Ketamine-induced Neuroapoptosis in the Fetal and Neonatal Rhesus Macaque Brain

• Previous work has shown that sustained exposure to ketamine (e.g ., 9–24 h) produces neurodegeneration in the developing rhesus macaque brain.

• Shorter exposure intervals (e.g ., 5 h) also produce neurodegeneration. Patterns of neuroapoptosis are different in fetal versus  neonatal rhesus macaque brain, with the number of damaged cells being greater in the fetal brain.

IN recent years, several classes of drugs, including those that block N -methyl-D-aspartate (NMDA) glutamate receptors, those that activate γ-aminobutyric acid receptor type A (GABA^A), and ethanol (which has both NMDA antagonist and GABA-mimetic properties), have been shown to trigger widespread apoptotic death of neurons in the developing animal brain.1,–,5Vulnerability to the apoptogenic mechanism reaches its peak during the developmental period of rapid synaptogenesis,1,2also known as the brain growth spurt period, which in rodents occurs primarily during the first 2 weeks after birth but in humans extends from approximately midgestation to several years after birth.6 

Drug-induced developmental neuroapoptosis has been shown at both light and electron microscopic levels to have classic morphologic characteristics of apoptosis.1,3,5,7,8The cell death process involves translocation of Bax protein to mitochondrial membranes and extra-mitochondrial leakage of cytochrome c, followed by activation of caspase-3.9,–,11Because commitment to cell death occurs before the caspase-3 activation step,11activated caspase-3 (AC3) immunohistochemistry reliably identifies cells that have progressed beyond the point of cell death commitment. Therefore, AC-3 immunohistochemistry has been used extensively for marking dying neurons in recent studies focusing on drug-induced developmental neuroapoptosis.3,8,–,27 

Neuroapoptosis induced in the developing brain by NMDA antagonist and GABA-mimetic drugs potentially is relevant in a public health context because there are many agents that the developing human brain may be exposed to that have NMDA antagonist or GABA-mimetic properties, including drugs that are sometimes abused by pregnant mothers (ethanol, phencyclidine, ketamine, nitrous oxide, barbiturates, and benzodiazepine28,–,30‡‡) and many drugs used worldwide in obstetric and pediatric medicine as anticonvulsants, sedatives or anesthetics. Of particular concern is mounting evidence from animal studies that even single or brief exposure to clinically relevant doses of commonly used anesthetics (ketamine, midazolam, propofol, isoflurane, sevoflurane, and chloral hydrate) may trigger a significant neuroapoptosis response in the developing brain3,13,–,18,22,23,26and that exposure of the developing rodent brain to these agents can result in long-term neurobehavioral impairment.5,26,31,–,34In addition, there is evidence potentially linking anesthesia exposure in infancy with long-term neurobehavioral deficits in both human35,–,39and nonhuman40primates.

The important question of whether anesthetic drugs can trigger neuroapoptosis in the developing nonhuman primate (NHP) brain was first addressed by Slikker and colleagues, who reported20,21that IV infusion of ketamine triggered neuroapoptosis in the rhesus macaque brain at postnatal day 5 if the infusion was continued for 9 or 24 h but not if it was stopped at 3 h. In addition, it was shown that a 24-h ketamine infusion triggered neuroapoptosis in the gestational day 120 (G120) fetal rhesus macaque brain,20but briefer periods of fetal exposure were not studied. More recently, we have shown27that a neuroapoptosis response is induced in the 6-day-old rhesus macaque brain by 5-h exposure to the potent volatile anesthetic isoflurane. To further clarify the clinical relevance of ketamine's apoptogenic potential, we administered ketamine by IV infusion for 5 h to pregnant (G120) or postnatal day 6 rhesus macaques and examined the fetal or neonatal brains 3 h later for histologic evidence of apoptotic neurons.

