Main

Cocaine use in America has continued despite widespread concern about its detrimental effects. Cocaine abuse among women of childbearing age has been particularly concerning(1). In some hospitals as many as 25% of women admitted to labor and delivery have positive urine toxicology screens for cocaine(2). Cerebral infarction and hemorrhage after maternal cocaine use during pregnancy has been described in newborn infants(3, 4). Recent studies have suggested that cocaine may cause brain injury by altering cerebral vascular reactivity(5–14), but results have been conflicting. Although some investigators have shown that cocaine causes cerebral vasoconstriction(5, 6, 11, 12), others have shown vasodilation(7–9, 13, 14). These conflicting results have been attributed to species, anesthetic, and methodologic differences. In addition, we and others have suggested that differences in cocaine metabolites may be partially responsible(7–9). Cocaine is rapidly metabolized to two major metabolites: EME and BE. Madden and Powers(6) and Schreiber et al.(12) have demonstrated that EME dilates cerebral blood vessels, whereas BE constricts them. EME and BE have been detected in urine and/or plasma after cocaine administration in sheep, baboons, and humans(7–9, 15–17). We have previously shown that, in sheep, EME is the primary cocaine metabolite(7–9). The cocaine metabolic profiles of other species have not been systematically studied. No studies have examined cerebrovascular effects of cocaine metabolites in vivo. We used unanesthetized, chronically prepared neonatal sheep to avoid potential fetal effects associated with maternal injection; e.g. alterations in uteroplacental blood flow(18, 19). We tested the hypothesis that EME has cerebral vasodilatory effects in vivo and may be responsible, at least in part, for cocaine's cerebrovascular effects in sheep.

METHODS

Subjects. Eight newborn mixed-breed sheep (4 ± 2 d old) were used for this study. Animals weighed 4.3 ± 0.8 kg. All surgical and experimental procedures were approved by our Institutional Animal Care and Use Committee.

Surgical preparation. Newborn sheep were brought to the animal care facility with their ewes on the day before study. Each animal was given 450,000 U of procaine penicillin intramuscularly before surgery. All cutaneous sites of entry were prepared with alcohol and betadine solution. The sheep were anesthetized with pentobarbital (15-20 mg/kg), via a catheter placed percutaneously in the external jugular vein. Additional pentobarbital (1-2 mg/kg) was administered as needed.

Polyvinyl chloride catheters were then placed into the left ventricle and brachiocephalic artery (via the axillary arteries), and into both femoral arteries and a femoral vein. No evidence of arterial insufficiency to the limbs was noted either at the time of surgery or on the next day (the day of study). The catheters were flushed and filled with heparinized saline (10 U/mL), sutured to the skin, and placed into a pouch attached to the abdomen.

The sagittal, coronal, and lambdoid sutures were identified, and a shallow burr hole was drilled over the sagittal suture approximately 0.5 cm anterior to the lambda. The sagittal sinus was identified, and the overlying dura was punctured with a 19-gauge needle. A polyvinyl chloride catheter was then inserted into the sagittal sinus, and its tip was positioned anteriorly to the confluence of the sinuses to minimize contamination from extracerebral venous blood. The catheter was then flushed and filled with heparinized saline and sutured securely to the scalp. The sheep was then weighed and allowed to recover in the laboratory until it could stand and suckle, and then it was returned to its mother. The next day the sheep was brought back to the laboratory. Full recovery from anesthesia and surgery was assessed by clinical examination, suckling behavior, and arterial blood gas and pH values.

Measurements. Blood flow was measured using radiolabeled microspheres(20). Approximately 1 × 106 microspheres (0.4 mL) labeled with either 153Gd, 114In,113 Sn, 103Ru, 95Nb, or 46Sc (Dupont NEN, Boston, MA) were injected into the left ventricular catheter over 30 s followed by 5 mL of 0.9% saline. Reference samples were withdrawn from the axillary artery at a rate of 2.65 mL/min, beginning 30 s before microshpere injection and continuing for 1.0 min after the injection was complete. This technique provides complete trapping of microspheres by organs with comparable distribution of spheres with the reference sample(20). The brain was dissected, weighed, and placed directly into counting vials. Blood and tissue samples were counted on a gamma counter (Packard model 5530), and organ blood flows were calculated. All samples had greater than 400 microspheres. Microsphere injections were not associated with changes in heart rate or blood pressure.

