Propofol and Thiopental Suppress Amyloid Fibril Formation and GM1 Ganglioside Expression through the γ-Aminobutyric Acid A Receptor

SOME previous reports have suggested that surgery during general anesthesia may be associated with an increased risk of Alzheimer disease (AD).^1–3  However, a retrospective study conducted to evaluate the association between exposure to anesthesia and AD found no association between the risk of AD and exposure to anesthesia in 1 or 5 yr preceding disease onset or between the risk of AD and the number of surgical operations.⁴  Similarly, several other studies have concluded that it is unlikely that multiple exposures to general anesthesia increase the risk of AD.^1,2,5  Various other studies have suggested a potential association between anesthesia/surgery and AD; however, there exist other reports with differing opinions. Perioperative factors such as hypoxia,^6–8  hypocapnia,⁹  and anesthetics¹⁰  may contribute to AD neuropathogenesis. Thus, the relation between general anesthesia and AD has not yet been established.¹¹ 

One of the pathological hallmarks and possible causes of AD is aggregation and accumulation of extracellular amyloid β-peptides (Aβ) in the brain. Aβ, which is produced from amyloid precursor protein by sequential endoproteolytic cleavages by β- and γ-secretase, is originally monomeric, soluble, and nontoxic, but becomes cytotoxic on aggregation and accumulation.^12,13  Furthermore, it is suggested that early pathological changes associated with AD are caused by the formation of a complex of GM1 ganglioside (GM1) and Aβ, GAβ, in the brain.¹⁴  The unique molecular characteristics of GAβ allow it to act as a seed for Aβ fibrillogenesis in AD brains.¹⁴  Aβ assembly is initiated by GM1 accumulation and clustering at presynaptic neuritic terminals in AD brains.^15,16 

Understanding the relation between anesthetics and Aβ aggregation is important to completely elucidate AD pathogenesis after surgery with general anesthesia. Here we investigated the effects of propofol and thiopental on Aβ and GM1 expression in PC12 cells and cultured neurons. Moreover, the involvement of a γ-aminobutyric acid A receptor (GABAAR), one of the molecular targets of propofol and thiopental, was investigated.

Thiopental, (+)-bicuculline, and Dulbecco’s modified Eagle’s medium were purchased from Wako (Osaka, Japan). Propofol was obtained from MP Biomedicals LLC. (Illkirch, France). Cholesterol, sphingomyelin, cholera toxin B subunit conjugated to horseradish peroxidase, thioflavin T (ThT), and thioflavin S (ThS) were purchased from Sigma-Aldrich (St. Louis, MO). Nerve growth factor was obtained from Millipore (Billerica, MA). Synthetic Aβ (1–40) was obtained from the Peptide Institute (Osaka, Japan). Horse serum and N2 supplements were obtained from Invitrogen (Carlsbad, CA). Fetal bovine serum was obtained from Thermo Fisher Scientific (Waltham, MA). The lactate dehydrogenase (LDH) assay toxicity kit was purchased from Promega (Madison, WI).

PC12 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated horse serum and 5% fetal bovine serum. For differentiation, PC12 cells were plated onto poly-L-lysine-coated 90-mm dishes at a density of 4×10⁵ cells/dish and incubated in Dulbecco’s modified Eagle’s medium containing 10ng/ml nerve growth factor for five days. The resulting differentiated cells are referred to as PC12N cells. The cells were cultured in a humidified 5% CO₂ atmosphere at 37°C. Cerebral cortical neurons were prepared from Sprague–Dawley rats on embryonic day 17, as described previously.¹⁷  The dissociated single cells were plated onto poly-L-lysine-coated dishes and incubated in feeding medium for three or 21 days in vitro in a humidified 5% CO₂ atmosphere at 37°C. N2 medium, the feeding medium, consisted of Dulbecco’s modified Eagle’s medium/F-12 with 0.1% bovine serum albumin fraction V solution (Invitrogen) and N2 supplements. The cells were pharmacologically treated with propofol, thiopental, bicuculline, or a vehicle or were left as untreated controls.

