In Vivo Monitoring of Sevoflurane-induced Adverse Effects in Neonatal Nonhuman Primates Using Small-animal Positron Emission Tomography

CLINICALLY used general anesthetics typically exert their action by the activation of inhibitory central nervous system (CNS) receptors, such as γ-aminobutyric acid (GABA) receptors, and/or the blockade of CNS excitatory receptors, such as N-methyl-d-aspartate (NMDA) receptors.^1–4  Many animal studies have shown that the developing brain is vulnerable to NMDA antagonists and/or GABA agonists: exposure to anesthetics during the period of rapid neuronal growth and synaptogenesis can induce massive neuroapoptosis.^3,5–7  These effects of general anesthetics are associated with subsequent, long-term behavioral deficits in both rats⁶  and nonhuman primates.⁸  Recently, several retrospective epidemiologic studies have indicated a relationship between general anesthesia in childhood and subsequent development of cognitive abnormalities and learning deficits in adolescence.^9–13  Therefore, it is necessary to study both the acute and long-term effects of early exposure to anesthetics on the developing CNS.

Sevoflurane is a commonly used inhalational anesthetic and has been frequently administered to pediatric patients during general anesthesia and sedation.^14–16  Sevoflurane is believed to activate glycine and GABA receptors type A and inhibit NMDA receptors.^17–20  In a neonatal rat model, a single episode of sevoflurane-induced general anesthesia (2 to 2.5% for 4 to 6 h) leads to widespread neuronal apoptosis in several brain regions and causes long-term behavioral impairments and memory dysfunction.^4,15,21–23  Repeated treatments with sevoflurane during gestation in the rat induced neuronal apoptosis in the brains of offspring.²⁴  In the mouse, exposure to sevoflurane on postnatal day (PND) 7 induced neuroapoptosis in the hippocampal region.^{3,20,23,25–28} l-Carnitine (LC) has been shown to attenuate brain injury associated with mitochondrial dysfunction^29,30  and has been reported to provide neuroprotective benefits in a variety of neurodegenerative and aging situations.^31–35  In the current study, acetyl-l-carnitine (ALC) was utilized to evaluate its protective ability against adverse reactions associated with exposure to sevoflurane.

Positron emission tomography (PET) is a leading molecular imaging approach that provides quantitative information about biologic, biochemical, pathologic, and pharmacologic processes in living tissue and organs.^36–38  We have developed a microPET protocol to investigate possible neurotoxic effects associated with developmental exposures to sevoflurane in vivo using imaging approaches.^7,39–43 

In the CNS, translocator proteins (TSPOs, 18 KDa), previously known as peripheral benzodiazepine receptors, are primarily located in glial cells, and their levels are believed to increase in areas of neuronal injury after exposure to neurotoxicants.^44,45  Therefore, they are widely recognized as important targets for PET imaging.^45–47  In this study, ¹⁸F-labeled fluoroethoxybenzyl-N-(4-phenoxypyridin-3-yl) acetamide ([¹⁸F]-FEPPA) was utilized as a radiolabeled ligand for TSPOs/peripheral benzodiazepine receptors in the CNS of neonatal monkeys.

Nonhuman primate models are thought to be more relevant than rodent models because of their similarities to human beings in regard to the length of pregnancy, brain neuroanatomical organization, and pharmacodynamics after the administration of anesthetics, and we have used these models in previous studies.^41,48–50  In the current study, noninvasive monitoring of physiologic parameters in neonatal monkeys was utilized to evaluate whether homeostasis in the exposed animals was within the normal range during anesthesia.

All animal procedures were approved by the National Center for Toxicological Research (NCTR) Institutional Animal Care and Use Committee (Jefferson, Arkansas) and conducted in accordance with the Public Health Service Policy on Humane Care and Use of Laboratory Animals.

All nonhuman primates were born and housed at the NCTR nonhuman primate research facility. Animal procedures were designed to minimize the number of animals required and any pain or distress associated with the experimental procedures. Neonatal rhesus monkeys (Macaca mulatta) were obtained from the NCTR breeding colony as previously described.⁵⁰  Briefly, breeder monkeys were housed in separate cages under a 12:12-h light/dark cycle, provided with water ad lib and fed high-protein jumbo monkey diet supplemented routinely with fresh fruit. All births occurred via natural delivery, and the day of birth was designated as PND 0. Neonatal monkeys stayed with their mothers except during anesthetic exposure or control sequestrations, until they were weaned at 6 months of age.