Blood samples for pH, arterial Pco2 and Po2, oxygen saturation, Hb concentration, and hematocrit were withdrawn into heparinized syringes. Blood gas values and pH were measured using model ABL 30 blood gas analyzer and electrode (Radiometer America; Westlake, Ohio). Oxygen saturation and oxygen content were measured using a Hemoximeter (OSM-3, Radiometer America; Westlake, OH). Arterial hematocrit was measured by the microhematocrit technique. Arterial blood pressure and heart rate were continuously monitored (model 2400 Gould instruments; Oxnard, CA) throughout the experiment.

For measurement of cocaine and metabolite levels, 3 mL of arterial blood were collected in tubes containing 0.1 mL of enzyme inhibitor (equal parts of saturated sodium fluoride solution and a 10% solution of glacial acetic acid). The plasma was stored frozen (-70 °C) until analyzed. Cocaine, BE, and EME levels were measured by electron impact gas chromatography-mass spectrometry(multiple ion monitoring) after extraction with solid-phase extraction cartridges. Deuterated internal standards were used for quantitation. The assay gives a linear response across a concentration range of 3.1 to 1000μg/L. The limit of quantitation by this assay is 1.0 μg/L for each analyte(21).

Experimental protocol. On the day of the study (24 h after surgery), the animal was brought to the laboratory, placed in a specially designed study cart, and allowed 1 h to become accustomed to surroundings. Six measurements were made during each study. For each measurement, blood samples(3 mL) were drawn first, and then microspheres were injected into the left ventricle while a reference sample was withdrawn from the brachiocephalic artery. Two baseline measurements were obtained 15 min apart. Then, 15-30 min after the second baseline, pure EME (2.5 mg/kg; Sigma Chemical Co., St. Louis, MO), dissolved in 5 mL of 0.9% saline, was injected over 10 s into the femoral venous catheter followed by 5 mL of 0.9% saline flush. Four additional measurements were obtained at 30 s, and 2, 5, and 60 min after EME injection.

Samples for determination of EME, cocaine, and BE were drawn before administration of EME and at 30 s, and 2, 5, 60, and 180 min after EME injection.

At the end of the study, the neonatal sheep were anesthetized with pentobarbital and killed with an overdose of saturated KCl solution. Catheter positions were verified. The brain was removed, weighed, and dissected for analysis of regional blood flow.

Four additional animals were studied using the identical protocol except that, instead of EME, they received an injection (5 mL) of 0.9% saline. These animals were used to assess the stability of this awake, unanesthetized preparation over the time course of the study.

Data analysis and calculations. Brain blood flow was calculated as [cpmbrain/cpmref] × reference withdrawal rate (mL/min), where cpmbrain and cpmref represent radioactive counts/min in brain and reference samples, respectively. CBF represented blood flow to the cerebral hemispheres. Arterial oxygen content (Cao2) was calculated as: Cao2 = 1.35 × Hb × O2 saturation. Venous oxygen content (Cvo2) was calculated similarly. CMRo2 was calculated as CMRo2 = [Cao2 - Cvo2] × CBF where Cao2 and Cvo2 represent arterial and venous O2 content, respectively. Vascular resistance was calculated as MAP/organ blood flow. Cerebral oxygen transport was calculated as Cao2 × CBF. Cerebral fractional oxygen extraction was calculated as CMRo2/oxygen transport.

Differences within groups were analyzed by repeated measures analysis of variance. If the F test was significant, specific differences were tested with the Newman-Keuls test. An unpaired t test was used to identify differences between groups. Significance was considered at p < 0.05. Data are reported as mean ± SEM.

RESULTS

CVR, CBF, and MAP at baseline and after i.v. administration of EME are shown in Figure 1. Baseline values are consistent with previous studies in the newborn sheep. Within 30 s of EME administration, there was a 21 ± 1.4% decrease in cerebral vascular resistance. CVR remained decreased at 2, 5, and 60 min. Cerebral blood flow increased by 20± 0.55% at 30 s, remained elevated at 2 and 5 min, and returned toward baseline by 60 min. There were no changes in MAP.

