The Common Inhalational Anesthetic Isoflurane Induces Apoptosis via  Activation of Inositol 1,4,5-Trisphosphate Receptors

Cells of wild-type (WT) chicken B lymphocyte (DT40) cell line, and its total IP³receptor knock-out (DT40 IP³R TKO) type, were cultured in RPMI 1640 with 10% fetal calf serum, 1% chicken serum, 50 μm 2-mercaptoethanol, 4 mm l-glutamine, and antibiotics in a 95% air, 5% CO²humidified atmosphere at 38°C as previously described.26Rat pheochromocytoma (PC12) cells transfected with WT presenilin 1 (PS1) or point mutated PS1 (L286V) were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% heat-inactivated horse serum, 5% fetal calf serum, 200 μg/ml G418, and antibiotics in a 95% air, 5% CO²humidified atmosphere at 37°C as described.1,27The transfection of the WT and mutant PS1 has been described and confirmed in detail previously.28,29Mouse HD knock-in striatal cells (STHdh  ^Q111/Q111) and their WT control cells (STHdh  ^Q7/Q7) were generated and cultured as described previously.30,31Briefly, cells were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal calf serum, 400 μg/ml G418, and antibiotics in a 95% air, 5% CO²humidified atmosphere at 33°C.

DT40 cells (WT and IP³R TKO), rat PC12 cells (WT and L286V), and mouse striatal cells (WT and HD) were exposed to isoflurane at different concentrations for various durations in a gastight chamber inside the culture incubator (Bellco Glass, Inc., Vineland, NJ), with humidified 5% CO²–21% O²–balanced N²(AirGas East, Bellmawr, NJ) going through a calibrated agent-specific vaporizer as described previously.1Gas phase concentrations in the gas chamber were verified and maintained at the desired concentration throughout the experiments using an infrared Ohmeda 5330 agent monitor (Coast to Coast Medical, Fall River, MA). In a pilot study, the cell medium was aspirated and extracted into hexane for high-performance liquid chromatography measurement (System Gold; Beckman Coulter, Fullerton, CA) to verify that the various anesthetic concentrations in the medium in millimolars are equivalent to the minimal alveolar concentration (MAC) in the gas phase inside the gas chamber using the concentration correlation previously described.32 

Translocation of membrane phospholipid phosphatidylserine from the inner to the outer leaflet of the plasma membrane is an early indication of cell damage. Annexin V, a phospholipid binding protein with a high affinity for phospholipid phosphatidylserine, can bind to phospholipid phosphatidylserine once it is exposed to the extracellular environment. Propidium iodide (PI) can bind to nucleic acid after penetrating a breached plasma membrane, as occurs in the later stages of cell damage. We treated DT40 cells, grown floating in the medium, with different concentrations of isoflurane (0.6, 1.2, and 2.4%) for 24 h, as well as with 2.4% isoflurane for different times (6, 12, and 24 h). Immediately after treatment, we determined annexin V– or PI–positive cells by the methods described previously.26Cells were dropped onto 25-mm cover slips and stained with annexin V or PI. The stained cells were visualized and counted by two persons blinded to the treatments. The percentage of annexin V– or PI–positive cells averaged from four areas on each cover slip was then calculated and compared.

Increased caspase-3 activity is a typical marker for apoptosis. The assay is based on the ability of the active enzymes to cleave the fluorogenic substrates Ac-DEVD-AFC (caspase 3; Calbiochem, San Diego, CA) and was performed as per instructions and as described previously.26DT40 cells grown on six well plates were treated with 2.4% isoflurane for 24 h and then were harvested via  trypsinization and washed with phosphate-buffered saline. The cell pellet was gently resuspended in CelLytic M lysis buffer and protease inhibitor cocktail (Sigma, St. Louis, MO), lysed, and centrifuged; the supernatant was used for the assay. Caspase substrates were added to a final concentration of 50 μm, and the samples were incubated at 37°C for 45 min in caspase assay buffer. Incubated samples were measured at an excitation of 400 nm and an emission of 505 nm in a multiwavelength-excitation dual wavelength-emission fluorometer (Delta RAM; photon Technology International, Birmingham, NJ). We determined the caspase-3 activity immediately after treating DT40 cells with 2.4% isoflurane for 24 h.

