General anesthetics, such as isoflurane, are used in millions of children around the world.1 However, preclinical studies demonstrating increased brain cell death after anesthetic exposure in developing animals have raised serious concerns about their safe use in the very young.2–8
Anesthesia-induced brain cell death is widespread and at least some of the dying cells are neurons. However, whether and to what extent neurons versus non-neuronal cells are affected by isoflurane exposure is unclear. Although previous studies have convincingly demonstrated the presence of increased brain cell death in neonatal animals using a variety of cell death markers, such as caspase 3, Fluoro Jade B, TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling), and cupric silver stains, these markers were not specific to a particular brain cell type and have not been routinely combined with phenotypic markers for specific cell types.2,3,5,9 Accordingly, previous studies have relied on cell morphology alone to identify cells vulnerable to anesthesia-induced degeneration such as neurons. Most of these studies have localized cell death to cortical and subcortical gray matter, where the predominant cell types are neurons and glia. Determining the relative extent to which these 2 cell types are affected by anesthesia exposure will be particularly important for predicting the functional consequences of anesthesia-induced cell loss because the different cell types possess vastly different regenerative capacities. Whereas glial cell proliferation can occur throughout adulthood, neuronal proliferation becomes increasingly restricted as the brain matures.10
In the case of neuronal cell loss, it is also unclear which neuronal subtypes are most vulnerable. Answering this question is of particular importance for understanding anesthesia-induced cell death because of the phenomenon’s particular pattern of distribution. In contrast to other types of brain injury, such as ischemia in which a majority of cells in an affected region are destroyed, anesthetics induce widespread, scattered cell loss. Dying cortical brain cells are found immediately adjacent to apparently unaffected cells, suggesting that intrinsic differences among brain cells confer distinct vulnerabilities; however, the nature of these differences is not known.
Accordingly, the present study quantitatively and qualitatively characterized the cellular phenotype of susceptible brain cells in the superficial layers of neocortex, a brain region substantially affected by isoflurane-induced apoptotic cell death in neonatal mice. This was accomplished by using specific immunohistochemical markers for neurons, inhibitory interneurons, and astrocytes, and by comparing these findings with naturally occurring apoptosis in fasted, unanesthetized littermates, in order to provide insights into the selectivity, potential mechanisms, and consequences of anesthesia-induced cell death.
All procedures were approved by the Institutional Animal Care and Use Committee and conformed to the guidelines for ethical treatment of animals. Efforts were made to minimize the number of animals used. Breeding pairs of male CD1 and female C57BL/6 mice were housed in a 12/12-hour light-dark cycle at 22°C with free access to food and water. This hybrid was selected because they exhibit robust anesthesia-induced apoptosis with acceptable survival.3
For caspase 3 immunohistochemistry, 7-day-old CD1 and C57BL/6 hybrid littermates (n = 14) were randomly assigned to a 6-hour exposure to 1.5% isoflurane (approximately 0.6 minimum alveolar concentration in these mice) in 30% oxygen (anesthesia, n = 8) or to 6 hours in room air (control, n = 6). Immediately after treatment, animals were euthanized with an overdose of ketamine, acepromazine, and xylazine. Brains were immersion-fixed in 4% paraformaldehyde in phosphate-buffered saline (pH 7.4), postfixed overnight at 4°C, and cryopreserved in 25% sucrose. Brains were snap frozen and 40-μm coronal sections were cut on a cryostat (Thermo Electronics, Kalamazoo, MI). Sections were mounted to charged slides and stored at −80°C until use.
For protein analyses, a separate set of animals (n = 22) was treated and euthanized as described above. The left hemispheres were cut into 4 coronal sections, frozen in liquid nitrogen, and stored at −80°C until use. At a later date, sections of neocortex around bregma −3 mm were separated with a razor blade on dry ice and then homogenized twice in cell lysis buffer solution for approximately 10 seconds each time at 4°C. The homogenate was then centrifuged at 13,000 rpm using a refrigerated microcentrifuge (Fresco centrifuge, Sorvall, Buckinghamshire, UK). The supernatant was removed and used for testing for the specific proteins.
Slide-mounted brain sections were blocked for 1 hour in normal goat serum, followed by incubation in rabbit antiactivated caspase 3 polyclonal antibodies (1:100, 9661L; Cell Signaling, Danvers, MA) for 18 hours at −4°C, combined with one of the following antibodies: (1) mouse anti–Neuronal Nuclei (NeuN) monoclonal antibodies (NeuN, 1:500, Chemicon, MAB377; Millipore, Billerica, MA), (2) mouse antiglutamate decarboxylase isoform 67 (antiglutamic acid decarboxylase [GAD]67, 1:2000, MAB5406; Chemicon), (3) mouse anti-S100β (1:500, CB1040; Millipore), or (4) chicken antiglial fibrillary acidic protein (GFAP) (1:500, AB5541; Chemicon). Sections were then rinsed in blocker and incubated in Alexa Fluor 488 goat antirabbit secondary antibodies (1:200, A11034, Molecular Probes Inc.; Invitrogen, Carlsbad, CA) for 4 hours at 20°C, combined with either Alexa Fluor 594 goat antimouse (1:250, A11032, Molecular Probes) or Alexa Fluor 594 goat antichicken (1:250, A11042, Molecular Probes) secondary antibodies, as appropriate for the primary antibody species. After immunostaining, sections were dehydrated in an ascending ethanol series, cleared in xylenes, and mounted with Krystalon (EMD, Gibbstown, NJ).
