THE REGULATORY PROPERTIES OF α5 SUBUNIT- CONTAINING γ-AMINOBUTYRIC ACID SUBTYPE A RECEPTORS IN LEARNING AND SYNAPTIC PLASTICITY
by
Loren John Martin
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Institute of Medical Science University of Toronto
© Copyright by Loren Martin, 2009
The Regulatory Properties of α5 Containing γ-Aminobutyric Acid Subtype A Receptors in Learning and Synaptic Plasticity
Loren Martin Doctor of Philosophy Institute of Medical Science University of Toronto 2009
Abstract
Synaptic plasticity, which is thought to represent the neuronal substrate for learning and memory is influenced by the degree of GABAergic inhibitory tone. In particular, γ-aminobutyric acid subtype A receptors (GABAARs), which mediate the majority of inhibitory neurotransmission in the mammalian brain regulate learning and plasticity. In these studies I examined a subpopulation of α5 subunit-containing GABAA receptors (α5GABAARs), which are preferentially expressed in the hippocampus, to determine whether they have a specific role in memory processes. I hypothesized that α5GABAAR-activity constrains hippocampus-dependent learning and CA1 synaptic plasticity. The main research objective of this thesis was to investigate the electrophysiological changes within the hippocampus that accompany genetic and pharmacological targeting of α5GABAARs and how these changes impact behaviour.
I found that the general anesthetic etomidate enhanced a tonic inhibitory conductance generated by α5GABAARs, and this action correlated with an impairment of long-term potentiation (LTP) and hippocampus-dependent memory performance for fear-associated memory and spatial navigation. Mice with a genetic deletion of the α5 subunit gene (Gabra5–/–) were resistant to the
LTP- and memory-impairing effects of etomidate. Additionally, the LTP- and memory-impairing effects of etomidate were rescued by pharmacologically inhibiting α5GABAARs. Genetic and
ii
pharmacological inhibition of α5GABAARs enhanced associative learning in trace fear but not contextual fear conditioning tasks. Interestingly, genetic deletion and pharmacological inhibition of α5GABAARs did not result in the common adverse side-effects associated with non-selective inhibition of GABAARs such as anxiogenesis or seizures. Further, I found that blocking the tonic inhibition generated by α5GABAARs lowered the threshold for LTP, such that lower stimulation frequencies enhanced LTP. Synaptic changes within this frequency band were modified independently of phasic GABAAR inhibition. Inhibiting the α5GABAAR-dependent membrane conductance was associated with an increase in the depolarizing envelope during 10 Hz stimulation. These experiments provide new insights into the in vitro and in vivo physiology of
α5GABAARs and suggest that a tonic inhibition generated by α5GABAARs constrains learning and glutamate plasticity through regulation of the membrane’s electrical properties.
iii
Acknowledgments
I thank Dr. Beverley Orser for her continuous support and guidance over the last 5 years. Her infectious enthusiasm for science and unparalleled work ethic have truly been an inspiration. I also thank my thesis and examining committee members Drs. Mike Salter, John Roder, Zhengping Jia, Harold Atwood, Paul Frankland and Michael Fanselow (UCLA). Thanks also to Drs. John Macdonald and Mike Jackson for helpful discussions and technical contributions.
I am grateful for the contributions of the Orser lab. I thank Ella Czerwinska, Rob Bonin, Agnieszka Zurek and Paul Whissell for making the lab a fun place in which to work. Their participation in theoretical discussions and grammatical corrections were greatly appreciated.
A special acknowledgment must also go to those whose emotional support never wavered – my family. First my parents, Carol and Wayne for enduring my 11 year academic adventure, my sister Lori-Ann and my wife Marianne for her love, tolerance, understanding and help.
List of Contributions
Several investigators assisted with the experiments reported in this thesis. In Chapter 4, V.Y. Cheng assisted with the data collection for the Morris water maze, and the drug injections for the fear conditioning, rotarod and open field experiments. In Chapter 5, G.H. Oh assisted with the data collection for the Morris water maze, and drug injections for the fear conditioning experiments. In Chapter 6, Dr. William Ju contributed the Western Immunoblots for the glutamate receptors. In Appendix 1, R.P. Bonin completed the tonic inhibitory recordings in cultured hippocampal neurons.
This work could not have been completed without the financial support provided to me by the Canadian Institutes of Health Research and the Ontario Graduate Scholarship program.
iv
Table of Contents
Abstract...... ii
Acknowledgments ...... iv
Table of Contents ...... v
List of Figures...... xi
List of Abbreviations...... xiii
Chapter 1. Thesis Structure ...... 1
1.1 General overview...... 1
1.2 General hypothesis and thesis questions ...... 4
1.3 Thesis structure...... 5
Chapter 2. General Introduction...... 7
2.1 The hippocampus...... 7
2.1.1 Gross structural overview...... 7
2.1.2 Circuitry...... 8
2.1.3 Laminar organization and neurotransmitter systems...... 9
2.1.4 Learning, performance and hippocampus-dependent tasks...... 18
2.2 Modifications of synaptic connections...... 23
2.2.1 Long-term potentiation: Properties and characteristics...... 24
2.2.2 Long-term depression...... 27
2.2.3 Bidirectional modification of synaptic plasticity ...... 29
2.2.4 Role of GABA receptors in the induction of NMDA-dependent LTP...... 33
2.3 GABA and GABAA receptor-mediated inhibitory neurotransmission...... 34
2.3.1 Synthesis and metabolism of GABA...... 34
2.3.2 GABAA receptor-mediated hyperpolarization and shunting inhibition ...... 35
2.3.3 General overview of GABAA receptors ...... 39
v
2.3.4 Location and physiological role of α5GABAAR subunits ...... 47
2.3.5 Pharmacology of α5GABAAR subunits...... 54
2.3.6 α5GABAA receptors as a candidate target for the modification of hippocampus-dependent learning and synaptic plasticity ...... 57
2.3.7 A role for α5GABAARs in neurological disorders...... 63
2.4 Anesthesia...... 68
2.4.1 In vitro actions of general anesthetics ...... 69
2.4.2 Action of anesthetics on LTP and memory ...... 71
2.4.3 Problems associated with anesthetic-induced amnesia ...... 72
2.4.4 Intraoperative awareness and post-operative cognitive dysfunction...... 73
2.5 Summary...... 75
Chapter 3. General Methods and Materials...... 77
3.1 Experimental animals ...... 77
3.2 Electrophysiology...... 78
3.2.1 Preparation of hippocampus tissue slices...... 78
3.2.2 Extracellular synaptic stimulation ...... 79
3.2.3 Extracellular field recordings ...... 79
3.2.4 Current clamp and voltage clamp recordings in hippocampal slices ...... 80
3.2.5 Recording electrodes ...... 83
3.2.6 Pharmacological agents...... 83
3.3 Behaviour ...... 86
3.3.1 Morris water maze...... 86
3.3.2 Fear conditioning...... 87
3.3.3 Pharmacological agents...... 88
3.4 Statistical analysis ...... 88
Chapter 4. α5GABAA receptors mediate the amnestic but not sedative-hypnotic effects of the general anesthetic etomidate ...... 90 vi
4.1 Introduction ...... 90
4.2 Specific Methods...... 92
4.2.1 Tonic currents...... 92
4.2.2 Long-term potentiation recordings...... 92
4.2.3 Contextual fear conditioning ...... 92
4.2.4 Water maze...... 93
4.2.5 Rotarod ...... 94
4.2.6 Open field ...... 94
4.2.7 Loss of righting reflex ...... 94
4.3 Results ...... 95
4.3.1 Etomidate enhanced tonic conductance in CA1 pyramidal neurons...... 95
4.3.2 Etomidate reduced LTP in WT but not Gabra5–/– brain slices...... 101
4.3.3 Amnestic effects of etomidate are mediated by α5GABAARs...... 104
4.3.4 Sedative-hypnotic effects of etomidate are not mediated by α5GABAARs...... 108
4.4 Discussion...... 112
Chapter 5. The LTP and memory impairing effects of etomidate are prevented by pre- treatment with the α5GABAA receptor selective inverse agonist L-655,708 ...... 117
5.1 Introduction ...... 117
5.2 Materials and Methods ...... 120
5.2.1 Synaptic plasticity in hippocampus slices ...... 120
5.2.2 Voltage clamp recordings...... 121
5.2.3 Fear-conditioned learning...... 121
5.2.4 Water maze learning...... 123
5.2.5 Elevated plus maze...... 124
5.2.6 Statistical analysis ...... 125
5.3 Results ...... 126
vii
5.3.1 L-655,708 reverses etomidate blockade of long-term potentiation...... 126
5.3.2 Etomidate potentiation of the CA1 tonic conductance is blocked by L-655,708 application ...... 129
5.3.3 L-655,708 reverses etomidate memory blockade...... 134
5.3.4 α5GABAAR activity impairs performance for trace fear conditioning...... 135
5.3.5 Etomidate impairment of spatial memory is reversed with L-655,708...... 139
5.3.6 α5GABAARs do not contribute to the anxiety-like behaviours ...... 140
5.4 Discussion...... 146
Chapter 6. Mechanism of α5 subunit-containing γ-aminobutyric acid type A receptor- mediated inhibition of learning and synaptic plasticity ...... 156
6.1 Introduction ...... 156
6.2 Specific methods ...... 158
6.2.1 Fear conditioning...... 158
6.2.2 Extracellular recordings...... 159
6.2.3 AMPA / NMDA Ratios...... 161
6.2.4 Spontaneous inhibitory recordings...... 161
6.2.5 Synaptic block of GABAAR-mediated IPSPs with SR-95531 ...... 162
6.2.6 Current clamp recordings and input resistance measurements...... 162
6.2.7 Preparation of protein samples and western blot analysis...... 163
6.3 Results ...... 165
6.3.1 Learning and memory co-vary with α5GABAAR activity in hippocampus- dependent tasks...... 165
6.3.2 α5GABAARs regulate synaptic plasticity within a narrow range of stimulus frequencies...... 167
6.3.3 Synaptic transmission, excitability and glutamate receptor expression are normal in Gabra5–/– mice ...... 173
6.3.4 Blockade of tonic but not phasic inhibitory neurotransmission enhances submaximal LTP ...... 181
viii
6.3.5 α5GABAARs regulate synaptic plasticity via membrane regulation ...... 192
6.4 Discussion...... 200
Chapter 7. General Discussion...... 205
7.1 Overview ...... 205
7.1.1 α5GABAAR-activity correlates with cognitive performance...... 207
7.1.2 α5GABAARs do not possess intrinsic anxiety, sedative or hypnotic properties .211
7.2 In vitro regulation of learning and memory by α5GABAARs on hippocampal synaptic plasticity ...... 213
7.2.1 α5GABAARs shift the threshold for LTP...... 214
7.2.2 α5GABAARs regulate membrane properties to limit LTP...... 215
7.3 Conclusion...... 220
7.4 Future Directions...... 222
List of Appendices ...... 225
Appendix 1. α5GABAARs do not regulate the cognitive impairing properties of ethanol...... 226
Appendix 2. Inclusion of thesis-relevant work published or submitted by candidate ...... 252
Appendix 3. Additional publications resulting from my doctoral studies ...... 253
References ...... 254
ix List of Tables
Table 3.1. The primary pharmacological agents used in this thesis...... 84 Table 4.1. Properties of spontaneous IPSCs recorded in CA1 pyramidal neurons of hippocampal slices...... 99 Table 5.1. Effects of etomidate and L-655,708 on spontaneous mIPSCs...... 132 Table 6.1. mIPSC kinetics of WT and Gabra5–/– CA1 pyramidal neurons in the presence or absence of L-655,708...... 186
List of Figures
Figure 2.1. Schematic representation of the hippocampal trisynaptic circuit and the connections of the hippocampus...... 14 Figure 2.2. Feedforward and feedback inhibitory neurotransmission and the primary location of α5GABAARs on the distal dendrites of pyramidal neurons...... 16 Figure 2.3. Morris water maze and fear conditioning experimental set-up...... 21 Figure 2.4. A theoretical plot showing the BCM model of synaptic plasticity...... 31
Figure 2.5. Schematic structure of the GABAA receptor...... 45 Figure 2.6. A schematic drawing illustrating the methods and drug applications used to determine the presence of the tonic conductance...... 52 Figure 3.1. The whole cell configuration equivalent circuit is shown...... 81 Figure 4.1. A low concentration of etomidate increased the holding current in CA1 pyramidal neurons from WT but not Gabra5–/– slices...... 97 Figure 4.2. Etomidate reduced the LTP of fEPSPs recorded in the CA1 region of hippocampal slices prepared from WT but not Gabra5–/– mice...... 102 Figure 4.3. Etomidate impaired contextual fear conditioning in WT but not Gabra5–/– mice. 106 Figure 4.4. Etomidate impairment of motor coordination, spontaneous motor activity and loss of the righting reflex (LORR) was not increased in Gabra5–/– mice...... 110
Figure 5.1. The intrinsic activity of α5GABAARs does not enhance LTP in hippocampal slices but reverses etomidate-induced LTP impairment...... 127 Figure 5.2. L-655,708 attenuated the etomidate-induced increase in the holding current in CA1 pyramidal neurons...... 130
Figure 5.3. The expression and activity of α5GABAARs modify contextual and trace fear conditioning...... 137 Figure 5.4. Normal acquisition of the matching to place version of the Morris water maze but impaired recall with etomidate-treatment...... 142 Figure 5.5. L-655,708 and etomidate do not contribute to anxiety-like behaviors in the elevated plus maze...... 144 Figure 5.6. A model illustrating the regulation of synaptic plasticity in the hippocampus by extrasynaptic α5GABAARs and blockade of LTP by etomidate...... 154
Figure 6.1. α5GABAARs physiologically regulate the acquisition of weak hippocampus- dependent associative fear memory tasks...... 169
Figure 6.2. α5GABAARs critically regulate the threshold for LTP within narrow range of stimulus frequencies...... 171 Figure 6.3. Normal synaptic transmission and excitability in Gabra5–/– slices...... 175 Figure 6.4. Glutamate receptor currents and expression were not altered in Gabra5–/– neurons...... 177 xi
Figure 6.5. LTP and LTD observed with 10 Hz stimulation is dependent on NMDA signalling pathways...... 179
Figure 6.6. Pharmacologic studies confirm that α5GABAARs are critical for the induction of LTP following submaximal but not high-frequency stimulation...... 184 Figure 6.7. The effect of SR-95531, L-655,708 and bicuculline on the tonic conductance in CA1 pyramidal neurons...... 188
Figure 6.8. Blockade of synaptic GABAA receptors does not enhance plasticity with 10 Hz stimulation...... 190 Figure 6.9. Decreased membrane resistance accompanies 10 Hz induced LTD in WT neurons...... 194
Figure 6.10. α5GABAARs regulate membrane properties, depolarizing envelope and input resistance with 10 Hz stimulation...... 198
xii
List of Abbreviations aCSF = Artificial cerebral spinal fluid
AMPA = α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate
ANOVA = Analysis of variance
APV = (2R)-amino-5-phosphonovaleric acid
D-AP5 = D(-)aminophosphopentanoic acid
α5GABAARs = Alpha5 subunit-containing GABAA receptors
BAPTA = 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid
BCM = Bienenstock Cooper and Munroe
CA1 = Cornu Ammonis area 1
CA3 = Cornu Ammonis area 3
CaCl2 = Calcium chloride
CGP = CGP-35348
CNQX = 6-cyano-7-nitroquinoxaline-2,3-dione
CO2 = Carbon dioxide
CS = Conditioned stimulus
CsCl = Cesium chloride
CsGluc = Cesium gluconate
CsOH = Cesium hydroxide
δGABAARs = Delta subunit-containing GABAARs
EC = Entorhinal cortex
ECF = Extracellular fluid
EGTA = Ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid
xiii fEPSP = Field excitatory postsynaptic potential
GABA = γ-Aminobutyric acid
GABAA = γ-Aminobutyric acid subtype A receptors
Gabra5 = α5-γ-Aminobutyric acid subtype A gene
Gabra5–/– = α5-γ-Aminobutyric acid subtype A gene deletion
GTP = Guanosine triphosphate
HEPES = 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
Hz = Hertz
ICF = Intracellular fluid i.p = Intraperitoneal
IPSC = Inhibitory postsynaptic current
IPSP = Inhibitory postsynaptic potential
ISI = Inter-stimulus interval
KCl = Potassium chloride
K-Gluc = Potassium gluconte
KMeSO4 = Potassium methylsulphate
KOH = Potassium hydroxide
LORR = Loss of righting reflex
LTD = Long-term depression
LTP = Long-term potentiation
MAC = Minimum alveolar concentration
Mg-ATP = Magnesium-adenosine triphosphate
MgCl2 = Magnesium chloride
xiv mGluRs = Metabotropic glutamate receptors mIPSC = Miniature inhibitory postsynaptic current
NaCl = Sodium chloride
Na-Gluc = Sodium gluconate
NaHCO3 = Sodium bicarbonate
NaH2PO4 = Sodium dihydrogenphosphate
NMDA = N-methyl-D-aspartic acid
NMDAR1 = N-methyl-D-aspartic acid receptor subunit 1
O2 = Oxygen
P = Postnatal day
POCD = Post-operative cognitive dysfunction
PPF = Paired pulse facilitation
TBS = Theta burst stimulation
TEA = Tetraethylammonium
US = Unconditioned stimulus
VGCC = Voltage-gated calcium channel
WT = Wild-type
X = Times or By
xv 1
Chapter 1. Thesis Structure
1.1 General overview
The brain is not hard wired and it is clear that afferent stimuli have the ability to modify existing synaptic connections (Bear and Kirkwood, 1993; Bear and Malenka, 1994; Malenka and Bear,
2004; Nicoll and Schmitz, 2005). The creation or modification of neural pathways is thought to represent our memories and experiences. This type of modification is often referred to as synaptic plasticity and more specifically long-term potentiation (LTP) and long-term depression
(LTD). LTP is the activity-dependent strengthening of synaptic contacts between neurons and is presumably the normal mechanism through which learning and experience occur (Bliss and
Collingridge, 1993; Malenka and Bear, 2004). LTD is the weakening of neuronal synapses and within the hippocampus, it results from persistent weak stimulation (Malenka and Bear, 2004).