The cells were lysed in Triton X-100-containing Tris buffer (5 mm Tris–HCl [pH 7.4], 2 mm EDTA, 1% Triton X-100, and Complete protease inhibitor cocktail [Roche Molecular Biochemicals, Penzberg, Germany]). The protein concentration of each sample was determined using the bicinchoninic acid protein assay kit (Thermo Fisher Scientific). The samples were separated on a 10–20% gradient polyacrylamide gel (Wako) and then electrotransferred to polyvinylidene difluoride membranes (Millipore). Blotted membranes were blocked with Block Ace (Yukijirushi, Sapporo, Japan) and then probed with cholera toxin B subunit conjugated to horseradish peroxidase. The bands were visualized by reaction with the SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific). Quantitative scanning of the blots was performed using National Institutes of Health Image version 1.59 (National Institutes of Health, Bethesda, MD).

The detergent-resistant membrane microdomains (DRMs) were isolated from neurons or synaptosomes, as described previously.¹⁸  In brief, neurons and synaptosomes were homogenized in 2-(N-morpholino)ethane sulfonic acid-buffered saline containing 1% Triton X-100 to obtain a lysate concentration of 1mg protein/ml and 200 µg protein/ml, respectively. The sucrose concentration of the extract was adjusted to 40% by adding 80% sucrose in 2-(N-morpholino)ethane sulfonic acid-buffered saline. The extract was then overlaid with a 5%/35% discontinuous sucrose gradient in 2-(N-morpholino)ethane sulfonic acid-buffered saline without Triton X-100 and centrifuged at 188,000g for 20h using an S120AT2 rotor (Hitachi, Tokyo, Japan). After centrifugation, 1-ml fractions were harvested from the top to the bottom of the gradient.

Male mice aged 6 months (SLC, Shizuoka, Japan) were housed in a room maintained at 24±1°C and illuminated daily for 12h (08:00–20:00). Free access was granted to food and water. All animal procedures were performed in accordance with a protocol approved by the Institutional Animal Care and Use Committee of Ritsumeikan University (Kyoto, Japan).

Synaptosomes were isolated from 6-month-old mice, as described previously.^18,19  In brief, a whole mouse brain was homogenized in 0.32 m sucrose buffer containing 0.25 mm EDTA (buffer A). The homogenate was centrifuged at 580g for 8min. The supernatant was centrifuged at 14,500g for 20min. The resulting pellet was suspended in 0.32 m sucrose buffer without EDTA (buffer B), overlaid with Ficoll in sucrose buffer, and centrifuged at 87,000g for 30min. The synaptosome-rich interface was removed and recentrifuged to remove any remaining Ficoll.

Seed-free solutions of Aβ (1−40) were prepared as described previously.²⁰  The cells treated with 50 µm seed-free Aβ were fixed for 10min in 10% formic acid and washed in phosphate-buffered saline. Thereafter, the cells were incubated for 10min in 0.25% potassium permanganate and then incubated in 2% potassium bisulphate and 1% oxalic acid until they became colorless. The cells were washed in water, treated for 10min with a solution of 0.015% ThS in 50% ethanol, and washed in 50% ethanol and water. Finally, the cells were mounted in Vectashield mounting medium (Vector Laboratories Inc., Burlingame, CA).

To prepare liposomes, cholesterol, sphingomyelin, and GM1 were dissolved in chloroform/methanol at a molar lipid ratio of 40:40:20. The mixtures were stored at −80°C until use. Immediately before use, the lipids were resuspended in Tris-buffered saline to yield a ganglioside concentration of 2.5 mm, and the suspensions were subjected to freezing, thawing, and sonication. Seed-free 50 µm Aβ solutions were incubated at 37°C, unless otherwise indicated, in the presence or absence of GM1 liposomes, as described previously.²¹  The ThT fluorescence intensity of the mixtures was determined using a spectrofluorophotometer (RF-5300PC; Shimadzu Co., Kyoto, Japan). The optimum fluorescence intensity of amyloid fibrils was measured at excitation and emission wavelengths of 446 and 490nm, respectively.