Identification of Cellular Phenotype
To determine the phenotype of degenerating cells, brain sections from anesthesia-treated and control animals, corresponding to Bregma −2.46 to −2.70 (figures 51–53 in the mouse brain atlas by Paxinos and Franklin11) and double-immunostained for caspase 3 and NeuN or triple stained for caspase 3, S100β, and GFAP, were examined by an observer unaware of group assignment. NeuN and S100β stains cannot be combined in the same section, because both secondary antibodies are raised in the same species.
Caspase 3 immunostaining was excited using the 488-nm laser line, and emission wavelengths between 510 and 540 nm were collected to identify caspase-positive cells in layers II/III from retrosplenial cortex to piriform cortex using an SP5 confocal microscope set up on a DMI6000 stand (Leica Microsystems, Wetzlar, Germany) equipped with a 63× objective (NA 1.4). This region was selected because it has repeatedly demonstrated increased numbers of apoptotic cells in immature rodents.2,3 Immunostaining for NeuN, S100β, or GFAP was excited using the 543-nm laser line, and emission wavelengths between 600 and 650 nm were collected. Confocal optical sections were collected through the midpoint of the caspase 3–positive cell (pinhole = 1 Airy unit). Data are expressed as the percentage of caspase 3–immunoreactive cells that were also NeuN- or GFAP-positive, respectively.
Quantification of Apoptotic Cells Using the Optical Dissector Method
Further quantification of the effects of isoflurane exposure on cortical neurons and on GABAergic interneurons was performed as previously described.3,8 Briefly, confocal image stacks of caspase 3/GAD67 double labeling were collected at 1-µm increments through the entire Z-depth of the tissue (40 μm) using 1× optical zoom. Six image stacks were collected from layers II/III of visual cortex, corresponding to figures 51 to 53 in the mouse brain atlas by Paxinos and Franklin,11 from each animal, as follows: for each hemisphere, 3 adjacent confocal image stack frames were collected beginning 750 μm from the midline and moving laterally (Leica SP5, 63× 1.4 NA objective, 1-μm steps). Image stacks, which were 120 × 120 µm in dimension for NeuN and 240 × 240 µm for GAD67, because of the significantly lower cellular density for the latter stain compared with NeuN, were transferred to Neurolucida software (v7.50.4; MBF Bioscience, Williston, VT) for analysis. Using the optical dissector method, an observer unaware of group assignment quantified the respective numbers of NeuN-positive or GAD67-positive cells, the corresponding number of caspase 3–positive cells, and the number of caspase 3/GAD67 or caspase 3/NeuN double-positive cells in each field.12,13 Cells were considered positive if their fluorescence intensity was 2 times or greater than the background intensity. Counts from all 6 respective image stacks were averaged for each animal.
Quantification of GAD67 and GAD65 Expression Using Competitive Enzyme-Linked Immunosorbent Assay
We used a competitive enzyme-linked immunosorbent assay to quantify the expression of the two γ-aminobutyric acid A (GABAA) synthesizing enzymes, GAD67 and GAD65. Rat antiglutamate decarboxylase isoform 67 (Anti GAD67, 1:5000, 671-C; Alpha Diagnostics Inc., San Antonio, TX) and goat antiglutamate decarboxylase isoform 65 (Anti GAD65, 1:32,000, Ab67725; Abcam, Cambridge, MA) antibodies were incubated overnight with the homogenized cortical tissue samples. These bound antibody/antigen complexes were then added to a GAD67 or GAD65 antigen-coated well blocked with 5% bovine serum albumin. Rabbit antirat and rabbit antigoat secondary antibodies were added to GAD67 and GAD65 complexes, respectively. The secondary antibodies were covalently bound to horseradish peroxidase, an enzyme that cleaves the peroxide in the chromophore 3,3′,5,5′-tetramethylbenzidine. This enzyme activation turned on the chromophore and emitted a blue signal, which when treated with 2 M sulfuric acid turned to a yellow color, which was measured at 450 nm using a spectrophotometer (Jenway Genova Life Science Spectrophotometer; Bibby Scientific Limited, Staffordshire, UK). Absorbancy was then compared with a standard curve allowing for the determination of the isoforms’ concentrations.
All sample sizes for group assignment were made a priori. For each animal, the total NeuN-positive cells were counted over the 6 fields. The number of caspase 3/NeuN double-positive cells was defined as an event. The data were normalized to events (caspase 3/NeuN double-positive cells) per 400 NeuN-positive cells counted, the lower end of cells encountered in each animal, to avoid extrapolation. Gross inspection of the raw data revealed that caspase 3 activation in NeuN-positive cells was a rare event with a mean incidence of 2.4% and a maximal incidence of 3.6% in the anesthesia-treated animals. This event rate met the criteria for analysis using the Poisson distribution.
The Poisson mean event rate, λ, and its 95% CI were determined using the MATLAB® function [lambdahat, lambdaci] = poissfit(data, alpha). The vector “data” represented the number of events per 400 counted NeuN cells for each animal in the group of interest and α = (1 − CI). The mean event rates, λ, derived from the MATLAB function, were used to construct probability distribution function curves for the 2 groups (see Appendix).
The raw event counts were used to compute the ratio of events in the anesthesia-treated group to the control group using equations 6 and 7 in Graham et al.14 This method was used as an independent means to assess the mean event ratio and to determine the 95% CI for the event ratio.