LTD is thought to be important for processing information concerning object-place configurations (Kemp and Manahan-Vaughan, 2004; Massey and Bashir, 2007).
Within the CA1 region of the hippocampus, both LTP and LTD have been shown to depend on the activation of N-methyl-D-aspartate (NMDA) receptors (Dudek and Bear, 1993; Bear and
Malenka, 1994; Cavus and Teyler, 1998; MacDonald et al., 2006). These processes are thought to result from Ca2+ influx through the NMDA receptor and depend on the timing and frequency of the synaptic input (Dudek and Bear, 1993; Abraham et al., 2001). The hypothesis has been proposed that a low level of calcium influx leads to LTD, whereas Ca2+ entry above a certain threshold leads to LTP (Cormier et al., 2001). The threshold for LTP versus LTD is on a sliding scale and depends on the history of synaptic stimulation. If the synapse has already undergone
1 2
LTP, the threshold for initiating LTP is raised, increasing the probability that a Ca2+ influx will yield LTD rather than LTP. This “plasticity of plasticity” is referred to as metaplasticity, a term originally coined by Abraham and Bear (Abraham and Bear, 1996). Metaplasticity is analogous to the “sliding threshold” model of experience-dependent synaptic plasticity described by
Bienenstock, Cooper, and Munro (Bienenstock et al., 1982), a model that was developed to describe visual cortex plasticity during development but was adapted to describe plasticity in other regions, including the CA1 hippocampus. Activity above the modification threshold (θM) leads to LTP, whereas low-threshold activity leads to LTD, where θM can be regulated by a number of factors, including prior activity at synapses and the relative contribution of NMDAR subtypes (Bear, 1995; Abraham et al., 2001). Increasing or decreasing the amount of synaptic plasticity (i.e. LTP or LTD, respectively) has been referred to as the bidirectional (up or down) regulation of synaptic plasticity (Bear, 2003), and will be used throughout this thesis. Further, the threshold for LTP, as referred to in this thesis is the frequency at which LTP was induced over
LTD. The mechanisms and theories surrounding synaptic plasticity and LTP will be discussed in further detail in Chapter 2.2.
The focus of this thesis was to examine the effects of a subpopulation of γ-aminobutyric acid subtype A receptors (GABAARs) on cognitive processes and synaptic plasticity. Specifically, α5 subunit-containing GABAARs (α5GABAARs) were examined for their role in hippocampus- dependent learning and CA1 synaptic plasticity. α5GABAARs have a dense distribution in the hippocampus (Sur et al., 1999; Pirker et al., 2000) and have been shown to improve learning when their activity is genetically or pharmacologically limited (Collinson et al., 2002; Crestani et al., 2002; Collinson et al., 2006; Dawson et al., 2006). However, it is still not certain whether or how this receptor population enhances learning and memory. Additionally, the role that
2 3
α5GABAARs may play in LTP is murky at best, as genetically deleting or reducing the expression of α5GABAAR does not improve LTP (Collinson et al., 2002; Crestani et al., 2002), but selective pharmacological inhibition of α5GABAARs enhances some forms of LTP (Atack et al., 2006b; Collinson et al., 2006). With respect to LTP, the differing results in various studies may be the due to the vastly different LTP induction protocols that have been used (Collinson et al., 2002; Crestani et al., 2002; Atack et al., 2006b; Dawson et al., 2006). Here I examined a wide range of electrophysiological protocols and behavioural assays to address the hypothesis that α5GABAAR-activity constrains hippocampus-dependent learning and CA1 synaptic plasticity.
Broadly, the objectives of this thesis were to investigate the relationship between α5GABAAR- activity and learning and memory and synaptic plasticity. I investigated electrophysiological changes within the hippocampus that accompany genetic and pharmacological manipulation of
α5GABAARs and determined how such changes impact behaviour. Specifically, the aims are:
1. To investigate the role of the α5GABAAR-mediated tonic conductance in modulating
excitatory synaptic plasticity and behaviour by using a combination of complementary
genetic and pharmacologic tools.
2. To determine whether the shunting conductance or membrane hyperpolarization
mediated by α5GABAARs modulates the plasticity of excitatory postsynaptic potentials
in the hippocampus.
3. To determine the role of α5GABAARs in different paradigms and stages of learning and
memory behaviour.
3 4
1.2 General hypothesis and thesis questions
In the hippocampus, α5GABAARs are heavily expressed and their contribution to modifying learning and memory has been documented by a few select laboratories (Collinson et al., 2002;
Crestani et al., 2002; Atack et al., 2006b; Cheng et al., 2006; Dawson et al., 2006).
Behaviourally, reducing the activity of α5GABAARs enhances hippocampus-dependent performance for memory tasks, but it is still not certain how this receptor alters memory.
Previous studies have loosely examined its role in general learning tasks and LTP, but none have systematically examined the in vivo and in vitro effects of its role in different forms of learning and LTD and LTP. I hypothesized that α5GABAAR-activity constrains hippocampus-dependent learning and CA1 synaptic plasticity. This section outlines the specific thesis questions that were generated to examine the role of α5GABAARs in learning and memory.
1. Does increasing α5GABAAR activity with the general anesthetic etomidate impair
learning and memory and synaptic plasticity?
2. Does pharmacological inhibition of α5GABAARs reverse etomidate-induced impairment
of memory and LTP?
3. How do α5GABAARs modify learning and memory and synaptic plasticity?
These questions will help me to achieve my aims by allowing me to specifically explore the electrophysiological responses of α5GABAARs and how these changes manifest at the behavioural level.
4 5
1.3 Thesis structure
The results of this thesis are presented in Chapters 4 through 6. In Chapter 4, I examined the actions of the anesthetic etomidate on the tonic current, the regulation of LTP, hippocampus- dependent and -independent memory, sedation, and motor coordination. I found that etomidate significantly potentiated the tonic current generated by α5GABAARs and this action correlated with an impairment of learning and memory and LTP. The α5GABAAR-dependent impairment of memory was found to be independent of the motor effects caused by increasing GABAergic activity with etomidate. These results were published in the Journal of Neuroscience (Cheng et al., 2006), and I served as the co-first author of this manuscript. The behavioural data described in this publication were collected with the assistance of V.Y. Cheng, who performed all drug injections.
In Chapter 5, I found that pharmacological inhibition of α5GABAARs with L-655,708 reversed the memory and LTP-impairing effects of etomidate. These results are important because they represent the first demonstration that a reversal agent exists for the memory-impairing effects of a general anesthetic. These data are also clinically relevant because they show that a reduction in
α5GABAAR-activity did not influence motor behaviours, anxiety levels, or seizure activity. The results presented in Chapter 5 have been accepted for publication in Anesthesiology. The behavioural data were collected with the assistance of G.H. Oh, who performed all drug injections.
The goal of Chapter 6 was to establish an in vitro mechanism that correlated with the behavioural memory phenotype associated with altering α5GABAAR activity. There has not been an established link between LTP and α5GABAARs to explain the in vivo phenotype of Gabra5–/–
5 6 mice. I based these experiments on the BCM model of bidirectional synaptic plasticity and saturated the frequency-response plot to determine: 1) whether a link between α5GABAARs and
LTP exists; and 2) whether the activity of α5GABAARs is important for the induction of LTP. I found that α5GABAARs regulate synaptic plasticity independently of synaptic GABAARs and that α5GABAARs act to regulate the neuronal membrane such that depolarization and input resistance are reduced during LTP. These results are currently in preparation for re-submission to the Journal of Neuroscience.
6 7
Chapter 2. General Introduction
2.1 The hippocampus
The link between the hippocampus and memory is an idea that arises from clinical and experimental observations of patients who sustained brain damage to the hippocampus and other regions of the medial temporal lobes or midbrain (Scoville and Milner, 1957). The integrity of the hippocampal formation is causally involved in many but not all forms of memory including declarative, spatial and episodic memories (Eichenbaum, 2000, 2004). It is now clear that hippocampal synapses undergo structural plasticity, which is often regarded as the main cellular mechanism underlying learning and memory (Bliss and Collingridge, 1993). The hippocampus, one of the most widely studied regions of the brain, is of interest to a wide spectrum of neuroscientists, ranging from those who study its normal structure and function to others who study its malfunctions in various diseases and pathological conditions. The hippocampus was examined in this thesis as it provides a robust model for examining the connections between learning and memory behaviours, and the underlying changes in plasticity
2.1.1 Gross structural overview
The hippocampus or more formally the hippocampal formation is located inside the medial temporal lobe of the cerebral cortex and is therefore considered part of the telencephalon
(Carpenter and Sutin, 1983; Suzuki, 1999), and more specifically part of the limbic system
(Gloor, 1997). The hippocampus as a whole, has the shape of a curved tube, a seahorse or a ram’s horn (Andersen, 2007). It consists of a ventral and dorsal portion, both of which share similar composition but are parts of completely different neural circuits (Czerniawski et al.,
7 8
2008; Howland et al., 2008). This general layout occurs across the full range of mammalian species, ranging from rodents to humans (Andersen, 2007). One of the most captivating features of the hippocampus is its neuronal circuitry. It has a relatively simple organization consisting of the principal cell layers and a highly organized laminar distribution of many inputs, which provides an excellent preparation for functional studies. The hippocampus proper has three subdivisions: CA3, CA2 and CA1. The other regions of the hippocampal formation include the dentate gyrus, subiculum, presubiculum, parasubiculum, and entorhinal cortex. The most abundant connections are with the entorhinal cortex (EC), which lies next to the hippocampus in the temporal lobe. The superficial layers of the EC provide most of the input to the hippocampus and the deep layers of the EC receive most of the outputs. For this thesis the CA1 region was examined because α5GABAARs are highly expressed in this region (Sur et al., 1999; Pirker et al., 2000), α5GABAARs are predominantly located to the extrasynaptic membrane in these neurons (Brunig et al., 2002), α5GABAARs are highly sensitive to neurodepressive agents including benzodiazepines and general anesthetics (Caraiscos et al., 2004a; Rudolph and Mohler,
2004), and α5GABAARs are involved in CA1-dependent spatial memory tasks (Collinson et al.,
2002).
2.1.2 Circuitry
A common organizational feature of connections between regions of the neocortex is that they are reciprocally linked (Wilson, 1987). However, this is not the case for the connections that link the various parts of the hippocampus (Carpenter and Sutin, 1983) (see Figure 2.1), which are often referred to as unidirectional. For convenience, the EC can be considered the first structure in the intrinsic hippocampal circuit. This is based on the fact that much of the neocortical input that reaches the hippocampus must first enter the EC (Ino et al., 1998). The projections from the
8 9
EC largely project to the dentate gyrus and form the major input pathway called the perforant pathway. Likewise the granule cells of the dentate gyrus give rise to axons called mossy fibers that project to the pyramidal cells of the CA3 subfield of the hippocampus. The CA3 pyramidal cells, in turn, are the source of the Schaffer collateral axons that project to the CA1 pyramidal cells. This loop of connections is referred to as the trisynaptic circuit of the hippocampus
(Andersen, 2007). The CA1 subfield then projects to the subiculum, providing its major excitatory input. Once the input has travelled to the CA1 region and the subiculum, the pattern of connections begins to become considerably more elaborate. For example, CA1 neurons not only project to the subiculum but also to the EC. Further, while the subicular neurons project to the presubiculum and parasubiculum, the major projection remains the deep layers of the EC
(Ino et al., 1998). Although this brief overview of the hippocampal circuitry leaves out many of the facts that make the system somewhat more complex, it does highlight the basic neuronal connections.
