INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE BLOOD-BRAIN BARRIER is composed of a continuous layer of cerebrovascular endothelial cells connected by intercellular tight junctions (3, 4, 9). This specialized barrier serves as an interface between circulating blood and brain interstitium and parenchyma, isolating brain tissue from blood constituents. Therefore, the blood-brain barrier maintains central nervous system homeostasis by preventing the entry of substances that might alter neuronal function. We have previously shown that, when ovine fetuses are examined in utero under stable homeostatic conditions, the blood-brain barrier is relatively impermeable to a small hydrophilic molecule, -aminoisobutyric acid (32).

It is well known that the majority of premature infants who develop intraventricular hemorrhage have been exposed to mechanical ventilation (15, 25, 35, 36). Although the exact mechanism(s) of the associations between ventilation and the development of intraventricular hemorrhage has not been completely delineated, it is plausible that ventilation-related changes in the microvasculature might predispose the endothelium to rupture.

Increases in venous pressures have been implicated in the disruption of the blood-brain barrier (7, 16). During acute hypertension, elevations in sagittal sinus pressures are associated with increases in blood-brain barrier permeability (7). Disruption of the blood-brain barrier during superior venae cavae occlusion and hypertension has been attributed to increases in pial venular pressure and diameter at the site of disruption (16). Moreover, venules and veins appear to be more susceptible than arterioles and capillaries to disruption of the blood-brain barrier (16).

Taken together, the findings that 1) the majority of infants with intraventricular hemorrhage have been exposed to mechanical ventilation, 2) the vascular origin of intraventricular hemorrhage has been localized to both the venous and arterial microvasculatures (5, 21), and 3) increases in pial venous pressures can be of sufficient magnitude to account for disruption of the blood-brain barrier suggest that ventilation-related changes might affect endothelial vascular function (7, 8, 15, 16, 24, 33, 35, 36).

Increases in mean airway pressure during positive-pressure ventilation have been shown to selectively alter blood-brain barrier function in newborn pigs (20). We have been examining the ontogeny of blood-brain barrier function in ovine fetuses and lambs (31, 32). During these studies, we have inadvertently found that the duration of ventilation also influenced blood-brain barrier function in ventilated premature lambs.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

This study was conducted after approval by the Institutional Animal Care and Use Committees of Brown University and Women and Infants' Hospital of Rhode Island and according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Animal preparation. Nine hysterotomy-delivered surfactant-treated premature lambs were studied at 90% or 135-139 days of gestation. For the purpose of this report, lambs at this gestational age are designated as premature lambs. As previously described in detail (30, 32), surgery was performed on nine mixed-breed ewes under 0.75-2.0% halothane and oxygen anesthesia at 131-133 days of gestation. Briefly, in the fetuses, polyvinyl catheters were placed into a brachial vein for isotope administration and a brachial artery and advanced to the thoracic aorta for blood sample withdrawal and blood pressure monitoring. After insertion, the catheters were closed and attached to the skin of the fetus. After the forelimb had been replaced into the uterus, the head and neck were exposed through the same incision. An endotracheal tube was then placed into the trachea to facilitate immediate suctioning and ventilation at delivery. The endotracheal tube remained open in the amniotic cavity to allow for lung liquid egress. The uterus and abdomen of the ewe were then closed.

Experimental protocol. After 2-4 days of recovery from surgery at 90% or 135-139 days of gestation, a hysterotomy was performed under ketamine anesthesia (15-40 mg/kg iv) to the ewe, and the fetuses were delivered. Two of the premature lambs were from singleton pregnancies and six were from twin pregnancies. In another twin pregnancy, only one premature lamb was studied. The ewe was then killed with an overdose of pentobarbital sodium (100-200 mg/kg). Immediately after delivery, the premature lambs were suctioned via the endotracheal tube, treated with surfactant (100 mg/kg, Survanta, Beractant, Ross Products, Columbus, OH), hand ventilated, stabilized, and placed on a positive-pressure ventilator (Bio-Med Devices, Flow-Disc MVP-10, pediatric respirator, Stamford, CT). Ventilation was begun on room air or oxygen as needed with a respiratory rate of 18-71 breaths/ min, a peak inspiratory pressure of 14-43 mmHg, and a positive-end expiratory pressure of 2-10 mmHg to achieve an initial arterial oxygen tension of 50-190 Torr and a carbon dioxide tension of 23-36 Torr. After the lambs had received surfactant and were stabilized on the ventilator with blood-gas values as outlined above, only minor changes in ventilator rates and airway pressures were necessary to maintain the blood-gas values similar to the lamb's own initial values.