Aβ mixtures incubated in the presence or absence of GM1 liposomes with or without propofol and thiopental for 24h were mixed with N2 medium (1:1). These mixtures were then applied to cultured neurons (3 days in vitro) for 48h.

The LDH release assay of Aβ mixture-treated neurons was performed using the LDH assay toxicity kit. The degree of LDH release in each sample was determined by measuring the absorbance at 490nm using a microplate spectrophotometer (Benchmark Plus; Bio-Rad, Hercules, CA). The background absorbance, assessed using a cell-free well, was subtracted from the absorbance of each test sample. The absorbances measured from three wells were averaged, and the percentage of LDH released was calculated by dividing the absorbance measured from each test sample by the absorbance measured after treatment with 1% Triton X-100 to induce the release of intracellular LDH, according to the manufacturer’s instructions.

Data are shown as the mean ± SD of at least four or five independent experiments. Statistical analysis was performed by one- or two-way factorial ANOVA combined with Scheffe test for all paired comparisons (StatView software for Macintosh; SAS Institute, Cary, NC). P < 0.05 was considered to be statistically significant.

To investigate the effects of propofol and thiopental on GM1 levels, PC12N cells were incubated with 20 µm propofol or 100 µm thiopental for 48h. GM1 expression, which was determined by the specific binding of cholera toxin B subunit conjugated to horseradish peroxidase, was significantly decreased by propofol or thiopental treatment (fig.1A). GM1 expression was not affected by treatment with 0.03% and 0.1% dimethyl sulfoxide (DMSO) solutions as controls for 20 µm propofol and 100 µm thiopental, respectively (fig. 1A). To determine whether the effects of propofol and thiopental on GM1 expression were time dependent, PC12N cells were incubated with or without 20 µm propofol or 100 µm thiopental at the indicated times. GM1 expression was significantly decreased 12h after propofol or thiopental treatment, and the decreased expression was sustained for at least 48h. The effects of propofol and thiopental were similar (fig. 1B). To evaluate whether the effects of propofol and thiopental on GM1 expression were concentration dependent, PC12N cells were incubated with the indicated range of propofol and thiopental concentrations for 24h before harvesting. GM1 expression was significantly decreased in a concentration-dependent manner by both the anesthetics (fig. 1C). In addition, the expression levels of β-actin (the internal control) were similar in all the samples (fig. 1, A–C).

To investigate the mechanism of propofol and thiopental action, PC12N cells were pretreated with bicuculline, a GABAAR inhibitor, in the presence or absence of propofol or thiopental. Bicuculline, at the indicated concentrations, was added 1h before the addition of propofol (20 µm) or thiopental (100 µm). After a 24-h incubation, bicuculline inhibited propofol- or thiopental-induced decreases in GM1 expression in a dose-dependent manner (fig. 2). Next, we investigated these phenomena in cultured neurons. Propofol- or thiopental-induced decreases in GM1 expression were also inhibited by bicuculline in cultured neurons (fig. 3). These results suggest that propofol- or thiopental-induced decreases in GM1 expression require GABAAR.