2.1.3 Laminar organization and neurotransmitter systems
The hippocampus, aside from being conveniently wired in a unidirectional manner is also favourably organized in a laminar fashion, such that the apical and basal dendrites and the axons and cell bodies are arranged in distinct layers, which make its use in slice electrophysiology uncomplicated (Gazzaniga, 2004). The pyramidal cell layer is tightly packed in CA1 and more loosely packed in CA2 and CA3. The narrow relatively cell free layer located deep to the pyramidal cell layer is called the stratum oriens. This layer contains the basal dendrites of the pyramidal cells and several classes of interneurons (Ganter et al., 2004; Mercer et al., 2007). In the CA3, but not the CA2 or CA1 regions a narrow acellular zone, the stratum lucidum is located just above the pyramidal cell layer and is occupied by the mossy fibers (Lim et al., 1997). The
9 10 stratum radiatum is located superficial to the stratum lucidum in CA3 and immediately above the pyramidal cell layer in CA2 and CA1. The stratum radiatum is the suprapyramidal region that contains the CA3 to CA3 autoassociational connections and the CA3 to CA1 Schaffer collateral connections. The most superficial layer of the hippocampus is the stratum lacunosum- moleculare. It is in this thin layer that fibers from the entorhinal cortex terminate. Due to the known laminar organization of the hippocampus electrophysiologists are able to stimulate a fairly homogeneous population of axons and record their population or monosynaptic responses.
Further, because the synaptic architecture of the hippocampus is well preserved in vitro, the hippocampal slice preparation is ideally suited for studies of synaptic transmission and plasticity.
Within this circuit, fast excitatory postsynaptic neurotransmission (i.e. activation time ~ 100–600
µs; deactivation time constant ~ 5–10 ms) is mediated by inward excitatory postsynaptic currents
(EPSCs) flowing through α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate (AMPA) type glutamate receptors (Mosbacher et al., 1994; Andersen, 2007). AMPAR subunit distribution varies among hippocampal cell types resulting in different properties between excitatory pyramidal cells and inhibitory interneurons. The traditional nomenclature for AMPA receptor subunits has been GluR1–4 or GluRA–D, but the currently accepted nomenclature is GluRA1–
A4 (Collingridge et al., 2009). For the purpose of this thesis, the traditional AMPA receptor nomenclature will be used (GluR1–4). In the hippocampus the AMPA receptor assemblies are predominantly GluR1/2, with some GluR2/3 and GluR1 homomeric channels (Nicoll et al.,
2006). Specifically, in CA1 pyramidal neurons the majority of AMPARs contain the Ca2+ impermeable GluR2 subunit with the majority of these neurons conducting AMPA currents with low Ca2+ permeability (Toth and McBain, 1998). In contrast, hippocampal interneurons express mainly GluR1 and GluR4 receptor subunits and do not possess high levels of GluR2 subunits
10 11
(Toth and McBain, 1998). Since interneurons possess low expression levels of GluR2, their synaptic AMPA currents show voltage-dependence with high permeability to Ca2+ ions (Geiger et al., 1995).
Slower hippocampal EPSCs (activation time ~ 7 ms; deactivation time ~ 200 ms) are the result of activating N-methyl-D-aspartic acid (NMDA) receptors and certain subtypes of metabotropic glutamate receptors (Ruppersberg et al., 1993; Andersen, 2007). The NMDARs are expressed in dentate granule cells, all pyramidal cells and in many interneurons (Laurie and Seeburg, 1994;
Monyer et al., 1994). It should be mentioned that the traditional nomenclature for NMDA receptor subunits has been NR1, and NR2A–NR2D, but the currently accepted nomenclature is
GluN1, and GluN2A–GluN2D (Collingridge et al., 2009). For the purpose of this thesis, the traditional NMDA receptor nomenclature will be used (NR1, and NR2A–NR2D). The NMDA receptors are tetrameric complexes composed of the obligatory NR1 subunit and one of the
NR2A-D subunits, with two copies each of the NR1 and NR2 subunits (Furukawa et al., 2005).
In CA1 pyramidal neurons, NR2B containing receptors are reported to be selectively targeted to the apical versus the basal dendrites (Kawakami et al., 2003). It has also been reported that in
CA1 pyramidal neurons, the NR2A subunit is expressed in the postsynaptic density of glutamatergic synapses, while the NR2B subunit is located extrasynaptically (Tovar and
Westbrook, 1999). The NR2C subunit is not significantly expressed in the hippocampus but
NR2D subunit expression is restricted to hippocampal GAD67-, parvalbumin-, and somatostatin- positive interneurons in the stratum radiatum CA1 and CA3 regions (Standaert et al., 1996). In
Chapter 2.2.1, the integrative and ion channel functions of NMDARs will be discussed in more detail.
11 12
The activity of hippocampal pyramidal neurons is regulated and synchronized by an inhibitory system mainly composed of GABAergic interneurons (Whittington and Traub, 2003). Fast inhibitory postsynaptic currents (IPSCs) are the result of ionotropic GABAAR activation, while slower IPSCs are due to the activation of GABAB receptors (Kandel and Schwartz, 2000).
Interneuronal activation typically results in two distinct forms of inhibition: feedforward and feedback inhibition. Feedforward inhibition is the result of axon collaterals from excitatory afferent fibers contacting local interneurons to inhibit the principal cell. This type of inhibition accounts for the majority of inhibition from extrinsic sources. The extra delay associated with the additional synapse ensures that the feedforward inhibitory event does not impinge on the extrinsic excitatory synaptic event in the principal neuron (Kandel and Schwartz, 2000). In turn, if a principal cell fires, a number of interneurons receive synaptic excitation such that they provide feedback inhibition to the local circuit. Thus the generation of an action potential from a hippocampal subregion is followed rapidly by a period of marked local feedback inhibition
(Andersen, 2007). A schematic representation illustrating feedforward and feedback inhibition is shown in Figure 2.2A.
All hippocampal interneurons release GABA as their primary neurotransmitter, but the diversity and classification of interneurons is complex (Maccaferri and Lacaille, 2003). A study that examined the different types of interneurons in the hippocampus in terms of morphology, physiology and receptor expression revealed that there are at least 16 different interneuron types based on cell morphology alone (Parra et al., 1998). Interneurons are classified by a number of descriptive features including the morphology of dendritic or axonal processes (eg. basket cell, horizontal cell or stellate cell), the two cell layers that contain the soma and axonal processes
(e.g. oriens-lacunosum moleculare) and the presence of neurochemical markers (e.g.
12 13 parvalbumin-positive and cholecystokinin-positive interneurons) (Maccaferri and Lacaille,
2003). Specific interneurons are also known to innervate subregions regions of the hippocampus.
For instance, bistratified, horizontal and oriens-lacunosum-moleculare interneurons specifically target the distal dendrites of excitatory pyramidal cells, whereas other interneurons including basket cells exclusively target the perisomatic regions (see Whittington and Traub, 2003 for a review).
Interneuronal connectivity is also correlated with specific GABAAR subtypes expressed at distinct target sites. For example, α5GABAARs are enriched in the dendritic regions of hippocampal subfields and are highly associated with the dendrite-targeting bistratified interneurons (Sperk et al., 1997; Thomson et al., 2000). There is even greater specificity for interneuronal targeting in the pyramidal cell layer, as α1 subunit-containing GABAARs are innervated by parvalbumin-positive basket cells and α2 subunit-containing GABAARs are targeted by cholecystokinin-positive basket cells (Nyiri et al., 2001). Another study showed that
IPSPs mediated by bistratified cells were not enhanced by zolpidem, an α5 subunit-sparing benzodiazepine type 1 agonist that has preference for α1 subunit-containing subtypes, but were enhanced by diazepam (Thomson et al., 2000). Conversely, zolpidem enhanced IPSPs are mediated by fast-spiking basket cells, which target the soma of pyramidal cells. These results suggest that the distal synapses of principal neurons contain α5GABAARs, but α5GABAARs are absent from the perisomatic and somatic region. As shown in Figure 2.2B the proposed location of α5GABAARs to the distal dendrites of the principal neuron is shown. This figure also illustrates the location of action for the different GABAAR blockers that were used in this thesis.
13 14
Figure 2.1. Schematic representation of the hippocampal trisynaptic circuit and the connections of the hippocampus.
(A) An illustration showing the trisynaptic circuit in a transverse section of the hippocampus.
Input enters the circuit from the entorhinal cortex via the perforant pathway and synapses onto the granule cells of the dentate gyrus. These cells in turn give rise to the mossy fiber pathway, which synapses onto CA3 pyramidal cells. These cells are the source of the Schaffer collateral pathway and project to the CA1 pyramidal cells. (B) A schematic illustrating the different compartments of the hippocampal circuitry. The projections are along the transverse axis of the hippocampal formation, the dentate gyrus is located proximally and the entorhinal cortex is located distally. This illustration highlights some of the important additional hippocampal connections aside from the trisynaptic circuit. α5GABAARs are primarily located in the stratum radiatum on the distal dendrites of CA1 neurons and receive the majority of inputs from Schaffer collateral or perforant pathway inputs. DG = dentate gyrus, Sub = subiculum, Pre = presubiculum, Para = parasubiculum, EC = entorhinal cortex.
14 15
15 16
Figure 2.2. Feedforward and feedback inhibitory neurotransmission and the primary location of
α5GABAARs on the distal dendrites of pyramidal neurons.
(A) Feedforward inhibition: Axon collaterals from excitatory afferent fibers contact local interneurons (large grey circle). The additional synaptic delay compared to direct afferent excitation onto principal cells (black triangle) provides a time-dependent sequence of excitation and inhibition from single afferent inputs (dual EPSP-IPSP sequence). Feedback inhibition:
Axon collaterals from local principal cells contact local interneurons, providing a period of inhibition of principal cell activity following the generation of an output. Excitatory synapses are represented by the small black circles and inhibitory synapses are represented by the small open circles. (B) α5GABAARs (red squares) are primarily located to the distal dendrites of pyramidal neurons in the stratum radiatum of the CA1 region. These receptors are targeted by distal dendrite-preferring interneurons (i.e. bistratified interneurons) and are primarily located extrasynaptically but can also be found synaptically. In contrast perisomatic- or somatic- targeting interneurons (i.e. basket cells) solely innervate synaptically expressed GABAARs (blue ellipses) and not α5GABAARs. The site of action for the different GABAAR inhibitors used in this thesis are shown. L-655,708 preferentially inhibits α5GABAARs, SR-95531 inhibits synaptically expressed GABAARs, while bicuculline inhibits all GABAARs. s.l.m., stratum lacunosum moleculare; s.rad., stratum radiatum; p.c.l., pyramidal cell layer; s.o., stratum oriens.
16 17
17 18
2.1.4 Learning, performance and hippocampus-dependent tasks
One important theme that often arises when discussing the function of the hippocampus is the distinction between learning and performance (Andersen, 2007). In a behaving animal, performance is used as an objective measure of achievement of a pre-defined task, such as freezing in fear conditioning or entering the correct arm of a maze. In contrast, learning reflects the underlying processes that are presumed to be taking place in the brain. These underlying processes are responsible for the acquisition of knowledge, while performance is the expression of this knowledge. The concept of learning itself is complex because the brain has evolved to express a number of qualitatively distinct forms of learning. The integrity of the hippocampus is essential for the processes specifically underlying declarative, episodic and spatial memories
(Eichenbaum, 2004). However, it is not possible to easily test declarative and episodic memories in rodents, and neuroscientists often rely on assessing the performance of mice and rats using spatial hippocampus-dependent tasks (Olton and Samuelsson, 1976; Barnes, 1979; Morris et al.,
1982).
Widely used tasks to study memory in rodents include the T maze and cross maze, foraging for food in a radial maze, or swimming to a hidden platform in the Morris water maze. The water maze is a spatial navigation task that requires a rodent to locate a hidden platform in a circular tub of opaque water using visual cues surrounding the room (Morris et al., 1982). It is often regarded as an attractive hippocampus-dependent task because it is free from the stress of food deprivation that is often associated with land-based mazes (Olton and Samuelsson, 1976; Adams et al., 2002) and does not use painful stimuli such as foot shocks (Fanselow and Helmstetter,
1988). In some of his original studies, Richard Morris (Morris et al., 1982) trained rats that had either been subjected to “hippocampal” lesioning (damage to the hippocampus, dentate gyrus
18 19 and subiculum), control cortical lesions (damage to an equivalent volume of parietal cortical tissue) or no surgery controls. In the water maze, both the cortical control and control groups quickly learned to swim toward the hidden platform, starting from any quadrant within the water maze. Only the hippocampus-lesioned group was impaired in navigating the water maze to locate the platform. Performance during a post-training probe trial revealed that both the control and cortical lesioned groups swam consistently across the location where the platform was previously positioned during training. This behaviour is arguably reminiscent of “free recall of memory”. In contrast, the hippocampus-lesioned animals did not display a distinct navigational pattern and randomly swam around the pool. The experimental set-up for the water maze that was used for this thesis s shown in Figure 2.3A.
Aside from the variety of mazes used to study the memory performance of rodents, another widely used and exploited assay is the conditioned fear response. Certain fear conditioning protocols are sensitive to hippocampal lesions while others are not (Kim and Fanselow, 1992;
Phillips and LeDoux, 1992; Kim et al., 1993; Maren et al., 1997), suggesting that different fear conditioning protocols require different brain regions. Two of the more commonly used fear conditioning protocols are contextual conditioning and cued fear conditioning. The expression of fear learning requires the amygdala but depending on the type of conditioning additional structures (i.e. the hippocampus) may be required for the processing of information, including contextual stimuli (Kim and Fanselow, 1992; Phillips and LeDoux, 1992). In these tests, mice learn to associate a distinct context or auditory tone (conditioned stimulus, CS) with an aversive stimulus such as a foot shock (unconditioned stimulus, US). A typical cued fear conditioning protocol will pair the CS and US with no time delay. When test subjects are placed back in the same training context, they exhibit a range of conditioned fear responses, including freezing.
19 20
Recently, the use of trace fear conditioning has provided researchers with a protocol that is markedly dependent on the dorsal hippocampal processing (Quinn et al., 2002; Chowdhury et al.,
2005; Misane et al., 2005; Quinn et al., 2005; Wanisch et al., 2005). The protocol for trace fear conditioning is similar to cued fear conditioning except that an empty trace-interval of 15 to 30 sec is interposed between the tone and foot shock (Misane et al., 2005). In trace conditioning, the animal must keep the conditioned stimulus in short-term memory for a period of time and has to actively process the stimulus to do so. Thus, successful trace conditioning relies on elaborate information processing and as mentioned, the hippocampus is necessary to build this trace. These fear memory traces need to be activated to access the fear memory system via the hippocampus.
In contrast, during cued conditioning, startle potentiation occurs regardless of whether the contingencies were effectively processed, suggesting that the CS is able to directly activate the fear system (i.e. amygdala) (Hamm and Weike, 2005). The experimental fear conditioning protocol used for this thesis is presented in Figure 2.3B.