All other data are presented as means ± SEM. Group comparisons were made using the Mann-Whitney U test. Statistical calculations were analyzed using Stata/IC 10.1 for Mac OS X (Stata Corp., College Station, TX). Statistical significance was accepted at P < 0.05.
Neonatal Isoflurane Exposure Increases Apoptosis Throughout the Developing Mouse Brain
In 7-day-old mice, isoflurane exposure led to widespread qualitative increases in apoptotic cell death relative to controls, as measured by caspase 3 immunolabeling. Although cellular degeneration was observed in many brain regions, such as thalamus, striatum, and hippocampus, superficial cortical cell layers II and III comprised the highest density of apoptotic cells (Fig. 1), consistent with previous findings.3,8 Accordingly, quantitative analyses focused on this region in an effort to characterize the cellular phenotype and to quantify apoptotic cell death after isoflurane exposure in an area of “maximal” injury.
Neurons Are Preferentially Lost After Isoflurane Exposure
To reveal the cellular specificity of isoflurane-induced apoptosis, caspase 3 labeling was combined with either NeuN immunohistochemistry, which labels postmitotic neurons, or S100β and GFAP immunohistochemistry, which primarily label glial cells, the 2 predominant cell classes in the neocortex. Confocal analyses in cortical layers II and III revealed that 98% ± 0.6% of all caspase 3–labeled cells colocalized with NeuN in isoflurane-treated mice (Figs. 2 and 3). In contrast, 0.3% ± 0.26% and 6.6% ± 1% of degenerating cells were GFAP- and S100β-positive, respectively (Figs. 3 and 4), suggesting that isoflurane overwhelmingly affects cortical, postmitotic neurons immediately after exposure.
Although far fewer cells were caspase 3 immunoreactive in control animals, similar relative percentages of the dying cells expressed NeuN (98% ± 2% [95% CI 92.4, 103.5] or 98% ± 0.6% [95% CI 96.9, 99.8] for control or anesthesia, respectively) or GFAP (0% [95% CI 0, 0] or 0.3% ± 0.26% [95% CI −0.3, 0.8], for control or anesthesia, respectively). These percentages followed the rare event rate, as outlined below and were indistinguishable from anesthesia-treated animals in terms of neuronal versus astrocytic cells (P = 0.18 or P ≈ 1 for NeuN or GFAP colocalization, respectively, comparing anesthesia with control). This suggests that, although cell loss was substantially higher after anesthesia exposure, the cell type being lost, predominantly neurons, was similar to naturally occurring cell death, as observed in control animals.
Isoflurane Exposure Increases Apoptosis at Least 11-Fold in Superficial Cortical Neurons
Because the majority of dying cells in superficial cortical layers were identified as neurons, based on colocalization with NeuN, we further quantified isoflurane-induced neuronal apoptosis among cells in the visual cortex, using the optical dissector method. In unanesthetized, fasted animals, 0.08% ± 0.001% of NeuN-positive layer II/III cortical neurons were immunoreactive for caspase 3, undergoing natural apoptosis. By contrast, the rate of apoptotic NeuN-positive neurons increased to 2.0% ± 0.004% of all postmitotic neurons immediately after isoflurane exposure (P = 0.0017 isoflurane versus control; Fig. 3). The average number of cells counted on a group basis was 590 ± 47 for the control group and 582 ± 39 for the anesthesia group; these values were not significantly different. Two caspase 3/NeuN double-positive cells were observed in the control group for an event rate of 0.23 per 400 NeuN-positive cells. A total of 95 double-positive cells were observed in the anesthesia-treated animals, yielding an event rate of 8.77 per 400 NeuN-positive cells. The mean (95% CI) for the ratio of events was 38.8 (10.5, 143). Using the Poisson statistics, the 95% CIs for λ ranged from 0.015 to 1.05 (mean = 0.23) caspase 3/NeuN double-positive cells per 400 NeuN-positive cells for the control animals, and from 6.64 to 10.83 (mean = 8.55) in the anesthesia-treated animals. The predicted ratio of events, comparing anesthesia-treated animals with controls, on the basis of mean λ was 37.2. Conversely, the probability of observing no events per 400 NeuN-positive cells in the control group was P = 0.79, whereas the probability of observing 1 event was P = 0.18. For the anesthesia-treated group, 8 events has the maximal probability of being observed (P = 0.137) (see Appendix). The central 50% of the probability mass for the event ratio lies between 24 and 62; thus, 50% of the time the observed ratio of caspase 3/NeuN double-positive cells per X (an arbitrary large number) NeuN-positive cells in anesthesia-treated to control animals will be between 24 and 62. The limits were narrower than, but close to, the limits that would be computed using the Wald large number approximation.15
Isoflurane Exposure Increases Apoptosis in GABAergic Cortical Neurons
Although many caspase 3/NeuN double-positive neurons in anesthesia-treated animals appeared to be principal cells based on morphological criteria (pyramidal structure with radially oriented apical and basal dendrites; Figs. 1 and 2), caspase 3–immunoreactive cells with bipolar or stellate morphologies were occasionally observed, suggesting that interneurons are also affected during isoflurane exposure. To determine whether this was indeed the case, caspase 3 immunolabeling was combined with immunohistochemistry for the GABAergic interneuron marker GAD67. In isoflurane-treated animals, 2.9% ± 0.02% of all caspase 3–positive neurons in superficial cortex also coexpressed GAD67 (Fig. 5). Expressed as the percentage of GABAergic interneurons that were affected by the anesthetic exposure, almost 28% of all GAD67-positive cells were also immunoreactive for caspase 3, implying that a significant percentage of the GAD67 population underwent apoptosis after isoflurane exposure. Further analysis, however, led us to interpret these numbers cautiously. Specifically, an estimation of the number of GAD67-immunoreactive somatic profiles revealed a significant reduction in the density of neurons with labeled soma in anesthesia-treated animals relative to controls. Accordingly, the reduced denominator in anesthesia-treated animals would bias the observed percentile toward overestimating the effect on the total neuronal population, but underestimating the effect on GABAergic neurons.
