Abstract
Studies in developing animals have shown that anesthetic agents can lead to neuronal cell death and learning disabilities when administered early in life. Development of human embryonic stem cell–derived neurons has provided a valuable tool for understanding the effects of anesthetics on developing human neurons. Unbalanced mitochondrial fusion and fission lead to various pathological conditions including neurodegeneration. The aim of this study was to dissect the role of mitochondrial dynamics in propofol-induced neurotoxicity.
Terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling staining was used to assess cell death in human embryonic stem cell–derived neurons. Mitochondrial fission was assessed using TOM20 staining and electron microscopy. Expression of mitochondrial fission-related proteins was assessed by Western blot, and confocal microscopy was used to assess opening time of the mitochondrial permeability transition pore (mPTP).
Exposure to 6 h of 20 μg/ml propofol increased cell death from 3.18 ± 0.17% in the control-treated group to 9.6 ± 0.95% and led to detrimental increases in mitochondrial fission (n = 5 coverslips per group) accompanied by increased expression of activated dynamin-related protein 1 and cyclin-dependent kinase 1, key proteins responsible for mitochondrial fission. Propofol exposure also induced earlier opening of the mPTP from 118.9 ± 3.1 s in the control-treated group to 73.3 ± 1.6 s. Pretreatment of the cells with mdivi-1, a mitochondrial fission blocker rescued the propofol-induced toxicity, mitochondrial fission, and mPTP opening time (n = 75 cells per group). Inhibiting cyclin-dependent kinase 1 attenuated the increase in cell death and fission and the increase in expression of activated dynamin-related protein 1.
These data demonstrate for the first time that propofol-induced neurotoxicity occurs through a mitochondrial fission/mPTP-mediated pathway.
Abstract
In supratherapeutic doses, propofol increased mitochondrial fission; key proteins involved in fission were increased. Suppression of these proteins mitigated the increases in mitochondrial fission and reduced toxicity. Propofol accelerated mitochondrial permeability transition pore (mPTP) opening; blockade of fission rescued the mPTP opening time. The data indicate that propofol, when administered in supratherapeutic concentrations, may induce neurotoxicity via a mitochondria fission and mPTP pathway.
Preclinical studies of neonatal neurotoxicity have shown that anesthetics injure mitochondria in the developing brain. Morphologic changes include alteration in mitochondrial size and structure of the cristae.
How anesthetics injure mitochondria is not clear. Mitochondria undergo cycles of fusion and fission, and these cycles are essential to their proper function. It is possible that anesthetic might interfere with mitochondrial fusion and fission.
The effect of propofol on neurons derived from human embryonic stem cells was evaluated in vitro.
In supratherapeutic doses, propofol increased mitochondrial fission; expression of key proteins involved in fission was increased. Suppression of these proteins mitigated the increases in mitochondrial fission and reduced toxicity.
Propofol accelerated mitochondrial permeability transition pore (mPTP) opening; blockade of fission rescued the mPTP opening time.
The data indicate that propofol, when administered in supratherapeutic concentrations, may induce neurotoxicity via a mitochondrial fission and mPTP pathway.
WHEN administered early in life, anesthetics, including propofol, can lead to death of the neurons and neuronal supporting cells and have been associated with increased risk of learning, memory, and behavioral deficiencies.1–6 These detrimental effects have been well established in many animal models and have raised safety concerns regarding the use of anesthetics in children. Although millions of children are exposed to anesthetics every year, the use of these anesthetics in imaging or surgery is undeniably necessary.7 Despite the large-scale efforts of research initiatives such as SmartTots, an organization tasked with evaluating the safety of anesthetics on the developing human brain, the effects of anesthetics in children remain uncertain.8–10
The human epidemiological studies conducted thus far have produced widely variable results,11–13 and the discrepancy in the results of these studies highlights the importance of developing a better human model by which to study the effects of anesthetics on the immature human brain. Human embryonic stem cells (hESCs) are pluripotent cells that can replicate indefinitely and are capable of differentiating into any cell type.14,15 Generating neurons from hESCs provides us with an essentially endless supply of human cells by which to study the effects of anesthetics on developing human neurons and the mechanisms governing anesthetic-induced neurotoxicity.
The mitochondria of the cell are extremely important organelles involved in many cellular processes including energy production, cell signaling, and apoptosis.16 To maintain proper functioning, the mitochondria continuously undergo cycles of fusion and fission. Unbalanced fusion and fission can sometimes lead to various pathological conditions including neurodegeneration and has been linked to many neurodegenerative disorders.17–21 Dynamin-related protein 1 (Drp1) is a key regulator of mitochondrial fission. Phosphorylation of Drp1 by cyclin-dependent kinase 1 (CDK1) at the Serine 616 position induces mitochondrial fission.22–24 Mitochondrial fusion is regulated primarily by the proteins mitofusion 1 and 2 (MFN1 and MFN2) and optic atrophy 1 (OPA1). Previous studies have shown that exposure of neonatal rat pups to general anesthetics induces significant increases in mitochondrial fission.25,26 However, the role of mitochondrial dynamics and related pathways in propofol-induced neurotoxicity has yet to be investigated.