A total of 18 neonatal (PND 5 or 6, 460 to 500 g) rhesus monkeys were utilized. Ten male and 8 female monkeys were randomly assigned to experimental groups: control (n = 4, 3 male and 1 female), control + ALC (n = 4, 2 male and 2 female), sevoflurane alone (n = 4, 1 male and 3 female), and sevoflurane + ALC (n = 6, 4 male and 2 female). About 30 min before the initiation of anesthesia or sequestration (controls), the neonates were separated from their mothers and hand-carried to a procedure room. Control animals were maintained in an incubator with the temperature set at 32°C, with no water or food, and were not sedated for physiologic measurements or blood sample collections. Animals in the anesthetic exposure groups were exposed to 2.5% sevoflurane in oxygen for 8 h.

To examine the severity and the nature of sevoflurane anesthetic–induced neuronal damage, an additional six monkey infants (exposed on either PND 5 or 6) were randomly assigned to a sevoflurane-alone group (n = 3) and a control, room air exposure group (n = 3).

Neonatal rhesus monkeys were treated following procedures described previously.³⁹  Briefly, animals were exposed on PND 5 or 6 to sevoflurane within a clear anesthesia induction chamber (18 × 9 × 8 in; E-Z Anesthesia^®, USA). Sevoflurane (Webster Veterinary Supply, USA) along with oxygen (nexAir, USA) was delivered from a sevoflurane-specific vaporizer (Tec 7, Baxter, USA) at a concentration of 2.5% (v/v) into the chamber at a rate of about 0.5 to 1.0 l/min. The induction chamber was warmed from the bottom by a heat therapy pump (T/PUMP^®, Gaymar Industries, USA). For controls, room air was used in place of the sevoflurane. A relief valve on the anesthesia chamber allowed the escape of gases to avoid accumulation of carbon dioxide, and waste anesthetic gas was scavenged using an attached canister containing activated charcoal. For animals receiving ALC (Sigma-Aldrich, USA), it was administered intraperitoneally at doses of 100 mg/kg in a saline vehicle, 1 h before and 4 h after the start of exposure to room air or the anesthetics. Anesthetized monkeys were maintained in a laterally recumbent position.

During exposures, a stomach tube (5 ml) was used to administer 5% dextrose and 0.45% sodium chloride (Baxter) every 2 h to both treated and control monkeys to maintain blood glucose levels. All subjects were administered two doses of glycopyrrolate (American Regent, USA; 0.01 mg/kg, intramuscularly) at 0 and 6 h during exposure to reduce airway secretions. Animals were monitored in the incubator until complete recovery and then returned to their mothers (approximately 2 h after exposure). Control animals were kept in the infant incubator under room air for 8 h.

Physiologic parameters of all subjects were monitored following procedures previously described.^49,50  Briefly, noninvasive pulse oximetry (N-395 Pulse Oximeter, Nellcor, USA; MouseO_X Plus Vital Sign Monitor, Starr Life Sciences, USA), capnography (Tidal Wave Hand-held capnography, Respironics, USA), sphygmomanometry (Critikon Dinamap Vital Signs Monitor, GE Healthcare, USA), and a rectal temperature probe were used to verify the physiologic status of subjects during the exposures and sequestrations. Heart and respiration rates, arterial blood oxygen saturation levels, expired carbon dioxide concentrations, rectal temperatures, and systolic and diastolic blood pressures were recorded every 2 h in anesthetized and control animals. Venous blood (approximately 300 μl) was collected at 2-h intervals for measurement of plasma glucose concentrations, venous blood gases, pH values, and hematocrits (GEM^® Premier 4000, Instrumentation Laboratory, USA). During microPET scans, the peripheral oxygen saturation, heart rate (HR), respiration rate, and rectal body temperature were recorded every 15 min.

