The Inhalation Anesthetic Desflurane Induces Caspase Activation and Increases Amyloid β-Protein Levels under Hypoxic Conditions*

Perioperative factors including hypoxia, hypocapnia, and certain anesthetics have been suggested to contribute to Alzheimer disease (AD) neuropathogenesis. Desflurane is one of the most commonly used inhalation anesthetics. However, the effects of desflurane on AD neuropathogenesis have not been previously determined. Here, we set out to assess the effects of desflurane and hypoxia on caspase activation, amyloid precursor protein (APP) processing, and amyloid β-protein (Aβ) generation in H4 human neuroglioma cells (H4 naïve cells) as well as those overexpressing APP (H4-APP cells). Neither 12% desflurane nor hypoxia (18% O2) alone affected caspase-3 activation, APP processing, and Aβ generation. However, treatment with a combination of 12% desflurane and hypoxia (18% O2) (desflurane/hypoxia) for 6 h induced caspase-3 activation, altered APP processing, and increased Aβ generation in H4-APP cells. Desflurane/hypoxia also increased levels of β-site APP-cleaving enzyme in H4-APP cells. In addition, desflurane/hypoxia-induced Aβ generation could be reduced by the broad caspase inhibitor benzyloxycarbonyl-VAD. Finally, the Aβ aggregation inhibitor clioquinol and γ-secretase inhibitor L-685,458 attenuated caspase-3 activation induced by desflurane/hypoxia. In summary, desflurane can induce Aβ production and caspase activation, but only in the presence of hypoxia. Pending in vivo confirmation, these data may have profound implications for anesthesia care in elderly patients, and especially those with AD.

Isoflurane, sevoflurane, and desflurane are the most commonly used inhalation anesthetics. It has been reported that isoflurane enhances the oligomerization and cytotoxicity of A␤ (44) and induces apoptosis (48 -51). Our recent studies have shown that a clinically relevant concentration of isoflurane can lead to caspase-3 activation, decrease cell viability, alter APP processing, and increase A␤ generation in human H4 neuroglioma cells overexpressing human APP (45)(46)(47). Loop et al. (49) reported that isoflurane and sevoflurane, but not desflurane, can induce caspase activation and apoptosis in human T lymphocytes. However, effects of desflurane and desflurane plus other perioperative risk factors, e.g. hypoxia, on APP processing and A␤ generation have not been assessed.
In the present study, we set out to determine effects of desflurane, hypoxia, and the combination of the two (desflurane/hypoxia) on caspase-3 activation, APP processing, and A␤ generation in H4 human neuroglioma cells (H4 naïve cells) and H4 naïve cells stably transfected to express full-length (FL) APP (H4-APP cells). We also investigated whether the caspase inhibitor, Z-VAD, the ␥-secretase inhibitor L-685,458, and the A␤ aggregation inhibitor clioquinol could attenuate desflurane/hypoxiainduced caspase-3 activation and A␤ generation.

EXPERIMENTAL PROCEDURES
Cell Lines-We employed H4 human neuroglioma cells (H4 naïve cells) and H4 naïve cells stably transfected to express full-length (FL) APP (H4-APP cells) in the experiments. All of the cell lines were cultured in Dulbecco's modified Eagle's medium (high glucose) containing 9% heat-inactivated fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. Stably transfected H4 cells were additionally supplemented with 200 g/ml G418.