Figure 1
figure 1

CVR (A), CBF (B), and MAP (C) in eight neonatal sheep before and 0.5, 2, 5, and 60 min after administration of EME. Mean ± SEM. *p < 0.01 compared with baseline(0 min).

Full size image

Regional brain blood flow responses are shown in Figure 2. In general, blood flow responses to the other brain regions studied paralleled those of the cerebral hemispheres with some differences in timing. Cerebellar and caudate blood flow increased as early as 30 s with cerebellar blood flow remaining increased at 60 min. Brainstem blood flow did not increase until 60 min.

Figure 2
figure 2

Regional brain blood flow (mL/100 g/min) at baseline (0 min) and 0.5, 2, 5, and 60 min after a single dose of EME (2.5 mg/kg). Values expressed as mean ± SEM; n = 8 neonatal sheep. *p < 0.05 compared with baseline.

Full size image

Arterial blood gases, glucose and lactate concentrations, MAP, and heart rate at baseline and in response to EME administration are shown in Table 1. There were no changes in lactate levels until the end of the study, at 60 min. Glucose concentration remained stable, as did pH, arterial Pco2 and Po2. Hb and hematocrit decreased slightly by the end of the study (0.32 ± 0.2 to 0.29 ± 0.2; and 106 ± 15 to 102 ± 18 g/L, respectively).

Table 1 Arterial pH, glucose, lactate, arterial Pco2 and Po2, hematocrit(Hct), Hb, MAP, and heart rate before and after EME (2.5 mg/kg)
Full size table

Cerebral O2 metabolism data are shown in Table 2. Cao2 decreased slightly after EME injection primarily because of the decrease in Hb concentration. Cerebral oxygen transport and extraction did not change significantly, although there was a trend toward an increase in O2 transport and a decrease in O2 extraction after EME. There was no change in CMRo2 throughout the study. There were no changes in animal activity, i.e. no seizures, no excitability, and no changes in animal behavior.

Table 2 Cerebral O2 metabolism before and after EME injection
Full size table

Plasma EME, cocaine, and BE levels are shown in Table 3. Plasma EME levels peaked at 0.5 min (2365 ± 114 ng/mL) and remained unchanged through the first 5 min of the study. By 60 min, EME levels were approximately one-third of the 30 s level, and 50% of this level by 180 min. Neither cocaine nor any other cocaine metabolite was detected in any of the arterial plasma samples.

Table 3 Arterial EME, cocaine, and BE levels (ng/ml) before and after administration of EME
Full size table

Figure 3 depicts results in the four additional animals that were used to verify the stability of the preparation. Blood flow to cerebral hemispheres (CBF), as well as blood flow to brainstem, caudate, and cerebellum, remained unchanged during the entire study period compared with baseline. There were no changes in MAP, CVR, CMRo2, or oxygen transport. These animals also demonstrated a similar decrease in hematocrit from beginning to end of study. There was no detection of cocaine, BE, or EME in any of the arterial plasma samples of any of these additional animals.

Figure 3
figure 3

CVR (A), CBF (B), and MAP(C) in four additional animals before and at 0.5, 2, 5, and 60 min after administration of 5 mL of normal saline.

Full size image

DISCUSSION

The most important finding of this study is that EME, a major metabolite of cocaine, causes cerebral vasodilation in newborn sheep. EME causes an acute increase in CBF as early as 30 s after i.v. administration. This vasodilation persists for at least 60 min and is not accompanied by any significant increase in cerebral oxygen consumption or change in blood pressure or oxygenation. This finding supports the hypothesis that EME has cerebral vasodilatory effects in vivo and leads us to conclude that it may be responsible, at least in part, for the cerebral vasodilatory response to cocaine in sheep.