All animal procedures were approved by the Oregon National Primate Research Center (Beaverton, Oregon) and Washington University Medical School Institutional Animal Care and Use Committees (St. Louis, Missouri) and were conducted in full accordance with the Public Health Service Policy on Humane Care and Use of Laboratory Animals. Two studies were performed, one pertaining to fetal and the other to neonatal subjects.

To produce time-mated pregnant dams, cycling females were paired for 4 days with fertile males beginning on day 10 ± 2 of the menstrual cycle (first day of menses = day 1; gestation length was counted from the first day of pairing, first day = day 0). Pregnancy was confirmed by blood progesterone concentrations of greater than 1 ng/ml at 29 ± 2 and 42 ± 3 days after the menses before mating. Blood samples were collected from conscious animals using minimal restraint. The pregnant dams were not exposed to any sedative agents throughout the pregnancy. Pregnant female macaques at G120 were exposed for 5 h to no anesthesia (n = 4) or to an IV infusion of ketamine (n = 3). Full-term gestation for these macaques is 165 days.

At the initiation of the experiment, the pregnant females were gently restrained according to standard methods (in the same apparatus used for blood sampling, as described, so the animals are well accustomed to the procedure) to facilitate placement of an IV catheter for administration of ketamine. Anesthesia was induced by a bolus of ketamine (10 mg/kg), which was followed by a continuous ketamine infusion at 10–85 mg · kg⁻¹· h⁻¹for 5 h, augmented by additional IV boluses as needed to maintain an intermediate surgical plane of anesthesia, defined as no movement and not more than 10% increase in heart rate or blood pressure in response to a profound mosquito-clamp pinch at hand and foot (checked every 30 min). Animals assigned to ketamine anesthesia were intubated tracheally (conventional direct laryngoscopy, 4.0 ID endotracheal tube; Mallinckrodt, Hazelwood, MO), mechanically ventilated using a conventional anesthesia machine ventilator (Hallowell EMC, Pittsfield, MA) to control oxygenation and gas exchange, and maintained using extended physiologic monitoring (continuous end-tidal carbon dioxide, FIO², electrocardiogram, peripheral oxygen saturation, noninvasive blood pressure [every 15 min], body temperature [esophageal temperature probe, Smiths Medical ASD, SurgiVet Monitor; Smith Medical, Wankesha, WI], blood gases, metabolic profile including pH, base excess, blood urea nitrogen, hematocrit, hemoglobin, sodium, chloride, potassium, and serum glucose and lactate concentrations [every 2 h; i-STAT, Princeton, NJ]). IV fluids (lactated Ringer's solution) and glucose were substituted as needed, and physiologic temperature was maintained using a warming device and forced warm air (Bair Hugger; Arizant Healthcare Inc., Eden Prairie, MN). Ultrasonography was performed hourly to monitor fetal heart rate and health of the fetus. After 5 h, the ketamine infusion was stopped and the animals were observed for 3 h. The trachea was extubated when appropriate, and the animals were then returned to climatized cages and offered biscuits and water as tolerated for the remainder of the observation period. At 8 h after time 0, ketamine anesthesia was induced again (10 mg/kg IV, followed by a continuous infusion at 50–85 mg · kg⁻¹· h⁻¹), and a cesarean section was performed to deliver the fetus, which was euthanized for tissue collection (high-dose phenobarbital followed by transcardial perfusion according to National Institutes of Health guidelines). After recovery from anesthesia, the dam was given appropriate analgesic and was returned to her home cage. Dams that were randomized to receive no anesthesia (control group) were measured at baseline, received an IV injection of saline by the same method used in the ketamine-treated animals, and were returned to their cages, where they had access to light food and water ad libitum . Eight hours later, a cesarean section was performed to remove the fetus for transcardial perfusion under deep anesthesia as described, and the dam was allowed to recover according to the paradigm described.