Analytic methods. Blood-brain barrier function was measured in the lambs with -[¹⁴C]aminoisobutyric acid (DuPont-NEN, Boston, MA) as previously described (6, 22, 29, 32). Briefly, after physiological determinations were obtained, -aminoisobutyric acid was rapidly injected intravenously, and the arterial plasma concentrations were obtained at fixed times before and after injection as follows: at 1, 0.25, 0.5, 1, 2, 3, 5, 7, 9, 15, 25, 35, 45, 55, and 60 min and at termination within 4-5 min after the end of the study. Brain vascular volume was determined in separate premature lambs treated under the same protocol (n = 3) with [¹⁴C]polyethylene glycol (59-83 µCi, Amersham) injected intravenously 2 min before the end of the experiment. The brain vascular volume values were similar to our previous values in late-gestation fetal sheep (32). -Aminoisobutyric acid is a synthetic amino acid that is not present in mammalian tissues. This amino acid has been used extensively to measure accurately the total and regional blood-brain barrier permeability in a variety of mammals, including fetal, newborn, and adult sheep (2, 32, 34). The premature lambs received 62-83 µCi of -[¹⁴C]aminoisobutyric acid. Arterial blood samples were withdrawn from the thoracic aorta as outlined above. Intermittent samples were withdrawn rather than a single integrated sample so that the plasma radioactivity profile could be compared among the premature lambs to ensure the accuracy of the methodology. The pattern of the plasma concentration profiles in our lambs was similar to that in our previous report (32).

1

1

Calculations and statistical analysis. The mean airway pressure (MAP, cmH₂O) was calculated by the formula
where PIP is the peak inspiratory pressure and PEEP is the positive end-expiratory pressure (cmH₂O). TI and TE are the inspiratory and expiratory times (s), respectively.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The premature lambs weighed 3.46 ± 0.77 kg. The mean duration of ventilation was 306 min with a range of 162-474 min. Calculated mean airway pressures available in two of nine lambs in this report were 7.9 and 8.3 cmH₂0. To ascertain that mean airway pressure did not increase with the duration of ventilation, mean airway pressures were calculated from a group of similarly studied lambs (n = 17) at the same gestational age (31). In this group, the mean airway pressure was 14 cmH₂O with a range of 7-20 cmH₂O. Calculated mean airway pressures did not correlate with the duration of ventilation [r = 0.27, n = 17, not significant (NS)].

Arterial pH, oxygen tension carbon dioxide tension, base excess, hematocrit, heart rate, mean arterial blood pressure, plasma osmolality, and glucose and lactate concentrations in our premature sheep were within normal physiological ranges (14, 23). Minor decreases in arterial pH and hematocrit and increases in carbon dioxide tension and base excess were in the physiological range and probably not of major physiological consequence (Table 1).

The K_i (µl · g brain¹ · min¹) values demonstrated direct linear correlations with the duration of ventilation (min) in the cerebral cortex (Figs. 1 and 2, r = 0.76, n = 9, P < 0.05), cerebellum (r = 0.83, n = 9, P < 0.05), medulla (r = 0.67, n = 9, P < 0.05), caudate nucleus (r = 0.67, n = 9, P < 0.05), thalamus (r = 0.73, n = 9, P < 0.05), hippocampus (r = 0.80, n = 9, P < 005), pons (r = 0.81, n = 9, P < 0.05), superior colliculus (r = 0.71, n = 9, P < 0.05), and cervical spinal cord (r = 0.69, n = 9, P < 005). Similar findings were observed in the inferior colliculus (K_i = 0.02 min + 0.40, r = 0.81, n = 9, P < 0.05) and choroid plexus (K_i = 0.02 min + 3.23, r = 0.86, n = 9, P < 0.05) but not in the pituitary gland (r = 0.38, n = 9, NS).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Blood-to-brain transfer constant (Ki) values in the cerebral cortex, cerebellum, and medulla plotted against duration of ventilation. Cerebral cortex: r = 0.76, n = 9, P < 0.05; cerebellum: r = 0.83, n = 9, P < 0.05; medulla r = 0.67, n = 9, P < 0.05.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Ki values in the caudate nucleus, thalamus, hippocampus, pons, superior colliculus, and cervical spinal cord plotted against duration of ventilation. Caudate nucleus: r = 0.67, n = 9, P < 0.05; thalamus: r = 0.73, n = 9, P < 0.05; hippocampus: r = 0.80, n = 9, P < 005; pons: r = 0.81, n = 9, P < 0.05; superior colliculus: r = 0.71, n = 9, P < 0.05; cervical spinal cord: r = 0.69, n = 9, P < 005.