It is important to bear in mind that there is rarely a simple one-to-one mapping of a specific behavioural task and presumed type of learning. Most behavioural tasks require effective sensory processing in higher order brain structures and efficient motor processing. Thus, caution must be exercised when interpreting results that may be caused by hippocampus-independent effects.
Researchers will typically integrate multiple behavioural measurements into their studies in order to gain perspective on the larger physiological function. In this thesis I used a combination of water maze training and contextual, cued and trace fear conditioning to ascertain the role of
α5GABAARs in different aspects of behavioural memory performance.
20 21
Figure 2.3. Morris water maze and fear conditioning experimental set-up.
(A) The set-up that I used for the water maze experiments is shown. For the experiments described in this thesis the water maze consisted of a 120 cm pool that contained water tainted with nontoxic white paint. Four equally spaced points around the edge of the pool were designed.
A clear coloured platform was placed 0.5 cm below the surface of the water in middle of a quadrant, 15 cm away from the pool wall. The mice could climb onto the platform to escape from the necessity of swimming. (B) The protocol for fear conditioning is shown. The fear conditioning paradigm used for the experiments described in this thesis consisted of a 3 day protocol. On day 1, the mice were exposed to a training protocol that consisted of three tone- shock (CS-US) pairings, separated by 60 sec. On day 2, the mice were re-exposed to the context in the absence of tone or foot shock. On day 3, the mice were exposed to a modified context in the absence (first 3 min) and presence (last 5 min) of tone.
21 22
22 23
2.2 Modifications of synaptic connections
The brain is an ever changing, modifiable, network of cells. The ability of neurons to modify their connectivity has been described in many brain structures. These underlying changes are thought to represent a cellular and physiological mechanism to explain cognitive, emotional and perceptual behaviour. For instance, in the ventral tegmental area (VTA), circuit modification in the form of LTP has been studied to explain addictive behaviours. Repeated cocaine exposure in vivo facilitates the LTP of excitatory synapses of the VTA (Liu et al., 2005). The facilitated LTP is the result of cocaine-induced reduction of GABAAR-mediated inhibition of VTA dopaminergic neurons and is thought to be important for the formation of drug-associated memory. Additionally, neuronal plasticity in the visual cortex is thought to be important for the shaping of ocular dominance during development (Kirkwood et al., 1996) and perceptual learning in adulthood (Schoups et al., 2001). Further, spatial learning is thought to result from the LTP of synaptic connections in the hippocampus.
In the hippocampus the type of plasticity can vary depending on the subregion, receptors, ion channels, transmitters or transporters that are involved. Stark examples of these differences are the mossy fiber-CA3 synapses and the Schaffer collateral CA1 synapses. While the former is dependent on the activation of metabotropic glutamate receptors (mGluRs) (Ito and Sugiyama,
1991), the latter is dependent on NMDAR activation (Collingridge et al., 1983). Interestingly, in the CA1 region, the depression and potentiation of synapses both require NMDAR activation, probably owing to the Ca2+ permeable nature of this receptor (Sweatt, 2003). The strengthening and weakening of CA1 synaptic connections was investigated in Chapter 6, as a basis for the enhanced learning that occurs when α5GABAARs are genetically or pharmacologically
23 24 inhibited. This next section will briefly highlight the features of synaptic plasticity and its bidirectional features.
2.2.1 Long-term potentiation: Properties and characteristics
The LTP of synapses was first described by Bliss and Lomo in the dentate gyrus of the anesthetised rabbit (1973). Early studies of LTP revealed that tetanizing stimuli used to induce potentiation were input specific, meaning that only the input pathways that were tetanized were potentiated (Bliss et al., 1973; Andersen et al., 1977). Hippocampal synapses also express the property of coopertivity (McNaughton et al., 1978). Coopertivity is a process that requires the coactivity of a considerable number of fibers when the stimulation intensity is weak; the simultaneous input of these weak stimuli is sufficient to potentiate synapses. The third property commonly associated with LTP is associativity (McNaughton et al., 1978), whereby a weak input could be potentiated if its activation coincided with a tetanus to another input.
Experimental analysis of the properties and mechanisms of LTP has concentrated on a Hebbian form of synaptic plasticity as exhibited by the perforant pathway-dentate gyrus granule cells and
Schaffer commissural CA1 pyramidal neurons (Bliss and Collingridge, 1993; Bliss et al., 2004).
Tetanic stimulation of both of these pathways results in a rapid, persistent increase in synaptic strength. LTP can be recorded for many hours in the slice preparation (Frey et al., 1993) and anesthetized animals (Bliss et al., 1973) and for days, weeks and months with chronically implanted electrodes (Komaki et al., 2007). In CA1 pyramidal neurons the induction of LTP is blocked by the NMDA receptor antagonist D(-)aminophosphopentanoic acid (D-AP5) highlighting the requirement for NMDA receptor activation in CA1 LTP (Collingridge et al.,
1983). Further, the use-dependent non-competitive NMDA antagonist MK-801 (Coan et al.,
24 25
1987), and the glycine site antagonist 7-chlorokynurenic acid (Bashir et al., 1990) also block the induction of LTP in CA1 pyramidal neurons, reinforcing the importance of NMDARs for LTP induction. It has been shown that NMDAR-activity does not contribute to baseline synaptic transmission at low stimulation frequencies. Particularly convincing evidence comes from studies in which the cre-lox technique was used to create mice harbouring a deletion of the
NMDAR1 subunit (the obligatory NMDA receptor subunit) in area CA1 but not the dentate gyrus. Slices from these mice showed normal synaptic transmission but no LTP in CA1 neurons and normal LTP in the dentate gyrus. Further, these mice exhibit impaired spatial but unimpaired nonspatial learning (Tsien et al., 1996).
It is now clear that two conditions must be met for an NMDAR-mediated response to be generated: presynaptic activity to release glutamate and strong depolarization in the postsynaptic neuron to relieve Mg2+ block of the NMDA receptor (Coan and Collingridge, 1985). Further, the
NMDAR is highly permeable to Ca2+ and NMDAR-dependent LTP is blocked with injections of the Ca2+ chelator ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA)
(Lynch et al., 1983). The importance of Ca2+ in LTP is reinforced by studies that showed thapsigargin, which prevents the refilling of internal Ca2+ stores and prevents the induction but not expression of LTP (Harvey and Collingridge, 1992). Internal Ca2+ is released from stores via the generation of inositol triphosphate (IP3) and/or Ca2+ induced activation of ryanodine receptors (Kandel and Schwartz, 2000). In the CA1 region NMDA-dependent LTP is blocked by dantrolene, an inhibitor of ryanodine receptors (Obenaus et al., 1989) and thapsigargin (Alford et al., 1993). It is also possible for Ca2+ entry to mediate a NMDA-independent form of LTP through the activation of voltage-gated Ca2+ channels (VGCC), most likely though the L-type
25 26
VGCCs, although this form of LTP normally results from unusually high stimulation frequencies
(> 200 Hz) (Grover and Teyler, 1990).
LTP is not a single phenomenon; rather, there are various forms of LTP with distinct time courses and with different underlying biochemical mechanisms (Madison and Schuman, 1991;
Malenka, 1994; Malinow, 1994; Huang et al., 2001). The best-studied form of LTP is NMDAR- dependent. This form of LTP tends to dissipate between 1 to 2 h post-induction, and it is sensitive to blockers of Ca2+/calmodulin-dependent protein kinase (CaMK) but insensitive to inhibitors of protein synthesis (Frey et al., 1993; Huang et al., 1994). The process of LTP exhibits at least two distinct temporal components: early LTP (E-LTP) and late-LTP (L-LTP).
Despite the requirement for NMDAR activation for the induction of LTP in CA1 pyramidal neurons, NMDARs are not required to maintain LTP following tetanic stimulation as the application of AP5 after tetanic stimulation does not decrease LTP (Izumi et al., 1991). L-LTP requires protein synthesis (Frey et al., 1993) and is a long-lasting form of synaptic potentiation that often is used as an allegory of long-term memory. This form of LTP lasts for several hours and is thought to alter long-term gene expression (Sweatt, 2003). L-LTP has been suggested to be important for the consolidation of short-term memory to long-term memory. Since
α5GABAARs are reportedly not important for the consolidation of hippocampus-dependent memory (Collinson et al., 2006), L-LTP was not examined in this thesis.
A review of the published E-LTP literature suggests that CA1 LTP underlies hippocampus- dependent spatial learning (Silva et al., 1992b; Silva et al., 1992a; Chen and Tonegawa, 1997).
The form of LTP that was examined in this thesis is E-LTP, which is expressed for up to 60 min following tetanic stimulation and is dependent on protein kinases but does not require protein
26 27 synthesis (Frey et al., 1993). In E-LTP several kinases are involved but none are obligatory.
Inhibitors of CaMKII, protein kinase A (PKA) or protein kinase C (PKC) have no effect on the induction of LTP when applied alone, and similarly, concurrent inhibition of PKA and PKC has no effect (Wikstrom et al., 2003). However, a CaMKII inhibitor and either a PKA or PKC inhibitor applied together, fully block the induction of LTP in young animals (Wikstrom et al.,
2003). This suggests that LTP is mediated by parallel kinase pathways that involve CaMKII and either PKA or PKC activation. Interestingly, mice that express a mutation in the α-CaMKII are deficient in their ability to produce LTP (Silva et al., 1992a), and are impaired in hippocampus- dependent tasks (Silva et al., 1992b).
Additionally, the cAMP responsive element binding protein (CREB), a nuclear protein that modulates the transcription of genes is required for the cellular events underlying long-term but not short-term memory (Silva et al., 1998). Increases in Ca2+ driven by the activation of synaptic
NMDARs turn on CaMK, which may also phosphorylate and activate CREB in neurons (Bito et al., 1996). Evidence suggests that increases in nuclear Ca2+ can also activate CREB, indicating that nuclear kinases may have a direct role in the modulation of CREB activity to regulate memory behaviours (Hardingham et al., 1997). Additionally, protein tyrosine kinases appear to be necessary for LTP as they phosphorylate the NMDAR and increase its function (O'Dell et al.,
1991; Lu et al., 1998). Apart from the induction of E-LTP, the expression of LTP has been largely attributed to the insertion of AMPA receptors from intracellular compartments (Lu et al.,
2001) or the lateral diffusion from the extrasynaptic regions (Tardin et al., 2003).
2.2.2 Long-term depression
The changes in synaptic strength that are thought to underlie the encoding of new information
27 28 are also represented by a depression of synaptic transmission, in the form of LTD (Dudek and
Bear, 1992, 1993; Kirkwood et al., 1993). Similar to LTP, it has now become clear that LTD has the same properties that are believed to be necessary for memory formation: cooperativity, associativity and input specificity (Hebb, 1949; Bliss and Collingridge, 1993). Recent evidence suggests that hippocampal LTP and LTD encode very different aspects of spatial memory.
Whereas novel empty space enhances the expression of LTP in the CA1 region (Kemp and
Manahan-Vaughan, 2004), an environment that contains novel objects facilitates the expression of LTD (Manahan-Vaughan and Braunewell, 1999; Kemp and Manahan-Vaughan, 2004). It was originally thought that LTD served as a forgetting mechanism (Tsumoto, 1993). This theory has since been revised when it became apparent that LTD is not the molecular mirror image of LTP.
Specifically, different phosphorylation sites on the AMPA receptor are affected differently in
LTP and LTD (Lee et al., 2000), and LTP and LTD require the activation of different NMDA receptor subunits (Liu et al., 2004). In the hippocampus, electrically induced LTD is generated by low frequency stimulation, in the range of 1–3 Hz, applied for prolonged durations (5–15 min) (Dudek and Bear, 1992; Fox et al., 2006). Further, the effects of LTD have been shown to persist for days and weeks in vivo (Manahan-Vaughan and Braunewell, 1999; Kemp and
Manahan-Vaughan, 2004).
In the CA1 region, LTD is homosynaptic and depends on NMDAR activation (Dudek and Bear,
1992; Manahan-Vaughan and Braunewell, 1999; Manahan-Vaughan et al., 2000). In particular,
LTD has been shown to depend on the activity of the NR2B subunit of the NMDAR (Liu et al.,
2004), but these results have not been consistently replicated (Morishita et al., 2006). Convincing evidence indicates that the process of LTD in the CA1 region activates NMDARs (Dudek and
Bear, 1992; Mulkey and Malenka, 1992), which leads to a rise in the postsynaptic Ca2+
28 29 concentrations (Mulkey and Malenka, 1992). However, the magnitude of the induced Ca2+ transients determines whether synapses will undergo LTP or LTD; larger Ca2+ transients are usually attributed to LTP and smaller transients are associated with the induction of LTD
(Cormier et al., 2001). Further, increasing evidence indicates that the mechanisms governing
LTD are complex, much like the mechanism(s) responsible for LTP. While LTP was found to be associated with the phosphorylation of the ser-831 residue of the AMPA receptor, LTD is associated with phosphorylation of the ser-845 residue (Lee et al., 2000). LTD is also associated with a physical loss of AMPA receptors, and it has been shown that AMPA receptors are rapidly internalized in response to LTD-inducing stimuli via a dynamin- and clathrin-dependent mechanism (Beattie et al., 2000). There is also much evidence to suggest that LTD-inducing stimuli reduce NMDAR-mediated transmission (Gean and Lin, 1993; Selig et al., 1995;
Montgomery and Madison, 2002), which strengthens the importance of LTD for mechanisms of learning and memory.
2.2.3 Bidirectional modification of synaptic plasticity
Given that most synapses in the brain can undergo LTP and LTD, it is natural to assume that the bidirectional properties of synaptic plasticity is an important clue to the understanding of memory formation. A wealth of literature suggests that there is a critical threshold of synaptic activity that is necessary to induce LTP versus LTD in an activity-dependent manner (Dudek and
Bear, 1993; Heynen et al., 1996; Castellani et al., 2001; Fox et al., 2006). Bienenstock, Cooper and Munro (Bienenstock et al., 1982) developed a mathematical model, now referred to as the
BCM model, to describe the development of the visual cortex. The BCM model identified a sliding threshold for the induction of both LTP and LTD as illustrated in Figure 2.4. In the BCM model, a previous history of high levels of synaptic activation produced a rightward-shift in the
29 30 threshold that made it less likely for input to elicit LTP and more likely to evoke LTD.
Conversely, a history of low level of synaptic activation caused a leftward-shift in the threshold that favoured the induction of LTP. The BCM model has been used to describe plasticity in other brain regions, including CA1 hippocampal neurons (Dudek and Bear, 1993) and the mammalian cortex (Castellani et al., 2001). Specifically, in CA1 neurons the BCM model has been related to the size of Ca2+ transients through NMDARs, with large Ca2+ transients initiating LTP and smaller Ca2+ transients resulting in LTD (Cormier et al., 2001).