Neonates were delivered naturally from dams that had been group housed or were specifically paired with males as part of the time-mated breeding program. Pregnancy was determined by ultrasound under light sedation of the dam using ketamine (5–10 mg/kg) during the first or early second trimester. On postnatal day 6, neonatal rhesus macaques were exposed for 5 h to no anesthesia (n = 5) or to IV ketamine (n = 4). Neonates were gently hand restrained; after IV access was established and blood drawn for baseline measurements, the animals assigned to receive general anesthesia were given ketamine (20 mg/kg bolus IV) followed by a continuous IV infusion (20–50 mg · kg⁻¹· h⁻¹for 5 h) and additional IV boluses for maintenance of an intermediate surgical plane of anesthesia, defined as no movement and not more than 10% increase in heart rate or blood pressure in response to a profound mosquito-clamp pinch at hand and foot (checked every 30 min). The animals were tracheally intubated (semirigid fiber-optic endoscope; Karl Storz America, El Segundo, CA; 2.0 ID endotracheal tube; Mallinckrodt), mechanically ventilated (small animal ventilator; Harvard Apparatus, Holliston, MA), and maintained using extended physiologic monitoring (continuous end-tidal carbon dioxide, FIO²[Capnomac; Datex Ohmeda, Madison, WI], electrocardiogram, peripheral oxygen saturation, noninvasive blood pressure [every 15 min; SurgiVet Advisor V9200; Smith Medical], rectal temperature [handheld rectal thermometer, Vicks, V965F; KAZ Inc., Proctor and Gamble, Cincinnati, OH]), blood gases, metabolic profile, including pH, base excess, blood urea nitrogen, hematocrit, hemoglobin, sodium, chloride, potassium, and serum glucose and lactate concentrations [every 2 h; i-STAT]). After 5 h, the ketamine infusion was stopped and the animals were observed for 3 h. The trachea was extubated when appropriate, and the animals were then kept in an infant monkey incubator (Snyder ICU cage; Snyder Mfg., Centennial, CO) and were provided formula as tolerated. At the end of the 3-h observation period (8 h after time 0), and after final measurements, the animals received ketamine (20 mg/kg IV), followed by high-dose phenobarbital to secure a deep level of anesthesia, and were then transcardially perfusion-fixed to prepare the brain for histopathologic analysis. Animals that were randomized to the control group (no anesthesia) were measured at baseline, received an IV injection of saline to mimic the stress to which the experimental animals were exposed before being anesthetized, and were then returned to their mothers until final measurements and euthanasia 8 h later.

The current study pertaining to ketamine is part of a larger research program aimed at comparing the potential toxic impact of several anesthetic drugs on the developing NHP brain. To make such a comparison, each agent must be administered in a manner that provides the same level of anesthesia for the same duration of time. Thus, the 5-h ketamine perfusion protocol used in this study was chosen to match a specific depth and duration of anesthesia that is used for each anesthetic drug being tested. The authors recognize that ketamine is not commonly used as a stand-alone drug in either pediatric or obstetric anesthesia to maintain patients in a fully anesthetized state for an extended period of hours, and this is a caveat that must be taken into consideration in interpreting the human significance of the findings.

For histopathologic analysis, the perfusion-fixed neonatal and fetal brains were cut on a vibratome into 70-μM serial coronal sections across the entire rostrocaudal extent of the forebrain and midbrain and into serial sagittal sections across the brainstem and cerebellum. Sections were selected at 2-mm intervals and stained for detection and quantification of apoptotic profiles by AC3 immunohistochemistry, as described previously.3,8,–,27Before the AC3 stain was applied, sections were processed for antigen retrieval to reduce background nonspecific staining and maximize the AC3-specific signal. For antigen retrieval, sections were immersed in a citrate buffer, pH 6.0, and subjected to heat in a pressure cooker for 10 min.