When the three sets of premature twin lambs were compared by ANOVA, the K_i values across the brain regions were higher in the second twin, which had been studied after a longer duration of ventilation (ANOVA interactions F = 2.8, P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

This study examined the effects of the duration of ventilation on blood-brain barrier permeability to -aminoisobutyric acid in hysterotomy-delivered surfactant-treated premature lambs. The novel finding of our study was that regional blood-brain barrier permeability increased significantly in most brain regions as the length of ventilation increased. It is important to emphasize that these changes in barrier permeability occurred over relatively short time intervals (minutes to hours).

Lambs delivered prematurely lack pulmonary surfactant and develop respiratory distress syndrome similar to that observed in premature infants (14, 27). Although our premature lambs were examined relatively late in the ovine gestation, premature lambs at this gestational age are known to develop severe respiratory distress (27, 28). Moreover, surfactant treatment of lambs at birth before the initiation of ventilation results in a relatively homogeneous distribution of surfactant (11, 13, 14). Therefore, in this study, the hysterotomy-delivered premature lambs were suctioned and given surfactant before being placed on nonsynchronized positive-pressure ventilation (Bio-Med Devices, Flow-Disc MVP-10, pediatric respirator). After surfactant treatment and ventilation, the blood-gas values of the lambs were within the physiological range (14, 23) and remained similar to the baseline values during the studies. The small decreases in pH and increases in carbon dioxide tension during the studies were also within the physiological range. Therefore, the increases in regional brain K_i values with increasing lengths of ventilation cannot be attributed to the changes in arterial blood-gas values.

After surfactant treatment, pulmonary compliance improves in premature lambs (12, 13). Intermittent positive-pressure ventilation of compliant lungs has been reported to impede venous return to the heart (18). In addition, during intermittent positive-pressure ventilation, sagittal sinus pressures increase when mean airway pressure is elevated (19). Therefore, it remains plausible that, in our surfactant-treated premature lambs with relatively compliant lungs (10, 12, 13), intermittent positive-pressure ventilation might have been associated with elevated cerebral venous pressures. However, we did not measure sagittal sinus pressures during these studies because the lambs in this study were a part of a larger investigation of the ontogeny of blood-brain barrier permeability in fetuses and lambs and the findings reported here were not expected (31, 32).

In our study, it was important to establish that increasing lengths of positive-pressure ventilation were not associated with elevations in mean airway pressure, because previous work had demonstrated that increases in mean airway pressures selectively alter blood-brain barrier function in newborn pigs (20). Although we were unable to calculate mean airway pressure in all of the nine premature lambs in the present study, we were able to calculate it in 17 similarly studied premature lambs (31). Because the premature lambs were ventilated for various time intervals before study, it was plausible that, with increasing durations of ventilation, increases in mean airway pressure might have been required to maintain arterial blood-gas values similar to the baseline values. However, it appears that treatment with surfactant at birth was sufficient to reduce to pulmonary compliance in our premature lambs for the duration of these studies, because increasing lengths of ventilation were not associated with increases in mean airway pressures.