The depth of anesthesia was assessed by monitoring the alteration of cortical electrical activity using quantitative electroencephalogram (EEG) technology. A SEDLine system (Masimo Corp, USA) was utilized to monitor neonatal monkeys during exposure to sevoflurane, according to the operator’s manual provided by the manufacturer.⁵¹  Briefly, the four-channel EEG sensor comprising 6 pregelled electrodes was connected to the animal’s head to collect analog EEG data. The analog EEG data were converted to digital data by the patient module connected to the sensor. Then the digital data were transferred to a computer and monitor, where they were processed, analyzed, and displayed. A patient state index (PSI), which is indicative of the level of consciousness/anesthesia, was generated using an algorithm developed on the basis of extensive EEG recordings.⁵²  The PSIs range from a scale of 0 to 100, where 100 represents the fully awake state. The optimal level of anesthesia was represented by PSIs within the range of 25 to 50. Real-time PSI data, including the associated .csv log files, were later transferred to external storage medium for statistical analysis and plotting. Background noise from muscular electrical activity and nonbrain electrical artifacts can interfere with the waveform of the EEG. Thus, the PSI values were generated only when the contributions from background nonbrain electrical activities were sufficiently low, as indicated by electromyographic data. The electromyographic data, a record of muscle activity, such as jaw clenching, ranged from a value of 0 to 100% and was plotted as a bar graph for review.

[¹⁸F]-FEPPA was prepared by 3D Imaging LLC (USA) according to the procedure of Wilson et al.⁴⁷  from the corresponding tosylate as described therein with minor modifications. [¹⁸F]-FEPPA was produced in more than 98% purity at a specific activity at end of synthesis of 1.1 to 3.7 TBq (30 to 100 Ci)/μmol, as compared to the previously reported value of 11 to 37 GBq (0.3 to 1 Ci)/μmol.⁵³  Specific activity at the time of use was one tenth to two thirds of the end of synthesis value.

A commercial high-resolution small-animal PET scanner (Focus 220, Siemens Preclinical Solution, USA) was used to quantitatively record images of the monkey brain. The scanner has 96 lutetium oxyorthosilicate detectors and provides a transaxial resolution of approximately 1.35-mm full-width at half-maximum. Data were collected in a 128 × 128 × 95 matrix with a pixel width of 0.475 mm and a slice thickness of 0.815 mm.

Monkeys were also imaged in a newly developed mobile neurologic CereTom CT scanner (Neurologica Co., USA). The computed tomography (CT) gantry of the CereTom scanner moves while the subject remains externally supported and fixed in space, therefore allowing it to be physically connected to the microPET scanner. Exposure settings for each CT scan were 120 kVp, 5 mAs, and scan time = 120 s. Data were collected in a 512 × 512 matrix with a pixel width of 0.49 mm and a slice thickness and space of 2.5 mm.

The first microPET and CT scans of each monkey brain were taken on PND 6 or 7, the day after experimental exposures on PND 5 or 6. Follow-up microPET/CT scans occurred at approximately 1 week (on PND 14), 3 weeks (on PND 30), and 2 and 6 months of age. For collection of the microPET/CT images, animals were positioned on a modified external bed controlled by the microPET manager. Instead of the standard microPET bed, this modified bed allowed for enough travel to move animals through the microPET and CT fields of view (over an axial range of greater than 50 cm). Throughout the microPET/CT imaging sessions (120 to 130 min), monkeys were anesthetized with 1.2 to 1.5% isoflurane gas alone delivered through a custom facemask, and an electronic heating pad was used to maintain body temperature at approximately 37°C. A transmission scan (10 min) with a rotating ⁵⁷ Co point source was performed using software supplied by the scanner manufacturer. For each imaging session, [¹⁸F]-FEPPA (56 MBq) was injected into the lateral saphenous vein of each animal (anesthetized). Immediately after the injection, a set of serial microPET images was recorded to assess the influx and accumulation of the tracer over the next 2 h (24 frames, 5 min each). Immediately after microPET imaging, microCT brain images of the animal were obtained in about 2 min for the purpose of coregistering anatomical (CT) with molecular (PET) data. Emission data were reconstructed as 3D volumes including scatter and attenuation correction using 3D reconstruction software (ASIPro, Concorde Microsystems, Inc., USA) installed in the CereTom CT controller unit. Once scanning was complete, animals were monitored in an incubator within a shielded isolation area until complete recovery and then returned to their mothers.