The inhibition of isoflurane-induced cytotoxicity by xestospongin C, a potent antagonist of IP³receptor, was assessed by the lactate dehydrogenase (LDH) release assay in PC12 cells and MTS (3-(4,5-dimethylthiazol-2-yl)- 5-(3carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) reduction assay in mouse striatal cells. LDH release reflects plasma membrane integrity, is a relatively late event in cytotoxicity, and is shared by both apoptosis and necrosis pathways, and its measurement is well described.1The MTS reduction assay reflects mitochondrial function and is considered a middle event in both apoptosis and necrosis. The tetrazolium compound is bioreduced by normal mitochondria into a colored formazan product measured by absorbance. We followed the standard experimental protocol for MTS reduction assay from Promega (Madison, WI). We treated PC12 and striatal neurons, grown on 24 well plates, with 2.4% isoflurane for 24 h. Immediately (PC12 cells) or 24 h (striatal neurons) after treatment, LDH or MTS assay was performed, respectively. Xestospongin C at 100 nm was used to inhibit IP³receptor activity. The results of LDH release and MTS reduction assays were expressed as a percentage of the control without anesthetic treatment.

The method used was the same as described previously.26DT40 cells (WT and IP³R TKO) grown on 25-mm cover slips were loaded with 2 μm rhod-2/AM in cell medium containing 2.0% bovine serum albumin in the presence of 0.003% Pluronic acid at 37°C for 50 min. Cells loaded with rhod-2 dye were washed and then reloaded with fluo-4/AM for an additional 30 min at room temperature. Cells were placed on a stage and exposed to 2 MAC (0.7 mm) isoflurane dissolved in the perfusion buffer. The images were recorded using the Radiance 200 imaging system with excitation at 488 and 568 nm for fluo-4 and rhod-2, respectively.

Cytosolic calcium concentration was measured using fura-2 fluorescence (Molecular Probes, Eugene, OR) with a photometer coupled to an Olympus 1Χ70 inverted microscope (Olympus America Inc., Center Valley, PA) and IPLab version 3.7 imaging processing and analysis software (Biovision Technologies, Exton, PA). The protocol to determine [Ca²⁺]^cwas similar to that previously described, with some modifications.33Briefly, cells grown on 25-mm round glass cover slips were washed three times with Krebs-Ringer’s buffer without addition of calcium and then loaded with 2.5 μm fura-2/AM (Molecular Probes) for 30 min at room temperature. The cells were then placed in a sealed chamber (Warner Instrument Inc., Hamden, CT) connected with multiple inflow infusion tubes and one outflow tube, which provided constant flow to the chamber. The cells were first washed with Krebs-Ringer’s buffer through one inflow tube for the baseline measurement of [Ca²⁺]^c, and then were exposed to isoflurane via  a separate inflow infusion tubes driven by a syringe pump (Braintree Scientific Inc., Braintree, MA). The fluorescence signals were measured with excitation at 340 and 380 alternatively and emission at 510 nm for a period up to 18 min for each treatment. A pilot study confirmed that the cells were still viable at the end of experiments for calcium measurement. The fluorescence measurements were calibrated by bathing cells in the HEPES buffer containing ionomycin 20 mm for maximum or 20 mm EGTA for minimum calcium values. The intracellular calcium concentration ([Ca²⁺]ⁱ) was calculated by the ratio method of Grynkiewicz et al. ,34using 224 nm as the Kd of fura-2. The final result of [Ca²⁺]^cwas averaged from the cells of at least three separate experiments. We used 0.7 mm (2 MAC) isoflurane to elevate [Ca²⁺]^cin all cells. Xestospongin C at 1 μm was used to inhibit isoflurane-mediated calcium release from the ER.

We used GraphPad Prism 4 software (GraphPad Software, Inc., San Diego, CA) and STATA version 7 Software (StataCorp LP, College Station, TX) for all statistical analysis. Annexin V and PI staining, LDH release, and MTS reduction were all expressed as percentage of those in the control. Peak [Ca²⁺]^cin PC12 and HD striatum cells was expressed as a percentage of its own baseline. We analyzed the data with one-way analysis of variance followed by Newman-Keuls multiple comparison tests using GraphPad Prism. We also confirm our analysis of dose response and time response of image analysis of cell damage marker annexin V and PI in DT40 cells with a mixed mode of regression using STATA version 7. A significantly increased odds ratio risk was determined if the estimate for the greater time point or dose exceeded the upper limit of the 95% confidence interval for the lower time point or dose. P < 0.05 was considered statistically significant.