Western Blots Analysis-The cells were harvested at the end of the experiments and were subjected to Western blots analyses as described by Xie et al. (52). Antibodies A8717 (1:1,000; Sigma) and C66 (1:1,000, generous gift of Dr. Dora Kovacs at Massachusetts General Hospital and Harvard Medical School) were used to visualize FL-APP (110 kDa) and APP-CTFs (10 -12 kDa). Antibody anti-␤-actin (1:2,000, Sigma) was used to detect ␤-actin (42 kDa). A caspase-3 antibody (1:1,000 dilution; Cell Signaling Technology, Inc., Beverly, MA) was used to recognize FL-caspase-3 (35-40 kDa) and caspase-3 fragment (17-20 kDa) resulting from cleavage at asparate position 175. Rabbit polyclonal anti-BACE-1 antibody (1:1,000; Abcam, Cambridge, MA) was used to detect protein levels of BACE (65 kDa). The quantification of Western blots was performed as described by Xie et al. (52). Briefly, intensity of signals was analyzed by using the National Institutes of Health image program (NIH Image 1.62). We quantified Western blots using two steps. First, we used levels of ␤-actin to normalize (e.g. determining ratio of FL-APP amount to ␤-actin amount) levels of FL-APP, APP-CTFs, FL-caspase-3, caspase-3 fragment, and BACE to control for any loading differences in total protein amounts. Second, we presented changes in levels of FL-APP, APP-CTFs, FL-caspase-3, caspase-3 fragment, and BACE in treated cells as a percentage of those in cells treated with controls.
Quantification of A␤ Using Sandwich Enzyme-linked Immunosorbent Assay-Secreted A␤ was measured with a Sandwich enzyme-linked immunosorbent assay by using an A␤ measurement kit (Invitrogen), and by the A␤ enzyme-linked immunosorbent assay core facility at the Center for Neurological Dis- eases, Harvard Institute of Medicine, Harvard Medical School, Boston, Massachusetts, as described by Xie et al. (53). Specifically, 96-well plates were coated with mouse monoclonal antibodies specific to A␤40 (2G3) or A␤42 (21F12). Following blocking with Block Ace, wells were incubated overnight at 4°C with test samples of conditioned cell culture media, and then an anti-A␤ (␣-A␤-HR1) conjugated to horseradish peroxidase was added. The plates were then developed with TMB reagent, and well absorbance was measured at 450 nm. A␤ levels in test samples were determined by comparison with signal from unconditioned media spiked with known quantities of A␤40 and A␤42. Statistics-Given the presence of background caspase-3 activation and cell death in cells cultured in serum-free medium, we did not use absolute values to describe changes in caspase-3 activation. Instead, 100% caspase-3 activation, FL-APP, APP-CTFs, and BACE in the manuscript refer to control levels for the purpose of comparison to experimental conditions. The data were expressed as the means Ϯ S.D. The number of samples varied from 3 to 10, and the samples were normally distributed. We used a two-tailed t test to compare differences between experimental groups and control groups. p values less than 0.05 (* or #) and 0.01 (** or ##) were considered statistically significant.

Desflurane Does Not Cause Caspase-3 Activation, APP Processing, or Secreted A␤ Levels in H4-APP Cells-We previously
reported that isoflurane can induce apoptosis and increase secreted A␤ levels in H4-APP cells (45)(46)(47). The effects of other commonly used inhalation anesthetics, including sevoflurane and desflurane, on cellular apoptosis, APP processing, and A␤ generation have not been reported. We therefore set out to determine these effects in H4-APP cells. Because caspase-3 activation is one of the final steps of cellular apoptosis (54), we assessed effects of desflurane on caspase-3 activation by quantitative Western blot analyses. In the present experiment, treatment with 12% desflurane for 6 h only induced modest caspase-3 cleavage (activation) (Fig. 1A), as assessed by determining the ratio of cleaved (activated) caspase-3 fragment (17-19 kDa) to FL-caspase-3 (35-40 kDa). Quantification of the Western blots, based on the ratio of caspase-3 fragment to FL-caspase-3, revealed that the desflurane treatment did not significantly induce caspase-3 activation as compared with control conditions: 100% versus 127% (p ϭ 0.14) (Fig. 1B). APP immunoblotting revealed no significant differences in levels of FL-APP and APP-CTFs between desflurane-treated and control H4-APP cells (Fig. 1, C-E). Finally, the desflurane treatment did not increase secreted levels of A␤40 or A␤42 as compared with control conditions (Fig. 1F). These results indicate that desflurane alone neither induces apoptosis nor increases A␤ generation.
