysfunction. We recently reported that a twofold elevation in plasma TNF-α produces significant reductions in renal function and hypertension in pregnant rats. The purpose of this study was to determine the role of the nitric oxide (NO) system in TNF-α–induced hypertension in pregnant rats.

Methods:

Tumor necrosis factor–α (50 ng/day) was chronically infused starting at day 14 of gestation. Mean arterial pressure, 24-h urinary nitrite/nitrate excretion, and renal nitric oxide synthase (NOS) protein expression by Western blot analysis was measured at day 19 of gestation.

Results:

A twofold increase in plasma TNF-α levels in pregnant rats resulted in a significant increase in arterial pressure (97 ± 3.6 v 116 ± 2.1 mm Hg, pregnant versus TNF-α pregnant, respectively, P < .05), but no significant change in urinary nitrite/nitrate excretion (22.0 ± 1.9 v 20.8 ± 2.5 μmol/24 h, pregnant versus TNF-α pregnant, respectively), a measure of whole body NO production. As abnormalities in renal production of NO would not be reflected in the measure of whole body NO production, changes in renal NOS protein levels were determined. The protein expression of both neuronal (nNOS) and inducible (iNOS) nitric oxide synthase were significantly decreased in the medulla of TNF-α pregnant rats (nNOS: 10.6 ± 0.7 v 8.2 ± 0.8 densitometric units, P < .05; and iNOS: 19.2 ± 0.9 v 15.4 ± 0.8 densitometric units, P < .05, pregnant versus TNF-α pregnant, respectively).

Conclusion:

The hypertension associated with a chronic twofold increase in TNF-α in pregnant rats is associated with significant decreases in renal nNOS and iNOS protein production.

Pregnancy-induced hypertension, preeclampsia, tumor necrosis factor–α, nitric oxide

Preeclampsia is a multisystemic disorder of pregnancy that is estimated to affect 5% to 10% of all pregnancies in the United States.^1,2 One hypothesis concerning the pathogenesis of preeclampsia involves inadequate trophoblast invasion of uterine spiral arteries, decreased placental perfusion, and endothelial dysfunction.^3,4 As improper remodeling of the spiral arteries leading to placental ischemia is an important initiating event, placental release of circulating factors, such as inflammatory cytokines, may serve as an important initiator of maternal endothelial activation and/or dysfunction.⁵ Tumor necrosis factor–α (TNF-α), an inflammatory cytokine capable of producing endothelial cell dysfunction, is one such factor that may be released from the placenta and impact the maternal circulation.⁶ Several lines of evidence support a potential role for TNF-α in mediating endothelial cell activation/dysfunction in preeclampsia. Both TNF-α mRNA and protein have been shown to be expressed by the placenta during normal pregnancy.^7,8 Furthermore, hypoxia has been shown to produce a twofold increase in TNF-α production in placental tissues placed as explants into culture.⁹ In women with preeclampsia, expression of placental cytokines, such as TNF-α, are significantly increased¹⁰ and plasma levels of TNF-α are elevated twofold.¹¹ Support for a role of cytokine mediated endothelial activation/dysfunction is suggested in the study by Rodgers et al, in which serum from preeclamptic women was shown to result in selective activation of endothelial cells.¹²

We recently reported that a twofold elevation in plasma TNF-α in pregnant rats was characterized by an increase in mean arterial pressure (MAP) and a decrease in both glomerular filtration rate (GFR) and renal plasma flow (RPF); features resembling preeclampsia.¹³ The purpose of this study was to examine the role of nitric oxide (NO) in mediating TNF-α–induced hypertension in pregnant rats. NO has been suggested to play an important role in mediating renal and systemic hemodynamic changes that occur during normal pregnancy.^14,15 The importance of NO in mediating renal and systemic changes during pregnancy is supported by studies in which inhibition of systemic NO synthesis attenuates the decrease in MAP and increases in RPF, GFR, and regional blood flow observed during normal pregnancy.^16,17 We and others have previously shown that during normal pregnancy in the rat, changes in renal and systemic hemodynamics are associated with increases in whole body NO production.^18,19 In addition, normal pregnancy is also associated with increases in renal protein expression of both neuronal and inducible nitric oxide synthase (NOS) isoforms.¹⁸ Therefore, as NO may play an important in normal pregnancy, NO deficiency may play an important role in preeclampsia. Thus, the overall goal of this study is to assess both whole body NO production and the renal NO production that are associated with the hypertension produced in response to a twofold elevation in plasma TNF-α.

