Main

Hypoxic-ischemic encephalopathy, a common neurologic sequela of perinatal asphyxia, is a significant cause of neonatal mortality and morbidity. In human neonates, perinatal asphyxic/ischemic or hypoxic injury may result from different conditions with varying degrees of insults and therefore with variable outcomes. Similarly, in experimental asphyxia, cerebral hemodynamic alterations and manifestations appear dependent upon the species, the model of asphyxia, and the kind of interventions that were made during insult or recovery(1–5). Thus, reports differ as to changes in postasphyxic CBF. Increases in CBF up to 150% and“luxury perfusion” have been described(6, 7). Others have reported on decreases in CBF and the “no-reflow phenomenon” after asphyxia(8–11). Experiments in primates have suggested that, to cause disability, asphyxia may need to be prolonged and nearly fatal(12). The most severe outcome is observed with prolonged partial asphyxia(12), especially with concomitant hypotension(13–16). We have found that in total asphyxia, i.e. respiratory arrest of short duration(1) or with partial asphyxia of 30-min duration(17), piglets recovered from asphyxic insult without systemic hypotension. In addition, during recovery, cerebrovascular reactivity was preserved in total asphyxia but attenuated in partial asphyxia(1, 5, 17).

Postasphyxic hypotension is a frequent finding in babies with severe asphyxic insult. The purpose of this study was to develop a model of“partial” asphyxia in newborn pigs, severe enough to produce systemic hypotension. We hypothesized that prolonged, partial asphyxia with hypotension causes loss of cerebrovascular reactivity and altered cerebral hemodynamics during recovery. In an effort to better understand the effects of prolonged partial asphyxia on the cerebral circulation, we monitored the changes in periarachnoid CSF, cAMP and cGMP concentrations, pial arteriolar BF(qualitatively) and diameters, regional and hemispheric CBF, and cerebrovascular reactivity.

METHODS

The surgical and experimental procedures were reviewed and approved by the Animal Care and Use Committee of The University of Tennessee, Memphis.

Piglets 2-5 d of age (1-2 kg) of either sex were anesthetized and instrumented before the experiments. Piglets were initially anesthetized with intramuscular ketamine hydrochloride (33 mg/kg) and acepromazine (3.3 mg/kg) and maintained on i.v. α-chloralose (50 mg/kg initially plus 3 mg/kg/2 h). The animals were ventilated with room air through a tracheostomy tube using a Bournes BP200 infant ventilator. Catheters were inserted into the femoral vein for maintenance of anesthesia and fluid administration (100 mL/kg/d of 5% dextrose in water), into the femoral arteries to record BP and heart rate, for blood gas sampling, and also for reference withdrawal in microspheres studies. A fourth catheter was inserted into the ventricle through the right carotid artery to inject the microspheres. In piglets, ligation of one carotid artery has no detectable effect on hemispheric CBF(4, 18). A closed cranial window was inserted over the parietal cortex for measurements of pial arteriolar diameters, qualitative assessments of pial arteriolar flow, and collection of cortical periarachnoid CSF for measurement of cAMP and cGMP. Body temperature was maintained between 37 and 38 °C.

Protocol. Two groups of piglets were studied, a time-control group and an asphyxia group. In the time-control group (n = 8), pial arteriolar diameters, periarachnoid CSF cyclic nucleotides, and CBF were measured at intervals under normoxic and normotensive conditions maintained for 2 h. In the asphyxia group (n = 11), 30 min after stabilization from surgery, asphyxia was produced for 60-min duration by ventilating the piglets with a gas mixture of 10% CO2, 10% O2, and 80% N2 and by decreasing the minute volume. Respiratory support and the gas mixture were manipulated to maintain the systemic BP around 4 kPa and the heart rate between 50 and 80 beats/min. At the end of 60 min of asphyxia, the minute volume was increased, and the piglets were ventilated with room air. Piglets were evaluated during 60-90 min of recovery. Arterial blood was collected every 10 min for blood gases and pH measurement using an Instrumentation Blood Gas Analyzer (Instrumentation Laboratory, Lexington, MA). BP and heart rate were monitored continuously during the experiments. Cortical periarachnoid CSF was collected every 10 min for later measurements of cAMP and cGMP. Diameters of two to four pial arterioles were recorded every 2 min during the experiments. Vascular reactivity to topically applied isoproterenol(10-4 M) and histamine (10-4 M) was evaluated before asphyxic insult and repeated during 60-90 min of recovery. Measurements of pial arterioles after application of each compound were taken every minute for 10 min. At the end of 10 min, the space under the cranial window was flushed with artifical CSF for 15-30 min before application of the second compound. CBF and cardiac output determinations were made at baseline, at 10, 30, and 60 min of asphyxia, and at 10 and 60 min of recovery.