It can be considered a limitation of the current study that only a single posttreatment survival interval was used and only a single method (AC3 staining) was used to detect apoptotic profiles. We chose the AC3 staining procedure for several reasons: (1) It has been shown in numerous previous studies3,8,–,26to mark faithfully cellular profiles that are undergoing apoptotic degeneration after exposure to various apoptogenic drugs, including ketamine13,15,16,23; (2) It has been demonstrated that cellular profiles marked by the AC3 stain are also marked by the cupric silver stain, which selectively stains cells that are dead or dying1,–,5,12; (3) It has been confirmed by electron microscopy that these cells display all of the classic ultrastructural characteristics of cells undergoing cell death by apoptosis1,–,3,7,8,12; and (4) It has been found9,27that the AC3 stain is valuable, not only for identifying cells that are undergoing apoptosis, but also for revealing whether the cell is in an early or advanced stage of degeneration and for revealing what type of cell is undergoing degeneration. The AC3 stain can provide this information because, in the early stages of apoptotic degeneration, brain cells generate copious amounts of AC3 protein, which fills the cell body and its processes, so that the full microanatomy of the cell can be visualized by immunohistochemical staining with AC3 antibody in this early stage. In the ensuing several hours, the processes shrivel and become fragmented, the cell body becomes condensed or rounded up into a spherical shape, and these changes signify that the cell is progressing from early to advanced stages of cell death.27 

AC3-positive cellular profiles were counted by an investigator who was blinded to the treatment condition. Each stained section was comprehensively scanned by light microscopy using a ×10 objective lens and a computer-assisted stereoinvestigator system (Microbrightfield Inc., Williston, VT) with an electronically driven automatic stage to plot the number and location of each AC3-stained neuronal profile in each section. The total area scanned and the total number of stained neurons per section and per brain were computer-recorded. The total area was converted to volume by multiplying by the thickness of the section (70 μM), and the total number of neurons was divided by the total volume of tissue within which counts were made to yield a density count (number of stained profiles per mm³) for each brain. Density counts were converted to total numbers of apoptotic neurons in a given brain by calculating the mean volume of tissue per counted section and multiplying by the number of serial sections cut from the entire brain. The density of apoptotic neurons (number per mm³) times the total brain volume yielded an accurate estimate of the total number of apoptotic neurons per brain.

There are important ethical and logistical considerations regarding NHP experimentation. It is important to include as few animals as are needed to answer research questions with as much precision as possible. Previous experiments in NHP and rodents have shown that, when neonatal animals are exposed to drugs that consistently and potently induce neuroapoptosis, the apoptosis is extensive and can be demonstrated compellingly with a small number of animals, typically between three and five per group.13,14,16,20,–,22,23,–,25,27For statistical analysis, an unpaired Student t  test with Welch correction, where appropriate, was used, a two-sided P  value < 0.05 was judged significant, and the 95% CI for the mean difference provided a measure of precision. Statistical analysis was performed with Analyze-It® Statistical Software (Leeds, United Kingdom) for Microsoft Excel® (Redmond, WA).

The pregnant dams were assigned on postconception day 120 ± 2 to receive general anesthesia (ketamine group; body weight range 6.15–7.45 kg; n = 3) or not (control group; 5.40–7.45 kg; n = 4). Induction (10 mg/kg, IV) and maintenance of ketamine anesthesia was generally well tolerated, and in all animals, the trachea was intubated for airway control. The animals retained adequate spontaneous ventilation and were connected to the breathing circuit of a conventional anesthesia machine to control oxygenation and gas exchange. Blood gases, acid-base status, laboratory variables, glucose and lactate concentration, as well as hemodynamic variable and body temperature were controlled within species-specific physiologic limits throughout the experimental period. Fetal heart rates remained within physiologic limits and varied in concert with the maternal heart rate. The animals were extubated without problems within approximately 90 min after the ketamine infusion was stopped, and the dams remained somnolent and were kept in a climatized cage under direct observation until the scheduled cesarean section. The cesarean sections were conducted using ketamine general anesthesia and standard intraoperative monitoring according to established protocols at the Oregon National Primate Research Center surgical center. The animals tolerated the procedure well and retained stable vital signs, blood gases, and acid-base levels within physiologic limits (table 1). The time from induction of general anesthesia to clamping of the umbilical cord of the fetuses was less than 10 min in all cases, and fetal heart rate monitoring before surgical incision and umbilical arterial and venous blood samples confirmed fetal well-being until the cord was clamped (table 2), which was followed immediately by transcardial perfusion fixation as described in Materials and Methods. Fetal body weight (ketamine group 215–237 g; control group 129–252 g) and other standard body dimensions were age appropriate for the macaque G120 time point.

