INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DOPAMINE IS COMMONLY USED in neonatal intensive care units to support blood pressure. However, there are discrepancies in published blood pressure nomograms for preterm and full-term infants (14, 24), and the clinical responses to dopamine can also be quite variable (16, 20, 23). Consequently, dopamine may transiently cause arterial blood pressure to exceed its "normal level" and could disrupt the blood-brain barrier (BBB). Developmental changes in barrier function have been described prenatally in humans (17) and sheep (8) and postnatally in rat (4, 5) and rabbit (31). Although the BBB forms during early development (7) and, in humans, cerebral endothelial tight junctions are evident at 8 wk of gestation (17), barrier function may be more susceptible to injury in the immature brain. For example, hypoxia-ischemia causes a greater increase in BBB permeability in 7- than in 21-day-old rats (18). Positive-pressure ventilation predisposes preterm lambs to increases in regional BBB permeability, a phenomenon not observed in adults (25). Furthermore, periventricular hemorrhage, which may be due in part to disruptions in BBB permeability, is more prevalent in preterm human newborns than in full-term newborns. Therefore, the state of maturation appears to influence the BBB susceptibility to injury.

In the present study, BBB permeability was evaluated in fetal sheep at 0.6 (93 days) and 0.9 (132 days) gestation during dopamine-induced hypertension under conditions of normal intrauterine blood gases and normovolemia. Experiments were performed on fetal sheep in utero because of the difficulty in establishing adequate pulmonary ventilation and in maintaining cardiovascular and thermal stability in prematurely delivered lambs. The fetal sheep brain at 93 days of gestation is thought to be comparable functionally to a human brain at ~26 wk of gestation, although the brain growth spurt occurs earlier in sheep (6). Permeability was assessed with the tracer aminoisobutyric acid (AIB). Because of its small size (104 Da), AIB is considered to be a sensitive tracer for detecting increases in permeability across tight endothelial junctions in brain (2). In fetal sheep, permeability to sucrose and albumin markedly decreases between 50 and 70 days of gestation, whereas there is little further decrease between values at 70 days of gestation and postnatal values (8). However, barrier function is not completely mature, because uptake of AIB is greater at 87 and 137 days of gestation than postnatally in lambs and sheep (26), and dexamethasone is more effective in reducing basal AIB uptake in preterm fetuses (28). Lack of complete maturation of barrier function could render the barrier more susceptible to disruption by arterial hypertension. We tested the hypothesis that dopamine-induced hypertension increases regional brain uptake of AIB and is associated with increased brain water content in fetal sheep at 0.6 and 0.9 gestation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments were approved by the Animal Care and Use Committee of The Johns Hopkins University. Time-dated pregnant ewes were anesthetized with halothane by facemask, and the trachea was orally intubated. Anesthesia was maintained by mechanical ventilation with 1.5% halothane-30% nitrous oxide-balance O₂. Under sterile conditions, a midline abdominal incision was made to expose the uterus (13). Each limb of the fetus was individually withdrawn through a single uterine incision for catheterization of two axillary arteries in the forelimbs and two pedal veins in the hindlimbs. A catheter was also placed in the amniotic fluid. All incisions were sutured. Catheters were routed subcutaneously to the flank of the ewe and secured in a plastic pouch stitched to the skin. Ampicillin (500 mg) was injected into the amniotic fluid. Procaine penicillin (1,200,000 U) was injected intramuscularly into the ewe.

Experiments were conducted 2 days after surgery, with the ewes standing in a cart. On the day of the experiment, the midgestational fetuses averaged 93 ± 1 conceptional days and the near-term fetuses averaged 132 ± 2 conceptional days. Arterial blood pressure was measured in reference to amniotic fluid pressure. Preductal arterial blood samples were obtained from the axillary artery catheter. Arterial pH, PCO₂, and PO₂ were measured with an electrode system (model 248, Chiron Diagnostics). Hemoglobin concentration and O₂ content were measured with a hemoximeter (model OSM3, Radiometer). Glucose was measured with a glucose analyzer (model 2300, Yellow Springs Instruments).