The mitochondrial permeability transition pore (mPTP) is a pore that spans the outer and inner mitochondrial membranes and is opened by oxidative stress.27 When the mPTP opens, there is a large influx of solutes and water into the mitochondria. This can lead to swelling and eventual rupture of the mitochondria and cell death.28 Inhibition of mPTP opening has been shown to attenuate ethanol-induced neurotoxicity in mice.29 However, the role of mPTP opening and its connection to mitochondrial fission in propofol-induced neurotoxicity has yet to be studied. The aim of this study was to dissect the role of mitochondrial dynamics and mPTP opening in propofol-induced neurotoxicity. We hypothesized that propofol would induce hESC-derived neuronal cell death though CDK1-mediated activation of Drp1, increased mitochondrial fission, and mPTP opening.
Materials and Methods
hESC Culture and Differentiation into Neurons
All human cell experiments described were approved by the institutional review board at the Medical College of Wisconsin (Milwaukee, Wisconsin). hESCs (H1 line; WiCell Research Institute Inc., USA) were cultured as previously described by our laboratory on a feeder layer of mitotically inactivated mouse embryonic fibroblasts (MEFs).30–32 In brief, hESCs were plated onto MEFs that had been cultured on matrigel-coated 60-mm dishes. The hESCs were cultured in hESC media consisting of Dulbecco's Modified Eagle's medium/F12 supplemented with 20% knockout serum (Life Technologies, USA), 1% nonessential amino acids, 1 mM l-glutamine, 1% penicillin–streptomycin (Life Technologies), 4 ng/ml human recombinant basic fibroblast growth factor (bFGF; Life Technologies), and 0.1 mM β-mercaptoethanol (Sigma-Aldrich, USA) in the hypoxic incubator. The cells were mechanically passaged when confluent and the media were changed daily. Cells at passages 55 to 70 were used in these studies.
A four-step directed differentiation protocol was used to generate neurons from the hESCs as previously described by our laboratory.30–32 In brief, the hESCs were dissociated from the MEF cultures, the pellet of hESCs was resuspended in hESC media without bFGF, and the cells were cultured in a normoxic incubator (20% O2/5% CO2, 37°C) on ultralow attachment six-well plates (Corning Inc., USA). Embryoid bodies were present in the cultures 1 day after digestion, and the media were changed daily. Five days after dispase digestion, the cultures were switched to neural induction media consisting of Dulbecco's Modified Eagle's medium/F12 supplemented with 1% N2 (Life Technologies), 1% nonessential amino acids, 1 mg/ml heparin (Sigma), and 5 ng/ml bFGF, and the media were changed daily for an additional 4 days. The embryoid bodies were then plated down to 35-mm matrigel-coated dishes. The media were changed every other day until rosette-like structures were present in the cultures (within 5 days of plating). The rosettes were manually selected using a pipette tip and transferred to new, matrigel-coated, 35-mm dishes and cultured in neural expansion media containing Dulbecco's Modified Eagle's medium/F12 supplemented with 2% B27 without vitamin A, 1% N2 (Life Technologies), 1% nonessential amino acids, 20 ng/ml bFGF, and 1 mg/ml heparin. The neural stem cells (NSCs) present in the rosettes were passaged enzymatically every 5 days using Accutase (Innovative Cell Technologies, USA). To generate neurons, NSCs were cultured in matrigel-coated dishes to approximately 95% confluency. Once confluent, the cultures were switched to neuron differentiation media. The media were changed every other day, and neurons were used for these studies 2 weeks after initiation of differentiation media.
Immunofluorescence Staining and Confocal Microscopy
Human embryonic stem cells–derived neurons were cultured on matrigel-coated glass coverslips in neuron differentiation media for 2 weeks. Immunostaining was performed as previously described by our laboratory.30–32 In brief, cells were fixed for 30 min at room temperature in 4% paraformaldehyde, washed, and incubated in 0.5% Triton X-100 (Sigma-Aldrich) in phosphate-buffered saline (PBS) for 15 min. The cells were then incubated for 20 min in 10% donkey serum (blocking agent) at room temperature and placed in the primary antibodies (rabbit anti-β-tubulin III [neuron-specific marker], mouse antidoublecortin [immature neuron marker; Abcam, USA], mouse Tetramethylrhodamine isothiocyanante-conjugated phalloidin [F-actin marker; Millipore, USA), or rabbit anti-TOM20 [translocase of outer mitochondrial membranes 20 kDa; Santa Cruz, USA]) overnight at 4°C. Where appropriate, after incubation in the primary antibody, the cells were washed and incubated overnight at 4°C with Alexa Fluor 488 or 594 donkey anti-mouse or rabbit immunoglobulin G secondary antibodies (Life Technologies). The cells were incubated for 15 min at room temperature in Hoechst 33342 (Life Technologies) to stain the cell nuclei. The coverslips were mounted onto glass slides using mounting media and imaged using a laser scanning confocal microscope (Nikon Eclipse Ti-E; Nikon Inc., USA).