Desflurane/Hypoxia Induces Caspase-3 Activation, Affects APP Processing, and Increases Secreted A␤ Levels in H4-APP Cells-Because hypoxia has been reported to potentiate AD neuropathogenesis (43,55), we assessed the effects of desflurane/hypoxia on caspase-3 activation, APP processing, and secreted A␤ levels in H4-APP cells. We chose 18% O 2 to assess the effects of desflurane on apoptosis and A␤ generation under mildly hypoxic conditions. Caspase-3 immunoblotting showed visible increases in protein levels of caspase-3 fragment following treatment with 12% desflurane/hypoxia (18%) for 6 h as compared with control conditions (Fig. 2A). Quantification of the results by determining ratio of cleaved (activated) caspase-3 fragment (17-19 kDa) to FL-caspase-3 (35-40 kDa) revealed that desflurane/hypoxia treatment led to a 306% increase in caspase-3 cleavage (activation) as compared with control conditions (p ϭ 0.002). APP immunoblotting showed visible  decreases in protein levels of APP-CTFs (10 -12 kDa), but not FL-APP (110 kDa), following the desflurane/hypoxia treatment as compared with control conditions (Fig. 2C). Quantification of the results showed that the desflurane/hypoxia treatment led to a 49% decrease in levels of APP-CTFs as compared with control conditions (Fig. 2E, p ϭ 0.019). Finally, the desflurane/ hypoxia treatment increased secreted A␤40: 5.3 pg/ml versus 15.9 pg/ml, p ϭ 0.002 and A␤42: 1.9 pg/ml versus 6.7 pg/ml, p ϭ 0.018 (Fig. 2F). These findings suggest that the combination of desflurane and hypoxia induces apoptosis, alters APP processing, and increases A␤ generation.
Caspase Inhibitor Z-VAD Attenuates the Desflurane/Hypoxia-induced Caspase-3 Activation-Given the findings that desflurane/ hypoxia can induce apoptosis and increase A␤ generation, we next asked whether the desflurane/hypoxia-induced alteration in APP processing and A␤ generation is dependent on caspase-3 activation. For this purpose, we set out to assess the effects of the caspase inhibitor Z-VAD on desflurane/hypoxia-induced caspase-3 activation and A␤ generation. Z-VAD treatment attenuated caspase-3 activation induced by desflurane/hypoxia treatment, whereas Z-VAD alone did not induce caspase-3 activation as compared with control conditions (Fig. 3A). Quantification of the Western blots revealed that Z-VAD attenuated desflurane/hypoxia-induced caspase-3 activation: 304% versus 94%, p ϭ 0.009 (Fig. 3B). Z-VAD also reduced desflurane/hypoxia-induced increases in secreted A␤ levels in H4-APP cells, p ϭ 0.00038 (Fig. 3C).
Next, we asked whether desflurane/hypoxia can induce caspase-3 activation in the absence of APP overexpression in naïve H4 cells. Caspase-3 immunoblotting showed that desflurane/hypoxia induced caspase-3 activation as compared with control conditions in both H4 naïve and H4-APP cells (Fig. 4A). Quantification of the Western blot revealed that desflurane/hypoxia treatment led to a 306% and a 195% increase in caspase-3 activation as compared with control conditions in H4 naïve cells (Fig. 4B, p ϭ 0.025) and H4-APP cells (Fig. 4C, p ϭ 0.007), respectively. APP immunoblotting did not show detectable changes in levels of APP-CTFs and FL-APP between desflurane/hypoxia treatment and control conditions in H4 naïve cells (Fig. 4D). As a positive control, APP immunoblotting showed that desflurane/hypoxia (lane 4) reduced levels of APP-CTFs without significant changes in FL-APP levels as compared with control conditions (lane 3) in H4-APP cells (Fig.  4D). Quantification of the Western blots revealed that desflurane/hypoxia treatment did not change the ratio of APP-CTFs to FL-APP in H4 naïve cells (Fig. 4E) but led to a 24% decrease in the ratio of APP-CTFs to FL-APP in H4-APP cells (Fig. 4F, p ϭ  0.03). A␤ levels were too low to be detected in H4 naïve cells in  these experiments (data not shown). Taken together, these findings suggest that whereas desflurane/hypoxia can induce caspase-3 activation in the absence of alterations in APP processing and A␤ generation in H4 naïve cells, desflurane/hypoxiainduced increases in A␤ generation are dependent on the desflurane/hypoxia-induced caspase-3 activation in H4-APP cells.