After baseline measurements were obtained, a continuous intravenous infusion (3 ml/h) of saline or dopamine was started. Dopamine was infused at 75 µg · min¹ · 100 g¹ in 93-day fetuses and at 25 µg · min¹ · 100 g¹ in 132-day fetuses. A higher dose was required in the 93-day fetuses to achieve comparable percent increases in arterial pressure. In pilot experiments, further increases in the dopamine dose did not produce further increases in arterial pressure at either gestational age. Steady-state increases in arterial pressure were achieved within a few minutes of the start of dopamine infusion. A second arterial blood sample was obtained 9 min after the start of infusion.

At 10 min of saline or dopamine infusion, 2-amino-[1-¹⁴C]isobutyric acid with a specific activity of ~60 mCi/mmol (Amersham International) was infused into the second venous catheter. A nominal dose of 50 µCi/kg fetal body wt was used in 93-day fetuses. In 132-day fetuses, a lower dose of 25 µCi/kg was used to achieve a comparable arterial concentration profile (26). Fetal body weight was estimated at the time of surgery for determining the dose. The actual dose was calculated on the basis of the fetal body weight at autopsy. AIB is usually injected as a bolus. We chose to infuse AIB diluted in saline over 3 min to avoid a sharp peak in the arterial concentration profile which would be underestimated by arterial sampling at discrete times, particularly in the 93-day fetuses, in which the number of samples is limited by the relatively small blood volume. The AIB solution was infused at 1 ml/min in 93-day fetuses and at 2 ml/min in 132-day fetuses. The 3-min AIB infusion was followed by a 1-min saline infusion to clear the catheter dead space.

To determine the arterial concentration profile, 0.3-ml blood samples were collected in heparinized tubes from an axillary artery catheter 0.5 min before AIB injection and 1, 2, 3, 4, 5, 6, 7, 10, 15, 20, 30, 45, and 60 min after injection in 93-day fetuses. In 132-day fetuses, 0.4-ml samples were taken at 0.5, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 10, 15, 20, 25, 30, 40, 50, and 60 min after AIB injection. The total volume of blood withdrawn was approximately equal to the amount of saline delivered in the dopamine and AIB infusions. A relatively long circulation period (60 min) can be used without accounting for backdiffusion of this tracer, because AIB is sequestered in cells and not metabolized (2, 9). A 60-min circulation time permits the plasma concentration of AIB to decrease by >90% from the initial value and thereby minimize any residual effect of tissue plasma volume changes on tissue counts of the tracer.

Immediately after the 60-min blood sample was obtained, the ewe was anesthetized with an intravenous bolus of pentobarbital sodium (30 mg/kg) and the fetal heart was arrested by injection of KCl into the fetal venous catheter. After the ewe was anesthetized, the ewe was injected intravenously with KCl. The entire radioactive uterus and fetus were removed, and the fetal brain was rapidly harvested and placed on a chilled plate. Brain tissue samples (~50 mg) were quickly dissected to minimize drying and were weighed in glass scintillation vials. Six samples were taken each from the frontal, parietal, and occipital cortical gray matter on one side. Three samples were taken each from cerebellum, medulla, pons, thalamus, caudate, and cervical spinal cord. Tissue samples were digested overnight in 0.5 ml of Solvable (Packard Instruments) at 50°C. From each blood sample, 50 µl of plasma were obtained in triplicate. Scintillation fluid (Atomlight, Packard Instruments) was added to each plasma and tissue sample for counting on a beta-scintillation counter (model LS2800, Beckman).

With the use of the average counts in the triplicate plasma samples at each time point, the time integral of the plasma concentration of AIB was calculated over the 60-min circulating period. The transfer coefficient from plasma to brain was calculated as the counts in the tissue divided by the time integral of counts in the plasma. For each brain region, an average transfer coefficient was calculated from the six samples in each cortical region and from the three samples in the remaining regions.

Fractional water content was calculated from the wet weight and the weight after 48 h of drying at 80°C. Cortical gray matter samples (~100 mg) were obtained in duplicate from frontal, parietal, and occipital cortical regions contralateral to those obtained for beta-scintillation counting.