Propofol, Mdivi-1, and Roscovitine Exposure
Brain concentrations of propofol in humans during anesthesia are believed to range between 4 and 20 μg/ml.33,34 We reported previously that propofol induced cell death in hESC-derived neurons in a dose-dependent and exposure number–dependent manner.32 All experiments in this study were performed in 2-week-old hESC-derived neurons after a single exposure to 6 h of 20 μg/ml research-grade propofol. The propofol was prepared as a 40-mg/ml stock solution in dimethyl sulfoxide (DMSO; Sigma-Aldrich) and diluted to the working concentration in neurobasal media. An equal volume of DMSO was used as the vehicle control. The cells were cultured in matrigel-coated 60-mm culture dishes (500,000 cells per dish) or 12-mm coverslips (100,000 cells per coverslip), and all exposures were done in a humidified chamber at 37°C.
Mdivi-1 is a selective inhibitor of Drp1, a key protein responsible for inducing mitochondrial fission. We used mdivi-1 (Sigma-Aldrich) to understand the role of mitochondrial fission in propofol-induced neurotoxicity. hESC-derived neurons were pretreated for 1 h with 25-μM mdivi-1 in a humidified, 37°C incubator. The cells were washed once with neurobasal media and immediately exposed to 6 h of propofol or DMSO. A 50-mM stock solution of mdivi-1 was prepared in DMSO and diluted to the working concentration in neurobasal media. During the pretreatment period, control cells were exposed with an equal volume of DMSO (“mdivi-1 control”) or neurobasal media alone (“no-treatment” group). After the propofol exposure, the cells were immediately fixed for TOM20 or terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling (TUNEL) staining.
Roscovitine is a selective inhibitor of several CDKs including CDK1. Roscovitine has been shown in both in vivo and in vitro models to reduce neuronal cell death and neurological deficits after brain trauma, cerebral ischemia, and trophic deprivation.35–37 However, the mechanism by which roscovitine provides neuroprotection and the role of CDK1 inhibition in anesthetic-induced neurotoxicity have yet to be studied. hESC-derived neurons were pretreated for 1 h with 10 μM roscovitine (Sigma-Aldrich) at 37°C. The cells were washed with fresh neurobasal media and exposed for 6 h to propofol or DMSO. The stock solution of roscovitine was prepared in DMSO and diluted in neurobasal media to the working concentration. To ensure that the DMSO was not producing an effect, the control cells were exposed during the pretreatment to either neurobasal media alone (“no-treatment” group) or an equal volume of DMSO (“roscovitine control” group).
TUNEL Staining
To assess cell death in the hESC-derived neurons, a cell death detection kit (Roche Applied Bio Sciences, USA) was used following the instructions provided by the manufacturer. This kit uses a process known as TUNEL to label the free 3′-OH termini of single- and double-stranded DNA breaks. The polymerase terminal deoxynucleotidyl transferase adds modified nucleotides to the free 3′-OH termini, which are secondarily labeled with a fluorescent marker. This allows for the visualization of TUNEL-positive cells on a fluorescent microscope.
Human embryonic stem cells–derived neurons were cultured on glass coverslips and exposed to propofol or DMSO with or without mdivi-1 pretreatment. The cells were rinsed with PBS after an 18-h washout in fresh neurobasal media and fixed for 30 min with 4% paraformaldehyde. Fragmented DNA was labeled using the TUNEL staining kit, and the cells were stained for 15 min with Hoechst 33342 to label the cell nuclei. The coverslips were mounted to glass slides and imaged at 200× total magnification using a fluorescent microscope. A minimum of five to six images were taken per coverslip with each image containing approximately 100 cells. Therefore, approximately 500 cells were imaged for each coverslip, and five coverslips were imaged per group. Cell death was quantified by manually counting the number of TUNEL-positive cells in the field and the total cell nuclei. The data are expressed as a percentage of TUNEL-positive cells present in the field. For these experiments, the coverslips were placed in separate wells of a 24-well plate. To eliminate the possibility of bias introduced in the plate, the coverslips were assigned randomly to the different groups across the entire plate. In addition, the experimenter was blinded to the groups being imaged and analyzed to eliminate the possibility of bias introduced by the experimenter.
Mitochondrial Morphology Analysis
To assess changes in mitochondrial morphology, hESC-derived neurons were exposed to mdivi-1 and/or propofol or DMSO and fixed in 4% paraformaldehyde. The cells were stained using the TOM20 primary antibody to label the mitochondria as described in the Immunofluorescence Staining and Confocal Microscopy section. The cells were imaged using the confocal microscope, and single-cell images were captured. The images were analyzed using the ImageJ 1.47v software (Wayne Rasband; National Institutes of Health, Bethesda, Maryland) as previously described by our laboratory.30 In brief, the fluorescent images were converted to binary images, and the length, area, width, and perimeter of the mitochondria were determined. To assess the mitochondrial morphology, the aspect ratio (AR) and form factor (FF) were calculated for each cell. The AR is a measure of the length of the mitochondria and is calculated as the ratio of the major to minor axes of the ellipse, whereas the FF is a measure of the degree of mitochondrial branching and is calculated using the equation: FF = Perimeter2/4π × area. The minimum value for both parameters is 1, which would represent a sphere. Healthy, elongated, and interconnected mitochondria have a high FF and AR, whereas fragmented and discontinuous mitochondria have low FF and AR values. For each group, 15 cells were imaged and analyzed from five coverslips (75 cells analyzed per group). The images from the 15 cells were averaged for each coverslip. The n is the number of coverslips imaged, which is 5 for each group. To minimize bias introduced by the experimenters, one experimenter was responsible for exposing the cells to the various conditions and another experimenter (who was blinded to the groups being studied) was responsible for imaging and data analysis.