Isoflurane Exposure Reduces GAD67 and GAD65 Expression in GABAergic Interneurons in Superficial Cortex
To confirm the observation of reduced GAD67 labeling, confocal image stacks were collected from visual cortex, and the optical dissector method was used to quantify the number of GAD67-positive cells, caspase 3–positive cells, and double-positive cells. After isoflurane exposure, the density of GAD67-immunoreactive cell bodies was reduced in anesthesia-treated animals by a factor of 3, relative to unanesthetized, fasted littermates (1.2 ± 0.4 · 103 vs 3.9 ± 0.7 · 103 cells/mm3, P = 0.011). However, in accordance with our other results, the density of all caspase 3–positive cells (both GAD67-positive and -negative) was increased 17-fold in these sections, comparing anesthesia animals with control (11.2 ± 2.1 · 103 vs 0.66 ± 0.03 · 103 cells/mm3, P = 0.01) (Fig. 6).
Enzyme-linked immunosorbent assay results mirrored the immunohistochemical findings. Isoflurane exposure caused a decrease in GAD67 expression compared with control (29.17 ± 11.10 vs 119.21 ± 11.23 ng/mg protein, P < 0.00001). The other isoform of the GABA-synthesizing enzyme, GAD65, was also decreased after anesthesia exposure compared with the control group (2.25 ± 0.74 vs 23.03 ± 8.47 ng/mg protein, P = 0.0008), whereas β-actin, a ubiquitous isoform of actin, was found to be 24.60 ± 4.60 ng/mg protein (CI 14.0, 35.2) in control animals versus 27.99 ± 3.63 ng/mg protein (CI 19.1, 36.9) in anesthetized animals (P = 0.8432 anesthesia versus control).
Prolonged exposure to the inhaled anesthetic isoflurane has been shown to trigger widespread brain cell death in several in vivo and in vitro neonatal animal models2–9,16–20 and to lead to subsequent long-term neurocognitive impairment,2,4,21 raising serious concerns regarding the safe use of isoflurane and similarly acting drugs in neonates.22–24 Several animal studies have localized cell death to cortical and subcortical gray matter, where the predominant cell types are neurons and glia, and have observed a predilection for the superficial cortical layers, peaking in 7-day-old rodents.2–4 However, the cellular phenotype of susceptible cells has not been immunohistochemically identified and the absolute extent of cell death has not been quantified; instead, previous studies have solely relied on morphological criteria for identification and have described apoptotic cell death as a percentage increase of physiological apoptosis.
In this regard, the present study in 7-day-old mice introduces 4 key findings. First, the great majority of cortical brain cells, 98%, eliminated immediately after a neonatal exposure to isoflurane were postmitotic neurons, as identified by NeuN expression. Second, isoflurane led to the demise of 2% of all cortical NeuN-positive neurons in layer II/III, a region consistently exhibiting substantial apoptosis, which represented an at least 11-fold increase over physiological apoptosis observed in unanesthetized, fasted littermates. Third, despite the disparate rates of apoptosis observed in anesthetized and unanesthetized animals, postmitotic neurons, and not astrocytes, were the predominant affected cell type in both groups. This observation suggests that anesthetic neurotoxicity may target the same cell population vulnerable to normal, developmentally regulated cell death. Finally, isoflurane led to neuroapoptosis in a segment of GABAergic interneurons and was associated with a decrease in the expression of the main GABA-synthesizing enzymes, suggesting that isoflurane may, at least transiently, interfere with proper inhibitory function in the developing brain.
Isoflurane Substantially Increases Apoptosis Among Postmitotic Neurons, but Not Astrocytes
NeuN is a neuron-specific protein that is found in the nuclei of neuronal cell types of the central nervous system, signifying postmigratory status and is absent from glial cells.25 In the murine neocortex, NeuN is first expressed in subplate neurons, the first cortical neurons to develop, starting at embryonic day 17.5 and does not reach adult levels until 16 days postnatally.25,26 Demonstrating that 98% of cells expressing the apoptotic marker caspase 3 also coexpressed the neuronal marker NeuN, the present study unequivocally identified the cortical cells primarily affected by cellular death immediately after an isoflurane exposure as postmitotic neurons. This finding is somewhat surprising, given the fact that susceptibility to anesthesia-induced, cortical cytotoxicity has historically been found to be limited to very young animals, peaking at postnatal day 7, and dramatically subsiding after 10 days of age in small rodents.27 However, our results extend observations in neonatal rats, demonstrating that isoflurane does not seem to induce cellular death in neuronal progenitor cells.28 Nevertheless, given that many susceptible cells were found in the superficial cortical layer II and that mammalian cortex forms in an inside-out pattern, whereby newly generated neurons migrate past the earlier generated cells to form more superficial layers,29 the present findings suggest that susceptibility for anesthesia-induced cell death is increased in relatively young, postmitotic neurons.