Among the 93-day fetuses, five received saline and five received dopamine. Among the 132-day fetuses, six received saline and seven received dopamine. Within each age group, comparisons between saline control and dopamine groups were made by t-test. Arterial blood data were compared before and after the start of the saline or dopamine infusion by paired t-test. Statistical significance was set at P < 0.05. Values are means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Infusion of dopamine increased mean arterial blood pressure from 38 ± 3 to 53 ± 5 mmHg in 93-day fetuses and from 55 ± 5 to 77 ± 8 mmHg in 132-day fetuses. The percent increase in arterial pressure in 93-day fetuses (39%) was similar to that in 132-day fetuses (40%). The increase was sustained throughout the 60-min period of AIB circulation (Fig. 1). In the groups receiving saline, arterial pressure was unchanged during AIB circulation.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Mean arterial blood pressure (mean ± SD) in 93- and 132-day fetuses receiving intravenous saline (control) and dopamine. Time 0, start of aminoisobutyric acid (AIB) infusion. Dopamine significantly increased pressure in both age groups.

Arterial pH, PCO₂, PO₂, and hemoglobin concentration were in the normal range for fetal sheep (Table 1). In the 93-day fetuses, infusion of dopamine increased arterial PO₂, hemoglobin concentration, and O₂ content.

The dose of AIB was based on a gross estimate of fetal body weight at the time of surgery. The actual dose calculated from the body weight measured at autopsy was similar in the saline control and dopamine-treated groups of 93-day fetuses (Table 2). The time integral of the plasma concentration of AIB also was similar in the saline control and dopamine-treated groups of 93-day fetuses. However, in the 132-day fetuses, the dose of AIB per actual body weight and the plasma AIB-time integral were greater in the dopamine-treated than in the saline control group, in part because of errors in estimating fetal body weight.

The AIB transfer coefficient (K_i) for each brain region is shown in Fig. 2. There were no significant differences between the control and the dopamine-treated group in any region at 93 days of gestation. Likewise, at 132 days of gestation, there were no regional differences in K_i between the control and the dopamine-treated group.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Regional AIB transfer coefficient (mean ± SD) in 93- and 132-day fetuses receiving intravenous saline (control) and dopamine. There were no significant differences between control and dopamine-treated groups at either age. Spinal C, spinal cord.

Fractional tissue water content in cortical gray matter was greater at 93 than at 132 days of gestation (Fig. 3). There were no significant differences between the control and the dopamine-treated group at either age.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Fractional tissue water content (mean ± SD) in cortical gray matter regions in 93- and 132-day fetuses receiving intravenous saline (control) and dopamine. There were no significant differences between control and dopamine-treated groups at either age.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The principal finding of this study is that dopamine-induced hypertension in fetal sheep at 0.6 and 0.9 gestation does not increase BBB permeability to the small tracer molecule AIB and does not increase water content of cortical gray matter. Thus dopamine administration to the healthy fetus does not acutely disrupt the BBB.

These experiments were conducted under conditions that were relatively normoxic and initially normotensive for the fetus. In fact, dopamine infusion actually increased arterial PO₂ and hemoglobin concentration in the 93-day fetuses. The increase in PO₂ may have been due to increased placental blood flow secondary to the increase in arterial pressure or to a decrease in fetal O₂ consumption secondary to peripheral vasoconstriction. The effect of dopamine on the BBB might be different during or after a hypoxic or hypotensive episode. In the rat, prolonged hypoxia-ischemia produces greater barrier disruption at postnatal day 7 than at postnatal day 21 (18). In the preterm nonhuman primate, respiratory acidosis potentiates bilirubin transport into brain and consequent toxicity (33). Therefore, the BBB in the immature brain appears to be more vulnerable to hypoxic-ischemic injury. Hence, we cannot exclude that dopamine-induced hypertension after resuscitation from a hypoxic-ischemic or hypotensive event could exacerbate BBB disruption and cause cerebral edema in the newborn.