Desflurane/Hypoxia Increases Levels of BACE in H4-APP Cells-We next asked whether desflurane/hypoxia can increase levels of BACE. Treatment with 12% desflurane/hypoxia (18% O 2 ) for 6 h caused visible increases in BACE levels as compared with control conditions (Fig. 5A). Quantification of the Western blot showed that desflurane/hypoxia treatment led to a 172% increase in BACE levels as compared with control conditions (p ϭ 0.006, Fig. 5B). These results suggest that desflurane/ hypoxia can activate caspase-3 and increase A␤ generation by enhancing BACE levels.
Hypoxia (18% O 2 ) Alone Does Not Affect Caspase-3 Activation, APP Processing, or A␤ Levels in H4-APP Cells-Finally, we assessed whether hypoxia (18%) alone can also induce caspase-3 activation, alter APP processing, and increase A␤ generation in H4-APP cells. Treatment with hypoxia (18% O 2 ) for 6 h did not induce caspase-3 activation (Fig. 7, A and B), did not alter APP processing (Fig. 7, C-E), and did not increase A␤ generation (Fig. 7F). Taken together, these findings suggest that the apoptosis, APP processing and A␤ generation induced by desflurane/hypoxia are most likely due to synergistic effects of desflurane/hypoxia.

DISCUSSION
The commonly used inhalation anesthetic isoflurane has previously been shown to promote A␤ aggregation and to enhance toxicity of A␤ (44). We have shown that isoflurane can induce cellular apoptosis and increase A␤ generation in H4-APP cells (45)(46)(47). Because desflurane is another commonly used inhalation anesthetic, we set out to assess the effects of desflurane on apoptosis, APP processing, and A␤ generation in H4-APP cells. We were able to show that a 6-h treatment with a clinically relevant concentration of desflurane (12%) did not induce caspase-3 activation and affected neither APP processing nor secreted A␤ levels in H4-APP cells. These results are consistent with findings of other studies, which illustrated that isoflurane and sevoflurane, but not desflurane, can induce apoptosis (49). However, it is still possible that different treatments of desflurane in other cell lines may lead to apoptosis and enhancement of A␤ generation.
Next, we found that treatment with 12% desflurane/hypoxia (18% O 2 ) for 6 h can induce caspase-3 activation, alter APP processing, and increase A␤ generation in H4-APP cells, whereas treatment with either 12% desflurane or hypoxia alone did not lead to similar effects in H4-APP cells. Collectively, these findings suggest that desflurane may promote AD neuropathogenesis only under hypoxic conditions. It is possible that treatment with 12% desflurane alone for 6 h had undetectable effects on apoptosis and APP/A␤, which were potentiated by hypoxic conditions. Hypoxia and cerebral blood supply insufficiency remain prevalent concerns in perioperative care for patients, because they can result from hypotension, shunting (e.g. one lung ventilation), and other pathological conditions. Therefore, these data, once confirmed in vivo, suggest that it

. Desflurane/hypoxia increases levels of BACE in H4-APP cells.
