INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

SUBSTANTIAL PHYSIOLOGICAL DATA in vivo point to a significant role for dilator prostanoids in the regulation of cerebral vascular tone and cerebral blood flow (CBF) in newborns (10). The physiological effects of arachidonate metabolites often have opposing actions. For example, PGE₂ and PGI₂ are potent vasodilators, whereas thromboxane (Tx) A₂ and PGF₂ are vasoconstrictors (15, 16). Arachidonic acid is liberated from damaged cell membranes during ischemia and is the source of vasoactive prostanoids, which are believed to influence the progression of cerebral ischemia.

The question as to whether the inhibition of cyclooxygenase has a beneficial effect on outcome after cerebral ischemia is controversial. Free indomethacin crosses the blood-brain barrier (2). It is generally assumed that the duration of action of indomethacin is long, but the plasma half-life is relatively short, and the cerebrospinal fluid (CSF) concentration is closely related to that of plasma (2). Indomethacin, a cyclooxygenase inhibitor, has been shown to prevent or attenuate the development of intraventricular hemorrhage (14). On the other hand, high doses of indomethacin are associated with a selective hypoperfusion to certain brain regions during normoxia (6). Indomethacin has been reported to have beneficial prophylactic effects in prevention and/or reduction of severity of intraventricular or periventricular hemorrhage in premature babies (1, 7, 14). High-dose indomethacin (5 mg/kg) has been found to decrease regional and total CBF for 120 min after administration in newborn pigs (18). In piglets, indomethacin decreases CBF at rest (13). These decreases in CBF occur concomitantly with decreases in CSF dilator PG levels (12).

No information exists with regard to the effect of prostanoids and indomethacin on CBF during cerebral ischemia in fetuses. Our previous study demonstrated that plasma concentrations of TxB₂ increased massively in sagittal sinus blood compared with arterial blood (21). In the present study, we hypothesized that during cerebral hypoperfusion, prostanoids produced in the microvasculature might cause vasoconstriction and reduce CBF, and that pretreatment with indomethacin may attenuate this reduction of CBF. To test these hypotheses, we measured regional brain blood flow in fetal sheep subjected to 10 min of cerebral ischemia with and without pretreatment of indomethacin (0.2 mg/kg), a nonselective inhibitor of both isoforms of cyclooxygenase (COX1 and COX2).

	METHODS AND MATERIALS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Eight chronically catheterized fetal sheep between 126 and 136 days gestation were used in this study. The pregnant ewes were of mixed Western and Florida native breeds. These experiments were approved by the University of Florida Institutional Animal Care and Use Committee and were performed in accordance with the Guiding Principles for the Care and Use of Animals of the American Physiological Society.

Surgical preparation. Aseptic surgery was performed at least 5 days before the start of experiments in each animal. The fetuses were chronically instrumented with tibial artery and saphenous vein catheters and amniotic fluid catheters as previously described. Fetal hindlimbs were identified and delivered through a small hysterotomy incision near the tip of one uterine horn. We introduced polyvinyl chloride catheters into the tibial artery (0.050 in. ID, 0.090 in. OD) and saphenous vein (0.030 in. ID, 0.050 in. OD) bilaterally and advanced the tips to the abdominal aorta and inferior vena cava, respectively. After closing the skin incisions in the fetal hindlimb, amniotic fluid catheters were sutured to the fetal skin, and the hindlimb returned to the amniotic space. After placement of these catheters and closure of the hysterotomy, the fetal head was located, the uterus was incised, and the head was delivered. After a single midline incision in the skin of the neck was made, lingual arteries were identified, ligated, and catheterized with polyvinyl chloride catheters (0.030 in. ID, 0.050 in. OD), with the catheter tips advanced retrograde to the lumen of the common carotid arteries. As previously described, this catheterization technique allows measurement of common carotid arterial pressure without interruption of carotid arterial blood flow (25). The carotid sinus denervation was performed using methods described by Wood (24). The carotid sinus nerves were identified bilaterally and cut. The walls of the common carotid arteries in this area, as well as the lingual arteries and common carotid arteries extending 0.5-1 cm rostral from the lingual-carotid arterial junction, were stripped of all visible nerve fibers. After performing these denervations in the fetal neck, the fetal skin was closed. An extravascular balloon occluder (In Vivo Metric, Healdsburg, CA) was placed around the brachiocephalic artery by means of a separate incision through the fourth intercostal space on the left side of chest. Then the head was returned to the amniotic cavity, and the uterus was closed. All catheters exited via a small incision in the flank of the ewes.