Another startling observation of the present study in postnatal day 7 mice was the dramatically lower rate of cell death, less than 0.5% and 7% of apoptotic cells expressed GFAP and S100β, respectively, observed among astrocytes, which are part of the glial family of brain cells and represent the other large portion of cortical cells. Astrocytes serve to protect, nurture, and support neurons by producing trophic factors, regulate neurotransmitters and ion concentrations, remove toxins and debris, mediate synaptogenesis, and contribute to synapse elimination as well as structural neuronal plasticity.30,31 This finding does not exclude that these cells are not involved in neuronal degeneration via proapoptotic signals, such as observed during brain ischemia.32 It is also conceivable that astrocytes may undergo more pronounced apoptotic cell death at a later time point compared with neurons, similar to experimental models of brain ischemia.33 Although GFAP is a widely accepted marker for astrocytes, it has also been found to be expressed in some neuronal progenitors.34 S100β is most frequently found in astrocytes, but has also been found in some neurons and oligodendrocytes.33,35 Moreover, some of the S100β-positive cells in our study may have exhibited neuronal morphology (Fig. 4); therefore, we cannot exclude that some of these cells may not have been astroglia, which may help explain the apparent overlap in the cell counts.
Previous studies in neonatal rats have demonstrated an up to 68-fold increase in apoptotic brain cell death after an isoflurane-based anesthetic exposure.2 However, the absolute percentage of cells affected by this phenomenon remained unknown. Apoptotic cell death is an integral part of normal brain maturation, eliminating 50% to 70% of neurons and progenitor cells during the extent of central nervous system development, which spans over several weeks in small rodents.36,37 Accordingly, at any given time point, only a small fraction of cells undergo physiological apoptotic cell death. In 5-day-old mice, apoptotic cell death has been found to occur in 0.07% of cortical neurons.38 Similarly, the present study observed caspase 3 labeling, a marker for apoptotic cell death, in <0.08% of fasted 7-day-old control mice. Conversely, immediately after a 6-hour exposure to isoflurane, the percentage of neurons undergoing apoptotic cell death increased at least 11-fold to approximately 2% of all neurons in the superficial layers of visual cortex. This region was selected for quantification because anesthesia-induced cell loss has repeatedly exhibited a substantial predilection for cells in superficial neocortex. Other less-susceptible brain regions not examined here would likely demonstrate a lower percentage of apoptotic cells.
The long-term effects of the elimination of up to 2% of neurons in neocortex and other brain regions on subsequent brain structure and function remain speculative. The temporary increase in neuroapoptosis observed here could possibly be offset by increased subsequent neurogenesis, similar to observations after postnatal hypoxia,39 or by subsequent decreases in naturally occurring apoptosis. The number of neocortical neurons peaks in 16-day-old mice and decreases by 30% thereafter,26 suggesting that the developing mouse brain may have sufficient reserve capacity to absorb a 2% neuronal loss. Consistent with this interpretation, a previous study by our group did not detect a significant decrease in adult cortical neuronal density after a similar neonatal isoflurane exposure,3 although a 2% reduction would be difficult to detect even with the most robust cell-counting techniques. Neurocognitive abnormalities were also absent in these animals, again suggesting that any deficits, if present, are subtle.3 However, the small quantity of the eliminated brain cells does not exclude the possibility of long-term network disruptions, because loss of even a small number of neurons with important function or during critical periods for brain development may have a significant impact on the subsequent development of neural networks.
Isoflurane Exposure Leads to Apoptotic Cell Death in GABAergic Interneurons and to Decreased Expression of GAD65 and GAD67
GABA, the main inhibitory neurotransmitter in the central nervous system, is synthesized by 2 isoforms of the enzyme GAD, which are located on 2 different genes and demonstrate dissimilar location and function.40 GAD67, the predominant isoform responsible for >90% of GABA production, is located in the cytosol, and thought to function as a trophic factor,41 whereas GAD65 is located at the nerve terminals, specifically responding to short-term demands for GABA, such as during neurotransmission.42 GAD67 immunohistochemistry revealed a 12-fold increase in the number of caspase 3/GAD67 double-positive cells in anesthesia-exposed animals. This finding suggests that isoflurane leads to the demise of a significant amount of GABAergic neurons immediately after exposure, as also recently observed in newborn rats.43 Similarly, exposure to other GABA agonists, muscimol and propofol have also been previously found to induce GABAergic neuronal death in immature rat telencephalon cultures.44 More mature cells treated with these GABA agonists, however, were less vulnerable to long-term effects, suggesting that susceptibility is age-dependent.44
In the present study, prolonged exposure to isoflurane led to a significant decrease in both isoforms of the GABA-synthesizing enzyme. These reductions in GAD expression could be explained by either isoflurane-induced neuronal cell death of GAD-containing (i.e., GABAergic) neurons or by a drug-induced downregulation or cleavage of the enzyme. Although the present study found a 12-fold increase in the number of apoptotic GABAergic neurons after isoflurane exposure, potentially explaining the observed reduction in GAD expression, the disparate reductions in the 2 GAD isoforms, by 50% for GAD67 and by >90% for GAD65, suggest that exposure to anesthesia might have additional functional effects, beyond cell death, on GABAergic interneurons. Previous findings of decreased GAD enzyme activity after treatment with the GABAA agonists vigabatrin, propofol, and muscimol support this hypothesis.44,45 This downregulation could be explained by an isoflurane-induced, excitotoxic cleavage of GAD65 and GAD67, as previously reported in cultured hippocampal GABAergic neurons.46 Regardless of the mechanism, given the fact that interneurons comprise 12% to 15% of all cortical neurons in adult rodents,47 the permanent loss of even a small number of GABAergic interneurons or prolonged interference with their function could potentially have profound effects on the normal balance of excitation and inhibition in the developing brain. Given the anesthesia-induced alterations in GAD expression, however, the absolute fraction of GABAergic neurons eliminated during isoflurane exposure is difficult to assess, because both the numerator, the number of surviving GABAergic neurons, and the denominator in this equation, GAD67 expression, changed.
