Anesthesia and Tau Pathology

Abstract

Alzheimer’s disease (AD) is the most common form of dementia and remains a growing worldwide health problem. As life expectancy continues to increase, the number of AD patients presenting for surgery and anesthesia will steadily rise. The etiology of sporadic AD is thought to be multifactorial, with environmental, biological and genetic factors interacting together to influence AD pathogenesis. Recent reports suggest that general anesthetics may be such a factor and may contribute to the development and exacerbation of this neurodegenerative disorder. Intra-neuronal neurofibrillary tangles (NFT), composed of hyperphosphorylated and aggregated tau protein are one of the main neuropathological hallmarks of AD. Tau pathology is important in AD as it correlates very well with cognitive dysfunction. Lately, several studies have begun to elucidate the mechanisms by which anesthetic exposure might affect the phosphorylation, aggregation and function of this microtubule-associated protein. Here, we specifically review the literature detailing the impact of anesthetic administration on aberrant tau hyperphosphorylation as well as the subsequent development of neurofibrillary pathology and degeneration.

Introduction

With a worldwide prevalence of 26.6 million cases as of 2006, Alzheimer’s disease (AD) remains the most common form of dementia affecting the elderly population, and it is estimated that without a significant therapeutic breakthrough, the prevalence of this disorder will quadruple by the middle of this century (Brookmeyer et al., 2007). Advancing age is considered the major risk factor for the development of sporadic AD, which constitutes the majority of AD cases (~99%) (Harman, 2002; Bufill et al., 2009). AD pathology is considered to be multi-factorial, whereby a genetic predisposition interacts with environmental or biological factors, resulting in the development and/or acceleration of this neurodegenerative condition (Iqbal and Grundke-Iqbal, 2008; Papon et al., 2011).

The two histopathological hallmarks of AD are extracellular amyloid plaques, which are composed mainly of beta-amyloid (Aβ) peptide, and intra-neuronal neurofibrillary tangles (NFT), which are primarily composed of aberrantly hyperphosphorylated tau protein assembled into paired helical filaments (Grundke-Iqbal et al., 1986). Tau is normally enriched in the axonal compartment, but in AD and other tauopathies, it becomes hyperphosphorylated then subsequently relocalizes and aggregates in the somatodendritic compartment of affected neurons (Buee et al., 2000; Avila et al., 2004).

Recent studies have demonstrated that anesthetics can have an impact on the neuropathogenesis of AD, and even possibly accelerate the clinical progression of this neurodegenerative disorder (Eckenhoff et al., 2004; Xie et al., 2006; Planel et al., 2009; Run et al., 2009; Tan et al., 2010; Tang et al., 2011a; Le Freche et al., 2012). Furthermore, both pre-clinical and clinical studies have recently supported the notion that anesthesia has a significant impact on tau pathogenesis (Planel et al., 2007; Planel et al., 2008; Planel et al., 2009; Run et al., 2009; Run et al., 2010; Tang et al., 2011a; Tang et al., 2011b; Whittington et al., 2011; Le Freche et al., 2012). In this article, we specifically review the literature on reported changes in tau pathology and function following anesthetic exposure in vivo, and provide tables summarizing the anesthetics used in these articles (Table 1), as well as the results obtained from different groups (Table 2).

Table 1.

List of anesthetics known to induce tau hyperphosphorylation in vivo, either directly or indirectly.

ANESTHETICS	NAME	FAMILY	ADMINISTRATION	PRIMARY MECHANISM	USAGE
Isoflurane	1-chloro-2,2,2-trifluoroethyl difluoromethyl ether	Halogenated ether	Volatile	Not well understood. Binds to GABA, glutamate and glycine receptors	Maintenance of general anesthesia
Sevoflurane	1,1,1,3,3,3-hexafluoro-2- (fluoromethoxy)propane	Halogenated ether	Volatile	Not well understood. Binds to GABA, glutamates and glycine receptors	Induction and maintenance of general anesthesia
Halothane	2-bromo-2-chloro-1,1,1- trifluorethane)	Halogenated ether	Volatile	Not well understood. Binds to GABA, glutamates and glycine receptors	Induction and maintenance of general anesthesia
Propofol	2,6-diisopropylphenol	Alkyl phenol	Intravenous	Potentiation of GABAA receptor activity	Induction and maintenance of general anesthesia Sedation for mechanical ventilation Procedural sedation
Pentobarbital	5-Ethyl-5-(1-methylbutyl)- 2,4,6(1H,3H,5H)-pyrimidinetrione	Barbiturate	Intravenous	Positive allosteric modulation of GABAA receptors, binding site is located in the Beta subunit. Also, targets glutamate, adenosine, and nicotinic acetylcholine receptors	Anesthetic in veterinary medicine
Dexmedetomidine	(S)-4-[1-(2,3-dimethylphenyl)ethyl]-3H-imidazole	α2-adrenergic receptor agonist	Intravenous	Agonist of α2-adrenergic receptors	Sedation without causing respiratory depression
Chloral hydrate	2,2,2-trichloroethane-1,1-diol	NA	Intravenous, oral	Enhances the GABA receptor complex.	Sedative, hypnotic drug

Table 2.

List of studies demonstrating in vivo tau hyperphosphorylation after anesthesia.