During the first 90 min, ketamine doses of 1.10–1.35 mg · kg⁻¹· min⁻¹were necessary to achieve the desired anesthetic depth (no movement and less than 10% increase in heart rate or blood pressure in response to profound mosquito-clamp pinch). The total doses of ketamine per animal were 260.4–432.6 mg/kg during the 5-h anesthesia period (48.1–86.5 mg · kg⁻¹· h⁻¹). One dam showed clinical signs of seizure activity (eye blinking, bilateral upper extremity jerks) at 85 min of ketamine anesthesia. This animal received 2.5 mg IV diazepam, which immediately stopped the clinical seizure-like activity. The subsequent clinical course of this animal was uneventful.

Animals assigned to the control group had unremarkable baseline measurements and underwent cesarean section 5 h later according to the same protocol and monitoring. At time of delivery, the measured physiologic variables (fetal heart rate, metabolic status, and blood gases of dam and fetus) were within physiologic limits and comparable with those in the ketamine group. One dam in the control group showed an anion gap-positive metabolic acidosis with increased lactate levels. Additional laboratory evaluation of this animal revealed mild aspartate aminotransferase elevation and low globulin levels. Differential diagnoses include subclinical chronic liver or renal disease, chronic infection, or a tumor disease. At the time of the experiment, the animal was clinically unremarkable as was the delivered fetus. The brain of this fetus had a low apoptotic profile count similar to the other brains in the control group.

Quantitative evaluation of AC3-stained brain sections from the fetal brains (G120) revealed that the mean (± SD) number of AC3-positive neuronal profiles per brain in the three ketamine-exposed fetal brains was 11.17 × 10⁵± 2.99 × 10⁵, and in the four control fetal brains was 2.26 × 10⁵± 1.37 × 10⁵(fig. 1), which amounts to a 4.9-fold increase in the number of apoptotic neurons in the ketamine-exposed fetal brains compared with the drug-naive controls. The difference between the means for ketamine-exposed versus  control brains was 8.92 × 10⁵(95% CI, 13.18 × 10⁵to 4.65 × 10⁵, P = 0.003).

The pattern of augmented neuroapoptosis in the ketamine-exposed fetal brains was widespread. Brain regions preferentially affected were the cerebellum, brainstem, inferior and superior colliculi, several thalamic nuclei, caudate nucleus, putamen, globus pallidus, nucleus accumbens, several neuronal groups in the hypothalamus and basal forebrain, and all divisions of the neocortex and several limbic cortical regions, with the notable exception of the hippocampus, which showed very little effect. The most severely affected regions were the cerebellum, caudate nucleus, putamen, and nucleus accumbens. The pattern of neuroapoptosis induced by ketamine in the fetal brain is illustrated in figures 2and 3. The histologic appearance of neurons undergoing apoptosis in the NHP fetal brain is illustrated in figures 4and 5.

On the day of the experiment, the infant macaques (ketamine group [n = 4], mean age 5.7 days, 449–680 g body weight; control group [n = 5], mean age 5.6 days, body weight 430–546 g) were assigned to receive general anesthesia or not. IV induction (20 mg/kg) and maintenance of ketamine anesthesia were well tolerated by all infant animals. At adequate anesthetic depth, the trachea was intubated in all animals and the lungs were mechanically ventilated, which allowed control of blood gases and acid-base status within physiologic limits throughout the 5 h of general anesthesia. Active warming and fluid and glucose application maintained basic homeostasis, documented by continuous intraoperative monitoring of hemodynamic variables and point-of-care laboratory results throughout the entire experimental period (table 3). Within 45 min after cessation of the ketamine infusion, the infant macaques were weaned from the ventilator, extubated without complications, and maintained in an animal incubator at physiologic condition. Animals remained drowsy and were observed directly for 3 h until transcardial perfusion.