A, desflurane/hypoxia (lanes 4 and 5) increases levels of BACE (65 kDa) as compared with control conditions (lanes 1-3). There is no significant difference in amounts of ␤-actin in control conditions or desflurane/hypoxiatreated H4-APP cells. B, quantification of the Western blot shows that desflurane/hypoxia (black bar) increases BACE levels as compared with control conditions (white bar) (**, p ϭ 0.006), normalized to ␤-actin levels. would be prudent to vigilantly avoid these pathological conditions in patients when desflurane is used as an inhalation anesthetic.
We employed 18% O 2 to assess the effects of desflurane on apoptosis and A␤ generation under mildly hypoxic conditions. Future studies will be necessary to assess whether severe hypoxic conditions, e.g. 1% O 2 , alone would have similar effects on apoptosis and A␤ generation.
Z-VAD, a broad caspase activation inhibitor, was able to attenuate both caspase-3 activation and A␤ generation induced by desflurane/hypoxia. These results suggest that the desflurane/hypoxia-induced alterations in APP processing and A␤ generation are largely dependent on the ability of desflurane/hypoxia to induce apoptosis. To further explore the mechanism by which desflurane/hypoxia induces apoptosis, affects APP processing, and increases A␤ generation, we assessed the effects of desflurane/ hypoxia on BACE. Desflurane/hypoxia treatment for 6 h increased protein levels of BACE in H4-APP cells. Previously, we showed that inhalation anesthetic isoflurane can induce apoptosis which in turn enhances BACE levels to facilitate APP processing and to increase A␤ generation (47). Recent studies have shown that subsequent to caspase activation during ischemia, reductions in protein levels of Golgi-localized ␥-ear-containing ARF-binding protein (GGA)-3, a protein that can alter trafficking and metabolism of BACE, is associated with increased levels of BACE and activity of ␤-secretase (56). It is interesting to speculate whether isoflurane and desflurane/hypoxia treatments might likewise reduce levels of GGA-3 to enhance BACE levels and ␤-secretase activity subsequent to caspase activation.
We have previously shown that an amyloid fibril-binding dye, Congo Red (32), a ␤-sheet breaker peptide iA␤5 (57), and a metal protein attenuation compound CQ (58) can attenuate, whereas A␤ can potentiate, apoptosis induced by isoflurane (46,47). We therefore assessed whether inhibition of A␤ aggregation and generation can attenuate desflurane/hypoxia-induced caspase-3 activation in H4-APP cells by determining the effects of CQ (an A␤ aggregation inhibitor) and L-685,458 (a ␥-secretase inhibitor) on caspase-3 activation induced by desflurane/hypoxia in H4-APP cells. L-685,458 inhibited both caspase-3 activation and A␤ generation induced by desflurane/ hypoxia. CQ also inhibited caspase-activation induced by desflurane/hypoxia. Collectively, these results suggest that reductions in A␤ generation and aggregation can attenuate desflurane/hypoxia-induced apoptosis. Conversely, treatment with A␤ potentiated desflurane/hypoxia-induced caspase-3 activation, suggesting a vicious cycle of apoptosis and A␤ generation that can be triggered by desflurane/hypoxia. In conclusion, we have found that the combination of desflurane and hypoxia, but neither alone, induces apoptosis, alters APP processing, and increases A␤ generation. Fig. 8 shows a scheme in which desflurane/hypoxia activates caspase, which then leads to increases in BACE and A␤ levels. Increased A␤ levels would then lead to further caspase activation, resulting in a vicious cycle. Further investigation, especially via in vivo studies, will be necessary to assess the potential role of desflurane and/or hypoxia in triggering or driving AD neuropathogenesis. These efforts should promote further attempts to determine the effects of anesthetics on AD neuropathogenesis, ultimately leading to provision of safer anesthesia care to patients, especially elderly patients, who are particularly susceptible to the incidence of post-operative cognitive dysfunction and risk for AD.