REFERENCE	ANESTHETICS USED	DOSES USED	MODEL USED	EFFECT OF HYPOTHERMIA	EFFECT OF NORMOTHERMIA	MICROTUBULE BINDING	COGNITIVE IMPAIRMENT	MECHANISMS	NOTES
Planel (2007)	Isoflurane Pentobarbital Chloral hydrate	Exposure to vapors 100mg/kg i.p. 500mg/kg i.p	C57BL/6 male mice	AT8, PHF-1, pS199, TG3, MC6, pS262, pS422 (+++)	Rescued	NA	NA	Inhibition of PP2A activity under hypothermic conditions	80% increase in tau phosphorylation for every degree Celsius drop below 37°C.
Planel (2008)	Pentobarbital	100 mg/kg i.p 100 mg/kg i.p.	C57BL/6J female and male mice hTau female and male mice	(MT)-free Tau (+++) (MT)-free Tau (+++)	NA NA	Impaired Impaired	NA	NA	No MT detachment or relocalization in non-transgenic mice Relocalization of 3R tau in the soma
Planel (2009)	Isoflurane	1.3% (1.0 MAC, 30% O2) for 4h	JNLP3 female mice	AT8, CP13, MC-6, PHF-1 (+++) Increase in insoluble tau levels	Rescued	NA	NA	Anesthesia-induced hypothermia inhibited PP2AC	Same results with repeated anesthesia in JNLP3 male mice
Run (2009)	Pentobarbital	50mg/kg i.p for 1h	C57BL/6J female mice	Thr181, Thr205, Thr212, 12E8 (+++)	Thr181, Ser199, Thr205, Thr212, Ser262, Ser404 (++)	NA	NA	Anesthesia-induced hypothermia leading to PP2AC inhibition Normothermic effect due to activation of MAPK and JNK	.
Tan (2010)	Isoflurane	1.5% (1.0 MAC, 30% O2) for 2h	Sprague-Dawley rats	Thr205, Ser396 (+++)	Rescued	NA	NA	Inhibition of PP2A under hypothermic conditions, rescued by normothermic conditions.
Whittington (2011)	Propofol	250 mg/kg i.p. (0.5h, 2h & 6h) 3μg/ml	C57BL/6 male mice Tau-SH-SY5Y cells	AT8, CP13, PHF-1 (+++) NA	AT8, CP13, PHF-1 (+) AT8, CP13, PHF-1 (+)	NA	NA	Normothermic effect likely due to PP2A inhibition.	Effect persisted up to 2h at AT8
Tang (2011)	Halothane or Isoflurane	0.9–1.1% (1.0 MAC, 40% O2), repeated exposure	3xTg-AD mice	NA	Thr231 (+++)	NA	NA	NA
Menuet (2012)	isoflurane	1.5% (1.0 MAC, 30% O2) for 4h	Tau P301L mice	AT8 (+++)	NA	NA	NA	NA	Upper airway dysfunction under hypothermic conditions, rescue effect by normothermia.
Liu (2012)	isoflurane	1.4% (1.0 MAC, 30% O2) for 2h	Sprague-Dawley rats	Thr231, Ser396 (+++),	Rescued	NA	YES	NA	Cognitive impairment not rescued by normorthermic conditions
Le Freche (2012)	Sevoflurane	1.5 or 2.5% (repeated exposure)	C57BL/6 female mice	NA	PHF-1, Thr181 (+++)	NA	YES	Normothermic effect likely due to activation of AKT and ERK kinases and inactivation of GSK-3.
Dong (2012)	isoflurane	1.4% for 2h	AD transgenic mice [B6.Cg-Tg (APPswe, PSEN1dE9) 85Dbo/J]	NA	PS262 (+++)	NA	NA	Normothermic effect may result from the isoflurane-induced caspase activation and Aβ generation	Effect persisted up to 24h after exposure
Whittington (personal communication)	Dexmedetomidine	300μg/kg i.p. (0.5h, 2h & 6h)	C57BL/6 male mice Tau-SH-SY5Y cells	NA	AT8, PHF-1, CP13 (+++)	NA	NA	NA	Effect persisted 2h later for all three epitopes and up to 6h at AT8

Anesthesia-Induced Tau Hyperphosphorylation: The Role of Hypothermia

Intracellular aggregates of abnormally hyperphosphorylated tau are present in a group of neurodegenerative diseases called tauopathies (Buee et al., 2000; Avila et al., 2004). Tau hyperphosphorylation can induce aggregation in vitro (Alonso et al., 2001; Sato et al., 2002), and is thought to induce NFT formation in the brain (Grundke-Iqbal et al., 1986; Trojanowski and Lee, 1994; Iqbal and Grundke-Iqbal, 2008). The distribution pattern of NFT in the brain of patients is highly hierarchical and has been divided into 6 histological stages (Braak and Braak, 1991, 1997; Braak et al., 2006) as well as 10 biochemical stages (Delacourte et al., 1999). Tau pathology correlates with dementia in AD as well as memory loss in normal aging and mild cognitive impairment (Wilcock and Esiri, 1982; Arriagada et al., 1992; Guillozet et al., 2003; Bretteville and Planel, 2008). In Alzheimer’s disease, the accumulation of NFT in neurons is preceded by the appearance of hyperphosphorylated tau.

This disruption of tau phosphorylation homeostasis can result from the dysregulation of the activities of both tau-related kinases and phosphatases. Environmental factors can also be critical in the development of altered signal transduction, which then disrupts the balance of tau phosphorylation homeostasis, ultimately leading to the development of neurofibrillary degeneration and neuronal cell death observed in AD (Iqbal and Grundke-Iqbal, 2005; Iqbal and Grundke-Iqbal, 2008).