Methods

All studies were performed in timed pregnant Sprague Dawley rats purchased from Harlan Sprague Dawley Inc. (Indianapolis, IN). Animals were housed in a temperature-controlled room (23°C) with a 12:12 h light:dark cycle. All experimental procedures executed in this study were in accordance with National Institutes of Health guidelines for use and care of animals and all protocols were approved by the Animal Care and Use Committee at the University of Mississippi Medical Center.

Experimental design

A mini-osmotic pump (model 2002, Alza Scientific Corporation, Palto Alto, CA) was implanted at day 14 of gestation into the jugular vein for chronic infusion of saline vehicle (control pregnant rats, n = 12) or recombinant rat TNF-α purified (BioSource International, Camarillo, CA) at a rate of 50 ng/day (TNF-α–treated pregnant rats, n = 11) for 5 days. At day 14 of gestation, rats were also surgically instrumented with a carotid catheter for subsequent arterial pressure measurement on day 19. Rats were housed in metabolic cages for 24 h starting on day 18 of gestation. On day 19 of gestation, after measuring arterial pressure, a blood sample was collected, kidneys were harvested, and litter size and pup weights were recorded.

Measurement of arterial pressure in conscious, chronically instrumented rats

Arterial pressure was determined in control pregnant (n = 12) and TNF-α–treated pregnant (n = 11) rats at day 19 of gestation. Pregnant rats were catheterized on day 14 of gestation under anesthesia using isoflurane (W.A. Butler Co., Memphis, TN) delivered by an anesthesia apparatus (Vaporizer for Forane Anesthetic, Ohio Medical Products, Madison, WI). A catheter of heat stretched PE-50 tubing was inserted into the carotid artery for blood sampling and blood pressure monitoring. After implantation, the catheter was tunneled to the back of the neck and exteriorized. On day 19 of gestation, pregnant rats were placed in individual restraining cages for arterial pressure measurements. Arterial pressure was monitored with a pressure transducer (Cobe III Transducer CDX Sema, Birmingham, AL). After achieving stabilization, arterial pressure was recorded continuously for a 2-h period.

TNF–α ELISA assay

A Rat Tumor Necrosis Factor–α Cytoscreen ELISA Kit (BioSource International, Camarillo, CA) was used for the quantitative determination of rat TNF-α in the range from 2 to 150 pg/mL. This assay displayed a sensitivity of 0.7 pg/mL with a 5% intra- and 7% interassay coefficient of variability. Recovery is estimated at 95% from pooled rat plasma.

Measurement of urinary nitrite/nitrate excretion

Urinary nitrite/nitrate excretion rates were determined in both control pregnant (n = 12) and TNF-α–treated pregnant (n = 11) rats fed a low nitrite/nitrate diet (AIN76, ICN Biomedicals Inc., Aurora, OH) for 5 days before collection. Echerichia coli was the source of nitrate reductase for conversion of nitrate to nitrite,²⁰ and sodium nitrate (Sigma-Aldrich, St. Louis, MO) was the standard to verify that all nitrate was converted to nitrite. The concentration of nitrite was measured colorimetrically using the Griess reagent. Sodium nitrite was used as the standard, and data were expressed as millimoles per liter of nitrate/nitrite excreted per 24 h by the rat.