Ampicillin trihydrate (Polyflex; Aveco Co, Fort Dodge, IA; 500 mg) was administered to the fetus via the amniotic fluid and to the mother (500-750 mg) intramuscularly at the time of surgery and again each time the fetus was studied or the catheters were flushed. Ampicillin (500-750 mg) was administered to the mother intramuscularly twice daily for 5 days after the surgery. All catheters were flushed and reheparinized at least once every 3 days.

Experimental protocol. Sheep were transported to the procedure room from their pens within the Health Center Animal Resources Department at least 1 h before the start of each experiment. Each fetus was studied twice. Experiments consisted of a 90-min preocclusion control period (90 to 0 min), a 10-min occlusion period (0-10 min) and a 10-min postocclusion recovery period (10-20 min). In one experiment on each fetus, the vehicle of indomethacin, 0.1 M phosphate buffer (PB) was injected intravenously, and in the other experiment 0.2 mg/kg indomethacin (an inhibitor of cyclooxygenase) was injected intravenously 90 min before the 10-min period of hypotension. PB and indomethacin were randomly administrated to the fetal sheep. In each experiment, the extravascular occluder was inflated for 10 min to produce arterial hypotension. Fetal arterial blood samples were drawn from tibial artery catheters (descending aorta) at 90 and 10 min before the start of the period of occlusion, at the end of the 10 min period of occlusion, and 20 min after the onset of occlusion. A 1-ml arterial blood sample was drawn before the experiment to measure blood gases. Blood samples (3 ml) were collected into chilled polystyrene tubes containing 150 µl of 0.5 M EDTA. Separate blood samples (1 ml) were collected into chilled polypropylene tubes containing 50 µl 0.5 M EDTA and 40 µg/ml indomethacin for measuring thromboxane (Tx) B₂ and PGE₂. Tubes were kept on ice until the end of the experiment and then centrifuged for 20 min at 3,000 g at 4°C. Plasma was separated and stored in separate aliquots at 20°C.

Plasma ACTH, cortisol, and arginine vasopressin (AVP) concentrations were measured by specific radioimmunoassay. These assays have been described in detail elsewhere (19, 25). PGE₂ and TxB₂, a stable metabolite of TxA₂, were measured with enzyme-linked immunoassay kits purchased from Cayman Chemical. Before assay, the prostanoids were extracted from acidified plasma with six volumes of ethyl acetate. The recovery with this protocol averages ~60%, and the extracted prostanoids dilute parallel to the standard curves.

Fetal arterial blood pressure and amniotic fluid pressure values were measured continuously during the 110-min experiments with a Grass recorder and Statham P23 ID pressure transducers. These hemodynamic variables were sampled, and analog-to-digital conversions were performed at 2-s intervals with an IBM PC computer. The data collection was achieved using ASYSTANT⁺ software (Asyst Technologies, Rochester, NY). All fetal intravascular pressure measurements were corrected by subtraction of amniotic fluid pressure.

Immediately before and 5 min after the onset of hypotension, colored microspheres (15 µm; Dye-Track; Triton Technologies, San Diego, CA) were injected through the venous catheter, and simultaneous reference blood samples were drawn from the lingual arterial catheters at the rate of 3 ml/min for 1 min, 30 s. Four colors of microspheres were used in each animal for measuring CBF. These techniques have been described in detail (4).

At the end of the second experiment, the pregnant ewes (and the fetus) were euthanized with an overdose of pentobarbital sodium. The fetal brains were dissected for microsphere extraction.