Because astrocytes are also a source for GABA and contain GAD67,48 we cannot exclude that glia may have had a role in the observed changes in GAD expression. However, converging lines of evidence lead us to believe that the anesthesia-induced reduction in GAD expression observed in our experiments was predominantly mediated by neurons, rather than astrocytes. First, we found that both GAD67 and GAD65 expressions were decreased after anesthetic exposure; the combination of GAD65 and GAD67 is only expressed in neurons, whereas astrocytes have only been shown to express GAD67.48 Second, although some GAD67–positively stained cells could have been astrocytes, it is more likely that they were GABAergic neurons, because cellular apoptosis overwhelmingly affected cells expressing the neuronal marker NeuN, and to a much lesser degree the astroglial markers GFAP or S100β, suggesting that alterations in GAD expression may have also predominantly occurred in neurons.
Human applicability of the present findings in animals remains unresolved. Although histopathological studies cannot be performed in healthy children after exposure to general anesthetics, several epidemiological studies have returned conflicting results. Some have detected an association between exposure to anesthesia and surgery early in life and subsequent behavioral or learning abnormalities,49,50 whereas others have not observed any deleterious effects in children exposed to anesthetics and sedatives during vulnerable periods in their brain development.51–53 The present study used clinically relevant doses of isoflurane, approximating 0.6 minimum alveolar concentration,8 for a relatively long exposure period, which may be outside of the normal clinical practice, to create a measurable effect, because the injury is exposure time–dependent and dose-dependent.9 The maturational stage of the human brain equivalent to the mice used in the present study remains controversial; older data suggest that postnatal day 7 mice are comparable to human infants,54 whereas more contemporary studies indicate mouse brain development at this stage to be closer to human fetuses at midgestation.55
In conclusion, a 6-hour exposure to clinical doses of isoflurane increased neuronal apoptotic cell death in 7-day-old mice, killing approximately 2% of cortical neurons, of which some were identified as GABAergic interneurons. Conversely, astrocytes were substantially less affected by isoflurane exposure at this age. Moreover, isoflurane exposure dramatically decreased expression of both isoforms of the GABA-synthesizing enzyme GAD, which indicates that the anesthetic drug may interfere with proper inhibitory function in the developing brain. The permanence of these findings, however, remains unknown. Additional studies will need to identify the phenomenon’s selectivity and molecular mechanisms to determine its applicability to pediatric anesthesia.
Name: George K. Istaphanous, MD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: George K. Istaphanous has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Christopher G. Ward, MD.
Contribution: This author helped conduct the study.
Attestation: Christopher G. Ward has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Xinyu Nan, BS.
Contribution: This author helped conduct the study.
Attestation: Xinyu Nan has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Elizabeth A. Hughes, BS.
Contribution: This author helped conduct the study.
Attestation: Elizabeth A. Hughes has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: John C. McCann, BS.
Contribution: This author helped conduct the study.
Attestation: John C. McCann has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: John J. McAuliffe, MD, MBA.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: John J. McAuliffe has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Steve C. Danzer, PhD.
Contribution: This author helped design the study and write the manuscript.
Attestation: Steve C. Danzer has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Andreas W. Loepke, MD, PhD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Andreas W. Loepke has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
This manuscript was handled by: Gregory J. Crosby, MD.