In vivo, the sliding threshold model of plasticity has been regarded as metaplasticity (Abraham and Bear, 1996). In line with the BCM theory, metaplasticity is dependent on the previous history of the synapse and is greatly influenced by synaptic modifications (Abraham, 2008). A stark example of metaplasticity occurs in the visual cortex of dark-reared rats. Following a period of reduced cortical activity caused by dark rearing, LTP is enhanced and LTD is reduced over a range of stimulation frequencies compared to light-reared animals (Czepita et al., 1994;
Maffei et al., 2006). Further, it has been shown that decreasing the NR2A/2B ratio promotes the induction of LTD over LTP in dark reared rats (Philpot et al., 2007). Despite the many different mechanisms that probably contribute to metaplasticity and synaptic plasticity, there are likely many overlapping processes. A greater understanding of metaplasticity might yield new insights into information storage and processing in the CNS that are distinct from the traditional views of
LTP or LTD.
30 31
Figure 2.4. A theoretical plot showing the BCM model of synaptic plasticity.
Metaplasticity is similar to the "sliding threshold" model of experience-dependent synaptic plasticity described by Bienenstock, Cooper and Munro (1982). This theoretical model of plasticity plots the relative change in EPSP amplitude from the baseline as a function of the stimulus strength. Activity above the modification threshold (θM) leads to LTP, whereas low threshold activity leads to LTD. θM can be regulated by a number of factors including the relative activity of NMDA subtypes and inhibitory neurotransmission.
31 32
32 33
2.2.4 Role of GABA receptors in the induction of NMDA-dependent LTP
It has been known for many years that blocking GABAergic inhibition can augment the induction of NMDAR-dependent LTP (Wigstrom and Gustafsson, 1983). The stimulation of the
Schaffer commisural pathway evokes an excitatory postsynaptic potential (EPSP) that is immediately followed by an inhibitory postsynaptic potential (IPSP). The IPSP is biphasic, consisting of a fast GABAAR-mediated component and the slower GABABR-mediated response, which curtails the EPSP. During tetanization GABAAR-mediated inhibition restricts the degree of postsynaptic depolarization, the firing of postsynaptic action potentials and the magnitude of
Ca2+ influx into postsynaptic neurons (van der Linden et al., 1993). Additionally, enhancing the
IPSP intensifies the Mg2+ block of the NMDA channel and greatly limits the extent to which
NMDARs participate in synaptic responses (Dingledine and Korn, 1985; Herron et al., 1985).
The release of GABA is important and differentially regulated during the various LTP induction protocols. For example, in a priming stimulus protocol (PSP), in which a single stimulus precedes a brief high-frequency burst, the priming stimulus causes a release of GABA that feeds back to activate GABAB autoreceptors. The subsequent burst of pulses then releases substantially less GABA due to activation of presynaptic GABABRs, which in turn facilitates LTP (Davies et al., 1991). Alternatively, a protocol such as theta-burst stimulation (TBS) does not have a priming stimulus but instead consists of 10 trains of 4 bursts of pulses delivered at 100 Hz separated by 200 ms for 1 sec. TBS has been shown to cause a substantial release of GABA when compared with a normal 100 Hz protocol (Chapman et al., 1998). This suggests that theta patterns of activity may be tuned to maximally depress GABA inhibition via the activation of autoreceptor mechanisms. However, the GABAB autoreceptor mechanism only operates for a few seconds and during conventional tetanus (i.e. 100 Hz), GABAAR-mediated inhibition
33 34
dominates (Davies and Collingridge, 1993).The role of α5GABAARs in terms of learning and
LTP will be discussed extensively in Chapter 2.3.6.
2.3 GABA and GABAA receptor-mediated inhibitory
neurotransmission
2.3.1 Synthesis and metabolism of GABA
GABA is the primary inhibitory neurotransmitter in the adult mammalian central nervous system. It is a small amino acid derived from glutamate by glutamic acid decarboxylase (GAD).
The metabolic pathway that synthesizes GABA is referred to as the GABA shunt. This pathway utilizes α-ketogluterate, which is formed from glucose metabolism via the Kreb’s cycle to synthesize GABA (Hertz et al., 1999). α-Ketogluterate is transaminated by α-oxoglutarate transaminase to form glutamate, the immediate precursor to GABA. Glutamate is decarboxylated by GAD, of which two distinct isoforms exist; GAD65 and GAD67 (Martin and Barke, 1998) and both isoforms require pyridoxal 5’-phosphate (PLP) as a cofactor (Erlander et al., 1991; Bu et al.,
1992; Bu and Tobin, 1994). The membrane associated GAD65 is primarily involved with the synthesis of GABA to be used as a neurotransmitter and is stored in the synaptic vesicles and released at inhibitory synaptic sites (Kaufman et al., 1991). In contrast, the cytosolic GAD67 is more widely distributed in the cytoplasm and is responsible for GABA to be used for other functions such as serving as a source of energy and a source of GABA to be released via non- vesicular mechanisms (Kaufman et al., 1991).
Immunoimaging and biochemical studies have shown that GAD65 is associated with the surface of synaptic vesicles and is responsible for both GABA synthesis and the packaging of GABA
34 35 into synaptic vesicles (Jin et al., 2003). The transportation of newly synthesized GABA into synaptic vesicles is further aided by the vesicular GABA transporter (vGAT), which is a 10- transmembrane helical protein (Gasnier, 2000). This is an active process that requires Mg2+- adenosine triphosphatase (Mg2+-ATPase) and depends equally on the electrochemical gradient and the pH gradient across the vesicular membrane (Hell et al., 1990). Vesicular GABA is released from synaptic vesicles following an action potential-dependent rise in intracellular Ca2+ and synaptic vesicular fusion with the cell membrane to undergo exocytosis, thereby releasing
GABA into the synaptic cleft (Moss and Smart, 2001). Aside from the activation of GABA receptors, the extracellular concentration of GABA in the synaptic cleft is decreased by diffusion away from synapses and reuptake into presynaptic terminals or surrounding astrocytes by the
GABA transporters (GATs) (Nelson, 1998). Reuptake of GABA is the primary mode of cessation for GABAergic release and occurs via one of at least four GATs, GAT1-4. Of these transporters, GAT1 and GAT4 mediate the largest amount of GABA uptake in axon terminals
(Radian et al., 1990) and astrocytic processes within the synapse, respectively (Borden et al.,
1995; Conti et al., 2004). Similarly, extracellular GABA is degraded by GABA transaminase
(GABA-T), which converts GABA to succinic semialdehyde (Tao et al., 2006). GABA uptake and degradation are never complete and it is estimated that there is on average between 0.1-0.4
µM of GABA in the extracellular space at any given time (Attwell et al., 1993; Richerson and
Wu, 2003).
2.3.2 GABAA receptor-mediated hyperpolarization and shunting
inhibition
The opening of an ion channel can have differential effects on the neuron depending on the activated conductance. Influx of cations or efflux of anions (i.e. an inward current) can result in a
35 36 depolarization, while anion influx or cation efflux (i.e. an outward current) hyperpolarizes the membrane, provided that the membrane resistance is sufficient. It is important to understand that according to Ohm’s law, stronger currents will be needed to produce significant polarization, if the membrane resistance is low (i.e. the membrane will shunt the current required to generate an action potential). Ohm’s law by definition is linear, and describes the effects of a single ion conductance, and can be written as:
I = gE where I = current (amperes), g = conductance (siemens) and E = electrical driving force (volts).
However, when there are multiple interacting conductances, a more rigorous model known as the
Goldman-Hodgkin-Katz (GHK) permeability model can account for the interaction of unequal concentrations of ions on either side of the membrane. According to the GHK model, the ions partition instantaneously and independently into and out of channels, which have identical partition coefficients on either side of the membrane (Hille, 2001). In this sense ion channels act as tiny batteries that can generate electrical currents and potentials across the cell membrane. The
GHK equation describes the relationship between a number of conductances based on ion concentration, and permeability, and can be written as:
where E is the membrane potential; R is the universal gas constant; T is the temperature (°K); F
+ + - is Faraday's constant, pK, pNa, and pCl are the membrane permeabilities for K , Na , and Cl ; [K]o,
+ + - [Na]o, and [Cl]o are the concentrations of K , Na and Cl in the extracellular fluid; [K]i, [Na]i, and
+ + - [Cl]i are the concentrations of K , Na and Cl in the intracellular fluid.
36 37
In this thesis I examined GABAARs, which directly control the gating of ion channels permeable
- - to Cl and HCO2 ions. The net effect of these ions on membrane potential is highly dependent on the electrochemical gradient of these anions. These receptors cluster at synaptic release sites where they produce synaptic conductances, with fast rise and decay kinetics and generate a phasic or fast inhibition (Prenosil et al., 2006). GABAARs are also found in the extrasynaptic or perisynaptic regions of the membrane where they display a high sensitivity to extracellular
GABA, generate a tonic inhibitory conductance and desensitize more slowly than synaptic
GABAARs (Yeung et al., 2003). The diversity of GABAergic inhibition largely depends the postsynaptic membrane location and the subunit composition of GABAARs (Bai et al., 2001;
Caraiscos et al., 2004b; Prenosil et al., 2006; Glykys et al., 2008). Slower neuronal inhibition is mediated by GABABRs, but this type of inhibition will not be discussed since GABAARs were the focus of this thesis.
In the mature brain, GABAAR-mediated events are distinguished as having either a hyperpolarizing or shunting effect on the neuronal membrane. The nature of GABAAR-mediated inhibition is determined by the concentration gradient for Cl- across the cell membrane, which can vary widely during development (Ben-Ari et al., 2007) and determines the synaptic reversal potential (Takeuchi and Takeuchi, 1971). As such, if the synaptic reversal potential is below the resting membrane potential, inhibition will be hyperpolarizing. In contrast, if the synaptic reversal potential is between the resting membrane potential and the threshold to generate an action potential, inhibition is referred to as having a shunting effect (Alger and Nicoll, 1979;
Gulledge and Stuart, 2003).
37 38
The GABAAR-mediated shunting inhibition acts to reduce the membrane depolarization effect of concurrent excitatory events. The activation of a shunting inhibitory conductance decreases the local input resistance and the amplitude of subsequent EPSPs is reduced. Shunting inhibition can only be studied in isolation when the membrane potential remains at a relatively stable subthreshold level that is the same value as the reversal potential for the inhibitory conductance
(Mann and Paulsen, 2007). This shunting effect lasts for only as long as the ion channels mediating the inhibitory event remain open, which might be as brief as a few milliseconds
(Bartos et al., 2001). Interestingly, shunting inhibition does not change the slope or gain of the input-output relationship when neurons are stimulated by nonfluctuating excitatory input
(Chance et al., 2002; Mitchell and Silver, 2003). We have recently shown in dissociated cultured that activation of α5GABAARs does not change the gain of the input-output relationship for hippocampal pyramidal neurons (Bonin et al., 2007).
The developmental stage, behavioural state or neuronal subtype can all determine whether a
GABAAR-mediated event will be depolarizing or hyperpolarizing. In immature cortical neurons, the GABAAR reversal potential is positive to the resting membrane potential due to high levels of intracellular Cl-, leading to depolarizing neuronal events. This effect acts to drive hippocampal networks (Sipila et al., 2005), modulate early neuronal migration (Wang and Kriegstein, 2009) and organize a heterogeneous population of immature neurons (Cossart et al., 2005). In contrast, in mature neurons the GABAAR reversal potential is typically in the range of -60 to -75 mV.
This is the result of the action of the K+/Cl- co-transporter KCC2, whose expression increases over the course of development (Rivera et al., 1999). Thus, during active network states,
GABAAR activation will typically lead to a hyperpolarizing and inhibitory response.
Interestingly, hyperpolarizing GABAergic neurons can paradoxically excite neurons through
38 39 their interaction with voltage-dependent conductances producing rebound excitation. Activation of membrane hyperpolarization can induce inward currents such as the hyperpolarization- activated current (Ih), deactivate outward currents such as IM or by deinactivating inward currents such as Na+ or Ca2+. These effects are most prominent in cortical and thalamic neurons
(Destexhe et al., 1998) and are thought to be important for the control of neuronal spike timing
(Cobb et al., 1995), as well as different states of vigilance and seizures (Amzica and Steriade,
1999; Steriade and Amzica, 1999).
2.3.3 General overview of GABAA receptors
GABAARs belong to a large Cys-loop family of fast ligand-gated ionotropic receptors that include the nicotinic acetylcholine receptors, glycine receptors, and ionotropic 5-HT3 (serotonin) receptors (Sine and Engel, 2006). GABAARs are structurally organized as a pentameric membrane-spanning protein surrounding a central pore that forms the ion channel. The various subunits that combine to form each GABAAR consist of a long N-terminal extracellular hydrophilic region, followed by four transmembrane (M1-4) α-helices with the M2 segment forming the lining of the ion channel and a large intracellular loop between M3 and M4 (Gurley et al., 1995) and each sequence ends with a relatively short extracellular C-terminal domain. The topology of the GABAAR is shown in Figure 2.5A and Figure 2.5B. The GABA binding site is formed at the N-terminal domains at the interface between the α and β subunits (Mehta and
Ticku, 1999). Several amino acids contribute to the GABA binding site with important contributors including the α1 Val178, Val 180 and Asp183 (Newell and Czajkowski, 2003) and
β2 Ser 204, Tyr205, Arg207 and Ser209 (Wagner and Czajkowski, 2001). Additionally, the N- terminal regions of the α and γ subunits are important for allosteric modification of the GABAAR and binding of benzodiazepines (Mehta and Ticku, 1999). The M1-4 segments are important for
39 40
modulation of GABAARs by general anesthetics (Belelli et al., 1997; Jenkins et al., 2001;
Jenkins et al., 2002), while the large intracellular loop between M3 and M4 is implicated in phosphorylation and binding of intracellular proteins involved in the targeting and clustering of these receptors (Swope et al., 1999).
Each GABAAR is assembled from a pool of 19 different subunits (α1–6, β1–3, γ1–3, δ, ε, π, θ,
ρ1–3), which combine to form a pentameric ion channel. The different subunit combinations contribute to the distinctive expression patterns and the pharmacological and biophysical characteristics of each GABAAR (Fritschy and Mohler, 1995; Nusser et al., 1998). The combination of five subunit partnerships that comprise each GABAAR are not random but rather quite specific (Laurie et al., 1992; Fritschy and Mohler, 1995). The arbitrary combination of 19 different subunits to form a pentameric ion channel would result in a large number of GABAARs.
However, there is a vast reduction in the combinatorial degree of freedom, which is made possible by limiting the subunit partners that can assemble together by imposing strict rules on the number of subunits in the same class of GABAAR that can co-assemble (Sieghart and Sperk,
2002; Whiting, 2003). The most prevalent of these combinations in the mammalian brain is 2 α1,
2 β2 and 1 γ2 subunits arranged around the central pore in a specific order. Specific GABAARs combined from different subunit arrangements have different developmental, physiological and pharmacological properties and are also localized to specific regions on the neurons (Mody,
2001; Farrant and Nusser, 2005; Glykys and Mody, 2007a).