To investigate whether isoflurane induces apoptosis in the absence of IP³receptors, we first studied its cytotoxic effects in the DT40 IP³R TKO cells and compared it with the corresponding WT cells. Isoflurane induced cell damage determined by both annexin V and PI staining concentration-dependently in WT but not in IP³R TKO cells (figs. 1A–C). The lowest concentration of isoflurane required to induce apoptosis in WT cells was 1.2% (figs. 1B and C), a clinically used concentration. Isoflurane also induced cell damage time-dependently in WT but not TKO cells (figs. 1D–F). In WT cells, 2.4% isoflurane induced early cell damage in as little as 6 h of treatment (fig. 1E), a duration often used for surgeries. To confirm that isoflurane induced cell damage in DT40 cells by apoptosis, we measured changes of caspase-3 activity after exposing these cells to 2.4% isoflurane for 24 h. Isoflurane caused a dramatic elevation (more than fourfold) of caspase-3 activity in DT40 WT but not IP³R TKO cells (fig. 2A).

To further examine the hypothesis that isoflurane induced apoptosis by disruption of intracellular calcium homeostasis, we first assayed the cytotoxic effects of thapsigargin, a selective inhibitor of the ER calcium adenosine triphosphatase, which can cause ER calcium to passively leak even without any contribution from IP³receptors.35Thus, in this positive control experiment, thapsigargin (100 nm for 2 h) induced similar apoptosis in both WT and TKO cells (fig. 2B), whereas 2.4% isoflurane for 24 h only induced apoptosis in DT40 WT cells. Isoflurane induced a sequential elevation of [Ca²⁺]^cand then mitochondria calcium concentration ([Ca²⁺]^m) only in DT40 WT but not TKO cells (figs. 2C and D), whereas thapsigargin produced this sequential calcium elevation in both WT and IP³R TKO cells (fig. 2E). These results show that the IP³receptor knock-out did not diminish the cells response to stress or calcium transients in general, and thereby strengthen the hypothesis that isoflurane induced apoptosis by causing calcium release from the ER via  overactivation of IP³receptors.

The mutation of PS1, a protein located primarily on ER membranes, appears in most cases of familial Alzheimer disease, and is associated with increased expression of ryanodine receptors27and increased activity of IP³receptors.36,37If isoflurane induces apoptosis by overactivation of IP³receptor, cells with enhanced expression or activity of this receptor should be more vulnerable to isoflurane-induced cytotoxicity and exhibit greater calcium release from the ER. Consistent with this hypothesis, isoflurane caused significantly more PI-positive cells in PC12 cells transfected with L286V PS1 than in its WT vector control (figs. 3A and B). To establish the linkage of these cytotoxic effects of isoflurane to IP³receptors, we coincubated cells with xestospongin C, a potent inhibitor of the IP³receptor.38,39Xestospongin C abolished the cell damage represented by elevated LDH (fig. 3C) and also nearly abolished the calcium release from the ER caused by isoflurane (figs. 3D and E) in L286V PC12 cells. Although xestospongin C may also inhibit calcium adenosine triphosphatase in some cell cultures,40it did not affect adenosine triphosphatase activity in this system because xestospongin C alone did not increase [Ca²⁺]^cin L286V PC12 cells (data not shown).

We also examined isoflurane effects in another model of elevated IP³receptor activity. Abnormal calcium release from ER via  elevated activity of IP³receptors is thought to play an important role in the neurodegeneration of HD.23,41,42Mutated Huntingtin protein with an enlarged polyglutamine repeat section causes excessive calcium release from the ER upon the activation of IP³receptors by an agonist.23Therefore, if isoflurane causes apoptosis via  activation of IP³receptors, the elevated activity of IP³receptors in mouse striatal cells with an overexpression of Q111 Huntingtin should not only render these cells more vulnerable to isoflurane-induced neurotoxicity and calcium release from the ER, but also xestospongin C should inhibit these effects. In accordance with this hypothesis, 2.4% isoflurane for 24 h induced more cell damage (PI-positive cells) in Q111 Huntingtin knock-in than in the corresponding WT control striatal cells (figs. 4A and B), and xestospongin C (100 nm) abolished this cytotoxic effect (fig. 4C). In the absence of extracellular calcium, 2.4% isoflurane (0.7 mm) induced significantly greater elevation of [Ca²⁺]^c(representing calcium release from the ER) in Q111 Huntingtin knock-in than the WT control striatal cells, an effect that was also nearly abolished by pretreatment of xestospongin C (figs. 4D and E).