Isolation of total cellular proteins and western blot analyses

Kidneys were removed from control pregnant (n = 8) and TNF-α–treated pregnant (n = 8) rats on day 19 of pregnancy, quick-frozen in liquid nitrogen, then stored at −80°C. Each kidney was ground using a liquid nitrogen chilled mortar and pestle and stored in a sterile tube at −70°C. Kidneys were homogenized 20% (w/v) in buffer containing 20 mmol/L HEPES, pH 7.5, 100 mmol/L pepstatin A, 100 mg/mL aprotinin, 10 mmol/L EDTA, 100 mg/mL leupeptin, 1 mmol/L phenanthroline, and 1 mmol/L E-64 (Sigma-Aldrich). Total protein concentration was determined using Sigma Protein Determination kit (P5656, Sigma-Aldrich). Equivalent amounts of total protein from each rat kidney sample were separated by electrophoresis using a 7.5% polyacrylamide resolving gel. Recommended positive controls were used for proper analysis (human endothelial, mouse macrophage, and rat pituitary, Transduction Laboratories, Lexington, KY). After transfer to nitrocellulose, membranes were probed with either the mouse monoclonal antibody ECNOS (Transduction Laboratories, Lexington, KY) for quantification of endothelial NOS or with the mouse monoclonal antibody nNOS (Transduction Laboratories, Lexington, KY) for quantification of both neuronal and inducible NOS. Actin (actin antibody, Amersham, Arlington Heights, IL) was used as an internal control. Horseradish peroxidase conjugated goat anti-mouse IgG (Amersham, Arlington Heights, IL) was used as a secondary antibody. Bound antibody was detected by chemiluminescence (ECL Plus kit, Amersham, Arlington Heights, IL) with quantification by densitometry (BioRad, Richmond, VA). For quantitation, renal protein expression for each specific NOS isoform was normalized after densitometry to the level of renal actin protein expression, which the authors have found not to change before, during, and up to 4 days post partum in the rat.

Statistical analysis

All data are expressed as mean ± SEM. Comparisons of control pregnant rats with TNF-α–treated rats were analyzed using factorial analysis of variance followed by Scheffé's test. A value of P < .05 was considered statistically significant.

Results

Mean arterial pressures in control and TNF-α–treated pregnant rats

Chronic infusion of TNF-α at a dose of 50 ng/day in pregnant rats resulted in significant increases in arterial pressure relative to control pregnant rats (Fig. 1). Mean arterial pressure (MAP) averaged approximately 116 ± 2.1 mm Hg (P < .05) in the TNF-α–treated pregnant rats at day 19 of pregnancy. This was a significant increase as compared with an average of 97 ± 3.6 mm Hg observed in control pregnant rats. In contrast, infusion of TNF-α at a dose of 50 ng/day for 5 days in virgin rats resulted in no significant effect on arterial pressure relative to vehicle-infused, control virgin rats.¹³

Comparison of mean arterial pressure at day 19 of gestation in recombinant rat tumor necrosis factor–α (TNF-α) treated (50 ng/day for 5 days) pregnant rats (n = 11) compared with saline vehicle–infused control pregnant rats (n = 12). All data are expressed as mean ± SEM. *P < .05 v control.

Figure 1.

No difference in pup weight (3.7 ± 0.2 g v 4.0 ± 0.1 g, TNF-α–treated pregnant versus control pregnant rats, respectively) or litter size (11 ± 0.0 pups v 12 ± 0.8 pups, TNF-α–treated pregnant versus control pregnant rats, respectively) was evidenced upon chronic infusion of TNF-α in pregnant rats. Nor was there a significant difference in body weight observed upon chronic infusion of TNF-α (311 ± 6.3 g v 327 ± 10.3 g, TNF-α–treated pregnant versus control pregnant rats, respectively).

TNF-α levels in control and TNF-α–treated pregnant rats

A significant elevation in plasma levels of TNF-α was observed in the TNF-α–treated pregnant rats as compared with the control pregnant rats (60.3 ± 8.7 pg/mL v 37 ± 5.9 pg/mL, TNF-α–treated pregnant versus control pregnant rats, respectively, P < .05).

Urinary nitrite/nitrate levels in control and TNF-α–treated pregnant rats

Urinary nitrite/nitrate excretion was measured to estimate whole-body production of NO. At day 19 of pregnancy, excretion of nitrite/nitrate in the TNF-α–treated pregnant rats (20.8 ± 2.5 μmol/24 h) did not differ significantly compared with control pregnant rats (22.0 ± 1.9 μmol/24 h) (Fig. 2).