To characterize the propofol- or thiopental-induced decrease in GM1 expression, DRM and non-DRM fractions were isolated from neurons that either had or had not been pretreated with bicuculline for 1h, and they were incubated for 24h in the presence or absence of propofol or thiopental. DRM fractions are thought to contain lipid rafts from cell membranes. The GM1 expression levels in each fraction were compared between the propofol- and thiopental-treated and nontreated neurons (fig. 4A). GM1 was mainly found in fraction 5, a DRM fraction, and fraction 11, a non-DRM fraction. GM1 expression in fraction 5 of the propofol- or thiopental-treated neurons was significantly decreased compared with that in fraction 5 of the nontreated neurons (arrows a and c in fig. 4, A1 and A2). The propofol- or thiopental-induced decrease in GM1 levels in fraction 5 was inhibited when cells were pretreated with bicuculline (arrows b and d in fig. 4, A1 and A2). GM1 expression in fraction 11 was similar in all samples (fig. 4, A3). In addition, the expression levels of flotillin-1, a DRM marker, and the transferrin receptor, a non-DRM marker, were similar in all the samples (data not shown). Moreover, we investigated whether GM1 expression was induced by the intraperitoneal administration of propofol (120mg/kg) and thiopental (50mg/kg) in the DRMs of synaptosomes that were prepared from mice brains. Twenty-four hours after the administration, GM1 levels were significantly lower in the DRM fraction (fraction 5) but not in the non-DRM fraction (fraction 11) that were isolated from the synaptosomes of propofol- or thiopental-treated mice compared with those isolated from the synaptosomes of control mice (arrows e and f in fig. 4B). The expression levels of flotillin-1 and the transferrin receptor were similar in all of the samples (data not shown).

We investigated whether Aβ assembly occurs in propofol- or thiopental-treated cultured neurons. Cultured cells were stained with ThS, which specifically binds to amyloid fibrils. ThS was visualized by fluorescence microscopy (fig. 5). Aβ assembly was observed in untreated neurons incubated with soluble Aβ, as reported previously.²²  The propofol or thiopental treatments markedly suppressed Aβ assembly. Furthermore, pretreatment with the GABAAR inhibitor bicuculline reduced the effects of propofol and thiopental.

We incubated soluble Aβ in the presence or absence of GM1 liposomes to investigate whether propofol and thiopental accelerate amyloid fibril formation in a cell-free system. The fluorescence intensity of ThT, which specifically recognizes amyloid structures, was measured in the incubation mixtures (fig. 6A). ThT fluorescence intensities of the solutions of Aβ alone did not increase during the 48-h incubation period, as reported previously (negative control of anesthetics).^15,21  Aβ assembly did not increase during the 48-h incubation period in the presence of DMSO (positive control of the anesthetics), propofol, or thiopental without GM1 liposomes. ThT fluorescence intensities in the incubation mixtures containing Aβ and GM1 liposomes without anesthetics were significantly increased after 48h, as reported previously (negative control of anesthetics).^15,21  Moreover, ThT fluorescence intensities in the incubation mixtures containing Aβ and GM1 liposomes without DMSO were significantly increased after 48h (positive control of anesthetics). Interestingly, both anesthetics significantly suppressed the GM1 liposome-induced Aβ assembly (fig. 6A). To examine the interfering effects of propofol or thiopental on ThT fluorescence intensity, we added propofol or thiopental at the indicated concentrations after incubating the solutions with or without Aβ and GM1 liposomes and measured ThT fluorescence intensities in the incubation mixtures. Neither of the anesthetics affected the increase in ThT fluorescence intensity in the incubation mixtures containing Aβ and GM1 liposomes (fig. 6B).

Previously, we have reported that GM1 liposomes induce the formation of toxic Aβ assemblies from soluble Aβ.¹⁵  We investigated whether propofol and thiopental inhibited the formation of toxic Aβ assemblies in the presence of GM1 liposomes. We incubated Aβ solutions in the presence or absence of GM1 liposomes with or without DMSO (positive control of anesthetics), propofol, or thiopental for 24h and then applied them to primary cultured neurons. To evaluate the extent of cell death, LDH release assays were performed 48h after the administration of the mixed solution (fig. 7). Significant neuronal death was observed after treatment with the mixture of soluble Aβ and GM1 liposomes, as reported previously.^15,23  Propofol and thiopental significantly suppressed the neurotoxicity induced by the mixture of Aβ and GM1 liposomes. Furthermore, the incubation mixtures of Aβ in the presence or absence (negative control of anesthetics) of DMSO, propofol, or thiopental without GM1 liposomes did not induce neurotoxicity.