We previously demonstrated in the brains of 4 to 6 month-old non-transgenic C57BL/6J mice that both intravenous (chloral hydrate and sodium pentobarbital) and inhalational anesthetics (isoflurane) rapidly induce pronounced tau hyperphosphorylation at several phosphoepitopes including AT8 (Ser²⁰²/Thr²⁰⁵⁾, PHF-1 (Ser³⁹⁶/Ser⁴⁰⁴), pS199 (Ser¹⁹⁹), TG3 (Thr²³¹), MC6 (Thr²³⁵), pS262 (Ser262), and pS422 (Ser⁴²²). This robust tau hyperphosphorylation was surprisingly reversed by the restoration of core body temperature to normal, thus demonstrating that anesthesia-induced hypothermia is a major mediator of tau hyperphosphorylation (Planel et al., 2007). The effect of hypothermia is very potent, with an 80% increase in tau phosphorylation for each degree Celsius below 37°C (Planel et al., 2004).

Tau phosphorylation is regulated by several protein kinases such as glycogen synthase kinase-3β (GSK-3β), mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK), calcium/calmodulin-dependent protein kinase II (CaMKII), c-Jun N-terminal kinase (JNK), cyclin-dependent kinase 5 (cdk5) and its specific activator p35, as well as AKT/PKB (protein kinase B) (Cross et al., 1995; Maccioni et al., 2001; Planel et al., 2002; Zhu et al., 2002; Tatebayashi et al., 2006; Sadik et al., 2009). Tau phosphorylation homeostasis is maintained through dephosphorylation mediated by protein phosphatase 2A (PP2A), protein phosphatase 2B (PP2B), and protein phosphatase 1 (PP1) (Goedert et al., 1992; Wang et al., 1996). PP2A is the major tau phosphatase in the brain, accounting for more than 70% of tau dephosphorylation (Liu et al., 2005). We determined that the hypothermia-induced tau hyperphosphorylation was not a consequence of tau kinase activation, but rather secondary to the direct inhibition of PP2A activity by the hypothermia itself (Planel et al., 2004; Planel et al., 2007).

Other groups have subsequently confirmed this observation using various anesthetics in different in vivo models. Indeed, Tan et al. demonstrated in 6-month old Sprague-Dawley rats that 1.5% isoflurane anesthesia for 2h without temperature control leads to a 10- and 2.9-fold increase in hippocampal tau phosphorylation at the Thr²⁰⁵ and Ser³⁹⁶ phosphoepitopes, respectively (Tan et al., 2010). No increase in tau phosphorylation was observed in rats following isoflurane administration when normothermic conditions were maintained. These investigators also observed the inhibition of hippocampal PP2A activity following 2h of isoflurane administration under hypothermic conditions, thus confirming our previous findings. Run et al. also showed a higher degree of tau phosphorylation at the Thr¹⁸¹, Thr²⁰⁵, Thr²¹², and 12E8 (Ser²⁶²/Ser³⁵⁶⁾ phosphoepitopes in 14–15 week-old C57BL/6J mice anesthetized for 1 hour with pentobarbital compared to mice sacrificed within 5 min of anesthetic administration. This observation correlated with the onset of hypothermia, which developed progressively during anesthesia (Run et al., 2009) and finally resulted in a drop in body temperature to a range of 26–30°C 1h after anesthesia.

Altogether, these data strongly suggest that anesthesia induced tau hyperphosphorylation is mediated through anesthesia-induced hypothermia and consequent inhibition of PP2A activity (Planel et al., 2004). They also suggest that hypothermia, which is common in the elderly, might enhance AD pathology (Whittington et al., 2010).

Anesthesia-Induced Hypothermia and Tau function

In the normal brain, tau is mainly found in the axonal compartment of neurons (Binder et al., 1985). In AD and other tauopathies, tau relocalizes to the somatodendritic compartment where classical NFT are located (Papasozomenos and Binder, 1987; Delacourte et al., 1990; Riederer and Binder, 1994). The appearance of hyperphosphorylated tau aggregates correlates with a loss of microtubules and the breakdown of normal axonal transport (Terry, 1996; Pollak et al., 2003). Thus, tau hyperphosphorylation has been proposed to impair its microtubule (MT) binding function, resulting in its detachment from MTs, which, in turn, destabilizes the MTs, disrupts MT-dependent axonal transport, and ultimately leads to the relocalization of tau to the somatodendritic compartment (Trojanowski and Lee, 1994; Mandelkow et al., 2003; Duff and Planel, 2005; Feinstein and Wilson, 2005; Mi and Johnson, 2006). However, it should be noted that it is still controversial whether tau hyperphosphorylation is the cause or consequence of MT collapse (Planel et al., 2008; Miyasaka et al., 2010).