Radioactive microsphere determinations Of CBF distribution. A known amount of radioactivity in 15-μm microspheres (≥1 000 000 microspheres) was injected into the left ventricle, and the injection line was flushed with 1 mL of saline. Withdrawal of reference blood samples (1.03 mL/min from the descending aorta) was begun 15 s before microsphere injections and continued for 2 min. Withdrawn blood was replaced with donor piglet blood. At the end of the experiments, the piglets were killed by injecting KCI into the left ventricle, and the brain was removed. The brain was subdivided into right and left hemispheres and major regions (cerebrum, white matter, gray matter, caudate nucleus, pons, medulla, diencephalon/mesencephalon, cerebellum, and choroid plexus). Samples were counted in a gamma-counter. The energy from each nuclide was separated by differential spectroscopy. Aliquots of the actual microsphere solutions injected were used for overlap calculations. The nuclides and energy windows used were (in keV) 46Sc 840-1240, 95Nb 710-810, 103Ru 480-510, 113Sn 370-440,57 Co 100-150, and 125I 36-70. The lungs were counted to detect extensive arteriovenous shunting of microspheres. “Lung blood flow” (comprised of bronchial flow and whole body arteriovenous shunt flow) averaged 2% of cardiac output, indicating that no extraordinary shunting of microspheres occurred. Cardiac output was calculated as (reference withdrawal rate) × (counts injected) × (counts in reference withdrawal)-1. BF to each brain region at the time the microspheres were injected was calculated by using the formula: Q = C× R × CR-1, where Q = organ blood flow in mL/min × 100 g, C = counts/100 g of tissue,R = rate of withdrawal of reference blood sample in mL/min, and CR = total counts in reference arterial blood sample.

Closed cranial window placements for measurements of pial arteriolar diameters and collection of periarachnoid CSF. To implant the cranial window the scalp was cut and reflected from the skull. A hole approximately 2 cm in diameter was made in the skull. Incision was made through the dura and arachnoid membranes, and these membranes then were reflected over the edge of the bone. A stainless steel ring with a premounted glass coverslip was inserted in the hole. The window was cemented in place with dental acrylic. Three needles pierced the ring and allowed injection of artificial CSF under the window and sampling of CSF from under the window. The space under the window was filled with artificial CSF (Na+ 150 mEq/L, K+ 3 mEq/L, Ca2+ 2.5 mEq/L, Mg2+ 1.2 mEq/L, Cl- 13 mEq/L, glucose 3.7 mM, urea 6 mM, HCO3- 25 mEq/L, pH = 7.33, Pco2 = 6.1 kPa, Po2 = 5.7 kPa). The volume directly below the window was approximately 500 μL. After implantation of the window, 30 min were allowed for equilibration of CSF under the window. Pial arterioles were observed with a Wild dissecting microscope. Pial arteriolar diameters were measured with a television camera mounted on the microscope, a video monitor, and a video microscaler. Briefly, the images of the arterioles were displayed on the television monitor, and the sides of the vessels were bracketed by parallel lines projected by the microscaler. As the arteriolar diameter changed, the lines were moved manually to correspond with the vessel walls. Precalibration of the distance between the lines allowed determination of arteriolar diameter. BF changes through the pial arterioles were also qualitatively monitored during the experiments.

Cyclic nucleotide assays. Cortical periarachnoid CSF (0.4 mL) collected from under the window was mixed with EDTA (5 mM) and stored at -60°C until assayed. cAMP and cGMP were measured in CSF samples using RIA procedures and commercial 125I-cAMP or 125I-cGMP scintillation proximity assay systems (Amersham International plc, Amersham, UK). Acetylation of CSF samples with a 2:1 mixture of triethanolamine and acetic anhydride was performed immediately before assay to increase the sensitivity of the method (the analysis range is 2-128 fmol of cAMP and 4-512 fmol of cGMP/sample). These RIAs used the simultaneous addition of sample,125 I-cAMP or 125I-cGMP, rabbit cAMP or cGMP antibody, and an anti-rabbit second antibody bound to scintillant-incorporated microspheres. Samples were mixed overnight on an orbital shaker (200 rpm) at room temperature. To determine the amount of 125I-cAMP or 125I-cGMP bound to the light-producing microspheres, the vials were counted using aβ-scintillation counter. All unknowns were assayed at two dilutions. The cyclic nucleotide concentration in the CSF sample was calculated from the standard curve.