Medical image analysis software (ASIPro, Concorde Microsystems, Inc.) was used for the anatomical/molecular data coregistration and analyses. Regions of interest (ROIs) were outlined and measured using tools provided by ASIPro. Radioactivity in different brain areas was also quantified using this software. All images were displayed using the same color scale. Tracer accumulations in the ROIs in the left thalamus, left frontal, and left temporal cortex were converted to standard uptake values (SUVs).

To examine the severity and the nature of sevoflurane anesthetic–induced neuronal damage, an additional six monkey infants (exposed on either PND 5 or 6) were randomly assigned to a sevoflurane-alone group (n = 3) and a control, room air exposure group (n = 3). The frontal lobe/cortex was chosen for pathologic analysis. Coronal sections (40 μm) of the frontal cortex were sectioned with a cryostat for Fluoro-Jade C (Histo-Chem, Inc., USA) staining as previously described.⁵⁴  Briefly, before staining, sections were mounted onto polylysinized slides. The mounted sections were bathed in a solution of 1% sodium hydroxide in 80% ethanol for 5 min and washed in 70% ethanol and distilled water for 2 min, respectively, followed by an incubation in 0.06% potassium permanganate solution for 10 min and another incubation in a 0.0001% solution of Fluoro-Jade C (Histo-Chem, Inc.) dissolved in 0.1% acetic acid vehicle for 10 min. A stock solution of 0.01% Fluoro-Jade C was prepared before being diluted to the 0.0001% working solution. The slides were rinsed through three changes of distilled water for 1 min per change. The air-dried slides were cleaned in xylene, and a cover-slip was applied with DPX nonfluorescent mounting media (Sigma, USA).

Radiotracer accumulation in the ROIs was converted to SUVs (SUVs = total radioactivity in ROI × body weight of monkey/injection dose of [¹⁸F]-FEPPA). All SUVs for the ROIs obtained from the microPET image were presented as mean ± SEM and compared between groups at a variety of time points using one-way ANOVA with Tukey post hoc analysis using SigmaPlot (Systat Software Inc., USA). A linear mixed model was used to evaluate the effect of treatment on physiologic parameters. A compound symmetry covariance structure in the context of repeated measures ANOVA was used to model covariance patterns of repeated measurements on the same animal assuming homogeneous variance/covariance. In the presence of any significant interactions, treatment comparisons were performed within each time point; otherwise, the mean comparisons were evaluated for each main effect. Planned treatment comparisons were performed between each sevoflurane-exposed group and the control group. The Dunnett test was performed in multiple pairwise comparisons between the fixed control group and the other treatment groups. The Bonferroni correction was used in the multiple pairwise tests on physiologic monitoring data sets acquired at multiple time points from the same set of animals. Student’s t test was used for the PSI comparisons between the sevoflurane-exposed groups with or without ALC and the comparison of Fluoro-Jade C staining between control and sevoflurane-exposed monkeys. All statistical tests in the analyses are two sided unless otherwise stated. Throughout this article, statistical significance was assessed at the 5% nominal level (α = 0.05). Multiplicity adjustments were performed to preserve the type I error rate at the specified nominal level.

Throughout the 8-h exposures or sequestrations, vital signs, including HR, blood pressure, respiration rate, end-tidal carbon dioxide (ETco₂), rectal temperature, and blood oxygen saturation, were monitored closely. Analysis of venous blood every 2 h also demonstrated that there were no significant alterations in pH, glucose, hematocrit, carbon dioxide or oxygen partial pressures, or hemoglobin oxygenation throughout the exposures (table 1).