The microtubule binding function of tau is regulated by two distinct mechanisms. The first level of regulation is assured by the establishment of a precise equilibrium between tau isoforms with 3 or 4 repeat binding domains: the inclusion of the alternatively spliced exon 10 leads to the presence of a fourth repeat domain and confers to the 4-repeat tau isoforms (4R) a higher capacity to bind and polymerize microtubules than tau isoforms containing only 3-repeat domains (3R) (Goedert and Jakes, 1990; Butner and Kirschner, 1991; Panda et al., 2003). The second level of regulation is by the homeostasis of tau phosphorylation levels established by a balance between tau kinases and phosphatases. Tau phosphorylation in and around the MT binding domain acts as a negative regulator of tau microtubule binding. Thus, dephosphorylated tau is more efficient to bind and polymerize microtubules than the phosphorylated form (Lindwall and Cole, 1984).

Recent work has begun to focus on the pathological consequences of aberrant tau hyperphosphorylation secondary to anesthesia-induced hypothermia. In non-transgenic mice, we observed that pentobarbital-induced hypothermia resulted in the robust hyperphosphorylation of microtubule (MT)-free tau, and the subsequent impairment of the ability of tau to bind and assemble MTs (Planel et al., 2008). In contrast to free tau, MT-bound tau was more resistant to hyperphosphorylation following pentobarbital-induced hypothermia, and MT detachment was not observed in non-transgenic mice. This lack of MT detachment is most likely secondary to the fact that murine adult tau solely consists of the 4-repeat (4R) tau isoforms (Janke et al., 1999), which are more tightly bound to microtubules than 3R tau and thus more resistant to detachment from MT. To verify this hypothesis, we used the htau transgenic mouse model (Andorfer et al., 2003) bearing a human tau minigene (Duff et al., 2000), which expresses both 3R and 4R human tau isoforms. This model also develops neurofibrillary pathology characterized by tau hyperphosphorylation, aggregation, somatodendritic relocalization, and cell death (Andorfer et al., 2005). Hence, in contrast to non-transgenic mice, a 4h exposure to pentobarbital-induced hypothermia resulted in detachment of 3R (but not 4R) tau isoforms associated with an accentuated relocalization of phosphorylated tau in the soma (Planel et al., 2008). However, this detachment of 3R tau did not result in the depolymerization of tubulin or dramatic reorganization of the axonal cytoskeleton (Planel et al., 2008), suggesting that, despite the pronounced hyperphosphorylation induced by anesthesia, tau isoforms residually bound to the MTs are sufficient for the maintenance of the MT network integrity.

Altogether, these data suggest that tau hyperphosphorylation induced by anesthesia affects negatively tau function by altering its capacity to properly bind and polymerize microtubules. However, anesthesia-induced tau hyperphosphorylation is not sufficient to detach enough tau from tubulin to result in the actual breakdown of the microtubule cytoskeleton (Planel et al., 2008).

Anesthesia-Induced Hypothermia and the Development of Tau Pathology

Based on our previous results showing that anesthesia-induced hypothermia could negatively impact tau’s physiological function, we addressed the question of whether anesthetics could accelerate tau pathology in a mouse model of tauopathy (Planel et al., 2009). We took advantage of the JNPL3 mice, an aggressive mouse model of tauopathy overexpressing the pro-aggregative Tau_P301L mutation and consequently developing neurofibrillary degeneration (Lewis et al., 2000). We specifically observed that a single exposure to 1 minimum alveolar concentration (MAC) isoflurane (1.3%) for 4h produced significant hypothermia, which immediately resulted in pronounced tau hyperphosphorylation in the brainstem of 4 month-old female mice. This hyperphosphorylation was accompanied by an increase in insoluble tau levels 1 week after exposure. It should be noted that the increase in tau aggregation, observed 1 week following anesthesia, occurred at a time when the mice were no longer hypothermic. Repeated anesthesia (1 MAC isoflurane for 4h twice a week for 2 weeks) led to the same observations in 8 month-old JNPL3 males, which are known to develop tau pathology at a slower rate than females. Of note, when the male JNPL3 mice were exposed to a similar dose of isoflurane with temperature control so as to avoid hypothermia, no increases in phosphorylated or insoluble tau were observed 1 week following exposure.

Other groups have further supported these findings. Indeed, using another transgenic mouse model overexpressing the same Tau_P301L mutation, following exposure to hypothermia induced by isoflurane, Menuet et al. have reported increased AT8 staining and tauopathy in several brainstem regions, including the nucleus ambiguus and Kolliker-Fuse nuclei, at 7 days post-exposure (Menuet et al., 2012). This mouse model of tauopathy has been previously shown to develop cognitive deficits as well as vocalization and respiratory abnormalities, the latter condition ultimately leading to premature death (Terwel et al., 2005; Dutschmann et al., 2010; Menuet et al., 2011b; Menuet et al., 2011a; Menuet et al., 2012). Moreover’, these transgenic mice develop evidence of tauopathy within the pontine Kolliker-Fuse nucleus (Dutschmann et al., 2010). Interestingly, these investigators observed in 4 month old Tau P301L mice that a single exposure to 4h 1MAC isoflurane, under hypothermic conditions, resulted in the premature development of upper airway defects, which are normally observed in this mouse tauopathy model at later stages of life. Indeed, measurements of upper airway function revealed that the administration of isoflurane in a hypothermic environment resulted in the development of respiratory compromise at 4 months, which was comparable to the respiratory pathology typically observed in these mice at 8 months (Menuet et al., 2012). However, the administration of a similar dose of isoflurane to these Tau P301L mice with temperature control prevented the development of the ventilation abnormalities. Forced ventilation also prevented the exacerbation of upper airway pathology in these mice following isoflurane-induced hypothermia, suggesting that other physiological factors occurring during hypothermia (besides lower temperature) may play a role in mediating the development of airway pathology following isoflurane. Interestingly, the pretreatment with the NMDA receptor antagonist, memantine (20 mg/kg i.p., 1h prior to anesthesia), also prevented the development of upper airway pathology following the administration of isoflurane without temperature control, suggesting that the NMDA receptor pathway is involved in the development of airway defects following hypothermic isoflurane exposure. Moreover, the degree of tauopathy correlated well with the early onset of upper airway dysfunction and pretreatment with memantine reduced the development of the tauopathy observed 7 days following hypothermic isoflurane administration. However, the mechanism by which this NMDA-R antagonist is mediating its beneficial effects, in this mouse tauopathy model, remains unexplained.