Dopamine increased mean arterial blood pressure ~40% in both age groups, but the infusion dose was threefold greater in the 93-day fetuses. In preliminary work, we found that greater doses of dopamine were required to increase arterial pressure at 90 days of gestation, consistent with maturation of peripheral vascular sensitivity to other adrenergic agonists in fetal sheep (1, 29). Higher doses of dopamine did not produce significantly greater increases in arterial pressure. Interestingly, the increase in arterial pressure to 53 mmHg in 93-day fetuses and to 77 mmHg in 132-day fetuses is close to the maximum attainable with a physiological stimulus at these ages. For example, evoking a Cushing response by increasing intracranial pressure to near baseline arterial pressure increased arterial pressure to 50 mmHg at 93 days and to 75 mmHg at 132 days (12). Ordinarily, the large placental vascular bed, which is under little adrenergic control, acts to buffer large increases in fetal arterial pressure. However, after birth, it is likely that dopamine could generate greater increases in arterial pressure. Therefore, the increase in mean arterial pressure achieved in the present study is near the maximum attained in a fetus for that age but is likely to be less than the maximum attainable soon after birth.

In adult animals, disruption of the BBB by hypertension generally requires an arterial pressure in excess of the upper limit of cerebrovascular autoregulation. This upper limit is difficult to determine in fetal sheep with intact reflexes. Cerebrovascular autoregulation is present up to an arterial pressure of at least 80 mmHg in sinoaortic-denervated fetal sheep at 120 days of gestation (19), whereas others describe a proportional increase in cerebral blood flow when perfusion pressure is increased from 40 to 57 mmHg after ganglionic blockade (15). Therefore, the lack of BBB disruption with dopamine infusion in the present study without autonomic blockade may be attributed to arteriolar vasoconstriction and an absence of autoregulatory breakthrough.

Cerebral water content is greater in immature than in mature brain. The values of 90% at 93 days of gestation and 85.5% at 132 days of gestation obtained in the present study agree with those of others (3). The lack of a significant increase in water content in the dopamine-treated fetuses is consistent with the lack of an increase in BBB permeability to AIB.

The ontogeny of BBB permeability to AIB in sheep has been reported by Stonestreet et al. (26). Uptake of AIB at 0.6 and 0.9 gestation was greater than postnatal values in most brain regions. The greatest uptake was seen in spinal cord, cerebellum, and brain stem at 0.6 gestation. Our results in 93-day fetuses are consistent with a greater AIB uptake in these caudal regions. Maternal treatment with dexamethasone for 48 h suppressed AIB uptake in spinal cord, brain stem, and other brain regions at 0.6 and 0.8 gestation, but not at 0.9 gestation, when endogenous cortisol concentration is increased (27, 28). Thus corticosteroids can modulate BBB function in the preterm fetus and could potentially influence any disruption with dopamine-induced hypertension after hypoxia-ischemia (18).

AIB is considered a sensitive tracer for barrier disruption because of its small size (104 Da), its sequestration in cells by the A transport system to limit backdiffusion, and the lack of metabolic breakdown and release of label (2, 9). This tracer is sufficiently sensitive to detect postischemic permeability increases of 60-70% in cortex and 90-107% in thalamus and brain stem of immature pigs (22). In the present study, the 95% confidence limits were within 20-25% of the mean transfer coefficient in cortical regions. Thus the AIB technique should have been able to detect smaller increases in uptake than that seen after ischemia. Interestingly, the increase in AIB uptake in piglets occurred 4 h after resuscitation and not immediately after resuscitation during the epinephrine-induced rebound hypertension (22). This delayed increase in permeability was prevented by treatment with superoxide dismutase (21). In mature brain, transient severe hypertension can cause prolonged superoxide production, vascular dysregulation ameliorated by superoxide dismutase, and increases in BBB permeability ameliorated by superoxide dismutase (32, 34). Thus we cannot exclude that dopamine-induced hypertension could cause a delayed injury to the endothelium.