Estimation of CBF with colored microspheres. The fetal brains were dissected into cerebral cortex, brain stem, hippocampus, hypothalamus, cerebellum, and pituitary. The dissected tissues were dissolved in 4 ml of 4 M potassium hydroxide and 0.2% Tween 80 for 48 h and filtered through an 8-µm-thick polyester membrane filter. The membrane (containing the microspheres) was placed in a microcentrifuge tube to air dry overnight. Dimethylformamide (150 µl) was added to the tube to leach the color microspheres from the membrane. Then the tube was centrifuged, and 80 µl of the supernatant was pipetted to a cuvette. The absorbance spectra was quantified with use of a spectrophotometer and the appropriate software for least-squares solution of absorbance of each dye. The reference blood samples were processed similarly, except that 16 M potassium hydroxide was used.

CBF was calculated by the following equation
where ABS is absorbance.

Statistical analyses. Change in mean arterial blood pressure over time was analyzed by t-test. Changes in the values of fetal hormonal and prostanoid variables over time and between groups were analyzed with two-way ANOVA. A multiple comparison of mean values was performed with Duncan's test. The hormonal data were not distributed normally. All ANOVAs performed on hormonal data were calculated after logarithmic transformation to correct heteroscedacity of the data. Fetal CBF was analyzed with three-way ANOVA (23) followed by Duncan's test. A significance level of P < 0.05 was used to reject the null hypothesis in all tests. Analyses were performed using SigmaStat software (Jandel Scientific, San Rafael, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Blood gases and pressure. The blood gases measured before PB and indomethacin injections indicated that all fetuses were healthy during the experiments (Table 1). Injection of indomethacin did not alter fetal blood gases. Fetal mean arterial blood pressure before, during, and after brachiocephalic occlusion are presented in Fig. 1. Lingual arterial blood pressure decreased significantly (P < 0.001) during the brachiocephalic occlusion in both PB- and indomethacin-treated groups. In the PB group, mean lingual arterial blood pressure decreased from 41.9 ± 3.1 to 16.5 ± 3.3 mmHg. In the indomethacin group, it decreased from 41.6 ± 2.5 to 14.6 ± 3.0 mmHg. Blood pressure returned to baseline level after the occluder was released. Femoral arterial blood pressure increased significantly (P < 0.05) during the brachiocephalic occlusion in PB-treated group, from 42.9 ± 3.7 to 64.9 ± 7.0 mmHg. In the indomethacin-treated group, mean femoral arterial blood pressure increased during occlusion to 51.1 ± 4.2 mmHg. This was higher (P < 0.05) than that after occlusion (39.3 ± 1.7 mmHg).

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Fetal mean lingual (top) and femoral (bottom) arterial blood pressure (MAP) before, during, and after 10-min period of brachiocephalic occlusion (BCO) in fetuses. There were no significant differences in MAP between phosphate buffer (PB)- and indomethacin-treated groups. Mean values are accompanied by vertical bars representing 1 SE. * P < 0.05 and ** P < 0.01 compared with baseline MAP before BCO.

Plasma prostanoid concentrations. The mean values of fetal plasma PGE₂ and TxB₂ concentrations are summarized in Table 2. There was no difference in the initial values of prostanoid concentrations in the two groups before PB and indomethacin injections. After indomethacin injection, plasma PGE₂ and TxB₂ concentrations decreased significantly (P < 0.05), indicating that indomethacin effectively inhibited cyclooxygenase activity. Compared with the preocclusion value, hypotension increased plasma PGE₂ and TxB₂ concentrations in the PB-treated group, but the apparent changes were not statistically significant.