In this study, we showed that propofol and thiopental might have a protective effect against Aβ aggregation, a hallmark of AD. First, we found that these anesthetics suppressed GM1 expression in PC12N cells and cultured neurons. This suppression was localized to DRM fractions and occurred through GABAAR. Second, the anesthetics caused suppression of Aβ assembly formation from soluble Aβ added to the medium of neuronal cell cultures. In addition, they caused suppression of GM1 liposome-induced Aβ assembly in a cell-free system. Finally, we showed that propofol and thiopental decreased the neurotoxicity of a mixture containing Aβ and GM1 liposomes. Collectively, these findings suggest that propofol and thiopental directly or indirectly inhibit Aβ fibrillogenesis.

GM1 is a physiological constituent of cellular membranes and plays important roles in the differentiation, functioning, and viability of cells, particularly neurons. We have previously reported that the biosynthesis and cell surface expression of GM1 are strictly dependent on the type of cell and its differentiation status²²  and that GM1 expression in DRMs of PC12N cells is regulated by insulin.²⁴  Previous studies reported that Aβ deposition starts at the presynaptic neuritic terminals in the AD brain^25,26  and that GM1 levels significantly increase in DRMs²⁷  and amyloid-positive synaptosomes isolated from AD brains.²⁸  Recently, we suggested that the age-dependent high-density GM1 clustering in DRMs at presynaptic neuritic terminals is a critical step for Aβ deposition in AD. However, the regulation of biosynthesis and cell surface expression of GM1 in neurons remains unclear. In the current study, we found that GM1 expression in DRMs of neurons and PC12N cells was inhibited by propofol and thiopental, acting through GABAAR. In addition, propofol and thiopental inhibited Aβ assembly from neuronal membranes by reducing GM1 expression (fig. 5). Thus, our results suggest that propofol and thiopental may indirectly inhibit the formation of Aβ fibrils from the cell surface of neurons by reducing GM1 expression in DRMs through GABAAR.

Several studies have suggested that propofol and thiopental, which are GABA-mimetic anesthetic agents, have a neuroprotective action.^29,30  In addition, several studies provide evidence supporting GABAergic dysfunction in AD.³¹  Studies of early-stage amyloid pathology using transgenic models have suggested that it progresses in a neurotransmitter-specific manner, in which cholinergic terminals appear most vulnerable, followed by GABAergic terminals.³²  A number of recent studies have associated GABAAR signaling with significant neuroprotection against Aβ-mediated toxicity.³³  However, the effects of propofol and thiopental on Aβ fibrillogenesis and neurotoxicity through a change in GM1 expression have not been previously reported.

The mechanism underlying propofol and thiopental inhibition against the formation of GM1 liposome-induced Aβ assembly has not been investigated. The Aβ assembly experiments were performed in a cell-free system. In addition, we used seed-free Aβ solutions and liposomes. This means that propofol and thiopental must directly inhibit the formation of GM1 liposome-induced Aβ assembly through molecular interactions (fig. 6). Further investigation is required.

A previous study reported that Aβ oligomerization was enhanced by high propofol concentration.^10,34  However, GM1 liposomes were not used, and the propofol concentration used was 100 µm, which is fivefold greater than that in our study. In the same report, the authors did not show a significant change in cytotoxicity in the presence of propofol in cultured pheochromocytoma cells. In our study, PC12N cells and primary cultured neurons were used for cell culture experiments. PC12N cells are nerve growth factor-treated PC12 cells and are used as a model system for neuronal differentiation. For Aβ oligomerization studies, we used GM1 liposomes. Our experimental conditions were designed to closely match those of the endogenous neural environment. Consequently, we were able to show a direct protective effect of propofol and thiopental on neurons (fig.7).

In conclusion, the current results suggest that propofol and thiopental are relatively safe with respect to the formation of Aβ assembly because they have direct and indirect inhibitory effects on Aβ fibrillogenesis. Our results may encourage further in vivo attempts to determine the effects of anesthetic agents on AD neuropathogenesis, ultimately leading to safer anesthesia care for patients, especially elderly patients who are particularly at risk of postoperative cognitive dysfunction and AD.