Respiration rate was moderately lower in the sevoflurane-exposed animals (57.4 ± 5.0) in comparison with the control animals (72.7 ± 5.8), without statistical significance (P = 0.10). Throughout the exposure, the differences in respiration rates between these two groups were not statistically significant (P > 0.05), with an exception at 4 h (P = 0.011). Overall, the differences in ETco₂ were not statistically significant (P = 0.07), except at 8 h when ETco₂ was significantly higher in the anesthetized animals, 25.6 ± 2.3 versus 14.3 ± 2.6 in controls (P = 0.005). Importantly, arterial blood oxygen saturation, carbon dioxide and oxygen partial pressures, as well as venous oxyhemoglobin, deoxyhemoglobin, blood glucose, and pH, did not differ between the sevoflurane-exposed monkeys and controls. Taken together, these data demonstrate that neither compromised ventilation and respiratory acidosis nor hypoglycemia occurred in the sevoflurane-exposed animals. Volatile and other general anesthetics are well known to cause myocardial suppression. In the current study, the HR of sevoflurane-exposed neonatal monkeys was moderately reduced in comparison with that of control animals (P = 0.022); however, there were no significant effects on blood pressure (P > 0.05). The decrease in HR in the sevoflurane-exposed animals occurred as early as 2 h after the start of exposure (P < 0.001) and remained steady thereafter. Sevoflurane-induced changes of HR and respiration rate were consistent with previous studies related to ketamine.^49,50  The sevoflurane-exposed animals also experienced a slight yet significant (P < 0.001) drop in core body temperature. As five blood samples were taken during the 8 h of observation, the hematocrit in all groups declined slightly over time (P < 0.05). The coadministration of ALC alone did not have any significant effects on any of vital sign parameters described in the Monitoring of Physiologic Parameters in Materials and Methods (P > 0.05).

To evaluate the effects of coadministered ALC on the depth of anesthesia, quantitative EEG analysis was used to determine the level of brain activity during sevoflurane exposure. In all 15 sevoflurane-exposed subjects, the contributions of muscular electrical activity were found to be without consequence. The numerous PSIs recorded throughout the sevoflurane exposures were averaged within each subject for comparison of the effects between the sevoflurane-alone group and the sevoflurane plus ALC group. The mean PSI for the sevoflurane-alone group (32.7 ± 4.8) was no different from that for the sevoflurane plus ALC group (32.9 ± 4.8). The similar levels of the PSIs for the two groups are evidenced by representative samples from several subjects recorded 1, 4, and 7 h after the start of sevoflurane exposure (fig. 1). In addition to brain electrical activity, several physiologic parameters also served as traditional indicators of depth of anesthesia. Here, there were no significant differences between the sevoflurane-alone and sevoflurane plus ALC groups with respect to HR (P = 0.829), mean arterial pressure (P = 0.113), or ETco₂ (P = 0.981) throughout the 8-h exposure.

Coronal CT images were obtained immediately after the microPET scans for the purpose of anatomical/molecular data coregistration. Figure 2 represents the microPET ([¹⁸F]-FEPPA)/CT brain images from a sevoflurane-exposed monkey: the PET/CT images were fused using ASIPro software. By using the Fusion Tool in the ASIPro software, multiple data sets (anatomical and functional) can be loaded in the same general orientation and viewed using either linked or fused views. As shown in figure 2, the anatomical (CT) images are displayed in the top row and the functional (microPET) images in the bottom row. Once the images were loaded, the anatomical images were transposed onto the functional images and “fused” images were created, as seen in the middle row. Coupling microPET ([¹⁸F]-FEPPA) with CT, the imaging data indicate that sevoflurane-induced neural damage was located primarily in cortical brain levels including the frontal and temporal lobes, as well as in the thalamus.

Eight hours of sevoflurane exposure produced significant increases in the accumulation of [¹⁸F]-FEPPA signals in neocortical areas, especially the frontal cortex (fig. 3), as evidenced by data collected from coronal scans obtained 1 day after sevoflurane exposure. All images were displayed using the same color scale shown in figure 3.

In addition, sevoflurane-induced neurotoxicity was assessed using Fluoro-Jade C staining. Fluoro-Jade C–stained neurons are characterized by high resolution and great contrast, and therefore, the degenerating neurons can be precisely localized.⁵⁴  The 8-h exposure to a clinically relevant concentration of sevoflurane (2.5%) produced extensive neuronal damage as indicated by increased Fluoro-Jade C-positive neuronal cells in the frontal cortex (fig. 4).

After the first microPET/CT scan was taken on PND 6 or 7, follow-up scans were obtained for each animal 7 days (PND 14), 21 days (PND 30), 2 months, and 6 months after sevoflurane exposure. For each microPET scan, images were obtained over 2 h after the injection of [¹⁸F]-FEPPA, and time activity curves from the frontal cortex (the most vulnerable brain area)^30,55  were obtained, as were those from the temporal lobe and thalamus. Tracer accumulations in the ROIs were converted to SUVs.