2.3.3.1 Synaptic GABAA receptors and phasic inhibition
The traditional form of GABAergic inhibition that arises from synaptic contacts is called phasic inhibition and transiently inhibits neurons for 10–100 ms (Semyanov et al., 2004; Krook-
40 41
Magnuson and Huntsman, 2007). This type of phasic inhibition is mediated by GABAARs that are known to cluster at synaptic release sites where they produce conductances, with fast rise and decay kinetics (Costa, 1998). In mature neurons the activation of these synaptically expressed
GABAARs typically results in membrane hyperpolarization. Synaptic GABAARs have a low affinity for GABA, but respond rapidly to fast increases in cleft GABA concentrations up to 1.5–
3 mM that decay within a few hundred microseconds (Mozrzymas et al., 2003). Accordingly, in the hippocampus fast synaptic inhibition is mediated, in part, by α1 subunit-containing
GABAARs on the dendrites and α2 subunit-containing GABAARs on the soma (Prenosil et al.,
2006). Additionally, in cerebellar granule cells the α1, α6, β2/3 and γ2 subunits have been found to be concentrated in GABAergic Golgi synapses and largely contribute to phasic inhibition in these cells (Nusser et al., 1998). Specifically, the α1β2/3γ2, α2β2/3γ2 and α3β2/3γ2 are thought to contribute to phasic synaptic inhibition (Farrant and Nusser, 2005). However, these receptors are also found extrasynaptically and it is interesting to note that a population of GABAARs that solely localizes to the synaptic density has not been found (Birnir and Korpi, 2007).
2.3.3.2 Extrasynaptic GABAA receptors and tonic inhibition
Whereas phasic inhibition lasts for milliseconds, a slower form of GABAergic inhibition known as tonic inhibition persists for seconds and longer (Semyanov et al., 2004; Krook-Magnuson and
Huntsman, 2007). Aside from the synaptically expressed GABAARs, a small population of
GABAARs are primarily expressed in the extrasynaptic membrane where they display a high sensitivity to ambient levels of GABA in the extracellular space, generate a tonic inhibitory conductance and display a slower desensitization than synaptic GABAARs (Brickley et al., 2001;
Yeung et al., 2003; Caraiscos et al., 2004b; Glykys et al., 2008). By definition the tonic conductance is “always on” and constantly activated by extracellular GABA, which acts to
41 42 persistently inhibit the neuronal network (Glykys and Mody, 2007a). Various sources have been proposed for the GABA found in the extracellular space. These include: astrocytic release (Liu et al., 2000; Kozlov et al., 2006), reversal of GABA transporter (Gaspary et al., 1998), non- vesicular release as well as action potential-mediated release (Attwell et al., 1993; Brickley et al.,
1996; Bright et al., 2007). It has been reported that the main source of ambient GABA responsible for the tonic conductance in CA1 and dentate gyrus granule cells is action potential- dependent vesicular release (Glykys and Mody, 2007b). Blocking action potentials and Ca2+ entry into the cells significantly reduced spontaneous GABA currents and the amount of tonic inhibition although both currents were not abolished (Glykys and Mody, 2007b). However, it seems unlikely that vesicular release is the primary contributor to extracellular GABA concentrations given that a sizeable tonic current is present in the presence of action potential blockers without supplementing extracellular GABA concentrations (Bai et al., 2001; Prenosil et al., 2006; Ivanov et al., 2008).
The tonic form of GABAAR inhibition largely depends on the location and subunit composition of GABAARs and can be exploited for developing highly specific drug targets for the brain
(Whiting, 2003). For instance, δ subunit-containing GABAARs (δGABAARs) are distributed at extrasynaptic regions over the cell surface of cerebellar granule cells (Laurie et al., 1992;
Yamashita et al., 2006), dentate gyrus granule cells (Nusser and Mody, 2002; Zhang et al.,
2007), and thalamic neurons (Harrison, 2007; Jia et al., 2008). A tonic current in dentate gyrus granule cells was first characterized in rat hippocampal slices (Brickley et al., 1996). This tonic conductance was sensitive to application of the GABAAR blocker, bicuculline, as application of the antagonist caused a decrease in the holding current. This current was also selectively enhanced by the GAT1 inhibitor NO711, but insensitive to zolpidem (Brickley et al., 1996).
42 43
Interestingly, our laboratory first identified pharmacological tools that can distinguish between the synaptic and extrasynaptic GABAAR currents. SR-95531 is a GABAAR antagonist that, at low concentrations preferentially blocks the synaptic currents but not the extrasynaptic tonic conductance in CA1 pyramidal neurons (Bai et al., 1999; Yeung et al., 2003). Additionally, penicillin has been shown to selectively block the synaptic currents, while no change in the tonic conductance was observed (Yeung et al., 2003).
The δGABAAR has a high affinity and a slower rate of desensitization to GABA and other agonists (Farrant and Nusser, 2005). The δGABAAR is inefficient in coupling GABA binding to channel gating, indicating that GABA is a low-efficacy agonist at these receptors (Wei et al.,
2003). However, it is not intuitively obvious why GABA is a low-efficacy agonist at δGABAARs while their affinity for GABA is very high but this property may subserve pharmacological manipulation of δGABAARs for clinical use. In particular, neuroactive steroids potently enhance
δGABAAR-activity and specifically enhance a tonic inhibitory conductance in central neurons at concentrations that occur in vivo (Stell et al., 2003; Smith et al., 2007). Further, the neurosteroid- induced augmentation of δGABAARs is known to regulate excitability (Stell et al., 2003) and may be clinically important for epilepsy (Peng et al., 2004; Zhang et al., 2007) and post-partum depression (Maguire and Mody, 2008). Additional GABAAR subunits such as the α4 subunit are expressed in the dentate gyrus and thalamus and the α6 subunit is expressed in the cerebellum.
These subunits localize to the extrasynaptic space, partner with the δ subunit, generate a tonic conductance and display a high neurosteroid sensitivity (Brickley et al., 2001; Wisden et al.,
2002; Chandra et al., 2006; Jia et al., 2007; Smith et al., 2007). The focus of this thesis is a subclass of extrasynaptic GABAARs containing the α5 subunit, which are sensitive to regulation by benzodiazepines and not low concentrations of neurosteroids. These receptors are typically
43 44 found in the mammalian brain in the conformational arrangement of 2 α5, 2 β3, and 1 γ2 subunits (Sieghart and Sperk, 2002).
44 45
Figure 2.5. Schematic structure of the GABAA receptor.
(A) Each GABAA receptor subunit consists of a long N-terminal extracellular hydrophilic region, followed by four transmembrane (M1-4) α-helices with the M2 segment forming the lining of the ion channel and a large intracellular loop between M3 and M4. Each sequence ends with a relatively short extracellular C-terminal domain. (B) The GABAA receptor-channels are pentamers composed of several related subunits. The most prevalent of these combinations in the mammalian brain is 2 α1, 2 β2 and 1 γ2 subunits arranged around the central pore in a specific order. However, the δ subunit can replace the γ subunit in extrasynaptic GABAA receptors.
45 46
46 47
2.3.4 Location and physiological role of α5GABAAR subunits
2.3.4.1 Regional expression patterns
Regionally the distribution of α5GABAARs is relatively sparse in the mammalian brain. Overall, the total number of GABAARs that contain the α5 subunit is approximately 5% (Sur et al.,
1999). In the hippocampus it is estimated that 20-25% of GABAARs contain an α5 subunit (Sur et al., 1999). This has lead to many studies exploring the unique possibility that α5GABAARs possess cognitive modifying properties (Collinson et al., 2002; Crestani et al., 2002; Yee et al.,
2004; Collinson et al., 2006). Specifically, in the hippocampus the α5 subunit is localized to the stratum radiatum and stratum oriens of the CA1 and CA3 regions (Pirker et al., 2000). These receptors are expressed on the apical and basal dendrites in both regions, where they may have slightly different functions (Haley et al., 1996). For instance, the threshold to induce LTP is considerably higher in the apical dendrites as compared with the basal dendrites. This difference has been attributed to the high density of inhibitory interneurons that primarily synapse onto the apical dendrites (Kaibara and Leung, 1993). In line with this observation, tonically active receptors are highly expressed on inhibitory neurons in the hippocampus and inhibition of these receptors could lead to disinhibition and increased LTP (Semyanov et al., 2003; Walker and
Semyanov, 2008).
Aside from its high expression in the hippocampus, the distribution of the α5 subunit is quite sparse in other brain regions, with the olfactory bulbs an obvious exception (Pirker et al., 2000).
Intense α5 subunit-immunoreactivity (IR) has been detected in the olfactory bulb (external plexiform layer and internal granular layer), inner layers of the cerebral cortex, endopiriform nucleus, subiculum and ventromedial hypothalamic nucleus (Pirker et al., 2000).
Correspondingly, the regional distribution of [3H]L-655,708, an isotope-labelled ligand that
47 48
predominantly labels α5GABAARs, strongly correlates with the immunocytochemical distribution of the α5 subunit (Sur et al., 1999). The precise function of the α5 subunit in these other brain regions has not been extensively studied. Strikingly, the number of GABAARs that contain an α5 subunit in the internal granule cell layer of the olfactory bulb is an incredible 35%
(Sur et al., 1999), yet we do not know their exact function in this region. It can be assumed that the basic physiologic function of α5GABAARs in this region is similar to other brain regions, with this receptor generating shunting inhibition, which regulates intrinsic neuronal excitability
(Caraiscos et al., 2004b; Bonin et al., 2007; Glykys et al., 2008). With this in mind, shunting inhibition has been shown to modulate gain during synaptic excitation (Mitchell and Silver,
2003) and α5GABAARs may be involved in boosting the signal-to-noise ratio of odor signals by silencing the basal firing rate of surrounding non-activated neurons. Thus, it is conceivable that the inhibition generated by α5GABAARs could inhibit weakly stimulated adjacent mitral cells to enhance discrimination or sensitivity between odours, and may also act to filter out background odours to enhance the transmission of a few select odours. For this view to be correct,
α5GABAARs would have to be located to extrasynaptic sites as suggested in other brain regions, which has not been determined, but this is a plausible explanation for the role of α5GABAARs in the olfactory system (Brunig et al., 2002).
2.3.4.2 Neuronal compartmentalization
There is conflicting evidence and differing views regarding the neuronal compartmentalization of α5GABAARs. Immunocytochemistry and in situ hybridization studies indicate that
α5GABAARs are localized primarily to extrasynaptic regions of pyramidal neurons and they may not be present at GABAergic synapses (Brunig et al., 2002). However, in the cultured hippocampal system using immunofluorescence techniques, Christie and de Blas (2002),
48 49
reported the occurrence of α5GABAARs not only extrasynaptically but also in GABAergic synapses on pyramidal cells. In this study, large α5 clusters co-localized with presynaptic
GABAergic terminals and large postsynaptic gephyrin clusters in approximately 50 % of neurons. In the same study, α5GABAARs were also expressed at the synaptic sites of interneurons, albeit to a lesser extent (< 10 % of neurons). It has also been shown by using electron microscopy immunogold labeling methods that in the intact hippocampus preparation,
α5GABAARs are located on the dendrites of pyramidal neurons in the CA1 region of the rat hippocampus and cerebral cortex, where they are expressed both extrasynaptically and synaptically (Serwanski et al., 2006). Consistent with data suggesting that α5GABAARs are expressed extrasynaptically, these receptors are known to localize at dendritic spines that receive glutamatergic input (Serwanski et al., 2006), which make them ideal targets for modifying LTP.
α5GABAARs much like putatively synaptic receptors appear in cluster formations on neuronal surface membranes and contain a γ2 subunit (Sur et al., 1998). Importantly, co-localization of the
γ2 subunit and the scaffolding protein gephyrin have been shown to be interdependent components of the same synaptic complex that is critical for the clustering of abundant GABAAR subtypes at inhibitory synapses (Essrich et al., 1998). However, the deletion of gephyrin in mutant mice does not alter the punctate staining of α5GABAARs (Kneussel et al., 2001). Instead the ERM (ezrin, radixin, moesin) protein radixin has been shown to be critical for α5GABAAR anchoring to the F-actin cytoskeleton (Loebrich et al., 2006). In this study both radixin and
α5GABAARs were found to be mainly located at extrasynaptic sites (> 75 %), but also, albeit to a lesser extent, expressed synaptically (< 25 %). It was determined by site-directed mutagenesis that the binding of radixin to F-actin and α5GABAARs is a phosphorylation dependent process.
The authors proposed that α5GABAAR clusters and/or radixin-α5GABAAR co-clusters might
49 50 shuttle between extrasynaptic and synaptic sites in a phosphorylation-dependent manner. This is a possibility, since, α5GABAARs are highly mobile and capable of moving between the synaptic and extrasynaptic space (Triller and Choquet, 2005; Jacob et al., 2008).
2.3.4.3 Electrophysiological characteristics
Despite the imaging and biochemical data, the independent electrophysiological examination of
α5GABAARs has yielded inconclusive results regarding the role of α5GABAARs in synaptic or extrasynaptic neurotransmission. For instance, in neocortical pyramidal cells the IPSPs elicited by bistratified dendrite-preferring, but not multipolar adapting and nonadapting interneurons are reduced by selectively inhibiting α5GABAARs with the benzodiazepine inverse agonist IAα5
(Ali and Thomson, 2008). This suggests that in the cortex, α5GABAARs are expressed synaptically but this may depend on input specificity and connectivity (i.e. dendrite preferring versus non-preferring). Correspondingly, miniature IPSCs resulting from the activation of synaptic α5GABAARs have been recorded in the cerebral cortex and the decay times of the
α5GABAAR-activated mIPSCs were slower than those receptors containing the α1 subunit
(Dunning et al., 1999). Further, in hippocampal slices prepared from α5 subunit point mutant mice (α5H105R), the amplitude of larger sIPSCs and eIPSCs are not enhanced by diazepam application, suggesting that α5GABAARs contribute to a small subset of larger IPSCs
(Zarnowska et al., 2009). The α5H105R mutant mice are interesting because they were generated to reduce benzodiazepine sensitivity of α5GABAARs, yet they exhibited a vast reduction in the number of α5GABAARs in the hippocampus (Crestani et al., 2002) with no obvious differences in baseline synaptic transmission (Zarnowska et al., 2009).