Previous studies describing the cerebrovascular effects of acute cocaine injection have yielded conflicting results. Although some investigators have demonstrated that cocaine causes cerebral vasoconstriction(5, 6, 11, 12), others have observed cerebral vasodilation(7–9, 13, 14). Species, developmental, and methodologic differences, including use of anesthetic agents, may have accounted for these conflicting results. Our results support another possibility; that is, that EME has cerebral vascular effects in vivo and therefore species or developmental differences in cocaine metabolism may also contribute to these reported differences in responses to cocaine injection. Cocaine is metabolized to its two major metabolites (EME and BE) in plasma and liver; EME is formed by enzymatic hydrolysis, whereas BE is formed by spontaneous hydrolysis(22). EME has been detected in urine and/or plasma after cocaine administration in humans(16, 21), baboons(15), rabbits(22), rats(23), cats(6), and piglets(5). We previously reported that in sheep (fetus, newborn, and pregnant ewe) cocaine is metabolized primarily to EME with only a small amount of BE detected in plasma at 5, 15, or 60 min after i.v. cocaine injection(7–9). This differs from previous studies in other species in which considerably more BE has been detected in plasma and/or urine after various forms of cocaine have been administered. The most comprehensive data available regarding cocaine metabolism is in humans; both BE and EME are found in human urine after nasal, oral or i.v. cocaine(pure or crack form)(16). Metabolites are found in urine for days after cocaine use and long after cocaine is eliminated from plasma(24). Currently available data regarding cocaine metabolism in different species is subject to methodologic differences in both cocaine administration and cocaine analysis, and there have been no systematic studies comparing cocaine metabolism (and percentage of various metabolites) in different species. Therefore, although we have noted a predominance of EME(over BE) in sheep, we can state only that this cocaine metabolic profile may be different from humans. Further studies are needed to compare cocaine metabolism in sheep with that in other species in which cocaine studies have been performed.

In the present study, EME caused cerebrovasodilation within 30 s of injection. Consequently, we can presume that EME must be metabolized rapidly from cocaine and has either an immediate direct or indirect effect on the cerebral circulation. In humans, cocaine can reach the brain in less than 10 s(25), with peak effect in approximately 7 min(26). Cocaine's lipophilic properties allow it to cross the blood-brain barrier quite easily(24). However, unlike cocaine, EME is a polar substance(24) and may not cross the blood-brain barrier well, if it does so at all.

The cerebral vasodilation caused by EME in this study is less than we previously reported after cocaine injection. There are several possible explanations for cocaine's greater cerebral vasodilatory effect. First, cocaine has systemic effects that EME does not have. For example, a single 4 mg/kg dose of cocaine caused a 55% increase in MAP in newborn sheep(7), whereas a 2.5 mg/kg dose of EME did not alter MAP at all. Acute hypertension may alter the integrity of the vascular endothelium even if cerebral autoregulation remains intact(27). Further, cocaine's systemic effects could disrupt the blood-brain barrier so that vasoactive compounds such as norepinephrine could alter cerebral metabolism and vascular reactivity. Second, there are cholinesterases synthesized and present in brain that may cause further metabolism of cocaine to EME(28). Since cocaine can cross the blood-brain barrier, it may be metabolized there as well. A combination of EME formed from cocaine in serum and EME formed from cocaine in brain might lead to a greater percentage increase in CBF after cocaine injection. Although it is not yet known, EME concentrations in sheep brain may be greater than in plasma; or, EME in the sheep brain may act differently or at specific receptor sites.

Regional brain blood flow responses paralleled those of the cerebral hemispheres with a few notable exceptions. Brainstem blood flow did not increase until 60 min. Cerebellar and caudate flows increased at 30 s and continued to increase through 60 min with the greatest increases in blood flow at 60 min compared with baseline. This may have occurred because of regional differences in oxygen metabolism, perfusion pressure, or a difference in the population of receptors potentially activated by EME. Recent studies have shown that there are maturational differences in regional blood flow that are related to regional differences in local capillary density, local energy demands for tissue growth, and possibly levels of myelination(29). There is relatively advanced maturity of the brainstem at birth compared with cortex or caudate in some species with better autoregulation of brainstem blood flow compared with cerebrum(30). Regional brain blood flow may also be affected differently by similar levels of cocaine or its metabolites. In rats, for example, Porrino et al.(31) showed that different i.v. cocaine doses caused selective increases in cerebral metabolic activity with certain brain regions becoming active only after higher doses. Stein and Fuller(32) measured cocaine-induced increases in regional CBF in rats and found that cocaine's duration of action varied heterogenously across both time and brain region.