Electron Microscopy
Human embryonic stem cells–derived neurons were cultured on matrigel-coated plastic coverslips and were exposed to propofol or DMSO with or without mdivi-1 pretreatment. The cells were immediately fixed at 4°C with 2% glutaraldehyde in 0.1 M sodium cacodylate buffer and postfixed for 1 h on ice with 1% osmium tetroxide. The cells were rinsed with distilled water and dehydrated using acetonitrile and graded methanol (50%, 20 min; 70%, 20 min; 95%, 20 min; and 100% 3×, 20 min). The cells were embedded in epoxy resin (EMbed-812; Electron Microscopy Sciences, USA) and polymerized at 70°C overnight. Thin (60 nm) sections were cut, and the sections were stained with lead citrate and uranyl acetate. The samples were imaged using a Hitachi H600 Electron Microscope (Hitachi, Japan).
mPTP Opening and Membrane Potential
Neural stem cells were cultured on glass coverslips for 2 weeks in neuron differentiation media. The hESC-derived neurons were exposed to propofol/DMSO with or without mdivi-1 pretreatment as described. The cells were then loaded for 20 min at 37°C with Hoechst 33342 and 50 nM tetramethylrhodamine ethyl ester (TMRE; Life Technologies), a lipophilic, membrane potential–dependent fluorescent dye that is readily taken up by mitochondria. The cells were washed with fresh neurobasal media and imaged using a confocal microscope. Rapid laser scanning with the 561-nm emission line laser of the confocal microscope was used to induce oxidative stress and opening of the mPTP by exposing a single cell to laser excitation every 3.5 s for 55 frames. Opening of the mPTP was visualized as a rapid loss of TMRE fluorescence. The arbitrary mPTP opening time was assessed by determining the time when the TMRE fluorescence had decreased by half from the baseline TMRE fluorescence intensity to the residual TMRE fluorescence intensity. An earlier opening of the mPTP is indicative of increased vulnerability of the cells to stress. When analyzing the images, a small area of mitochondria (the size of which was held constant across groups and images) was analyzed. This was intended to prevent any change in mitochondrial number or surface area across groups from influencing the results because we were only interested in the time it took the pore to open. Neurons loaded with TMRE and Hoechst 33342 after propofol or DMSO exposure were also imaged on the confocal microscope to assess the mitochondrial membrane potential. The fluorescence intensity of each cell (area of analysis held constant across groups) was measured using the ImageJ software, and the mitochondrial membrane potential was reported in arbitrary units. To minimize any potential bias in the results, the experimenter was blinded to the group being imaged and analyzed.
Western Blot
Human embryonic stem cells–derived neurons were exposed to propofol or DMSO with or without mdivi-1 pretreatment, rinsed with PBS, and immediately lysed in Radio-Immunoprecipitation Assay lysis buffer (Cell Signaling, USA) containing phosphatase inhibitor cocktail (Roche Diagnostics) as described previously by our laboratory.32 Total protein concentration was determined for each sample, and the samples were boiled for 5 min at 97°C. After the denaturation step, 25 μg of protein/lane was loaded to gels for sodium dodecyl sulfate polyacrylamide gel electrophoresis separation. After the gel separation, the protein was transferred to a nitrocellulose membrane, and the membranes were blocked. The membranes were then incubated at 4°C overnight with the primary antibodies: rabbit anti-phosphorylated Drp1 (pDrp1 Ser 616) (Cell Signaling), mouse anti-Drp1 (BD Technologies, USA), mouse anti-CDK1 (Santa Cruz), rabbit anti-OPA1, mouse anti-Mitofusion 1, or mouse anti-Mitofusion 2 (Abcam). The membranes were then incubated for 1 h at room temperature with secondary antibodies conjugated to horseradish peroxidase (Cell Signaling). The antibody-labeled proteins were then detected with chemiluminescence detection reagent (Cell Signaling) and exposed on x-ray film. To quantify the optical densities from the x-ray films, the ImageJ software was used. For all Western blots, an n of 5 was used, which represents five Western blots run from five separate preparations with each preparation containing five 60-mm culture dishes of neurons for each group run.
Statistical Analysis
All experiments were performed on samples from three to four independent neuronal differentiations. We chose to use a sample size of n = 5 coverslips per dishes for our experiments. This was based on previous findings by our group indicating that an n of 3 is sufficient in this model, but an n of 5 provides greater confidence in the results generated. The sample size was increased during the review process, but no adjustments were made for multiple comparisons. Values had normal distributions and were reported at mean ± SEM. When two groups were compared, a Student t test was used, and one-way ANOVA with a Tukey correction for multiple testing was used when more than two groups were compared. The SigmaPlot 12.5 software (Systat Software, Inc., USA) was used for all statistical analyses, and a two-tailed P value less than 0.05 was considered significant.