Statistical analysis. Values are reported as means ± SEM. Comparisons among different time periods were made using analysis of variance with repeated measures and followed by Fisher's protected least significant difference test. Comparisons between the two groups were made using a t test for planned comparisons (paired or unpaired, as appropriate). A p value <0.05 was considered significant. First-level Winsorization(19) was used for statistical analysis when a single value was extreme in comparison with the remainder of the sample.

RESULTS

Physiologic measurements in the time-control piglets are shown in Table 1. Arterial BP, blood gases, and pH were maintained at baseline levels over time. We observed a trend toward an increase, rather than a decrease, in pial arteriolar diameter, CBF, and cyclic nucleotides in 2 h.

Table 1 Arterial blood pressure, pH, and gases, CSF periarachnoid cyclic nucleotides, and CBF measurements in time-control piglets
Full size table

Four piglets did not tolerate the asphyxic insult and expired at 30-50 min of asphyxia in spite of aggressive resuscitation. The CBF measurements were not performed in one piglet because of technical difficulty; complete CBF data were obtained from six piglets. The cranial window data were available from seven piglets. In each piglet, two to four pial arterioles of different diameters (85-220 μm) were monitored every 2 min.

All piglets started with normal BP, pH, and blood gas values (Fig. 1). With the onset of asphyxia, BP decreased from 10± 0.7 to 4.7 ± 0.3 kPa and was maintained at that level by manipulating the degree of asphyxia (changes in inspired gas mixture). During recovery, BP increased only to 6 ± 0.7 kPa and remained at that level. With asphyxia, arterial pH decreased from 7.40 ± 0.06 to 6.90 ± 0.02, followed by a gradual increase to 7.30 ± 0.01 during recovery. During asphyxia, Paco2 increased from 4.5 ± 2 to 12.5 ± 0.5 kPa and Pao2 decreased from 11.7 ± 0.8 to 4.7 ± 0.3 kPa, respectively, both measurements quickly returning to baseline upon reventilation. During recovery, piglets were ventilated with room air, and no attempt was made to correct the metabolic acidosis or the systemic hypotension.

Figure 1
figure 1

Changes in arterial BP, pH, Po2 and Pco2 during asphyxia and recovery. x axis, time period (min); y axis, BP (kPa), pH, Pao2 (kPa), and Paco2 (kPa). n = 7 piglets. Values are means ± SEM. *p < 0.05 compared with baseline (0 min); mm Hg = kPa/0.1333.

Full size image

Figure 2 shows the cyclic nucleotides concentrations from the periarachnoid CSF collected every 10 min during asphyxia and recovery. The CSF cAMP increased 3-fold by 20 min of asphyxia; at 40 min, values gradually decreased to baseline level. During early recovery, the cAMP increased 2-fold, but, as recovery continued, the CSF cAMP decreased and stabilized at 50% below baseline values. The CSF cGMP increased over 3-fold by 40 min of asphyxia. With continuation of asphyxia, cGMP decreased toward baseline. A 2-fold increase in CSF cGMP was observed in early recovery, returning toward baseline as recovery continued.

Figure 2
figure 2

Percent changes in periarachnoid CSF cAMP and cGMP concentrations during asphyxia and recovery. n = 7 piglets. Values are means ± SEM. *p < 0.05 compared with baseline(0 min).

Full size image

Pial arteriolar diameters increased significantly with the onset of asphyxia (Fig. 3) and then decreased during the last 30 min of asphyxia, but measurements remained above baseline. These findings persisted throughout the recovery period. At the onset of asphyxia, red cell movement within the pial arterioles appeared to increase (qualitative assessment), and the vessels were engorged. As asphyxia continued, red blood cell movement decreased, and cells became clumped together inside the vessels. During recovery, initially forward flow was noted, but soon red blood cell movement decreased and, utimately, sludging resulted. Thus there was a lack of correlation between the pial arteriolar diameter and CBF.

Figure 3
figure 3

Changes in pial arteriolar diameters during asphyxia and recovery. x axis, time period (min); y axis, pial arteriolar diameter (μm). A total number of 23 pial arterioles were measured in seven piglets. Values are means ± SEM. *p< 0.05 compared with baseline (0 min).