Conversely, there are also impressive electrophysiological data that indicate that these receptors
50 51 generate an extrasynaptic tonic conductance. In CA1 pyramidal neurons the amplitude of the bicuculline-sensitive tonic conductance is significantly reduced in null mutant mice with a genetic deletion of the α5GABAAR (Gabra5–/– mice), while there are no differences in sIPSCs when compared to WT control slices (Caraiscos et al., 2004b). This tonic conductance is sensitive to midazolam (Bai et al., 2001; Yeung et al., 2003) but insensitive to zolpidem
(Caraiscos et al., 2004b) suggesting the presence of an α5 and a γ subunit (Pritchett and Seeburg,
1990). In line with this evidence, selectively inhibiting α5GABAARs with low concentrations of the benzodiazepine inverse agonist L-655,708 did not alter sIPSCs in the CA1 and CA3 regions of the hippocampus suggesting that phasic currents are not readily influenced by α5GABAAR modification (Caraiscos et al., 2004b; Glykys et al., 2008). At low concentrations, the general anesthetic etomidate does not alter the frequency or peak amplitude of sIPSCs in hippocampal pyramidal neurons but enhances a tonic conductance in WT but not Gabra5–/– neurons (Cheng et al., 2006), suggesting different underlying receptor subtypes. However, in the hippocampus, the amplitude of evoked IPSCs is reduced and decay times are slowed in slices prepared from
Gabra5–/– mice when compared to WT control slices indicating that deletion of α5GABAARs may reduce evoked synaptic responses (Collinson et al., 2002). Despite the overwhelming evidence indicating that α5GABAARs generate this tonic conductance in CA1 pyramidal neurons, there are reports indicating that a small component of this tonic conductance may be mediated by δGABAARs (Glykys et al., 2008). A schematic drawing illustrating the methods and drug applications used to determine the presence of the tonic conductance in pyramidal neurons is shown in Figure 2.6.
51 52
Figure 2.6. A schematic drawing illustrating the methods and drug applications used to determine the presence of the tonic conductance.
Spontaneous electrophysiological responses are recorded in voltage clamp mode. The downward deflections represent the action potential-independent mIPSCs. Application of bicuculline results in an outward shift from the baseline current (IHold). Note the blockade of the mIPSCs when bicuculline is applied. Removal or washout of bicuculline allows the current to return to baseline.
In α5GABAARs, benzodiazepines potentiate a tonic current and this is represented by an inward change in the holding current. Note that the mIPSCs are still present and are also enhanced with benzodiazepine application. The tonic current mediated by δGABAARs is insensitive to benzodiazepines and in this example would not be potentiated.
52 53
53 54
2.3.4.4 Oscillatory activity
Despite the high proportion of α5GABAARs in the CA3 region, the ability of these receptors to modify mossy fiber LTP in this area has not been studied. This is somewhat surprising given the high concentration of zinc in the synaptic vesicles of mossy fiber terminals (Ketterman and Li,
2008) and the propensity for zinc to reduce responses to GABA that are mediated by those containing the α5 subunit (Burgard et al., 1996). Instead, studies examining the α5 subunit in the
CA3 region have investigated their role in the oscillatory activity of neuronal networks (Towers et al., 2004; Glykys et al., 2008). It has been shown that the presence of a tonic GABAergic conductance regulates interneuronal firing properties in a rhythmic oscillatory manner
(Scanziani, 2000), such that the power of gamma frequency oscillations are reduced when the tonic GABAergic conductance is increased (Whittington et al., 1996). Specifically, in slices prepared from null mutant mice for α5GABAARs (Gabra5–/– mice) the frequency and power of gamma frequency oscillations are increased in the CA3 region (Towers et al., 2004; Glykys et al., 2008). Similarly, after application of L-655,708, an α5GABAAR-selective benzodiazepine inverse agonist (Quirk et al., 1996), gamma frequency oscillations emerged in WT slices that previously did not have detectable spontaneous activity (Glykys et al., 2008). These results suggest that in the absence of tonic inhibitory currents in CA3 pyramidal cells, spontaneous gamma oscillations are present. Still, it remains to be determined whether the primary function of
α5GABAARs in the CA3 region is to coordinate and regulate neuronal coherence or to actively modify CA3-mossy fiber LTP.
2.3.5 Pharmacology of α5GABAAR subunits
In this next section, the various pharmacological agents that have been shown to distinguish or differentiate the α5GABAARs from other GABAARs will be discussed. Full activation of the
54 55
GABAAR requires the binding of two molecules of GABA (Baumann et al., 2003). This activation causes the receptors to undergo conformational changes that result in transitions between various open, closed and desensitized states (Chupakhin et al., 2006). Traditionally,
α5GABAARs can be distinguished from other benzodiazepine sensitive α-containing GABAARs by the imidazopyridine, zolpidem (Pritchett and Seeburg, 1990; Faure-Halley et al., 1993).
Zolpidem has a high affinity for α1 subunit-containing GABAARs (Ki = 19 nM), lower affinity for α2 (Ki = 450 nM) and α3 (Ki = 398 nM) containing GABAARs and a very low affinity for
α5GABAARs (Ki > 15,000 nM) (Pritchett and Seeburg, 1990). The unique pharmacological response of α5GABAARs is considered below.
2.3.5.1 Benzodiazepines and inverse agonists
GABAARs are important sites for a number of clinically important drugs and their interaction with benzodiazepines has been the most thoroughly investigated. Benzodiazepine pharmacology is determined by the binding in the interface of the α and γ subunits. Classical benzodiazepines such as diazepam and flunitrazepam predominately mediate their effects through interactions with α1βγ2, α2βγ2, α3βγ2 or α5βγ2 containing GABAARs (Sigel, 2002). They have reduced activity at receptors containing γ1 or γ3 subunits and no activity at receptors containing α4 or α6 subunits (Hevers and Lüddens, 1998 ). In this regard, benzodiazepines such as midazolam and diazepam do not distinguish between the benzodiazepine sites of different GABAAR-subtypes and have non-specific actions (Pritchett and Seeburg, 1990). In contrast, benzodiazepines including quazepam and cinolazepam (Sieghart, 1989) and non-benzodiazepine ligands including zolpidem have been developed with a high specificity for α1βγ2 containing GABAARs (Korpi et al., 2002). Further, receptors containing α2 or α3 subunits have a moderate affinity for zolpidem, while α5GABAARs have low to no affinity for this drug (Olsen and Sieghart, 2008).
55 56
Interestingly, the benzodiazepine subunit target highly correlates with in vivo alterations for the behavioural actions of benzodiazepines (Rudolph et al., 1999). Specifically, results have shown that the α1 subunits contribute to the sedative and partly to the anticonvulsant effects of
GABAARs (Rudolph et al., 1999). The α2 subunits mediate the anxiolytic activity of diazepam
(Low et al., 2000), while the α3 subunits are implicated in anxiolysis (Dias et al., 2005) and anticonvulsant activity of GABAARs (Fradley et al., 2007). The α5 subunits likely contribute to the amnestic properties of GABAARs (Collinson et al., 2002; Crestani et al., 2002).
α5GABAARs are interesting drug targets, due to their low overall expression level and highly localized distribution pattern (Sur et al., 1999; Pirker et al., 2000). Specifically, compounds known as benzodiazepine inverse agonists that exhibit a high specificity for α5GABAARs have been developed and possess cognitive enhancing properties (Atack et al., 2006a; Atack et al.,
2006b; Atack et al., 2006c; Dawson et al., 2006; Ballard et al., 2009).
Inverse agonists are drugs that bind to the agonist-binding site, in this case the benzodiazepine site, and act to reduce the receptor’s basal activity (Kenakin, 2004; Vilardaga et al., 2005). Non- selective benzodiazepine inverse agonists act to reduce the overall GABAAR activity, thereby producing anxiogenesis and tonic convulsions and are thus not suited for clinical use (Petersen,
1983). However, inverse agonists with high affinity for α5GABAARs do not display these adverse properties and instead have cognitive enhancing effects. The compound L-655,708 binds to the benzodiazepine site of the α5GABAAR (Quirk et al., 1996), and acts to reduce GABA- evoked currents (Casula et al., 2001). L-655,708 is a partial inverse agonist with preference for
α5GABAARs that is 107-fold, 61-fold and 54-fold more sensitive to GABAARs containing the
α1, α2, or α3 subunits, respectively (Quirk et al., 1996). In vivo rat pharmacokinetic analyses has shown that L-655,708 has a relatively short half-life (0.5 h) with kinetics in the brain
56 57 mirroring those in the blood plasma (Atack et al., 2006c). In vivo binding experiments showed that plasma concentrations of around 100 ng/ml gave relatively selective in vivo occupancy of rat brain α5- versus α1-, α2-, and α3-containing GABAARs. Intraperitoneal injections of 1 mg/kg of
L-655,708 yield a 64 % receptor occupancy of α5GABAARs versus 18 % occupancy at α1, α2, or α3 subunits (Atack et al., 2006c). L-655,708 also displays memory-enhancing properties at concentrations that are reported not to be proconvulsant (Atack et al., 2006b) but it is reported to have mild anxiogenic properties (Navarro et al., 2002). L-655,708 acts to reduce the
α5GABAAR-mediated tonic conductance in hippocampal neurons, in vitro (Scimemi et al., 2005;
Glykys et al., 2008). L-655,708 has proven to be a useful molecular tool for labeling
3 α5GABAARs with its radioactive form [ H]L-655,708 (Quirk et al., 1996). α5IA is another benzodiazepine inverse agonist with high selectivity for α5GABAARs that enhances the performance of rats in hippocampus-dependent behavioural tasks (Chambers et al., 2003;
Chambers et al., 2004). Both L-655,708 and α5IA have been shown to enhance LTP in hippocampal slices at concentrations that are selective for α5GABAARs and do not overtly enhance slice excitability (Atack et al., 2006b; Collinson et al., 2006). These benzodiazepine inverse agonists selective for α5GABAARs may prove useful for treating memory-impairing disorders. Their cognitive enhancing properties will be discussed in detail in Chapter 2.3.6.
2.3.6 α5GABAA receptors as a candidate target for the modification of
hippocampus-dependent learning and synaptic plasticity
Behavioural assays, used together with genetic and pharmacological interventions have demonstrated a role for α5 subunit-containing GABAARs in a number of physiological processes. These include the modification of learning and memory (Collinson et al., 2002; Cheng
57 58 et al., 2006; Dawson et al., 2006), rewarding/seeking behaviours (McKay et al., 2004; Cook et al., 2005) and sensorimotor gating (Hauser et al., 2004) with no direct influence on anxiogenesis, sedation or seizure activity (Collinson et al., 2002; Atack et al., 2006b). The significant and advantageous role of α5GABAARs in learning and memory was initially investigated by generating Gabra5–/– mutant mice (Collinson et al., 2002) and a mouse strain that expressed a point mutation at position 105 of the α5 subunit (α5H105R) (Crestani et al., 2002). The
α5H105R mice display no overt phenotypic abnormalities, have a normal life span, breed normally and do not display overt epileptic activity. They also have normal motor performance and coordination, which has been further supported by data in other laboratories (Savic et al.,
2008). Since α5GABAARs are predominantly localized to the hippocampal stratum radiatum
(Sur et al., 1999), these mouse-models have been useful in determining the role of α5GABAARs in learning and memory.
The genetic deletion or reduction of α5GABAARs as observed in Gabra5–/– and α5H105R mutant mice, respectively correspond to enhanced hippocampus-dependent memory behaviours whereas non-hippocampus-dependent behaviours are unaffected (Collinson et al., 2002; Crestani et al., 2002). Specifically, Gabra5–/– mice display enhanced acquisition in the matching to place version of the spatially-mediated hippocampus-dependent water maze task (Collinson et al.,
2002). However, in the two-way active avoidance paradigm (Gray, 1982), there are no differences in performance between WT and Gabra5–/– mice (Collinson et al., 2002). Similar to
Gabra5–/– mice, α5H105R mutant mice express enhanced performance in the hippocampus- dependent trace fear conditioning paradigm but perform similar to WT littermates in the non- hippocampus-dependent cued fear conditioning protocol (Crestani et al., 2002). The procedure for trace fear conditioning is similar to cued fear conditioning in that a set number of tone-shock
58 59 pairings are presented. However, the two protocols differ because in trace fear conditioning an empty trace interval (i.e. time separation between auditory tone and aversive stimulus) is interposed between the tone and the foot shock, where as in cued fear conditioning the foot shock co-terminates with the cessation of the tone. It has been shown that in order to enhance the hippocampus-dependency of the trace fear conditioning protocol, the trace interval should be within the range of 15–30 s (Quinn et al., 2002; Chowdhury et al., 2005; Misane et al., 2005). In the Crestani (2002) study the trace interval used was brief (1 s), and it is possible that the enhanced performance of the α5H105R mutant mice may not be directly related to the hippocampus. With such a short trace interval, the amygdala or other extrahippocampal regions may play a more dominant role (Lee et al., 2001; Quinn et al., 2002; Chowdhury et al., 2005). A candidate brain region for the enhanced trace fear conditioning performance in the α5H105R mice may be the transition zone between the hippocampus and the amygdala and not necessarily the hippocampus. It is interesting that in the amygdalohippocampal zone, 19% of all GABAARs contain an α5 subunit (Sur et al., 1999). The amygdalohippocampal region is suggested as an area of interest due to the high involvement of the amygdala in fear responses (Kim et al., 1993;
Maren, 2001). Despite the enhanced learning and memory behavioural phenotypes in Gabra5–/– and α5H105R mutant mice, there have been no obvious differences for the induction of LTP in either genotype with tetanic stimulation protocols (Collinson et al., 2002; Crestani et al., 2002).
2.3.6.1 Memory and LTP enhancing properties of benzodiazepine inverse agonists
The concept that GABAARs modify learning and memory is not new, as bicuculline, a non- selective blocker of GABAARs enhances memory processes (Brioni and McGaugh, 1988).
Similarly, non-selective benzodiazepine inverse agonists have been shown to enhance cognitive performance in animal models (Bernston et al., 1996) but these drugs are anxiogenic (Dorow et
59 60 al., 1983), convulsant (Petersen, 1983), proconvulsant (Nutt et al., 1984) or may alter additional processing (Sarter et al., 2001). As mentioned in Chapter 2.3.5.1 the benzodiazepine class of inverse agonists selective for α5GABAARs has great clinical utility because of their relatively low occurrence of unwarranted side effects (Maubach, 2003). Selectively inhibiting
α5GABAAR-activity with the benzodiazepine-like inverse agonist, α5IA, improves water maze learning and LTP under certain patterns of stimulation (Dawson et al., 2006). This drug does not have convulsant, proconvulsant or anxiogenic properties with an in vivo benzodiazepine receptor occupancy of greater than 90 % (Dawson et al., 2006).
Further, a study characterizing the use of α5IA-II, an improved version of α5IA, demonstrated that α5GABAARs may be important for the encoding and recall but not the consolidation of spatial information (Collinson et al., 2006). In particular, injections of the α5GABAAR-selective benzodiazepine inverse agonist α5IA-II enhanced performance in the water maze when injected either before training or immediately before testing, suggesting an effect of α5GABAARs on the encoding and recall phases of memory. α5IA-II had no effect on performance when given immediately following training indicating that α5GABAARs are not important for the consolidation phase of memory. In contrast, the non-selective benzodiazepine agonist CDP impaired performance when given during the encoding and recall phase, whilst having no effect on consolidation (Collinson et al., 2006).