Overall, these data demonstrate that the capacity of isoflurane to exacerbate neurofibrillary pathology in Tau_P301L mouse models is dependent on physiological changes occurring during anesthesia-induced hypothermia, and not directly on this volatile anesthetic per se.

Anesthesia-Induced Hypothermia and cognitive deficits

It has been well established that hypothermia leads to memory disruption in numerous organisms including rodents (Riccio et al., 1968; Hamm, 1981; Richardson et al., 1983; Santucci et al., 1987). However, the mechanisms by which hypothermia produces memory disruption are not well understood. Some studies suggest that a reduction of protein synthesis resulting from a general decrease in energy-dependent cellular processes plays a role, whereas others, including our own studies, suggest a disruption of cellular signaling events positioned upstream of transcription or translation (Fulton et al., 2008; Whittington et al., 2013).

However, few studies have addressed the cognitive impact of changes in tau pathology following anesthesia-induced hypothermia. Recently, Tan et al., observed that the increase in phosphorylated tau following the administration of 1.5% isoflurane anesthesia for 2h without temperature control in 6 month old Sprague Dawley rats was associated with the impairment of working and reference memory as measured using the Y-maze test (Tan et al., 2010). However, it is very important to note that a direct, causal relationship between the changes in tau phosphorylation observed following anesthesia-induced hypothermia and memory impairment was not established in this study. In 18-month old Sprague Dawley rats, 1.4% isoflurane for 2h with temperature control resulted in the impairment of memory and learning as measured with the Morris Water Maze (Liu et al., 2012). However, these cognitive changes were not associated with any changes in total tau or tau phosphorylated at the Thr²³¹ and Ser³⁹⁶ phosphoepitopes. Hence, the restoration of normal body temperature can prevent the increase in tau phosphorylation observed following isoflurane anesthesia; however, this does not necessarily eliminate the cognitive impairment observed in aged rodents following exposure to this anesthetic (Culley et al., 2003; Liu et al., 2012). Of note, thus far, the studies examining the relationship between tau pathology and cognitive function following isoflurane-induced hypothermia have been somewhat limited by the fact that, in these studies, only a few phosphoepitopes have been analyzed for aberrant phosphorylation, they solely have examined non-transgenic animals without preexisting AD pathology, and they have been designed such that no causal relationship between pathological tau burden and the degree of cognitive decline has ever been definitively established.

In summary, anesthesia-induced hypothermia results in pronounced tau hyperphosphorylation, and the acceleration of tau pathology in transgenic mouse tauopathy models. With isoflurane, the maintenance of normothermia appears to mitigate the negative effects of this anesthetic on tau phosphorylation; however, cognitive impairment still has been reported under these conditions. Studies with memantine in a mouse tauopathy model suggest that the NMDA-R pathway may play a role in the development of tau pathology observed following isoflurane-induced hypothermia; however, other receptor pathways cannot be excluded at the present time. Furthermore, the mechanism by which memantine antagonizes the negative effects of isoflurane-induced hypothermia on tau pathology remains unclear and warrants further investigation. Although the impairment of memory observed following isoflurane-induced hypothermia has been associated with increases in tau phosphorylation in rodents, a true causal link between tau hyperphosphorylation and impaired cognition has not yet been established. Hence, future studies are required to address this potential relationship.

Anesthesia-Induced Tau Pathology and Normothermia

Although there is ample evidence suggesting that anesthetic-induced hypothermia possesses the capacity to produce aberrant tau hyperphosphorylation and even accelerate neurofibrillary pathology, recent studies suggest that certain anesthetics can induce tau hyperphosphorylation in the absence of hypothermia. For example, Run et al. demonstrated in C57BL/6J mice that following exposure to either 30 sec of ether vapor or 5 min after an intraperitoneal pentobarbital injection, an increase in tau phosphorylation was observed at multiple phosphoepitopes (Run et al., 2009), including Thr¹⁸¹, Ser¹⁹⁹, Thr²⁰⁵, Ser²⁶², Ser⁴⁰⁴, and Thr²¹². In terms of the mechanism underlying this normothermic tau hyperphosphorylation, these investigators also observed that this increased tau phosphorylation was associated with the activation of the stress-induced kinases mitogen-activated protein kinase (MAPK) and c-Jun N-terminal kinase (JNK), suggesting that these specific protein kinases may a play a role under these conditions.