In the current study, the accumulation of [¹⁸F]-FEPPA was clearly increased in frontal cortical ROI in sevoflurane-exposed monkeys 1 day after exposure (fig. 5A). Coadministration of ALC effectively blocked this increase in retention of the [¹⁸F]-FEPPA signal induced by sevoflurane. There was no significant effect of ALC on SUVs when given alone.

On PND 14, approximately 7 days after the anesthetic exposure, SUVs in the frontal area were still significantly higher in the exposed animals than in the controls at most of the time points, demonstrating a prolonged increased uptake and retention of [¹⁸F]-FEPPA in treated subjects (fig. 5B). Compared with controls, there was no significant difference in SUVs detected in animals treated with sevoflurane plus ALC.

On PND 30, about 3 weeks after anesthetic exposure, the uptake of [¹⁸F]-FEPPA in the anesthetic-exposed monkeys remained higher than that seen in the controls at most time points, but there were no statistically significant findings at any specific time point (fig. 5C) at that time. This is likely due to the small number (n) used to derive these data. At 2 and 6 months of age, the uptake of [¹⁸F]-FEPPA in the anesthetic-exposed monkeys returned to similar levels to that observed in controls at most time points (no significant differences were observed at any time; data not shown).

On PND 6/7 (1 day after exposure), the accumulation of [¹⁸F]-FEPPA in the ROIs in the temporal area and thalamus was also quantitated. At this time, the accumulation of [¹⁸F]-FEPPA was increased in the temporal cortical lobe (fig. 6) and thalamus (fig. 7) in sevoflurane-exposed brains over virtually all time points. However, these differences were only statistically significant in the temporal lobe for a few specific time points (fig. 6A). The elevated accumulation of [¹⁸F]-FEPPA induced by sevoflurane exposure was clearly attenuated by the coadministration of ALC, but since there were no or only a few significant differences to begin with, it was not possible to demonstrate a statistically significant effect of ALC here. No significant effects as compared to controls were detected in the animals exposed to ALC alone.

On PND 14, approximately 7 days after the anesthetic exposure, the uptake of [¹⁸F]-FEPPA in the temporal area and thalamus was still higher in the sevoflurane-exposed animals than in the controls at most of the time points, but again these effects were not statistically significant (figs. 6B and 7B).

By PND 30, or about 21 days after anesthetic exposure, SUV values in the temporal lobe were not different between any of the experimental groups, whereas in the thalamus, they were still highest in the sevoflurane-exposed group, but not statistically significant compared to the controls. The values for the sevoflurane plus ALC group were indistinguishable from the control group (figs. 6C and 7C).

Consistent with observations for the frontal lobe, at the age of 2 and 6 months, the uptake of [¹⁸F]-FEPPA in the temporal lobe and thalamus in anesthetic-exposed monkeys was similar to that of controls at most time points, and no significant differences were observed at any time (data not shown).

Little information is available about the effects of developmental exposure to sevoflurane.^56,57  However, preclinical studies have demonstrated that early exposure to general anesthetics, such as ketamine, isoflurane, nitrous oxide, and sevoflurane, induces widespread neuronal apoptosis in many brain regions in rodent models.^{5,6,17,21,25,49,50,58,59}  In a previous study,⁶⁰  we also provided evidence of sevoflurane-induced neurotoxicity and some observations on the temporal features of this effect in neonatal rats. The first report describing neuronal cell damage in nonhuman primates exposed perinatally to anesthetics was published in 2007.⁵⁰  Because the brain growth spurt in both human and nonhuman primates extends over a much longer time period than that observed in the rat, matching the timing of developmental events between humans and nonhuman primates is less problematic than matching the same between primates and rodents.^48,61,62 

In the current study, microPET/CT technology was employed to investigate whether a relatively long-term exposure (8 h) to clinically relevant doses of sevoflurane (2.5%) could induce toxic processes in the brain of the nonhuman primate and whether such effects could be attenuated by the antioxidant agent ALC. Our data demonstrated that an 8-h exposure of neonatal monkeys to sevoflurane at a clinically relevant concentration did not significantly disturb critical physiologic parameters compared with those of control, demonstrating that the sevoflurane-induced effect observed here was not attributed to compromised homeostatic mechanisms.