In the first 90 min, 0.52–1.28 mg · kg⁻¹· min⁻¹ketamine was applied to the neonatal macaques to achieve the desired anesthetic depth (no movement and less than 10% increase in heart rate or blood pressure in response to profound mosquito-clamp pinch). The total doses of ketamine per infant macaque were 92–280 mg/kg during the 5-h anesthesia (18.40–56.0 mg · kg⁻¹· h⁻¹). There were no adverse effects of ketamine anesthesia observed in the neonatal cohort.

Quantitative evaluation of AC3-stained sections from the postnatal day 6 neonatal brains revealed that the mean (± SD) number of apoptotic neuronal profiles per brain in the four ketamine-exposed neonatal brains was 5.44 × 10⁵± 1.89 × 10⁵and in the five control neonatal brains was 1.42 × 10⁵± 0.95 × 10⁵(fig. 1), which equals a 3.83-fold increase in the number of apoptotic neurons in the ketamine-exposed brains compared with the brains from drug-naive controls. The difference in the mean number of apoptotic neurons between the ketamine and control groups was 4.02 × 10⁵(95% CI, 6.29 × 10⁵to 1.74 × 10⁵, P = 0.004).

Ketamine caused a less widespread pattern and less dense concentration of neuroapoptosis in the neonatal brains than in the fetal brains. A major difference between the fetal and neonatal response to ketamine was that many of the caudal and subcortical brain regions that were affected preferentially in the fetal brains, including the cerebellum and brainstem, showed little or no neuroapoptotic response in the neonatal brains. A major exception to this rule was that the strongest neuroapoptotic response to ketamine in the neonatal brain was in the basal ganglia (caudate nucleus, putamen, nucleus accumbens, globus pallidus) and several thalamic areas (fig. 6). These same regions, although not necessarily the same neuronal populations within each region, also had a high neuroapoptotic count after ketamine exposure of fetal brains. The neuroapoptotic response in the cerebral cortex of ketamine-exposed neonates was distributed evenly across cortical regions and was increased only moderately compared with controls in most cortical regions. This distribution is different from that recently described in neonatal macaque brains exposed to isoflurane.27The apoptotic response after isoflurane exposure was most prominent in several divisions of the neocortex, especially the temporal cortex (layers II and IV) and the primary visual cortex (layers II and V). The striking difference in pattern and density of neuroapoptosis in the primary visual cortex after isoflurane exposure versus  ketamine exposure is illustrated in figure 7.

The rate of natural neuroapoptosis and of ketamine-induced neuroapoptosis in the fetal versus  neonatal period is illustrated in figure 8. The rate of neuroapoptosis declines between the fetal and neonatal period for both phenomena, and the decline is more rapid for ketamine-induced than for natural neuroapoptosis. However, because the magnitude of the drug-induced phenomenon is so much higher, the two curves perhaps will approach zero at roughly the same time in the late neonatal or early juvenile period. The dashed lines with arrows represent the amount of neuroapoptosis that can be attributed to ketamine treatment at each age (8.92 × 10⁵for fetuses and 4.02 × 10⁵for neonates). The value for fetuses is 2.22 times greater than the value for neonates.

Ketamine anesthesia was titrated clinically to achieve a moderate anesthesia plane (see Materials and Methods), and the total doses of ketamine per kilogram body weight varied between individuals and between the fetal and neonatal cohorts. Although individual biologic variability and the small number of animals preclude reaching firm conclusions regarding linearity, a linear dose-response relationship is suggested by the fact that at each age, the lowest dose was associated with the lowest apoptosis count and higher doses were associated with correspondingly higher counts, with the exception of one neonate that received a high dose and had a lower than would have been predicted neuroapoptotic score (table 4).