We recently examined the impact of the administration of propofol (2,6 diisopropylphenol), an intravenous sedative-hypnotic anesthetic commonly used in humans to provide general anesthesia as well as conscious sedation, on tau phosphorylation under normothermic conditions (Whittington et al., 2011). Compared to hypothermic propofol administration, normothermic propofol administration produced a lower degree of hippocampal tau phosphorylation; nevertheless, hyperphosphorylation still occurred in the absence of hypothermia. Specifically, 30 min following the administration of propofol 250 mg/kg i.p. to C57BL/6 mice, significant increases in hippocampal tau phosphorylation were observed at the AT8 (Ser²⁰²/Thr²⁰⁵), CP13 (Ser²⁰²), and PHF-1 (Ser³⁹⁶/Ser⁴⁰⁴) phosphoepitopes (Whittington et al., 2011). Interestingly, tau hyperphosphorylation persisted at the AT8 phosphoepitope 2h after propofol anesthesia, despite the recovery of the righting reflex in all of the mice. In the cortex, propofol also increased tau phosphorylation at the AT8, PHF-1, pSer²⁶², MC6, and pSer⁴²² phosphoepitopes 30 min following administration, and again this increase persisted at AT8 2h following propofol administration. Moreover, the mechanism of tau phosphorylation following normothermic propofol administration in vivo, was found to be similar to that observed during hypothermia-induced tau phosphorylation (Planel et al., 2007), in that the tau hyperphosphorylation under normothermic conditions was associated with a decrease in PP2A activity (to 77 ± 9%).

We also have determined that the tau hyperphosphorylation observed after normothermic propofol administration in vivo is not, in essence, a function of an anesthetic-induced physiologic change. To establish this, we examined tau hyperphosphorylation under normothermic condition in vitro using SH-SY5Y human neuroblastoma cells that were stably transfected to constitutively express human tau with 3 MT-binding domains (Tau-SH-SY5Y cells) (Mailliot et al., 2000; Delobel et al., 2002; Hamdane et al., 2003). We observed that a 1h exposure of a physiologic dose of propofol, 3 μg/ml (16.8 μM) at 37°C, in these Tau-SH-SY5Y cells resulted in increased phosphorylated tau levels at the AT8, CP13, and PHF-1 phosphoepitopes. Thus, tau phosphorylation following normothermic propofol administration is not the consequence of an anesthetic-induced physiologic change such as hypothermia, hypercarbia, or hypotension, but is indeed the consequence of a direct, pharmacologic effect.

Although propofol has the capacity to increase tau phosphorylation in the absence of hypothermia (Whittington et al., 2011), it is unclear whether this ability to induce tau hyperphosphorylation is specific to propofol, or a property shared by other commonly used intravenous anesthetics. Thus, we investigated the impact of a sedative dose of dexmedetomidine (300 μg/kg i.p), a commonly used sedative-analgesic agent, on levels of phosphorylated tau in non-transgenic mouse hippocampus under normothermic conditions (Whittington et al. personal communication). Thirty min following its administration, Dex significantly increased hippocampal tau phosphorylation at the AT8 (182 ± 36%), PHF-1 (183 ± 27%), and CP13 (170 ± 68%) phosphoepitopes. Interestingly, at 2h, Dex treatment was still associated with significant increases in hippocampal phosphorylated tau levels at all three phosphoepitopes, despite the return of the righting reflex in all of the mice at this time point. Furthermore, 6h post-dexmedetomidine this increase still persisted at AT8. Thus, these data demonstrate that, in terms of intravenous anesthetics, the ability to induce tau hyperphosphorylation under normothermic conditions is not a property unique to propofol. However, the mechanism by which two pharmacologically diverse anesthetics produce similar changes in tau phosphorylation remains to be determined.

Interestingly, tau phosphorylation under normothermic conditions does not appear to be a phenomenon solely observed with intravenous anesthetics. Indeed, Le Freche et al. have recently observed that acute exposure to the inhalational anesthetic sevoflurane resulted in significant, dose-dependent tau hyperphosphorylation in the hippocampus of nontransgenic mice (Le Freche et al., 2012). These same investigators have also examined the long-term effects of repeated exposure to sevoflurane on spatial memory and tau phosphorylation in mice. Specifically, starting at 6 months of age, the nontransgenic mice were repeatedly exposed to low-dose (1.5%) or high dose (2.5%) sevoflurane for 1h, under normothermic conditions, and this exposure was repeated monthly until 10 months of age (5 exposures in total). At 9 months of age, spatial learning and memory testing, as assessed using the Morris Water Maze (MWM), revealed significant impairment of acute memory retention following exposure to both doses of sevoflurane. One month following the last sevoflurane exposure, hippocampal tau phosphorylation was increased at the Ser³⁹⁶/Ser⁴⁰⁴ and Thr¹⁸¹ phosphoepitopes, without evidence of tau aggregation. Mechanistically, these same investigators concluded that repeated exposure to sevoflurane might involve the activation of AKT and ERK kinases as well as inactivation of GSK-3, suggesting that normothermic sevoflurane exposure may lead to the dysregulation of signaling along the MAPK and AKT/GSK3 pathways. The finding that tau kinases may be involved in the hyperphosphorylation observed following normothermic anesthesia exposure are consistent with the findings of Run et al. and suggest that PP2A inhibition may not be the dominant mechanism by which hyperphosphorylation occurs following this type of normothermic anesthesia exposure (Run et al., 2009). In summary, the findings of Le Freche et al. are critical in that they demonstrate clearly that repeated normothermic exposure to a volatile anesthetic can result in aberrant tau hyperphosphorylation, which is associated with memory impairment. Nevertheless, a causal relationship between the phosphorylated tau burden and cognitive impairment, under these same conditions, remains to be established, and further studies are thus indicated.