Our results are most consistent with the hypothesis that isoflurane, a commonly used inhalational anesthetic, induces cell apoptosis by excessive calcium release from the ER via  overactivation of IP³receptors. First, cells depleted of IP³receptors were resistant to isoflurane-induced apoptosis and calcium elevation, whereas cells with enhanced activity of the IP³receptor were more vulnerable to isoflurane-induced apoptosis. Finally, the IP³receptor antagonist xestospongin C significantly inhibited isoflurane-induced cell damage and calcium elevations in the vulnerable cell models.

Our results suggest a direct interaction between isoflurane and the receptor protein, but cannot prove it. Alternative possibilities are that isoflurane interacts with an associated protein or signaling system, or the ER lipid membrane itself. But it has been demonstrated that anesthetics similar to isoflurane can bind specifically to many membrane proteins,43so it is plausible that the effect on IP³receptors may be the result of a direct interaction.

Consistent with our previous study and others,1,44isoflurane induced apoptosis in DT40 WT cells concentration- and time-dependently. It is important to note that there exists considerable variation in cell sensitivity to isoflurane-induced apoptosis. For example, the minimal time to induce apoptosis in DT40 cells with 2.4% isoflurane was only 6 h, with a minimal concentration of 1.2%. In rat cerebral cortical neurons or PC12 cells, the minimal concentration and time needed to induce apoptosis was 2.4% isoflurane for 24 h.1Immortalized striatal neurons also needed a minimal exposure of 2.4% isoflurane for 24 h to induce modest cell damage in this study. In normal human peripheral lymphocytes, only 0.85% isoflurane was needed to induce apoptosis,44and it is likely that normal human neurons are similarly sensitive. Although this study suggests that the IP³receptor is an important target underlying isoflurane-induced apoptosis, this variability in sensitivity suggests that there may be other targets that contribute, or that downstream effects are considerably different. In addition, the shortest times for 2.4% isoflurane to induce apoptosis in chicken lymphocytes are 6 h (fig. 1E) and 12 h (fig. 1F), respectively, when techniques to detect early-phase (annexin V) apoptosis or late-phase (PI staining, primarily necrosis) cell damage were used. This demonstrates that the detection of vulnerability to isoflurane cytotoxicity will vary depending on the assays used to measure cell damage. Isoflurane may produce only modestly enhanced LDH release in PS1 mutated PC12 cells (fig. 3C) because the LDH method detects severe late cell damage, whereas the cell injury induced by these concentrations of isoflurane must be more subtle, given the lack of overt clinical sequelae. Further studies are needed to determine the features that underlie this differing vulnerability to isoflurane-induced toxicity.

Increased calcium release from the ER via  the IP³receptors may contribute to neurodegeneration in other conditions, such as Alzheimer disease and HD.36,37,45Various mutations of PS1, a protein located primarily in ER membranes, appear in most cases of familial Alzheimer disease. This PS1 mutation has been associated with increased activity of the IP³receptors,36,37which may render neurons more vulnerable to apoptosis induced by a variety of factors that trigger calcium release (e.g. , excitotoxicity). Our results are consistent with this hypothesis in that PC12 with L286V PS1 were more vulnerable to isoflurane cytotoxicity than its WT control, in a xestospongin C–dependent manner. It should be noted that many other PS1 mutations, such as PS1 delta 9, M146V, or just enhanced expression of full-length PS1, have been shown to contribute to cell apoptosis.46–48If these other mutations enhance IP³receptor activity, our data would predict enhanced vulnerability to isoflurane cytotoxicity. Isoflurane may contribute to pathogenesis via  other mechanisms as well. For example, isoflurane has been recently reported to increase β-amyloid production, aggregation, and neurotoxicity2,6,8and may be associated with tau hyperphosphorylation.49Thus, there is a growing concern regarding the use of isoflurane in patients with symptoms of, or vulnerabilities to, neurodegenerative disease. Although the connection between polyglutamine-containing proteins and IP³receptors is not yet clear, the observations with respect to isoflurane are analogous.