Urinary nitrite/nitrate excretion rates at day 19 of gestation in TNF-α–treated pregnant rats (n = 11) compared with saline vehicle–infused control pregnant rats (n = 12). Animals were maintained on a low nitrite/nitrate diet to minimize dietary intake of nitrite/nitrite. All data are expressed as mean ± SEM. Abbreviation as in Fig. 1.

Figure 2.

Renal protein expression of the nitric oxide synthase isoforms in control and TNF-α–treated pregnant rats

Comparable levels of renal endothelial (eNOS) nitric oxide synthase protein expression were found at day 19 of gestation in both medulla and cortex of TNF-α–treated pregnant and control pregnant rats (medulla: 17.8 ± 1.4 v 18.9 ± 0.9 densitometric units; cortex: 16.7 ± 0.8 v 17.5 ± 1.7 densitometric units; TNF-α–treated pregnant versus control pregnant rats, respectively) (Fig. 3). However, renal protein expression of both inducible (iNOS) and neuronal (nNOS) nitric oxide synthase were significantly decreased in the medulla at day 19 of gestation in the TNF-α–treated pregnant rats relative to control pregnant animals; while only nNOS was significantly decreased in the cortex (Fig. 3). In the medulla, renal iNOS expression decreased by 20% (15.41 ± 0.8 v 19.18 ± 0.9 densitometric units, TNF-α–treated pregnant versus control pregnant rats, P < .05) in the TNF-α–treated pregnant rats as compared with control pregnant rats while renal nNOS expression decreased by 22% (8.24 ± 0.8 v 10.62 ± 0.7 densitometric units, TNF-α–treated pregnant versus control pregnant rats, P < .05). In the cortex, renal nNOS expression decreased by 30% (5.8 ± 0.1 v 12.6 ± 0.7 densitometric units, TNF-α–treated pregnant versus control pregnant rats, P < .05).

Renal protein expression of the renal nitric oxide synthase (NOS) isoforms at day 19 of gestation in TNF-α–treated pregnant rats (n = 8) compared with saline vehicle–infused control pregnant rats (n = 8). Representative Western blots (A) and quantitation of renal endothelial (eNOS), inducible (iNOS), and neuronal (nNOS) nitric oxide synthase protein expression (B). Renal NOS protein expression levels are normalized relative to renal actin protein expression levels. The SE is calculated from at least three determinations from at least three independent experiments. All data are expressed as mean ± SEM. *P < .05 v control. Other abbreviation as in Figs. 1 and 2.

Figure 3.

Discussion

One of the leading theories for preeclampsia proposes decreased placental perfusion leading to hypoxia with subsequent release of endothelium-activating factors into the maternal circulation.^3,4 Inflammatory cytokines, such as TNF-α, may be one of the putative placentally derived circulating factors, as they are capable of producing endothelial activation/dysfunction.⁵ Several lines of evidence support a role for TNF-α in mediating endothelial dysfunction in preeclampsia. Endothelial cell function is selectively altered by factor(s) released from the placenta of preeclamptic women.¹² Furthermore, in vitro studies mimicking in vivo placental hypoxia have shown that reduced oxygen elicits a twofold increase in TNF-α production in placental tissue.⁹ As plasma levels of TNF-α are elevated twofold in women with preeclampsia,¹¹ we previously reported that a twofold elevation in plasma TNF-α in the pregnant rat is associated with significant hypertension.¹³ In this study, rat TNF-α infused at a rate of 50 ng/day for 5 days resulted in a twofold increase in plasma TNF-α and a significant increase in MAP of approximately 19 mm Hg.

As deficiency in NO production has been suggested to play a role in preeclampsia, the purpose of this study was to assess NO production in association with the hypertension produced in response to a twofold elevation in plasma TNF-α in the pregnant rat. In this study 24-h urine collections representing steady-state levels of NO production were measured in animals fed a low nitrite/nitrate diet in order to provide an accurate measure of endogenous whole body NO synthesis exclusive of dietary intake contributions. No change was observed in urinary nitrite/nitrate excretion in the TNF-α–treated pregnant rats as compared with the control pregnant rats. However, as abnormalities in renal production of NO may not be reflected in the measure of whole body NO production, changes in renal NOS isoform protein expression were determined.