Full size image

Regional CBF and cardiac output measurements at baseline and at 10, 30, and 60 min of asphyxia and again at 10 and 60 min of recovery are shown in Table 2. Although right and left hemispheric BF were measured, no differences were observed between hemispheres. Fluctuations of CBF during asphyxia and recovery had different patterns in different regions (Fig. 4). BF to the cerebrum, white matter, gray matter, and caudate nucleus progressively decreased during asphyxia, with“no-reflow” for the duration of reventilation. BF to the brainstem region (pons and medulla) had biphasic patterns during asphyxia and recovery; BF to these regions returned toward baseline by 60 min of recovery. During asphyxia, BF to the cerebellum and choroid plexus decreased significantly. However, during recovery, BF to these regions returned to about one-third of baseline. Cardiac output decreased during asphyxia. After a transient increase during early recovery, cardiac output decreased significantly below baseline level.

Table 2 Changes in CBF(mL·min-1 · 100g-1) and cardiac output(mL·min·kg) during baseline, asphyxia, and recovery*
Full size table
Figure 4
figure 4

Changes in blood flow (medulla, cerebellum, and cerebrum) during asphyxia and recovery. x axis, time period (min);(y) axis, blood flow (mL/100 g-1·min-1).n = 6 piglets. Values are means ± SEM. *p< 0.05 compared with baseline (0 min).

Full size image

As to pial arteriolar reactivity to topical vasodilators, the percent increase in pial arteriolar diameter with isoproterenol administration decreased from 45 ± 7% to 8 ± 3% (p < 0.05). Similarly, percent increase in pial arteriolar diameter with histamine decreased from 53 ± 5% to 5 ± 3% (p < 0.05).

DISCUSSION

Our findings in newborn pigs with prolonged partial asphyxia/hypotension include: 1) a significant decrease in CSF cAMP but not cGMP during recovery, 2) a no-reflow phenomenon that was observed in higher brain regions, and 3) the loss of cerebrovascular reactivity to vasodilator stimuli.

Partial asphyxia is a common cause of perinatal brain damage in human infants(7). In this experiment, we were able to develop a model of partial asphyxia/incomplete ischemia severe enough to produce systemic hypotension during asphyxia and recovery. This model differs from total asphyxia(20) wherein there is complete cessation of gas exchange and from complete ischemia when blood is almost entirely stripped away from the cerebral circulation. During partial asphyxia/incomplete ischemia the presence of intraluminal blood may cause a greater generation of lactic acid and oxygen free radicals(10, 21–25). In our model, the persistent hypotension may be more detrimental, because systemic arterial BP is a critical factor in maintaining CBF(13, 16, 26).

Cerebral cyclic nucleotides. The effects of asphyxia/ischemia on cerebral cyclic nucleotides have not been studied extensively. The limited studies show that the brain's response to acute ischemia and hypoxia is increased brain tissue and CSF cAMP(23, 27). Our study is unique because we made longitudinal measurements of CSF cAMP and cGMP every 10 min during asphyxia and recovery. In our piglets, during prolonged partial asphyxia, CSF cAMP increased significantly and returned to baseline with continuing asphyxia. After a transient rise in CSF cAMP during early recovery, the CSF cAMP levels fell and stabilized at 45-50% below baseline, as recovery continued. Our time-control piglets did not show decrease in CSF cAMP over time.

Ischemia and recirculation appear to be situations where the fluctuating levels of the brain's ATP contribute to the regulation of cAMP(27). When the brain's O2 supply is abolished even briefly, there is a precipitous fall in ATP and a sharp rise in concentration of cAMP(28). cAMP is synthesized from ATP by the plasma-membrane-bound enzyme, adenylyl cyclase, and it is rapidly and continuously destroyed by one or more cAMP phosphodiesterases, hydrolyzing cAMP to 5′-AMP(29).