1. DeFrances CJ, Cullen KA, Kozak LJ. National Hospital Discharge Survey: 2005 annual summary with detailed diagnosis and procedure data. Vital Health Stat. 2007;13:1–209
2. Jevtovic-Todorovic V, Hartman RE, Izumi Y, Benshoff ND, Dikranian K, Zorumski CF, Olney JW, Wozniak DF. Early exposure to common anesthetic agents causes widespread neurodegeneration in the developing rat brain and persistent learning deficits. J Neurosci. 2003;23:876–82
3. Loepke AW, Istaphanous GK, McAuliffe JJ III, Miles L, Hughes EA, McCann JC, Harlow KE, Kurth CD, Williams MT, Vorhees CV, Danzer SC. The effects of neonatal isoflurane exposure in mice on brain cell viability, adult behavior, learning, and memory. Anesth Analg. 2009;108:90–104
4. Stratmann G, May LD, Sall JW, Alvi RS, Bell JS, Ormerod BK, Rau V, Hilton JF, Dai R, Lee MT, Visrodia KH, Ku B, Zusmer EJ, Guggenheim J, Firouzian A. Effect of hypercarbia and isoflurane on brain cell death and neurocognitive dysfunction in 7-day-old rats. Anesthesiology. 2009;110:849–61
5. Brambrink AM, Evers AS, Avidan MS, Farber NB, Smith DJ, Zhang X, Dissen GA, Creeley CE, Olney JW. Isoflurane-induced neuroapoptosis in the neonatal rhesus macaque brain. Anesthesiology. 2010;112:834–41
6. Rizzi S, Ori C, Jevtovic-Todorovic V. Timing versus duration: determinants of anesthesia-induced developmental apoptosis in the young mammalian brain. Ann NY Acad Sci. 2010;1199:43–51
7. Lemkuil BP, Head BP, Pearn ML, Patel HH, Drummond JC, Patel PM. Isoflurane neurotoxicity is mediated by p75NTR-RhoA activation and actin depolymerization. Anesthesiology. 2011;114:49–57
8. Istaphanous GK, Howard J, Nan X, Hughes EA, McCann JC, McAuliffe JJ, Danzer SC, Loepke AW. Comparison of the neuroapoptotic properties of equipotent anesthetic concentrations of desflurane, isoflurane, or sevoflurane in neonatal mice. Anesthesiology. 2011;114:578–87
9. Johnson SA, Young C, Olney JW. Isoflurane-induced neuroapoptosis in the developing brain of nonhypoglycemic mice. J Neurosurg Anesthesiol. 2008;20:21–8
10. Henschel O, Gipson KE, Bordey A. GABAA receptors, anesthetics and anticonvulsants in brain development. CNS Neurol Disord Drug Targets. 2008;7:211–24
11. Paxinos G, Franklin KB The Mouse Brain in Stereotaxic Coordinates. 20012nd ed San Diego Academic Press
12. West MJ, Slomianka L, Gundersen HJ. Unbiased stereological estimation of the total number of neurons in the subdivisions of the rat hippocampus using the optical fractionator. Anat Rec. 1991;231:482–97
13. West MJ. Stereological methods for estimating the total number of neurons and synapses: issues of precision and bias. Trends Neurosci. 1999;22:51–61
14. Graham PL, Mengersen K, Morton AP. Confidence limits for the ratio of two rates based on likelihood scores: non-iterative method. Stat Med. 2003;22:2071–83
15. Rothman KJ, Greenland S Modern Epidemiology. 19982nd ed Philadelphia Lippincott-Raven
16. Wei H, Kang B, Wei W, Liang G, Meng QC, Li Y, Eckenhoff RG. Isoflurane and sevoflurane affect cell survival and BCL-2/BAX ratio differently. Brain Res. 2005;1037:139–47
17. Wise-Faberowski L, Zhang H, Ing R, Pearlstein RD, Warner DS. Isoflurane-induced neuronal degeneration: an evaluation in organotypic hippocampal slice cultures. Anesth Analg. 2005;101:651–7
18. Wei H, Liang G, Yang H. Isoflurane preconditioning inhibited isoflurane-induced neurotoxicity. Neurosci Lett. 2007;425:59–62
19. Ma D, Williamson P, Januszewski A, Nogaro MC, Hossain M, Ong LP, Shu Y, Franks NP, Maze M. Xenon mitigates isoflurane-induced neuronal apoptosis in the developing rodent brain. Anesthesiology. 2007;106:746–53
20. Head BP, Patel HH, Niesman IR, Drummond JC, Roth DM, Patel PM. Inhibition of p75 neurotrophin receptor attenuates isoflurane-mediated neuronal apoptosis in the neonatal central nervous system. Anesthesiology. 2009;110:813–25
21. Stratmann G, Sall JW, May LD, Bell JS, Magnusson KR, Rau V, Visrodia KH, Alvi RS, Ku B, Lee MT, Dai R. Isoflurane differentially affects neurogenesis and long-term neurocognitive function in 60-day-old and 7-day-old rats. Anesthesiology. 2009;110:834–48
22. Jevtovic-Todorovic V, Olney JW. PRO: anesthesia-induced developmental neuroapoptosis—status of the evidence. Anesth Analg. 2008;106:1659–63
23. Loepke AW, McGowan FX Jr, Soriano SG. CON: the toxic effects of anesthetics in the developing brain: the clinical perspective. Anesth Analg. 2008;106:1664–9
24. Loepke AW. Developmental neurotoxicity of sedatives and anesthetics: a concern for neonatal and pediatric critical care medicine? Pediatr Crit Care Med. 2010;11:217–26
25. Mullen RJ, Buck CR, Smith AM. NeuN, a neuronal specific nuclear protein in vertebrates. Development. 1992;116:201–11
26. Lyck L, Krøigård T, Finsen B. Unbiased cell quantification reveals a continued increase in the number of neocortical neurones during early post-natal development in mice. Eur J Neurosci. 2007;26:1749–64