No change in renal eNOS protein expression at day 19 of gestation was observed in the TNF-α pregnant rats as compared with the control pregnant rats. We have previously shown that renal eNOS protein expression decreases during normal pregnancy by approximately 40%.¹⁸ Thus, renal eNOS protein expression was decreased in the TNF-α pregnant rats to the same extent as previously observed in the normal pregnant rat.

The hypertension observed in response to a twofold elevation in plasma TNF-α levels in the pregnant rat was associated with significant reductions in renal expression of the nNOS isoform within both the cortex and the medulla. Renal nNOS protein expression was decreased by approximately 22% in the medulla and by 30% in the cortex of the TNF-α pregnant rats as compared with the normal pregnant rats. During normal pregnancy, nNOS protein expression is increased by approximately 22% at day 19 of gestation relative to the prepregnant level.¹⁸ Thus, the increase in renal nNOS protein expression observed in normal pregnant rats was not observed in TNF-α pregnant rats. Therefore, renal nNOS protein expression was decreased and this decrease was associated with significant increases in arterial pressure in the TNF-α pregnant rats.

Evidence from several studies indicates that NO generated by nNOS in the kidney may be involved in the long-term control of systemic and renal hemodynamic changes that occur during normal gestation. Inhibition of nNOS with 7-nitroindazole significantly increases MAP and reduces RPF and GFR in pregnant but not in virgin rats.²¹ Nitric oxide generated from a renal nNOS has been implicated in mediating the increase in renal hemodynamics observed during normal pregnancy as both renal nNOS protein expression and renal nNOS activity are increased in the rat.^18,22 Variations in renal nNOS expression may also contribute to changes in arterial pressure and renal function during preeclampsia. A significant reduction in renal protein expression of nNOS is associated with the significant hypertension and decreases in GFR produced in response to a chronic reduction in uterine perfusion pressure in the pregnant rat, thus suggesting that NO generated from nNOS may contribute to changes in systemic and renal hemodynamics in a rat model of preeclampsia.²³ The importance of nNOS in mediating the increases in arterial pressure during preeclampsia however is still unclear since pregnant nNOS knockout mice do not develop hypertension.²⁴

Whether the decrease in renal nNOS protein expression is involved in mediating the decrease in renal hemodynamics in pregnant rats infused chronically with TNF-α is unknown. Evidence from several studies indicates that NO generated by nNOS in the kidney may be involved in the long-term control of renal hemodynamic changes in response to changes in sodium intake,^25,26 thus indicating the importance of nNOS in mediating changes in systemic hemodynamics.

Within the medulla, a significant decrease of 20% was observed for renal iNOS protein expression of the TNF-α pregnant rats. However, no significant decrease in renal iNOS protein expression was observed in the cortex of the TNF-α pregnant rats as compared with the normal pregnant rats. Renal iNOS protein expression during normal gestation in the pregnant rat is significantly elevated¹⁸; however, this elevation in renal iNOS protein expression does not occur in the TNF-α pregnant rats as a significant decrease in iNOS is observed within the medulla. Thus, the hypertension induced in response to a twofold elevation in TNF-α is associated with a significant decrease in medullary iNOS protein expression. The importance of iNOS in mediating changes in arterial pressure has been suggested by studies in which inhibition of iNOS conveys salt sensitivity to normal Dahl resistant rats.²⁷ The importance of iNOS in mediating changes in renal function during pregnancy-induced hypertension, however, remains to be determined.

In summary, we found that the hypertension associated with a twofold increase in plasma TNF-α levels in the pregnant rat was associated with no significant reduction in whole body NO synthesis, as assessed by urinary excretion of nitrite/nitrate, but was associated with significant reductions in renal nNOS protein expression, within both the cortex and medulla, and renal iNOS protein expression, within the medulla. These changes in renal NOS expression may contribute to the hypertension observed in pregnant rats chronically infused with TNF-α to a level that mimics that found in preeclamptic women.