Regional CBF. The mean values of fetal regional CBF (rCBF) are reported in Table 3. In the PB-treated group, brachiocephalic occlusion significantly (P < 0.05) decreased rCBF in cerebral cortex, brain stem, hippocampus, hypothalamus, and cerebellum. Indomethacin pretreatment (n = 7) slightly decreased but did not significantly change basal rCBF. After indomethacin treatment, hypotension decreased rCBF in all of the regions studied, especially in cerebral cortex, brain stem, and cerebellum (P < 0.05). Hypotension produced somewhat smaller changes in rCBF after indomethacin treatment. Thus the induced changes in rCBF in cerebral cortex, brain stem, hippocampus, hypothalamus, and cerebellum were 49%, 30%, 36%, 18%, and 42% less than after PB treatment. Three-way ANOVA was performed revealed that there was a significant interaction between drug (PB or indomethacin) and blood pressure (±occlusion) treatments (F = 5.224, P < 0.05).

Plasma hormone concentration. Plasma ACTH concentration was increased significantly from 73.8 ± 14.2 to 1,811.3 ± 1,009.3 pg/ml in the fetuses treated with PB and from 200.4 ± 51.9 to 1,489.6 ± 482.0 pg/ml in those treated with indomethacin by the brachiocephalic occlusion. Plasma AVP concentration was increased significantly from 10.5 ± 2.41 to 47.0 ± 17.59 pg/ml in the fetuses treated with PB and from 20.0 ± 5.9 to 74.3 ± 32.45 pg/ml for those treated with indomethacin by the brachiocephalic occlusion. Indomethacin treatment did not alter ACTH and AVP responses to hypotension. Plasma cortisol concentration was increased from 11.3 ± 2.51 to 20.9 ± 5.28 ng/ml in the fetuses treated with PB and from 15.8 ± 4.43 to 28.2 ± 8.47 ng/ml treated with indomethacin by the hypotension. There was no significant difference between PB and indomethacin treatment (analyzed by 2-way ANOVA; Fig. 2). This is consistent with our previous studies that indomethacin did not further attenuate hormonal responses to hypotension in denervated fetuses (20).

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Fetal plasma ACTH, arginine vasopressin (AVP), and cortisol concentrations before, during, and after 10-min period of BCO in fetuses. Mean values of hormones in PB (, solid lines)- and indomethacin (, dashed lines)-treated fetuses are accompanied by vertical bars representing 1 SE. Two-way ANOVA did not show significant effect between groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

The present study tested the effect of a therapeutic dose of indomethacin (0.2 mg/kg) on the CBF response during cerebral ischemia. It has been observed that prostanoids are important in modulating perinatal cerebral hemodynamics in human beings and in other species (11). Our previous study demonstrated that TxB₂ secretions increased significantly in sagittal sinus blood during cerebral hypoperfusion, and PGE₂ concentrations increased significantly in fetal brain stem and hypothalamus interstitial fluids. We tested the hypothesis that indomethacin may compromise fetal CBF by inhibiting the production of these or other prostaglandins in the brain during cerebral ischemia.

Papile and co-workers (17) demonstrated CBF autoregulation in the preterm fetal lamb and reported that the range of autoregulation was narrowed in the preterm lamb. They speculated that autoregulatory mechanisms in the fetal brain under normal physiological conditions in utero may not protect against acute fluctuations in blood pressure, particularly hypotension (17). In the present study, brachiocephalic occlusion decreased carotid arterial blood pressure to the same degree in both PB- and indomethacin-treated fetuses. Although the induced decrease in arterial blood pressure was similar in both groups, the induced changes in rCBF were not. We cannot conclude that indomethacin shifted the autoregulation curve toward higher pressures, because we did not generate an autoregulation curve in the present study. However, we did find that the reduction in rCBF during the period of hypotension was ~30% less in indomethacin-treated fetuses compared with PB-treated fetuses. We believe that this difference was secondary to the inhibition of synthesis of constrictor prostanoids by indomethacin. Furthermore, we believe that the inhibition of prostanoids involved in the control of CBF is likely to be within the cerebral vasculature itself.