Results
Characterization of hESC-derived Neurons and Propofol-induced Cell Death
Human embryonic stem cells were taken through a four-step differentiation protocol using chemically defined media to generate neurons. The cells were characterized using immunofluorescence staining after 2 weeks in differentiation media. About 90 to 95% of the cells in culture stained positive for the neuron-specific marker, β-tubulin III (fig. 1A, a), indicating that the differentiation protocol was highly efficient in the generation of neurons. The cells also displayed a morphology characteristic of neurons with small cell bodies and long interrelating projections. Although this immunostaining confirmed the purity of the cultures, it did not establish the maturity level of the cells. To assess the maturity level of the generated neurons, the cells were immunostained using the doublecortin antibody, a marker of migratory and immature neurons. Most of the neurons generated (90 to 95%) stained positive for this marker of immature neurons (fig. 1A, b). In addition, the β-tubulin III and doublecortin staining colocalized (fig. 1A, c). Our laboratory has shown previously that these cultures display synapse-like structures and express additional neuronal markers as well as presynaptic and postsynaptic markers.30–32
We have shown previously that propofol induced cell death in hESC-derived neurons in a dose-dependent and exposure number–dependent manner.32 Exposure to 6 h of 20 μg/ml propofol induced a significant increase in TUNEL-positive cells compared with vehicle/control (DMSO)-treated cells, indicating a significant increase in cell death (fig. 1B). In the control-treated group, 3.18 ± 0.17% of the cells were TUNEL positive, whereas 9.6 ± 0.95% of the cells were TUNEL positive in the propofol-treated group. The cells were exposed to 6 h of 10 μg/ml propofol, 20 μg/ml propofol, or the vehicle control and then costained for the neuron-specific marker, β-tubulin III and phalloidin, an F-actin marker. Nearly all of the cells in culture in all groups stained positive for the neuron-specific marker, indicating high differentiation efficiency (fig. 1C). After exposure to 10 μg/ml propofol, the actin cytoskeleton of the β-tubulin-III-stained neurons appeared disorganized compared with control-treated cells as assessed by phalloidin staining. The cytoskeletal disruption was further exacerbated in the 20 μg/ml propofol-treated group (fig. 1D). These data ultimately suggest that the propofol-induced mitochondrial fission is detrimental to the formation of the neuronal network, and cytoskeletal disorganization has been linked to apoptosis.38–40
Propofol Exposure Increases Mitochondrial Fragmentation
To assess changes in mitochondrial dynamics (mitochondrial fusion and fission), hESC-derived neurons were exposed to propofol or the vehicle control for 6 h and immunostained for TOM20. The cells were exposed to 0, 5, 10, and 20 μg/ml propofol and imaged on the confocal microscope using a 488-nm laser to visualize the mitochondria. There was a dose–response to propofol of mitochondrial fission in the cells. In control-treated cells and 5 μg/ml propofol-treated cells, the mitochondria appeared elongated and interconnected. In contrast, the mitochondria appeared visibly smaller in the 10 μg/ml propofol-treated group and severely fragmented and discontinuous in the 20 μg/ml propofol-treated group. These results were quantified by assessing the FF and AR of the cells, and both measurements were reduced in the 10 μg/ml propofol-treated group with a greater reduction in the 20 μg/ml propofol-treated group (fig. 2). Cells were then exposed to 6 h of 20 μg/ml propofol or the vehicle control after a 1-h pretreatment with mdivi-1 or control (DMSO) and immunostained for TOM20. The mitochondrial in the control-treated groups appeared healthy and elongated. The mitochondria were severely fragmented in the groups treated with propofol after pretreatment for 1 h with media alone (no treatment) or the mdivi-1 control. The mitochondrial morphology appeared to return to control conditions in cells pretreated with mdivi-1 before the propofol exposure (fig. 3A, a).
To quantify the changes in mitochondrial morphology, ImageJ software was used to analyze the confocal images. After exposure to propofol, the cells had significantly lower AR and FF values compared with control-treated cells indicating that the mitochondria were significantly shorter and less branched, respectively (fig. 3A, b and c). The AR and FF of the control-treated cells were 4.11 ± 0.15 and 6.35 ± 0.38, respectively. In propofol-treated cells, the AR and FF were reduced to 2.27 ± 0.09 and 3.04 ± 0.06, respectively, indicating significant mitochondrial fission. The mitochondrial fission in the propofol-treated group was not rescued by pretreatment of the cells with the mdivi-1 control (AR: 2.49 ± 0.15; FF: 3.16 ± 0.09) but was rescued to control values when the cells were pretreated with 25 μM of the mitochondrial fission blocker, mdivi-1 (AR: 4.72 ± 0.18; FF: 6.85 ± 0.31).
To further assess changes in mitochondrial morphology, the cellular ultrastructure of hESC-derived neurons exposed to propofol or the vehicle control was imaged using an electron microscope. In the control-treated group, the mitochondria appeared elongated and branched with a long oval appearance and regular cristae. In contrast, when the cells were treated with propofol, the mitochondria became strikingly smaller and more spherical in shape. When the cells were pretreated for 1 h with 25 μM mdivi-1 before the propofol exposure, the cellular ultrastructure returned to control conditions with the mitochondria appearing elongated and healthy (fig. 3B).