It has been estimated that the minimum ketamine exposure duration for inducing neuroapoptosis in the developing NHP brain lies somewhere between 3 and 9 h20. The current findings clarify that a 5-h ketamine infusion is sufficient to induce a significant neuroapoptotic response in either the fetal or neonatal NHP brain. A comparison of the brain growth spurt curve for the rhesus macaque with that for the human6suggests that the rhesus brain at G120 may be at a developmental stage approximately equivalent to the brain of a mid- to late–third-trimester human fetus (or prematurely born infant), and the postnatal day 6 rhesus may be equivalent to that of a 4–6-month-old human infant. Therefore, the period in human development to which the current data are potentially most relevant would be an 8-month interval from the middle of the third trimester of gestation to approximately 6 months after birth.

The total number of neurons that showed AC3 positivity as a result of ketamine exposure was more than two times greater when exposure occurred in the fetal period. This suggests that human risk for ketamine-induced neuroapoptosis may be greatest at some point in the third trimester of prenatal development, which corresponds in general to the range of brain ages for human infants born prematurely. Such infants often are exposed to anesthetic drugs for procedural sedation and/or surgery. Because a G120 rhesus brain equates closely to a late, but not early, third trimester human brain, this points to the need for NHP studies examining ketamine neurotoxicity at additional gestational ages corresponding to a human brain age earlier in the third trimester. Similarly, whereas all of the recent human epidemiologic studies pertaining to developmental anesthesia neurotoxicity35,–,39have focused on full-term infants and children, the focus of future human research should be expanded to include third trimester fetuses and prematurely born infants.

Our data indicating that ketamine, an NMDA antagonist, may be more damaging to the fetal than neonatal NHP brain, has implications in both a clinical anesthesia and a drug abuse context. Maternal abuse of drugs is the most common cause of preventable developmental brain damage in modern society. Alcohol, the most widely and frequently abused drug in human experience, which has damaged more fetal brains (fetal alcohol spectrum disorder) than any drug in human history, is also an NMDA antagonist, as is phencyclidine (1-(1-phenylcyclohexyl)-piperidine [Angel Dust]), a dissociative anesthetic and notorious psychotomimetic drug of abuse. ‡‡In recent years, ketamine has become a popular drug of abuse in the United States and other regions of the world.28,–,30Our findings identify the third trimester of pregnancy as a developmental period when neurons may be at peak sensitivity to the apoptogenic action of NMDA antagonist drugs. This information is relevant to the use of ketamine to anesthetize pregnant women or premature infants41,–,43and also to the maternal abuse of ketamine, phencyclidine, or alcohol during this period of heightened fetal sensitivity.

A comparison of the fetal and neonatal patterns of damage induced in the macaque brain by ketamine suggests that in the fetal period, caudal and some subcortical brain regions transition through their window of vulnerability, so that a number of these regions are vulnerable in the fetal but not neonatal period. A similar finding was reported recently 44for fetal macaques exposed to alcohol at G120 versus  at G155 (1 week before term). However, it is noteworthy that several subcortical brain regions, especially the basal ganglia and thalamus, were severely affected in both the fetal and neonatal period. This observation has potentially important implications, in that a deficit in neuronal mass of the basal ganglia has been emphasized in several reports as a prominent finding in children who were exposed in utero  to alcohol,45,46and the same finding was reported recently in children exposed in utero  to antiepileptic drugs.47Alcohol and antiepileptic drugs have apoptogenic properties similar to those of ketamine.2,–,4,13,44 

The pattern of degeneration induced by ketamine in the fetal NHP brain differed substantially from the pattern induced by ketamine in the neonatal brain. This signifies that if anesthesia-induced neuroapoptosis in the primate brain gives rise to lasting neurocognitive disabilities, the nature of the disabilities will vary as a function of the time of exposure. This will complicate the task of establishing correlations between anesthesia exposure of the developing human brain with subsequent neurobehavioral disturbances, in that the disturbances will display wide-ranging heterogeneity correlated not with anesthesia per se , but with age at the time of exposure. In addition, the duration of exposure and type of anesthetic drug and many other interacting variables will complicate the analysis.