We designed our study to avoid the initial transient constriction produced by indomethacin injection. At the time of study, CBF was not altered in the normotensive condition. In one sense, our results contrast with those of Leffler et al. (12), who found that indomethacin decreased CBF in newborn piglets by inhibiting dilator prostaglandins. We acknowledge that our design ignores this transient phase of the response. However, indomethacin can also produce transient effects that are not directly related to prostanoid production (3, 9). Because the initial response to indomethacin likely includes both prostaglandin-dependent and -independent changes in CBF, we waited 90 min after injection of the indomethacin to allow establishment of a new steady state. Although we acknowledge that, after 90 min, other compensatory responses might be involved, we argue that prostanoid biosynthesis within the vasculature is still reduced, a conclusion consistent with the reduction in circulating prostanoid concentrations.

In a separate study, we found that brachiocephalic occlusion in carotid sinus-denervated fetuses caused the release of TxB₂ into the sagittal sinus blood (21). The appearance of TxB₂, the stable metabolite of TxA₂, in sagittal sinus blood suggests that during hypoperfusion the concentration of TxA₂ in the cerebral microvasculature is high. The source of the TxA₂ is unknown but could be platelets or vascular endothelium. Thromboxane synthase, the enzyme that converts PGH₂ to TxA₂, is found in both sites in addition to neurons within the brain. Although it is theoretically possible that the TxB₂ in the vasculature derives from neurons, we argue that in this case it is more likely to derive from a source on the blood side of the blood-brain barrier. We base this argument on our own previous observation that cerebral hypoperfusion decreases the concentration of TxB₂ in the extracellular fluid in hypothalamus and brain stem, whereas it massively increases the concentration of TxB₂ in sagittal sinus blood (21). Our results are consistent with a regulatory role of TxA₂ (or another constrictor prostanoid) within the cerebral vasculature and agree with those of Chemtob et al. (5), who demonstrated in newborn piglets that inhibition of TxB₂ synthesis was associated with an extension of the lower blood pressure limit of CBF autoregulation and an attenuation in the decrease in CBF below this blood pressure. Together, the results are consistent with the notion that TxA₂ production within the cerebral microvasculature might be protective because it delays reperfusion and therefore limits oxidative stress.

In this experiment design, all fetuses were carotid sinus-denervated to eliminate complicating variables introduced by the presence of carotid arterial baroreceptors and chemoreceptors. The fetal rCBF was comparable to that reported by other investigators in control animals. Itskovitz and Rudolph (8) demonstrated that sinoaortic denervation of the preterm fetal lamb does not alter resting heart rate or blood pressure and had no effect on CBF. This is consistent with the conclusion that the autonomic nervous system does not appear to have a major role in the control of cerebral circulation (22). Nevertheless, we cannot rule out the possibility that a portion of the change in CBF flow might have derived from residual reflex responses in these fetuses.

The hormonal responses to cerebral hypoperfusion of this study were consistent with our previous experiments. In other experiments, we found that indomethacin reduced the magnitude of the ACTH and AVP responses to hypotension in late-gestation fetal sheep (20). Interestingly, we found that the effect of indomethacin on the ACTH and AVP responsiveness to hypotension was dependent on the presence of intact baroreceptor and chemoreceptor afferent fibers. In the present experiments, in which fetuses were subjected to carotid sinus denervation and indomethacin treatment, the endocrine results were much the same. We are assuming that indomethacin reduces prostanoid biosynthesis on both sides of the blood-brain barrier, although this assumption is somewhat controversial. Indomethacin is highly lipid soluble (it is also highly protein bound in plasma). Indomethacin crosses the intact meninges by simple diffusion (2); it can be found in the brain within 30 min of systemic administration, and the concentrations of indomethacin in cerebrospinal fluid are similar to the unbound concentrations in plasma (2). Collectively, these studies suggest an effect of brain prostanoids as modulators of endocrine responses to hypotension that requires intact afferent fibers from arterial receptors.

In conclusion, this study demonstrated that indomethacin had a beneficial effect on fetal rCBF in response to cerebral ischemia in late-gestation fetal sheep. We therefore speculate that, during cerebral hypoperfusion, endogenous production of vasoconstrictor prostanoids limit reperfusion.

Perspectives