Possible causes for low CSF or brain cAMP include deficiency of ATP, increased phosphodiesterase activity, low adenylyl cyclase activity, and decreased adenylyl cyclase agonists, such as adenosine and catecholamines(30). Our observation of low levels of CSF cAMP during late recovery may be explained by sluggish ATP regeneration, which is likely to occur with associated mitochondrial damage after hypoxia/ischemia(15, 31) and hypotension(14). Irreversible alterations in mitochondrial function(3, 32) also could be induced by very low pH, leading to a permanent suppression of ATP production even when oxygen is present(3, 32). During late reperfusion, the period of“secondary ischemia,” generation of oxygen free radicals may further impair mitochondrial ATP production, leading to low levels of cAMP and neuronal death(23, 24, 33, 34).

cGMP, like cAMP, participates in intracellular signaling. However, catabolism and resynthesis of guanine nucleotides differs from the adenylate system(35). Guanylyl cyclase catalyzes the production of cGMP from GTP. Low concentrations of cGMP have been measured in the cerebral cortex of gerbils after ischemic insult(27, 30). In our piglets, at no time during asphyxia or recovery did the cGMP level decrease below baseline values.

Cerebral blood flow and pial arteriolar diameters. Cortical BF measurements by microspheres correlated well with qualitative changes in pial arteriolar BF; both showed severe decrease in flow, especially during late recovery. Term infants with evidence of hypoxic-ischemic encephalopathy have been reported to have lowered estimated CBF in the first few days of life(36, 37). Although in the lower normal range of cerebrovascular autoregulation for newborn pigs(26), BP in our piglets was not adequate to maintain cerebral perfusion pressure. Lou et al.(38) reported that in fetal lambs restoration of systemic BP by infusion of whole blood after severe hypotensive-asphyxia did not raise the CBF. Hypotension in our piglets may be explained by decreased cardiac output, probably secondary to myocardial ischemia. This postichemic hypotension was severe enough to have caused the no-reflow phenomenon. The no-reflow phenomenon, first described in adult rabbits by Ames et al.(10) and subsequently observed in several other species(11, 39), was thought to be due to cerebral vasospasm. However, we did not see any pial arteriolar vasoconstriction. Instead, we observed pial arteriolar dilation, whereas cortical CBF was minimal or absent. Thus, changes in pial arteriolar diameter were not correlated with cortical BF.

Cerebral microvascular obstruction secondary to increased blood viscosity, endothelial cell injury, polymorphonuclear leukocyte-endothelial adhesion, and/or platelet and erythrocyte aggregation could also contribute to no-reflow(10, 40). We did observe intravascular clumping of blood cells during the recovery period.

Other proposed mechanisms for the no-reflow phenomenon include microvascular compression due to perivascular edema(40), generation of oxygen free radicals(25, 34), and influx of Ca2+ ions(41). Krause et al.(41) showed that calcium antagonists inhibited postischemic brain hypoperfusion. Rosenberg et al.(34) treated newborn lambs with oxygen free radical-scavenging enzymes and observed an improvement in postasphyxial CBF.

Variations in regional CBF responses were observed in our piglets. Brainstem structures showed a biphasic BF pattern during asphyxia and recovery. Preservation of brainstem BF has also been observed with short periods of asphyxia(1, 42). We found no difference between hemispheric BF during asphyxia or recovery. A similar observation was made by Laptook et al.(4). Piglets, like human newborns, have functional collateral brain circulation.

Cerebrovascular reactivity. We observed a generalized pial arteriolar unresponsiveness to vasoactive stimuli (isoproterenol and histamine), suggesting vasomotor paralysis(9). In newborn lambs, Rosenberg(43) observed attenuation of cerebral vasodilation to hypoxia and an absent response to hypotension during recovery from partial asphyxia. Sankaran et al.(36) reported, a loss of cerebrovascular reactivity to hypercapnia in the first few days of life in term infants with hypoxic-ischemic encephalopathy. Mechanisms suggested for impaired cerebral vasoreactivity include obstruction of resistance vessels by cerebral edema, excessive production of cerebral vasoconstrictor prostanoids, or vascular endothelial injury secondary to oxygen free radicals(25, 43). However, high levels of brain oxygen free radicals have not been found to be consistently associated with loss of cerebrovascular reactivity(44). Vascular reactivity is preserved in acute asphyxia(5) but lost or reversed in total cerebral ischemia(45, 46), each condition generating similar levels of brain oxygen free radicals.

Conclusion. Prolonged partial asphyxia with hypotension in newborn pigs resulted in no-reflow during recovery. During recovery, pial arterioles were unresponsive to topical application of vasodilators, suggesting cerebral vasoparalysis. The CSF cAMP was decreased during recovery, which may indicate cerebral mitochondrial damage secondary to systemic hypotension, metabolic acidosis, or generation of oxygen free radicals.