The findings that exposure to a volatile anesthetic under normothermic conditions leads to remote tau hyperphosphorylation, has been further supported by recent work in a triple-transgenic (3xTg-AD) mouse model (Oddo et al., 2003), which harbors 3 mutant genes: human beta-amyloid precursor protein (APP_Swe), presenilin-1 (PS-1_M146V), and tau (tau_P301L) and develops Aβ and tau pathology as well as cognitive impairment (Filali et al., 2012). Using quantitative immunostaining, Tang et al. (Tang et al., 2011b) observed that repeated normothermic exposure to halothane or isoflurane (5h once a week for 4 weeks), at 6–8 months of age, was associated with significantly increased hippocampal tau phosphorylation at the Thr²³¹ phosphoepitope in the 3xTg-AD mice group. Furthermore, similar increases in hippocampal tau phosphorylation were observed following either halothane or isoflurane. Dong et al. have also recently shown that 2h exposure to 1.4% isoflurane in 5 to 8 month-old APPswe/PS1dE9 line 85 mice (Jankowsky et al., 2004), led to persistent phosphorylation at pS262 up to 24h after anesthesia (Dong et al., 2012).

Hence, repeated volatile anesthetic exposure possesses the potential to induce persistent tau hyperphosphorylation even in the absence of anesthesia-induce hypothermia. However, the mechanisms involved in mediating this effect and the long-term neuropathological and functional sequelae stemming from this aberrant hyperphosphorylation have yet to be fully elucidated. Furthermore, as previous studies with isoflurane have not observed such negative effects on tau pathology in a normothermic setting (Planel et al., 2009; Menuet et al., 2012), these recent studies by Le Freche et al. and Tang et al., which indeed demonstrate a pathological change in tau following normothermic volatile anesthetic exposure, strongly suggest that whether an anesthesia-induced detrimental effect on tau pathology will become manifest is possibly dependent on several factors: the anesthetic exposure paradigm, mouse strain and genotype, tau phenotype, and the actual time point at which the animal is examined for tau pathology.

Tau, anesthesia and synaptic plasticity

In addition to the potential degradation of normal cytoskeletal structure and the disruption of axonal transport, there is the possibility that the tau hyperphosphorylation, insolubility, as well as the impaired tau binding and assembly observed following anesthesia might ultimately impair processes such as synaptic plasticity, learning and memory. Although there is considerable evidence suggesting that Aβ plays a significant role in mediating synaptic dysfunction and neurotoxicity in AD (Kamenetz et al., 2003; Haass and Selkoe, 2007), the precise mechanisms by which hyperphosphorylated tau species and other tau-related pathologic changes result in synaptic dysfunction and cognitive impairment are unclear.

Nevertheless, recent studies have suggested that tau pathology may indeed impair hippocampal synaptic transmission. For example, Hoover et al. have shown that the relocalization of hyperphosphorylated tau to dendritic spines was associated with synaptic dysfunction via a mechanism involving the impairment of glutamate receptor trafficking (Hoover et al., 2010). Furthermore, Burnouf et al., demonstrated in hippocampal slices from THY-Tau22 mice, a transgenic mouse model associated with hippocampal tau pathology and concomitant memory impairment, that the enhanced synaptic transmission normally induced by brain-derived neurotrophic factor (BDNF) was impaired, and this change was associated with cognitive impairment (Burnouf et al., 2013). Mechanistically, this impaired response to exogenous BDNF was the consequence of N-methyl D-aspartate receptor (NMDA-R) dysfunction, which appears to be secondary to an abnormal interaction between tau pathology (hyperphosphorylated and insoluble tau forms) and the NR2B subunit as well as Src protein kinase, a major kinase regulating NR2B expression. As Src inhibition has been associated with decreased NR2B expression and synaptic plasticity (Sinai et al., 2010), the potential for pathological tau species to interact with this kinase, could theoretically contribute to memory impairment.

While the relationship between glutamatergic receptor and synaptic plasticity has been extensively studied, GABAergic (γ-aminobutyric acid) synapse development has been less scrutinized (Kuzirian and Paradis, 2011). Since many anesthetics bind GABA receptors (Fig. 1), their effects on tau could be mediated by this interaction. Indeed, it has been shown that GABA_A receptor activation could enhance tau phosphorylation through reduced binding of PP2A to tau, and with the involvement of cdk5 but not GSK-3β (Nykanen et al., 2012). Moreover, there might be a feedback loop between tau pathology and GABAergic synapses since it was recently discovered in Tau_P301L mutant mice that GABAergic neurons were in a hypermetabolic state leading to increased GABA levels (Hege Nilsen et al., 2013).

Despite these findings, further studies are definitely warranted to elucidate the precise mechanisms by which tau hyperphosphorylation and aggregation lead to synaptic dysfunction and neurotoxicity, and whether anesthetics can exacerbate such systems.