EME might cause cerebral vasodilation in neonatal sheep by several mechanisms. First, EME might act directly at an EME or cocaine receptor. Second, EME might cause the release of a neurotransmitter such as norepinephrine, dopamine, adenosine, or another intermediary agent(s). Third, EME might interact with another brain receptor. Several investigators have examined the cerebrovascular effects of cocaine metabolites in vitro. Madden and Powers(6) applied EME to isolated cat cerebral arteries and described mild relaxation. Schreiber et al.(12) applied EME to cannulated pressurized sheep cerebral arteries and noted vasodilation. Kurth et al.(5) topically applied EME to piglet cerebral arterioles and demonstrated that EME produced less vasoconstriction than BE. There have been no in vivo studies examining the cerebrovascular effects of cocaine metabolites. The present study is the first to demonstrate that EME has cerebral vasodilatory effects in vivo similar to effects previously described in vitro. However, the mechanism for this effect remains unknown.

EME did not have any significant systemic effects. Although there were no significant changes in arterial Po2, cerebral oxygen transport, or oxygen extraction, there was a small (although physiologically insignificant) decrease in Cao2. This decrease is explained by the small decrease in Hb that occurred (Table 1) due to withdrawal of blood for various analyses. Decreases in oxygen content cause cerebral vasodilation(33). However, Jones et al.(34) have shown that the small decrease in Cao2 noted in the present study would be expected to increase CBF by only 2-3%. There were no changes in heart rate, blood glucose, MAP, respiratory rate, or animal activity. In an in vitro study, Wang and Carpenter(35) noted a very mild effect of EME on the sinoatrial node of adult rats. Misra et al.(23) injected EME into the tail veins of male Wister rats and demonstrated no systemic effects at high doses. Therefore, although the cerebrovascular effects of EME and cocaine are similar, their systemic effects are different. One likely reason for this may be cocaine's known effects on catecholamine levels; cocaine blocks reuptake of norepinephrine and stimulates release of epinephrine from the adrenal medulla(36). We did not measure catecholamine levels in this study but indirect indicators of catecholamines were not present, e.g. no change in blood pressure, heart rate, or glucose concentration. Cerebral vasodilation without systemic effects occurs in response to stimuli such as hypercapnia and mild anemia; further, certain drugs can alter baseline CBF and cerebral vascular reactivity without systemic effects (e.g. indomethacin)(37). It is possible that EME's cerebral vascular effects occur directly or indirectly via similar mechanisms. A systematic evaluation of mechanisms for EME's vascular effects may therefore further our understanding of basic mechanisms of regulation of cerebral vascular tone.

There were some limitations to this study. First, we measured cerebrovascular effects of a single dose of EME and obtained measurements at the same time intervals after injection as in our previous study evaluating cerebral responses to cocaine(7). However, we may have missed a more delayed response to EME. In the present study, EME levels decreased but remained present in arterial blood samples for at least 180 min after injection (Table 3). Further, although cerebral vascular resistance returns to baseline 15 min after cocaine injection, CVR remained decreased 60 min after EME injection. It is possible that CVR might have decreased further or remained decreased compared with baseline if we had done more delayed measurements. Second, we studied only one dose of EME. We chose this 2.5 mg/kg EME dose because, after several pilot studies using different doses, we had obtained EME levels similar to those we had previously observed after a 4 mg/kg cocaine dose in the lamb. The present study did not examine a dose-response relationship. Further, we do not know if the results can be extrapolated to the fetus, either directly or after maternal injection. EME could affect uteroplacental blood flow differently than cocaine. In addition, although we have shown that both the sheep fetus(8) and the pregnant ewe(9) metabolize cocaine to EME, we do not know the percentage of EME which may cross the placenta. Further studies are needed to evaluate both cerebral vasoactivity and placental transfer of EME in fetuses.

In summary, we have shown that EME causes cerebral vasodilation in neonatal sheep. Vasodilation occurs immediately and persists for at least 60 min after injection. These cerebrovascular effects occur independently of systemic changes. EME may therefore be responsible, at least in part, for the cerebral vasodilation observed in developing sheep after cocaine injection.