27. Yon JH, Daniel-Johnson J, Carter LB, Jevtovic-Todorovic V. Anesthesia induces neuronal cell death in the developing rat brain via the intrinsic and extrinsic apoptotic pathways. Neuroscience. 2005;135:815–27
28. Sall JW, Stratmann G, Leong J, McKleroy W, Mason D, Shenoy S, Pleasure SJ, Bickler PE. Isoflurane inhibits growth but does not cause cell death in hippocampal neural precursor cells grown in culture. Anesthesiology. 2009;110:826–33
29. Rakic P. Mode of cell migration to the superficial layers of fetal monkey neocortex. J Comp Neurol. 1972;145:61–83
30. Aschner M, Sonnewald U, Tan KH. Astrocyte modulation of neurotoxic injury. Brain Pathol. 2002;12:475–81
31. Stevens B. Neuron-astrocyte signaling in the development and plasticity of neural circuits. Neurosignals. 2008;16:278–88
32. Anderson MF, Blomstrand F, Blomstrand C, Eriksson PS, Nilsson M. Astrocytes and stroke: networking for survival? Neurochem Res. 2003;28:293–305
33. Yu AC, Wong HK, Yung HW, Lau LT. Ischemia-induced apoptosis in primary cultures of astrocytes. Glia. 2001;35:121–30
34. Yang Q, Hamberger A, Khatibi N, Stigbrand T, Haglid KG. Presence of S-100 beta in cholinergic neurones of the rat hindbrain. Neuroreport. 1996;7:3093–9
35. Vives V, Alonso G, Solal AC, Joubert D, Legraverend C. Visualization of S100B-positive neurons and glia in the central nervous system of EGFP transgenic mice. J Comp Neurol. 2003;457:404–19
36. Oppenheim RW. Cell death during development of the nervous system. Annu Rev Neurosci. 1991;14:453–501
37. Rakic S, Zecevic N. Programmed cell death in the developing human telencephalon. Eur J Neurosci. 2000;12:2721–34
38. Hodge RD, D’Ercole AJ, O’Kusky JR. Insulin-like growth factor-I (IGF-I) inhibits neuronal apoptosis in the developing cerebral cortex in vivo. Int J Dev Neurosci. 2007;25:233–41
39. Fagel DM, Ganat Y, Silbereis J, Ebbitt T, Stewart W, Zhang H, Ment LR, Vaccarino FM. Cortical neurogenesis enhanced by chronic perinatal hypoxia. Exp Neurol. 2006;199:77–91
40. Buddhala C, Hsu CC, Wu JY. A novel mechanism for GABA synthesis and packaging into synaptic vesicles. Neurochem Int. 2009;55:9–12
41. Kanaani J, Kolibachuk J, Martinez H, Baekkeskov S. Two distinct mechanisms target GAD67 to vesicular pathways and presynaptic clusters. J Cell Biol. 2010;190:911–25
42. Erlander MG, Tobin AJ. The structural and functional heterogeneity of glutamic acid decarboxylase: a review. Neurochem Res. 1991;16:215–26
43. Zhou ZW, Shu Y, Li M, Guo X, Pac-Soo C, Maze M, Ma D. The glutaminergic, GABAergic, dopaminergic but not cholinergic neurons are susceptible to anaesthesia-induced cell death in the rat developing brain. Neuroscience. 2011;174:64–70
44. Honegger P, Pardo B, Monnet-Tschudi F. Muscimol-induced death of GABAergic neurons in rat brain aggregating cell cultures. Brain Res Dev Brain Res. 1998;105:219–25
45. Honegger P, Matthieu JM. Selective toxicity of the general anesthetic propofol for GABAergic neurons in rat brain cell cultures. J Neurosci Res. 1996;45:631–6
46. Baptista MS, Melo CV, Armelão M, Herrmann D, Pimentel DO, Leal G, Caldeira MV, Bahr BA, Bengtson M, Almeida RD, Duarte CB. Role of the proteasome in excitotoxicity-induced cleavage of glutamic acid decarboxylase in cultured hippocampal neurons. PLoS One. 2010;5:e10139
47. Beaulieu C. Numerical data on neocortical neurons in adult rat, with special reference to the GABA population. Brain Res. 1993;609:284–92
48. Lee M, Schwab C, McGeer PL. Astrocytes are GABAergic cells that modulate microglial activity. Glia. 2011;59:152–65
49. Wilder RT, Flick RP, Sprung J, Katusic SK, Barbaresi WJ, Mickelson C, Gleich SJ, Schroeder DR, Weaver AL, Warner DO. Early exposure to anesthesia and learning disabilities in a population-based birth cohort. Anesthesiology. 2009;110:796–804
50. DiMaggio C, Sun LS, Li G. Early childhood exposure to anesthesia and risk of developmental and behavioral disorders in a sibling birth cohort. Anesth Analg. 2011;113:1143–51
51. Rozé JC, Denizot S, Carbajal R, Ancel PY, Kaminski M, Arnaud C, Truffert P, Marret S, Matis J, Thiriez G, Cambonie G, André M, Larroque B, Bréart G. Prolonged sedation and/or analgesia and 5-year neurodevelopment outcome in very preterm infants: results from the EPIPAGE cohort. Arch Pediatr Adolesc Med. 2008;162:728–33
52. Bartels M, Althoff RR, Boomsma DI. Anesthesia and cognitive performance in children: no evidence for a causal relationship. Twin Res Hum Genet. 2009;12:246–53
53. Guerra GG, Robertson CM, Alton GY, Joffe AR, Cave DA, Dinu IA, Creighton DE, Ross DB, Rebeyka IMWestern Canadian Complex Pediatric Therapies Follow-up Group. . Neurodevelopmental outcome following exposure to sedative and analgesic drugs for complex cardiac surgery in infancy. Paediatr Anaesth. 2011;21:932–41
54. Dobbing J, Sands J. Quantitative growth and development of human brain. Arch Dis Child. 1973;48:757–67
55. Clancy B, Darlington RB, Finlay BL. Translating developmental time across mammalian species. Neuroscience. 2001;105:7–17