Our findings document important differences in the patterns of neuroapoptosis induced in the neonatal NHP brain by ketamine, as observed in the current study, versus  isoflurane, as observed in a previous study.27The most prominent difference is that the isoflurane pattern was heavily concentrated in the temporal and occipital lobes, where it featured a laminar pattern of degeneration affecting many neurons in layers II and IV of the temporal cortex and layers II and V of the primary visual cortex. The ketamine-exposed neonatal brains displayed a pattern more evenly and lightly distributed over various divisions of the cerebral cortex, and none of the ketamine-exposed brains showed a laminar pattern of dense degeneration selectively affecting large numbers of neurons in specific layers of the temporal and primary visual cortices.

Our findings corroborate those of Slikker et al.  regarding NHP susceptibility to ketamine's apoptogenic action, but some of our observations are different from theirs. They reported that in both neonatal and fetal rhesus macaques, ketamine triggered neuroapoptosis in only one brain region, the frontal cortex.20Our findings document a more widespread pattern of neuroapoptosis in both the fetal and neonatal macaque brain after 5-h exposure to ketamine. We believe these divergent results can best be explained by differences in experimental protocol. For example, our methods for sampling and counting apoptotic profiles are different from theirs, and their experiments, in which ketamine was found to induce neuroapoptosis, are based on a prolonged time frame and ours on a much shorter time frame. They maintained the ketamine infusion for either 24 or 9 h and allowed an additional 6-h survival time before obtaining the brains for histologic analysis. Thus, the total time for their experiments was either 30 or 15 h. We terminated drug exposure at 5 h and allowed an additional 3 h survival, for a total duration of 8 h. We geared our experiments to a shorter time frame because we have observed in numerous previous studies in rodents3,5,9,11,12,–,14,16,19and in NHPs exposed to alcohol44that evidence for neuroapoptosis (caspase 3 activation) is detectable in many brain regions within approximately 3–5 h after initiation of drug exposure. To reliably detect this rapid apoptosis response, we have found AC3 immunohistochemistry to be the best method. However, once the apoptosis cascade becomes evident by AC3 staining, it transpires rapidly to end-stage cell death. Within a time frame of 5–6 h, the apoptotic profiles begin to lose immunoreactivity, undergo degradation into small fragments, and are phagocytized.3,11Thus, if brains are examined at 15 or 30 h after initiation of anesthesia, it is possible to detect late-dying cells, but the most sensitive cells (those that died early) may not be detected.

Paule et al.  40have reported relatively robust and apparently permanent neurocognitive deficits in rhesus macaques after neonatal exposure to ketamine for 24 h. The human relevance of this finding is difficult to assess because of the long exposure time. Our demonstration that, within a more clinically relevant exposure time, both ketamine and isoflurane induce a significant neuroapoptosis response in the developing monkey brain clearly indicates the need for well-designed research to clarify whether these more clinically relevant anesthesia exposure conditions are associated with long-term neurobehavioral disturbances.

In conclusion, our data demonstrate that a 5-h exposure to ketamine, an anesthetic drug with NMDA antagonist properties, is sufficient to induce a significant neuroapoptosis reaction in the developing NHP brain at either a fetal or neonatal age. The apoptotic response at the fetal age (early third trimester) was 2.2 times greater than at the neonatal age, which signifies that the third trimester of pregnancy may be a period when the primate brain is particularly sensitive to the neuroapoptogenic action of NMDA antagonist drugs. This is potentially important in a public health context because at this developmental age, human fetuses and premature infants are exposed to NMDA antagonist drugs (ketamine, nitrous oxide), either alone or together with other neuroactive agents, for sedative, analgesic, or anesthetic purposes, and fetuses also may be exposed to NMDA antagonists (ketamine, phencyclidine, alcohol) by mothers who abuse these drugs for recreational purposes.