Propofol-induced Cell Death Is Attenuated by Pretreatment of the Cells with Mdivi-1
TUNEL staining was used to assess cell death in hESC-derived neurons after exposure to propofol or the vehicle control with or without mdivi-1 pretreatment. The number of TUNEL-positive cells appeared considerably higher in the propofol-treated group compared with the control-treated groups (fig. 4A). To quantify the results, the cells were manually counted to obtain a percentage of TUNEL-positive cells. Once again, propofol induced a significant increase in TUNEL-positive cells (7.54 ± 0.2%) compared with control-treated cells (1.81 ± 0.37%). The increase in TUNEL-positive cells was partially attenuated by pretreatment of the cells for 1 h with 25 μM of the mitochondrial fission blocker mdivi-1 (3.02 ± 0.23%) before propofol exposure, suggesting that mitochondrial fission plays an important role in propofol-induced cell death (fig. 4B).
Exposure of hESC-derived Neurons to Propofol Induces Alterations in the Expression of Mitochondrial Fission-related Proteins
To further dissect the mechanism of propofol-induced neurotoxicity in hESC-derived neurons, the expression of key mitochondrial fission- and fusion-related proteins was assessed by Western blot (fig. 4). We found that the level of activated Drp1 (pDrp1 Ser 616), a protein important in inducing mitochondrial fission, was significantly increased after propofol exposure (fig. 5A), confirming propofol-induced increases in mitochondrial fission through Drp1 activation. However, there was no significant change in the expression of any of the mitochondrial fusion-related proteins (MFN1, MFN2, or OPA1) after propofol exposure, suggesting that propofol alters mitochondrial dynamics through effects only on mitochondrial fission (fig. 5B).
CDK1 Inhibition Attenuates Propofol-induced Cell Death, Drp1 Activation, and Mitochondrial Fission
Cyclin-dependent kinase 1 is a kinase and has been reported to be involved in phosphorylation and activation of Drp1.24,41 The expression of CDK1 was increased after exposure to propofol as assessed by Western blot (fig. 6A). Inhibition of CDK1 with the chemical blocker roscovitine attenuated the propofol-induced cell death as assessed by TUNEL staining (fig. 6B). In addition, pretreatment of the cells with roscovitine reduced the expression of pDrp1 (Ser616) (fig. 6C), suggesting that the propofol-induced increase in CDK1 expression is inducing mitochondrial fission and cell death through activation of Drp1. Pretreatment of the cells with roscovitine also attenuated the propofol-induced mitochondrial fission, further confirming a role for CDK1 in the mechanism of propofol-induced neurotoxicity (fig. 6D).
Propofol Exposure Reduces the Mitochondrial Membrane Potential and Induces an Earlier Opening of the mPTP
To assess the downstream pathway of propofol-induced neuronal cell death, mPTP opening time was analyzed. hESC-derived neurons were loaded with TMRE, a cationic dye that is readily sequestered by the mitochondria. The mitochondria appeared elongated and healthy in all control-treated groups but appeared severely fragmented in the propofol-treated group on confocal microscopy images. Once again, the mitochondria appeared healthy in the propofol group when the cells were pretreated with mdivi-1 (fig. 7A). Opening of the mPTP was then induced using rapid laser scanning on the confocal microscope. The arbitrary mPTP opening time was calculated as the time at which the TMRE fluorescence intensity had decreased by half. Figure 7B depicts the quantified mPTP opening times for each group. The mPTP opened earlier in the propofol no-treatment and propofol mdivi-1 control-treated groups compared with all other groups, suggesting that mPTP opening is likely involved in the mechanism governing propofol-induced cell death. In addition, the mPTP opening was rescued by pretreatment of the cells with mdivi-1, suggesting an important connection between mitochondrial fission and mPTP opening. Decreased mitochondrial membrane potential has been reported to be related to matrix configuration and cytochrome c release during apoptosis.42 Exposure to 6 h of 20 μg/ml propofol induced a significant decrease in the mitochondrial membrane potential of the hESC-derived neurons, further indicating mitochondrial damage (fig. 7C).
Discussion
In this study, changes in mitochondrial dynamics after propofol exposure in hESC-derived neurons and the role that these changes play in propofol-induced neurotoxicity were assessed. Exposure of hESC-derived neurons to 20 μg/ml propofol for 6 h induced significant cell death, increased mitochondrial fission and led to earlier opening of the mPTP, and a decrease in the mitochondrial membrane potential. In addition, we showed that propofol altered the expression of CDK1 and Drp1, key proteins involved in mitochondrial fission, but did not significantly alter the expression of mitochondrial fusion-related proteins. Collectively, our data indicate, for the first time, that propofol induces neuronal cell death through CDK1/Drp1-induced mitochondrial fission and mPTP opening.