Recent studies have demonstrated that the up-regulation of TSPOs is related to numerous nervous system disorders, including neurotoxic brain damage.^45,63–66  In the CNS, the up-regulated expression of TSPO has been regarded as a biomarker of activated microglia.^66–68  Such microglial activation usually begins within several hours after toxic insult and lasts for several days.^69–71  [¹⁸F]-FEPPA is a specific TSPO ligand that can be efficiently synthesized at high radiochemical yields with high specific activity that has proven useful in nonhuman primate studies.^{40,41,43,47,72–74}  Due to the increased expression of TSPOs in areas of neuronal injury/neurotoxicity, [¹⁸F]-FEPPA can be used as a marker for microglial activation and an indirect marker for neuronal damage, injury, and neurotoxicity. TSPO is recognized as an important surrogate of associated neurotoxicity, which can be monitored using PET imaging using appropriate radioligands such as [¹⁸F]-FEPPA.^40,41,43 

In the current study, [¹⁸F]-FEPPA was used to repeatedly visualize and quantify, in the same animal, aspects of possible adverse neuronal effects associated with developmental exposure to sevoflurane. Increased accumulation of [¹⁸F]-FEPPA was interpreted as increased levels of TSPO expression, representing neuroinflammation induced by neuronal insults, such as sevoflurane exposure.

Previous data from rodent models have demonstrated that sevoflurane can be toxic to the developing brain, and some studies have reported cognitive deficits in later life after such exposures.^{23,27,28,75,76}  However, definitive studies that clearly delineate the association between the pediatric use of general anesthetics and sedatives and developmental neurotoxicity are highly desired. In the current study, [¹⁸F]-FEPPA was effectively taken up and concentrated into brain ROIs in both control and anesthetic-treated monkeys after intravenous injection. Signal levels of [¹⁸F]-FEPPA, and presumably TSPO, were significantly elevated in multiple brain regions within 1 day after sevoflurane exposure, an effect that lasted for at least 1 week. This observation suggests increased microglial activation and astrocytosis in response to sevoflurane-induced neuronal effects. This effect seemed to last for about 3 weeks after exposure, at least in the frontal lobe and thalamus, as evidenced by increased retention of [¹⁸F]-FEPPA, although by this time the effects were no longer statistically significant. These observations in the developing monkey are concordant with those observed in the developing rodent model.⁶⁰  On the other hand, sevoflurane-induced neural effects in the nonhuman primate were observed in multiple brain areas including the frontal cortex and temporal lobe and the thalamus, whereas in the rat, it was seen primarily in the frontal cortex. The current data indicate yet again that adverse effects associated with general anesthesia (in this case sevoflurane) are expressed in a time- and region-specific fashion. Although causality is presently unclear and probably multifactorial, the severity, as well as the pattern of gaseous anesthetic-induced neural damage, depends on the amount (dose) given, the duration of exposure, the route of administration, the receptor subtype activated, and the stage of development at the time of exposure. These factors are important because they can help identify thresholds of exposure for producing neurotoxicity in the developing nervous system. Toward this end, our high-resolution microPET/CT scanners, coupled with the appropriate radiotracer, have been able to provide in vivo molecular imaging at a sufficient resolution to resolve both major structures and neuronal activities in the nonhuman primate brain.

Consistent with the microPET/CT data, sevoflurane-induced neurotoxicity was confirmed and further characterized in the neonatal monkey brain by pathologic measurement. In this current study, Fluoro-Jade C staining was applied to confirm the presence of sevoflurane-induced neuronal damage at the cellular level. Fluoro-Jade C is a neuronal stain, which has a high affinity for degenerating neurons, regardless of specific insult or mechanism of cell death.⁵⁴  Pathologically, direct evidence of enhanced neuronal damage in the infant monkeys exposed to sevoflurane was obtained as indicated by increased numbers of Fluoro-Jade C–positive/neurodegenerative neurons in the frontal cortical area.