Clinical Implications

There is a paucity of information examining the impact of the clinical administration of anesthesia on tau pathology in humans. This is partly due to the fact that, unfortunately, in actual clinical practice, it is often difficult to separate the impact of anesthesia from surgery, when determining the effect of anesthesia on changes in AD-associated biomarkers. Also, these studies have been somewhat limited in that they have lacked proper control groups. Nevertheless, there are some studies that have examined changes in the levels of total and phosphorylated tau in the postoperative period. Palotás et al. found that levels of total tau were significantly elevated at 6 months in patients who underwent coronary bypass surgery without cardiopulmonary bypass (Palotas et al., 2010). Although the patients were no longer on sedative medication at the time of CSF sampling, given the remote changes in tau pathology observed in rodents following anesthesia (Planel et al., 2009; Tang et al., 2011b; Le Freche et al., 2012), a potential contributory effect of anesthesia on the increases in total tau observed in these patients could not be excluded. Hitherto, Tang et al. have published the only human study to date that has examined tau phosphorylation in otherwise healthy individuals following anesthesia for routine surgery in which the anesthesia type was actually taken into consideration (Tang et al., 2011a). These investigators measured levels of total tau and tau phosphorylated at Thr¹⁸¹ in patients undergoing surgery for idiopathic cerebrospinal fluid leak correction, and observed that both total tau and phospho-Thr¹⁸¹ steadily increased until at least 48h, which was the last sampling time point. As the authors noted, the increase in phospho-Thr¹⁸¹ postoperatively is particularly interesting, as increased levels of phosphorylation at this epitope in CSF has been observed to be the most sensitive indicator of cognitive decline in normal patients (Tang et al., 2011a).

As noted above, it is not possible to dissociate the effects of anesthesia and surgery in patients. However, a recent preclinical study in 3xTG-AD mice has attempted to tackle this problem by comparing desflurane alone versus desflurane and cecal ligation and found that surgery per se led to memory impairment, enhanced tau pathology and neuroinflammation (Tang et al., 2012). However, the study was not conclusive in terms of a causative link between cognitive impairment and tau pathology.

Nevertheless, although it is difficult to clinically dissociate the impact of anesthesia from surgery in studies designed to determine the impact of anesthesia on tau pathology in humans, there are instances in which patients receive anesthetics without undergoing a surgical procedure. For example, many of the intravenous anesthetics used during surgery are often used for sedative purposes in patients requiring mechanical ventilation in an intensive care unit setting. Furthermore, as opposed to a transient exposure for a surgical procedure, many of these patients are exposed to some of these anesthetics for days and, in some instances, for weeks. Given that there is preclinical evidence suggesting that the ability of anesthetics to exacerbate and accelerate tau pathology may indeed be a dose-dependent phenomenon and/or a consequence of repeated exposure, clearly future studies focusing on the impact of prolonged or repeated anesthesia exposure in humans, particularly those with or at risk for AD, are necessary.

Conclusion

In summary, both pre-clinical and early clinical studies suggest that anesthetics can accelerate tau pathology. Furthermore, the presence of hypothermia during anesthesia administration, a common clinical occurrence, appears to exacerbate the degree of phosphorylation and accelerate the development of neurofibrillary pathology in models of pre-existing tau pathology. However, although it may accelerate the process, hypothermia appears not to be an absolute requisite for the development of tau pathology following anesthesia exposure, as changes in tau pathology have been described during normothermic anesthesia administration.

Hence, the potential clinical impact of anesthesia on the development of tau pathology, and on AD pathology in general, should not be underestimated as recently it is estimated that 234.2 million people undergo surgery and thus anesthesia each year (Weiser et al., 2008). Furthermore, as life expectancy continues to increase and with estimates predicting that, without a significant therapeutic breakthrough, the number AD cases is expected to significantly increase (Brookmeyer et al., 2007), the potential impact of this exposure on patients with or at risk for AD should not be discounted or ignored. This is particularly true given the fact that patients with AD are at higher risk for an exacerbation of their pre-existing cognitive impairment post-operatively, a condition associated with significant medical costs (Xie and Tanzi, 2006).

Highlights.

• Exposure to anesthetics might enhance the risk of Alzheimer’s disease

• Alzheimer’s disease brains contain intraneuronal neurofibrillary tangles composed of hyperphosphorylated tau; tau pathology correlates with dementia

• Here we review evidences showing that anesthesia can enhance tau pathology

Abbreviations

3R

3-repeat tau isoforms

4R

4-repeat tau isoforms

3xTg-AD

triple transgenic Alzheimer’s disease mouse model

AD

Alzheimer’s disease

AKT/PKB

protein kinase B

APP

amyloid-β precursor protein

Aβ

amyloid-β

BDNF

brain-derived neurotrophic factor

CaMKII

calcium/calmodulin-dependent protein kinase II

cdk5

cyclin-dependent kinase 5

CSF

cerebrospinal fluid

Dex

dexmedetomidine

ERK

extracellular signal-regulated kinase

GABA

γ-aminobutyric acid

GSK-3β

glycogen synthase kinase-3β

i.p

intra peritoneal

JNK

c-Jun N-terminal kinase

MAC

minimum alveolar concentration

MAPK

mitogen-activated protein kinase

MT

microtubule

MWM

Morris water maze

NFT

neurofibrillary tangles

NMDA-R

N-methyl D-aspartate receptor

NR2B

subunit 2B of NMDA receptor

PP1

protein phosphatase 1

PP2A

protein phosphatase 2A

PP2B

protein phosphatase 2B

PS-1

presenilin-1

-Ser

serine

Src

short for sarcoma, a non-receptor tyrosine kinase

Tau-SH-SY5Y cells

human neuroblastoma cells constitutively expressing human tau with 3 microtubule-binding domains

Thr

threonine