The developing brain is most vulnerable to anesthetic exposure during the period of rapid synaptogenesis, which in humans ranges from the third trimester of pregnancy through the third year of life.43 It is difficult to accurately assess the maturity level of hESC-derived neurons due to a lack of information regarding neuronal characteristics during human brain development. However, most of the hESC-derived neurons in our cultures stained positive for doublecortin, a marker of immature neurons (fig. 1A). Although it is not possible to determine the exact developmental stage of the stem cell–derived neurons, they are immature and vulnerable to anesthetic-induced cell death. Our laboratory showed previously that these cells also express the presynaptic and postsynaptic markers, Synapsin I and Homer I, respectively, and display synapse-like structures upon electron microscopy imaging, indicating that they also form synapses in culture.30,32 Therefore, we have developed a culture system of pure, immature, and functional neurons, which allows us to study the mechanisms governing propofol-induced neurotoxicity in a human model. The mechanisms by which anesthetics induce neurotoxicity remain largely unknown. MicroRNAs, neuroinflammation, calcium signaling, and reactive oxygen species production all appear to play a role in the toxicity.30,32,44–46 In this study, we examined the role of mitochondrial dynamics in the mechanism of propofol-induced neurotoxicity.
In this study, we found that exposure of hESC-derived neurons to 6 h of 20 μg/ml propofol induced significant cell death and disorganization of the actin cytoskeleton (fig. 1, B–D). Increases in mitochondrial fission have been shown to induce apoptosis- and necrosis-mediated cell death in some studies.47,48 Additional studies have demonstrated that inhibition of Drp1 could prevent increases in mitochondrial fission and lead to increased neuronal cell survival, implicating Drp1 inhibition as a potential neuroprotective strategy.49 It is important to note that the process of mitochondrial fusion and fission is extremely complex, and increases in mitochondrial fission have been shown to be beneficial in certain cell types.50 However, the role of mitochondrial fission in anesthetic-induced cell death is just beginning to be understood. It has been shown that exposure to a combination of general anesthetics for 6 h in 7-day-old rat pups induced significant mitochondrial fission.25 In addition, our laboratory has shown previously that exposure of hESC-derived neurons to 24 h of 100 μM ketamine induced an increase in cell death and mitochondrial fission.30 However, the functional role of mitochondrial dynamics in anesthetic-induced neurotoxicity has yet to be studied. In this study, propofol dose-dependently induced mitochondrial fission in the hESC-derived neurons. The propofol-mediated increase in mitochondrial fission and cell death could be prevented by pretreatment of the cells with the mitochondrial fission blocker, mdivi-1 (figs. 3 and 4). This suggests that mitochondrial fission is involved in propofol-induced neurotoxicity. However, pretreatment of the cells with mdivi-1 did not return the propofol-induced cell death all the way back to control levels. This highlights the complexity of the mechanisms governing propofol-induced neurotoxicity. Although the pathway we have elucidated in this study does appear to play an important role in this process, the role of additional pathways cannot be excluded.
To understand the upstream effectors of mitochondrial fission after propofol exposure, we began by looking at the expression of Drp1. Drp1 is an important protein responsible for inducing mitochondrial fission, and it has been shown in primary cortical rat-cultured neurons that recruitment of Drp1 along with several proteins plays a role in mitochondrial fission and cell death under excitotoxic conditions.51 Drp1 is activated primarily by phosphorylation at the Serine 616 position.48 We found that the expression of activated Drp1 was substantially increased after exposure to propofol (fig. 5). Drp1 is phosphorylated at the Ser616 position and activated predominantly by CDK1.22 Taguchi et al.52 were the first to report that CDK1 is responsible for phosphorylating and activating Drp1 in rat cultures at the Serine 585 position, which is equivalent in humans to the Serine 616 position. They went on to show that activation of Drp1 by CDK1 leads to increased mitochondrial fission. More recently, it was shown by immunoprecipitation that CDK1 and Drp1 interact in cultured mouse cardiomyocytes and that blockade of CDK1 with roscovitine in these same cells decreases the expression of pDrp1 Serine 616.53 Drp1 regulation is complex and involves many additional proteins; for example, phosphorylation of Drp1 at the Serine 637 position by protein kinase A leads to inactivation of Drp1 and a decrease in mitochondrial fission, whereas dephosphorylation of Drp1 at the same position by calcineurin results in activation of Drp1.54 CDK1 remains a key protein in Drp1 activation and mitochondrial fission. It has been shown that pretreatment of primary neuron cultures with various CDK1 inhibitors can protect against neuronal cell death.35 We found that the expression of CDK1 was increased in the hESC-derived neurons after exposure to propofol (fig. 6A). Finally, we found that pretreatment of the cells with the CDK1 blocker, roscovitine, could rescue the propofol-induced cell death, activation of Drp1, and increase in mitochondrial fission, further suggesting a functional role of CDK1 in propofol-induced neurotoxicity. Nevertheless, mitochondrial fission and Drp1 activation are extremely complex and as such the potential role of additional unstudied components cannot be excluded.