The pharmacologic blockade of excitatory NMDA receptors during development is thought to result in the subsequent up-regulation of NMDA receptors, resulting in a subsequent overexcitation from physiologic levels of glutamate.^50,77,78  The activation of GABA receptors during early development is actually excitatory, whereas in adults, this is inhibitory. Thus, inhibition of NMDA receptors or the stimulation of GABA receptors can increase overall excitability of the CNS and lead to excessive neuronal damage in the developing rodent and nonhuman primate. Studies have shown that sevoflurane activates glycine and GABA receptors while antagonizing NMDA receptors.^17–20,79  Although the underlying mechanisms are largely unknown, increased oxidative stress associated with mitochondrial dysfunction seems to play an important role in anesthetic-induced neurotoxicity.^28,80,81 

From our lipidomic analyses,⁸²  we noted that alterations in mitochondrial membrane morphology and function were particularly significant, suggesting that mitochondria may be the most vulnerable initial target of anesthesia-induced developmental neurotoxicity. To further explore potential mechanisms underlying anesthetic-induced neuronal damage and to evaluate possible protective strategies, ALC, an antioxidant and mitochondrial membrane–stabilizing agent involved in fatty acid transport and metabolism,^30,83–86  was coadministered with sevoflurane. ALC is an esterified form of LC that exists in relatively high levels in the brain.^86–89  In contrast to LC, ALC is more easily transported across the blood–brain barrier and, therefore, likely has a more rapid and perhaps a greater effect on brain metabolism.^90–93  ALC acts to preserve cellular membrane integrity and plays a critical role in mitochondrial oxidation of long-chain fatty acids. It is thought to provide neuroprotection in neurodegenerative diseases and situations of metabolic stress such as ischemia, hypoxia, aging, and alcoholism and in traumatic injury.^{31–34,86,94–99}  Thus, increased levels of ALC could enhance the capability of the mitochondrial antioxidant system and decrease the incidence of free radical-induced damage.

In our previous pediatric anesthetic study, mitochondrial viability was significantly attenuated in a dose-dependent manner after prolonged exposure to propofol, suggesting that the elevated mitochondrial production of reactive oxygen species is involved in producing anesthetic-induced cell damage.¹⁰⁰  In our previous studies,^39,40  neuronal damage was suggested in PND 14 rat pups that were exposed to different general anesthetics, including sevoflurane, on PND 7. The anesthetic-induced neuronal apoptosis observed in those studies was also attenuated by the coadministration of ALC. In the current study, coadministration of ALC did not affect the level of consciousness during sevoflurane-induced general anesthesia. Coadministration of ALC effectively attenuated the increased uptake of the TSPO tracer in multiple brain regions of developing nonhuman primates exposed to sevoflurane. These data strongly suggest that gaseous anesthetics can cause oxidative damage and dysfunction, likely via disruption of mitochondrial function that subsequently manifests as damage to neural cells.

In summary, it has been demonstrated that a single 8-h exposure to sevoflurane did not cause significant alterations in critical homeostatic mechanisms. MicroPET data indicate that exposing the developing monkey brain to sevoflurane during a period of rapid brain development can induce adverse effects in several brain regions including the frontal cortex, temporal lobe, and the thalamus, as evidenced by increases in microglial activation. Coadministration of ALC did not have any significant effect on depth of anesthesia, although it blocked, at least partially, the anesthetic-induced neuronal effects. The levels of TSPO in the CNS are thought to be important biomarkers of pathogenic processes, and the ability to track such levels using microPET technology provides a minimally invasive approach for investigating the location, time course, and severity of neurotoxic insults. The use of a nonhuman primate model combined with molecular imaging tools would seem to provide the most expeditious approach toward decreasing the uncertainty in extrapolating preclinical data to the human condition.

This work was supported by the National Center for Toxicological Research, Jefferson, Arkansas, and the Center for Drug Evaluation and Research/U.S. Food and Drug Administration (FDA), Silver Spring, Maryland. This document has been reviewed in accordance with U.S. FDA policy and approved for publication. Approval does not signify that the contents necessarily reflect the position or opinions of the FDA nor does mention of trade names or commercial products constitute endorsement or recommendation for use. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the FDA.

The authors declare no competing interests.