It is not yet clear whether all anesthetics activate IP³receptors. Sevoflurane, a relatively new inhalational anesthetic, did not induce similar neuronal apoptosis as isoflurane at equipotent concentrations,1so there is hope that other anesthetic drugs might prove to be less neurotoxic, at least in in vitro  studies. Clinical studies are ultimately required to demonstrate whether these cell culture studies have any relevance to humans.

It should be noted that isoflurane has been long considered an agent for cardioprotection and neuroprotection.50–52It is likely that isoflurane is both neurotoxic and neuroprotective, depending on the concentration and duration of exposure, and the degree of patient vulnerability. Isoflurane may be inherently cytotoxic, but it provides cardioprotection or neuroprotection via  a preconditioning mechanism, much like hypoxia. Mild calcium release from the ER and moderate elevation of [Ca²⁺]^cby isoflurane at low concentrations and short duration may trigger the ER stress response, marked by the expression of genes characterizing the well-known “preconditioning” effect.53,54Longer exposures to isoflurane, producing extensive and prolonged calcium release from ER, may deplete ER calcium and shut down protein synthesis, leading to “cytotoxicity” effects.1,55,56Therefore, like ischemic preconditioning, isoflurane for short durations might provide cytoprotection via  preconditioning, whereas prolonged exposures produce cytotoxicity directly. This hypothesis was supported by our recent experiments, which demonstrated that 2.4% isoflurane preconditioning for 1 h abolished neurotoxicity induced by isoflurane itself for 24 h in cerebral cortical neurons.57These potential dual features of isoflurane should be considered in future studies, and perhaps in clinical application in vulnerable populations.

This study has several limitations that should influence in vivo  interpretations: (1) The cells with total IP³R knock-out are chicken B lymphocytes, not neurons. Although the ER calcium handling mechanisms are thought to be the same, downstream effects of the calcium transients might be considerably different. To our knowledge, there are no immortal neuronal cell lines with IP³R knock-out, but transgenic mice with type I IP³R58or type II and III59knock-out have been reported, from which cultures of primary neurons could be studied. Expression of IP³receptors could also be reduced via  small interfering RNA in normal neurons. (2) All of the results from this study are from cell lines, which are immortal transformed cells. Although clearly different from normal cells, in general, such cells are more resistant to stressors, making the results of our study somewhat more relevant to the in vivo  situation. Future in vivo  studies are needed to investigate whether IP³R also plays a role in isoflurane-mediated neurotoxicity in animals or patients.

Taken together, our findings suggest that the commonly used inhalational anesthetic isoflurane may induce cell apoptosis by triggering abnormal calcium release from the ER via  activation of IP³receptors. Preexisting genetic features may render some populations at increased risk from this effect.

The authors thank Mark P. Mattson, Ph.D. (Chief), and Sic L. Chan, Ph.D. (Senior Research Fellow, Laboratory of Neurosciences, National Institute on Aging, Baltimore, Maryland), for providing PC12 cells transfected with Presenilin-1 mutation; Marcy MacDonald, Ph.D. (Associate Professor, Molecular Neurogenetics Unit, Center for Human Genetic Research, Harvard University, Boston, Massachusetts), for providing striatal cells with knocked-in mutated Huntingtin protein; and Tomohiro Kurosaki, M.D., Ph.D. (Professor, Institute of Physical and Chemical Research, Research Center for Allergy and Immunology, Turumi-ku, Yokohama, Kanagawa, Japan), for providing total IP³receptor knock-out chicken lymphocytes. The authors appreciate the technical support from Qingcheng Meng, Ph.D. (Senior Research Scientist), statistical analysis by Adam Burkey, M.D. (Assistant Professor), and the editing contributions of Christopher Ward, M.D. (Anesthesia Resident, Department of Anesthesiology and Critical Care, University of Pennsylvania, Philadelphia, Pennsylvania). The authors also appreciate the discussions with Randall Pittman, Ph.D. (Professor at the Department of Pharmacology, University of Pennsylvania, Philadelphia, Pennsylvania).