Minimization of AIB backdiffusion by sequestration of label permits relatively long circulation times of the tracer to be used and allows the concentration in the plasma to decrease to relatively low levels. In the present study, we did not correct the tissue counts of AIB for the amount of tracer left in the plasma space, because we estimated the error to be small. By 60 min of circulation, the concentration of tracer in the plasma had decreased to ~30 dpm/µl. Counts in the tissue were in the range of 15-30 dpm/mg. With a plasma space of ~0.03 µl/mg in fetal sheep (26), the counts in the tissue attributed to plasma would be on the order of 1 dpm/mg, or 3-7% of the total tissue counts. Therefore, even if dopamine were to decrease cerebral plasma volume by 50%, which is highly unlikely, the extravascular AIB concentration would be underestimated by at most 1.5-3.5%, which is well within experimental error. Therefore, the 60 min of AIB circulation permitted low concentrations in the plasma, which did not require correction for potential changes in cerebral blood volume. Moreover, an increase in arterial and venous pressure with dopamine is more likely to increase, rather than decrease, cerebral blood volume and result in a slight overestimation of K_i. Furthermore, the control values of K_i obtained in each region in each age group closely matched the values obtained by Stonestreet et al. (26), who performed the plasma volume correction.

Changes in K_i are assumed to represent changes in the permeability-surface area product and to be blood flow independent. Blasberg et al. (2) estimated that K_i would be within 6% of the permeability-surface area product when K_i is <10% of plasma flow (if it is assumed that AIB is not distributed in red blood cells). Because forebrain blood flow is ~35-40 ml · min¹ · 100 g¹ at 0.6 gestation and ~120 ml · min¹ · 100 g¹ at 0.9 gestation (10) and because hematocrit is ~33-42%, K_i should be <25 µl · g¹ · min¹ at 0.6 gestation and <75 µl · g¹ · min¹ at 0.9 gestation for K_i to be essentially independent of blood flow. Forebrain K_i values of ~4 µl · g¹ · min¹ in the present study are well below these limits. Therefore, dopamine would have to cause an extremely profound decrease in blood flow to make K_i blood flow dependent.

Dopamine could reduce blood flow secondary to the observed increase in arterial O₂ content (from 4.2 to 5.9 mmol/l at 0.6 gestation). This increase would be predicted to decrease forebrain blood flow by 15% at 0.6 gestation (11). Such a small predicted decrease in blood flow should not have a significant effect on AIB uptake.

Another consideration is that a decrease in capillary surface could potentially mask an increase in AIB permeability. However, hypertension induced by dopamine is unlikely to decrease capillary recruitment and capillary surface area. In mature brain, all capillaries are thought to be perfused continuously with plasma, even when blood flow is substantially reduced during ischemia (30). This issue has not been well addressed in immature brain. It is conceivable that plasma flow in individual capillaries could be intermittent in immature brain. However, AIB was infused over a 3-min period and gradually decayed to within 10% of the 60-min value by 20 min. Thus plasma AIB concentration was changing on a time scale of minutes and not seconds. It is unlikely that the periodicity of any intermittent plasma flow for the majority of capillaries would be on the order of minutes. Hence, most capillaries are expected to be exposed to a similar plasma concentration profile of AIB.

In the 132-day fetuses, the AIB concentration-time integral was greater in the fetuses receiving dopamine because of errors in estimating fetal body weight at the time of surgery for calculating the dose of AIB. However, this difference in the time integral between the saline control and dopamine-treated groups should not cause an error in the calculated transfer coefficient, because the transfer coefficient remains constant over a wide range in values of the plasma concentration-time integral (2).

In conclusion, results of the present study imply that the BBB of fetal sheep is sufficiently robust to withstand 40% increases in arterial pressure without acutely sustaining increases in permeability to a small tracer amino acid, even by 0.6 gestation. Dopamine is often used to support arterial pressure in preterm human newborns. The present results demonstrating a lack of effect of dopamine-induced hypertension on BBB permeability and cerebral edema in utero are somewhat reassuring. However, it should be borne in mind that cerebrovascular autoregulation may be more effective in limiting BBB disruption in healthy fetal sheep than in sick human newborns, especially those that experience an asphyxic insult. In addition, the normal increases in arterial PO₂ and arterial blood pressure that occur after birth may render cerebral vessels far more susceptible to oxidative stress. Therefore, dopamine should be used judiciously in sick newborns, and care should be taken to avoid abrupt hypertension.