To understand the downstream components of the propofol-induced mitochondrial fission pathway, we evaluated changes in mPTP opening time. Several studies have implicated mitochondrial fission, opening of the mPTP, and loss of the mitochondrial membrane potential as a key cascade of events that leads to cell death after a stress.55,56 Mitochondrial fission has been shown to induce earlier opening of the mPTP, and mPTP opening induces cell death.22,56 Blockade of mPTP opening has been shown to attenuate ethanol-induced neurotoxicity. In vitro cultures of primary mouse neurons and astrocytes and 7-day-old mouse pups were exposed to multiple doses of ethanol. Ethanol exposure induced mPTP opening and cell death in the cultures and widespread neuronal cell death and behavioral abnormalities in vivo. The cell death and behavioral changes were attenuated by blockade of mPTP opening.29 This was of interest since ethanol and propofol both act on the γ-aminobutyric acid type A receptor in the brain. The role of mPTP opening and the link between mitochondrial fission and the mPTP have yet to be studied in anesthetic-induced neurotoxicity. We found that the mPTP opened earlier in cells treated with propofol or the mdivi-1 control compared with all control-treated groups. Opening of the mPTP was delayed in cells pretreated with mdivi-1, suggesting that increases in mitochondrial fission induced by propofol might lead to mPTP opening and, ultimately, cell death (fig. 7, A and B).
Although the culture of human neurons from stem cells is a useful and promising model for the study of neurotoxicity, the direct applicability of these data to developmental neurotoxicity in humans remains to be defined. This study alone cannot adequately address the role of mitochondrial abnormalities in propofol-induced neurotoxicity. Thus, future animal studies will be important to further assess the role of mitochondrial fission in propofol-induced neuronal death and cognitive deficits. In addition, the underlying mechanisms of anesthetic-induced developmental neurotoxicity are complex. Cell death may not be the only determinate of later neurocognitive deficits. Other mechanisms such as abnormal neurogenesis and synaptogenesis might also contribute to the cognitive dysfunction. In the current study, we observed disorganization of the actin cytoskeleton at the lower, more clinically relevant dose of 10 μg/ml propofol (fig. 1, C and D), suggesting that cellular abnormalities occur without cell death. The actin cytoskeleton can regulate many important cellular processes in the brain, including cell division, migration, cytokinesis, and differentiation.57,58 The potential contribution of disorganization of the actin cytoskeleton in propofol-induced developmental neurotoxicity will be explored in future studies.
One limitation of this study is that our in vitro model system contains only neurons. This is not truly representative of the human brain, which is composed of neurons and several types of glial cells. Interactions between neurons and glial cells may alter the signaling pathways we have identified in this study as well as the responsiveness of neurons to propofol. Nevertheless, it is critical to first understand the direct effects of propofol on human neurons, the mechanisms governing propofol-induced neuronal cell death and possible protective approaches before understanding how glial cells may alter these effects and pathways. Another limitation of this study is that the use of an in vitro culture system limits the data that can be obtained compared with the use of an in vivo animal model. Cultured neurons lack stimuli from the complex network of cells that is found in vivo and endpoints of interest such as cognitive functioning cannot be assessed. However, this human model allows us to carefully dissect out mechanisms of anesthetic-induced developmental neurotoxicity that can potentially be translated to an animal model to further develop protective strategies against anesthetic-induced developmental neurotoxicity.
An additional limitation of this study is the supratherapeutic dose of propofol used. Propofol is highly lipophilic and concentrates in lipid-rich tissues such as the brain.59 It is uncertain what concentration of propofol neurons in the developing human brain are exposed to during surgery. In addition, many animal and human epidemiologic studies have shown that anesthetic-induced neurotoxicity is dose, time, and exposure number dependent. This study focused on a single exposure to propofol with cell death used as the endpoint of interest. We have shown in the hESC-derived neurons that propofol induces cell death at lower doses after multiple exposures.32 However, it is difficult to study the mechanisms by which this occurs when multiple exposures are involved. The high dose of propofol used in this study may limit how these data connect to the clinic. However, the results of this study may still be translatable to the human brain in which the same mechanisms could induce toxicity at different doses due to the increased vulnerability in the whole brain. Future studies aimed at understanding the role of mitochondrial abnormalities and the neurotoxic effects of propofol at lower doses by coculturing hESC-derived neurons with glial cells and evaluating the effects of multiple exposures will be extremely important.
In conclusion, this study has established, for the first time, a role for mitochondrial fission and mPTP opening in propofol-induced toxicity in human neurons. These data suggest that propofol exposure in hESC-derived neurons leads to activation of Drp1 through CDK1, increases in mitochondrial fission, a decrease in the mitochondrial membrane potential, and mPTP opening, which together induce cell death. The detrimental effects of anesthetics on the developing brain have been well studied in animals. However, the mechanisms by which these anesthetics induce toxicity remain unclear. Our in vitro human model of hESC-derived neurons has allowed us to study the mechanism by which propofol induces cell death. Nevertheless, future studies confirming these results in whole-animal models will be important. Understanding the mechanisms by which propofol and other anesthetics induce toxicity in the developing brain is critical to develop potential protective strategies. The results of this study may one day contribute to the development of such approaches aimed at mitigating the neurotoxic effects of propofol in young children.
Acknowledgments
This study was supported by grants P01GM066730 and R01HL034708 from the National Institutes of Health (NIH), Bethesda, Maryland; FP00003109 from Advancing Healthier Wisconsin Research and Education Initiative Fund, Milwaukee, Wisconsin (to Dr. Bosnjak); and R01GM112696 from the NIH (to Dr. Bai).
Competing Interests
The authors declare no competing interests.