Increases in mean airway pressure have been shown to selectively alter blood-brain barrier permeability in newborn pigs (20). At elevated mean airway pressures (16.8 cmH₂O), the permeability of the blood-brain barrier to sucrose was increased significantly in the cerebrum and not in the lower brain structures (20), whereas barrier permeability to sodium decreased at elevated mean airway pressures (20). In the present study, barrier permeability, measured to a small-molecular-weight synthetic amino acid, increased in the cerebrum and in the lower brain structures as the length of ventilation increased.

Sucrose is a hydrophilic sugar (molecular weight 342) and -aminoisobutyric acid is a small-molecular-weight (103) synthetic hydrophilic amino acid. Therefore, it is likely that the blood-to-brain transport characteristics were relatively similar between these substances and differ substantially from the characteristics of other tracers such as sodium (20). Thus our findings, combined with the previous work (20), suggest that during intermittent positive-pressure ventilation both increases in mean airway pressure and the length of ventilation influence blood-brain barrier function. In contrast to the previous work, in which sucrose permeability increased at elevated mean airway pressures only in the cerebrum (20), our study showed that permeability increased as the length of ventilation increased in all brain regions except for the pituitary gland. The reason(s) that the duration of ventilation influences barrier permeability in most brain regions to -aminoisobutyric acid and mean airway pressure appears to affect sucrose transport only in the cerebrum remains to be determined. Although we did not examine the effects of the duration of ventilation on blood-brain barrier permeability to sucrose in the present study, it would be of interest to examine the permeability characteristics to sucrose or other small-molecular-weight molecules to be certain that our findings with -aminoisobutyric acid can be generalized to other small molecules. However, the dichotomy between the brain regions affected for sucrose with elevated mean airway pressures (20) and for -aminoisobutyric acid with constant mean airway pressures and increasing durations of positive-pressure ventilation might relate in part to the higher molecular weight of sucrose than -aminoisobutyric acid.

The mechanism(s) underlying the increases in regional blood-brain barrier permeability as the length of ventilation increased cannot be determined from our studies. However, several possibilities exist. Several lines of evidence suggest that increases in venous pressures might be the mode by which positive-pressure ventilation affects blood-brain barrier function (1, 7, 16, 18). Although we did not measure cerebral venous pressures in the present study, we speculate that, during intermittent positive-pressure ventilation, fluctuations (25) and/or increases in venous microvascular pressures (16) might have in part contributed to the increases in regional barrier permeability observed with the increasing lengths of ventilation in the premature lambs in this study. Alternatively, it has been previously shown that increases in mean airway pressures are associated with increases in cerebrospinal fluid prostanoids (6-keto-PGF₁, thromboxane-B₂, PGE₂, and PGF₂) (17). Oxygen-derived free radicals are generated during prostanoid production, and oxygen free radicals have been shown to alter blood-brain barrier permeability (37). Although we do not know whether prostanoid production increases with increasing lengths of ventilation, it remains possible that prostanoid-related increases in oxygen free radicals might have contributed to the increases in blood-brain barrier permeability in our study.

In summary, we conclude that increases in the length of positive-pressure ventilation predispose premature lambs to increases in regional barrier permeability to -aminoisobutyric acid. These alterations in barrier function occur over relatively short time intervals (minutes to hours) and are most likely not attributable to increases in mean airway pressure.

Perspectives. The majority of infants who develop intraventricular hemorrhage have been exposed to mechanical ventilation (15). Respiratory distress syndrome and the mechanics of ventilation are associated with fluctuating patterns of cerebral blood flow velocity and arterial blood pressure (25, 26). These fluctuating patterns have been reported to place infants at high risk for intraventricular hemorrhage (25, 26). Positive-pressure ventilation results in increases in sagittal sinus pressure (18). Increases in sagittal sinus and microvascular venous pressures appear to be associated with disruption of the blood-brain barrier (7, 16). Therefore, it is reasonable to speculate that, during intermittent positive-pressure ventilation, both increases in mean airway pressures (20) and the length of ventilation may render the delicate microvascular endothelium vulnerable to barrier breakdown and/or rupture. At least one report has suggested that bleeding in the germinal matrix was found to be perivenous (21). Therefore, it remains possible that ventilation places the premature infant at risk for intraventricular hemorrhage because of endothelial rupture as a result of increases in cerebral microvascular venous pressures.