It has been shown that a subcutaneous implantation of a slow release pellet formulation of L-
655,708 enhanced both the acquisition of the time required to find the platform in the water maze and also the time spent in the correct quadrant of the pool during the water maze probe trial
(Atack et al., 2006b). This is congruent with the importance of α5GABAARs for the encoding
60 61 and recall of spatial information. A concentration of 10 nM L-655,708, which preferentially binds to α5GABAARs in tissue slices, enhanced TBS LTP (Atack et al., 2006b). This is interesting because in Gabra5–/– mice there is no increase in LTP with TBS (Collinson et al.,
2002). However, it should be noted that in the study by Atack et al., (2006b) a conditioning stimulus of 10 pulses presented at 100 Hz was given 30 min before TBS. Following the conditioning stimuli there were no differences between the vehicle and L-655,708-treated slices despite the significant enhancement of synaptic responses. This result gives credence to the possibility that the type of synaptic stimulation may influence the responses or involvement of
α5GABAARs in LTP and possibly even learning.
The correlative studies examining the relationship between in vitro actions and behavioural consequences of inverse agonists are difficult to interpret and can often be misleading. RY-080 is a benzodiazepine inverse agonist that has 40 to 50-fold higher in vitro affinity for α5GABAARs compared with α1GABAARs (Liu et al., 1995) but it is reported to be proconvulsant in mice (Liu et al., 1996). However, follow-up studies have determined that the in vivo preference of RY-080 is only 7 to 10-fold higher than the α1 and α2/α3 subunits, suggesting that the proconvulsant effects of α5GABAAR selective inverse agonists may not be mediated by the α5 subunit (Atack et al., 2006a). Interestingly, inverse agonists may also prove to be useful for reversing pharmacologically-induced memory impairment. In human volunteers, α5IA-II has been shown in humans to reverse the memory-blocking effects of ethanol during word-list learning (Nutt et al., 2007). While inverse agonists including, L-655,708 and α5IA are not currently available for clinical use, these compounds may serve as prototypes for drug development to treat cognitive diseases and pharmacological memory impairments.
61 62
The findings regarding α5GABAARs are important because they suggest that a single population of GABAARs can be targeted to specifically alter learning and memory processes without the adverse consequences of decreasing overall inhibition (see Ben-Ari, 2006 for a review). This has led other laboratories, including our own to further investigate the role of α5GABAARs in learning and memory. General anesthetics including etomidate (Cheng et al., 2006) and isoflurane (Caraiscos et al., 2004a) potentiate an α5GABAAR mediated tonic conductance in
CA1 pyramidal neurons and classically disrupt memory processes (Simon et al., 2001; Cheng et al., 2006). As will be elucidated in Chapter 4, Gabra5–/– mice are resistant to the amnestic properties of etomidate but not the hypnotic or sedative effects of the drug (Cheng et al., 2006).
Additionally, scopolamine-induced memory impairment is reversed with BiRY-080, a novel inverse agonist with 130-fold selectivity for α5GABAARs (Harris et al., 2008). This finding has important implications for treating disorders such as Alzheimer’s disease, as scopolamine is a muscarinic antagonist, and the depletion of cholinergic cells is a major contributor to the cognitive-impairing behaviours of dementia (Geula et al., 2008). Recently, high efficacy agonists selective for α5GABAARs have been developed that enhance the GABA-evoked currents in α5 subunit expressing oocytes by greater than 300 % over control currents (Savic et al., 2008). In particular, the α5GABAAR-selective agonist SH-053-R-CH3 impairs contextual fear memory
(Savic et al., 2008), similar to the effects of etomidate (Cheng et al., 2006). It has also been shown that reducing α5GABAAR expression with an antisense oligodeoxynucleotide that targeted the α5 subunit did not reverse midazolam-induced contextual fear conditioning impairment (Gafford et al., 2005) although this result may simply be related to the involvement of α5GABAARs in the different stages of memory. In the aforementioned oligodeoxynucleotide experiments, midazolam was injected during the consolidation phase of memory and it has
62 63
already been demonstrated that α5GABAARs are not important for the consolidation of memory
(Collinson et al., 2006). These strong pharmacological data provide further “proof of principle” that compounds displaying selective influence on α5GABAAR isoforms have significant memory modifying effects and great clinical utility.
2.3.7 A role for α5GABAARs in neurological disorders
Despite the dominant role α5GABAARs play in the learning and memory process, these receptors are also attractive candidates for involvement in additional neurobehaviours. In particular, this section will briefly discuss the possible involvement of α5GABAARs in the addicting properties of ethanol and in pathological CNS disorders including schizophrenia, autism and epilepsy.
2.3.7.1 The neurobehavioural properties of α5GABAA receptors in ethanol-reward and
addiction
Congruent with data suggesting that α5GABAARs are involved in the memory-impairing properties of ethanol (Nutt et al., 2007), these receptors have also emerged as targets for the rewarding-reinforcing properties of ethanol (June et al., 2003; Cook et al., 2005). The benzodiazepine-like inverse agonist RY 023 (RY) exhibits both a high affinity (Ki of ~ 2.7 nM) and high selectivity (~ 75-fold) at recombinant GABAARs composed of α5β2γ2 subunits (Liu et al., 1996). RY has been shown to suppress ethanol-maintained responding in a lever-pressing task (June et al., 2001) and ethanol-induced sedation (Cook et al., 2005). Further, it has been confirmed that the benzodiazepine site mediates the ethanol effect as the non-specific competitive β-carboline benzodiazepine antagonist ZK 93426 reversed the RY-induced suppression of ethanol-maintained responding (Cook et al., 2005).
63 64
However, studies that used Gabra5–/– mice indicate that α5GABAARs are not essential for the rewarding properties mediated by ethanol (Stephens et al., 2005). It was demonstrated that
Gabra5–/– mice did not differ from WT littermates in operant responding for 10 % ethanol- containing solutions (Stephens et al., 2005). Whereas in the same study, the inverse agonist
α5IA-II decreased ethanol-maintained responding in WT mice at a 27-fold lower dose than
Gabra5–/– mice. Taken together these results suggest that benzodiazepine inverse agonists selective for α5GABAA receptors are instrumental for suppressing ethanol-mediated responding, although the role of these same receptors in signalling ethanol reward is less convincing.
2.3.7.2 Schizophrenia
Alterations in individual subunits of GABAARs in schizophrenia have been the focus of recent postmortem studies. The expression of the α1 and α2 subunits has been reported to increase in the prefrontal cortex of patients with schizophrenia (Ohnuma et al., 1999; Ishikawa et al., 2004) with no significant change reported for the expression of the α5 subunit (Akbarian et al., 1995)
(Impagnatiello et al., 1998). The ligand [11C]Ro15-4513 binds with high affinity to the benzodiazepine binding site and has been used to quantify the in vivo location and distribution of the α5 subunit using positron emission tomography (PET). There are no differences between control subjects and schizophrenic patients when the ligand [11C]Ro15-4513 is used to visualize the α5 subunit during PET scans. However, the expression of the negative symptoms in schizophrenia is negatively correlated with the expression of the α5 subunit in the prefrontal cortices and the hippocampus (Asai et al., 2008).
Further, animal models used to study schizophrenic-like behaviours have focused on the
64 65 behavioural paradigm, prepulse inhibition (PPI). PPI is a neurological phenomenon in which a weaker prestimulus (prepulse) inhibits the reaction of an organism to a subsequent stronger startling stimulus (pulse). Deficits of PPI manifest in the inability to filter out unnecessary information and they have been linked to abnormalities of sensorimotor gating (Swerdlow et al.,
2008). Such deficits are noted in patients suffering from illnesses such as schizophrenia (Kumari et al., 2007; Takahashi et al., 2008), obsessive-compulsive behaviour (Braff et al., 2001), and dementia (Perriol et al., 2005). Notably, the hippocampus has been found to play a role in sensorimotor gating and the reduction of α5 subunit containing GABAA receptors in the hippocampus is sufficient to attenuate PPI and enhance spontaneous locomotor activity in the open field (Hauser et al., 2004). Interestingly, excitotoxic neonatal lesions of the ventral hippocampal formation in rats reproduce numerous aspects of schizophrenia and messenger
RNA (mRNA) expression is increased for the α1, α5 and γ2s subunits in the cortices of the frontal pole (Mitchell et al., 2005). It still remains to be determined whether pharmacological interventions designed to activate α5GABAARs possess any antipsychotic-like properties. The benefits of such drugs may be useful given the above evidence that suggests that the expression of the α5 subunit is decreased in human patients, and animal models mimicking this disease.
2.3.7.3 Autism Spectrum Disorders:
Given the data supporting the role α5GABAARs play in learning and memory processes, these receptors have been investigated for their contributions to developmental disabilities such as autism and fragile X syndrome, which typically result in varying degrees of intellectual and cognitive impairment (York et al., 1999). Specifically, there is a strong correlation among autism, bipolar disorder and special abilities (typically increased intellect) in families of children with autism (Delong, 2007). Genetic studies implicate linkage of one autism subgroup to the
65 66
chromosome 15q11–13 region, especially to the GABAA receptor subunit genes Gabra3 and
Gabra5 (McCauley et al., 2004).
Interestingly, a subgroup of families with autism-like behaviours, show a high coincidence of major affective disorder, special talents and intellectual abilities that suggests the parent-of- origin may determine gain-of-function, such that maternal transmission may result in an increase in memory, talent or specific intellectual abilities (Delong, 2007). Alternatively, inheritance, as seen with Prader-Willi syndrome, is not expressed with maternal transmission and is fully expressed with paternal transmission (Horsthemke and Wagstaff, 2008). Further, any complete theory of autism must take into account aspects of enhanced memory and learning ability, to explain not only savant characteristics and superior visual-spatial abilities (Caron et al., 2004) but also the notable intellectual abilities found in family members (especially fathers) of individuals with autism (Dor-Shav and Horowitz, 1984; DeLong et al., 2002). The savant abilities of affected individuals takes on a special significance given that mutant mice with a genetic deletion corresponding to the human chromosome 15q11–13, display enhanced learning.
Gabra5–/– mice have enhanced learning and memory in the hippocampus-dependent water maze model of spatial learning (Collinson et al., 2002) and α5H105R mutant mice demonstrate facilitation of associative learning in trace fear conditioning (Yee et al., 2004). Further, ligand binding studies show that the GABAergic receptor system is significantly and specifically reduced in the hippocampus in autism (Blatt et al., 2001; Guptill et al., 2007), but it is not known whether this specifically affects the α5 subunit. The complexities of autism go beyond the scope of this chapter, but it is important to emphasize that a component of this disorder may involve
α5GABAARs.
66 67
2.3.7.4 Seizure disorder
Tonic neuronal inhibition has been shown to play a critical role in excitability (Semyanov et al.,
2003) and information processing (Chadderton et al., 2004). However, Gabra5–/– mice do not display overt convulsant activity (Collinson et al., 2002), and inverse agonists with a high preference for α5GABAARs are not convulsant or proconvulsant (Atack et al., 2006b; Dawson et al., 2006). These data suggest that α5GABAARs may not be intimately or immediately involved in regulating epileptogenesis or seizure activity. This is striking since the tonic conductance associated with δGABAARs is involved in the regulation of excitability and α5GABAARs have been shown to regulate the intrinsic excitability of pyramidal neurons (Bonin et al., 2007). Thus, the adaptive role of a tonic conductance generated by either δGABAARs or α5GABAARs may be to reduce excessive excitation and neuronal excitability (Stell et al., 2003; Glykys and Mody,
2006; Bonin et al., 2007). This has prompted speculation that reduced tonic inhibition of pyramidal neurons contributes to epileptogenesis, but it still remains unknown why there is no epilepsy phenotype when α5GABAARs are genetically or pharmacologically reduced.
In animal models of epilepsy, GABAARs are known to undergo a significant conformational change (Hu et al., 2008) and differential expression of GABAAR subunits (Fritschy et al., 1999;
Eugene et al., 2007). Specifically, in the pilocarpine model of temporal lobe epilepsy,
α5GABAAR immunoreactivity is decreased in the CA1, CA3 and entorhinal cortices (Fritschy et al., 1999; Houser and Esclapez, 2003; Scimemi et al., 2005) but this decrease correlates well with the generalized loss of pyramidal cells (Fritschy et al., 1999). Despite the downregulation of
α5GABAARs, the tonic current from epileptic rats was unchanged from non-seized rats in CA1 pyramidal cells (Scimemi et al., 2005). The tonic current in the CA1 region of pilocarpine-seized rats displays an increased sensitivity to low concentration of the δGABAAR-preferring
67 68 neurosteroid allotetrahydrodeoxycorticosterone (THDOC) (Stell et al., 2003) suggesting that during seizure activity, alternative GABAAR subunits can compensate for the loss of
α5GABAARs.
Interestingly, in pilocarpine-treated rats the normal diffuse labeling of δGABAARs in the dentate molecular layer was decreased four days following status epilepticus and remained low throughout the period of chronic spontaneous seizures. In contrast, the expression of the α4 subunit, which routinely partners with the δ subunit, was increased along with the expression of the γ2 subunit (Peng et al., 2004). This is interesting since it has also been reported that
α4βxδ partnership is replaced by an α4βxγ2 partnership following status epilepticus in rats (Hu et al., 2008). Additional compensation in the dentate gyrus during status epilepticus has been shown to be mediated in part by an up-regulation of α5GABAARs (Fritschy et al., 1999) suggesting that a significant restructuring of subunit specific GABAARs takes place in the brains of epileptic individuals. Although this brief section has only touched upon some of the subunit specific alterations involving extrasynaptic receptors, it should not be overlooked that alterations in synaptic GABAARs is also known to significantly contribute to the post-seizure phenotype
(Poulter et al., 1999; Nishimura et al., 2005; Zhang et al., 2007).
2.4 Anesthesia
In our laboratory, one of the main interests is the effects of general anesthetics on GABAergic neurotransmission. In this thesis, general anesthetics were used as pharmacological probes to enhance the tonic conductance in CA1 pyramidal neurons. Next, general anesthetics will be discussed with respect to their pharmacological properties, targeting of specific GABAAR
68 69 subtypes, role in learning and memory and the problems associated with anesthetic-induced amnesia.
2.4.1 In vitro actions of general anesthetics
General anesthetics produce widespread neuronal depression and a range of multiple behavioural endpoints, including sedation, amnesia, hypnosis, analgesia and immobility (Campagna et al.,
2003). These agents were once thought to mediate their actions through a non-specific disruption of the neuronal lipid bilayer causing global depression (Rudolph and Antkowiak, 2004). It is now clear that general anesthetics enhance neuronal inhibition and mediate their specific effects through membrane bound receptors; in particular, the GABAAR is considered a prime target for numerous general anesthetics (Campagna et al., 2003; Hemmings et al., 2005; Bonin and Orser,
2008). In vitro studies have shown that the intravenous general anesthetics etomidate and propofol mediate their actions at the GABAAR through allosteric interactions (Hill-Venning et al., 1997; Krasowski et al., 1998). Both of these drugs prolong the duration of GABAergic IPSCs in hippocampal neurons by increasing the Cl- channel open probability and frequency (Orser et al., 1994; Bai et al., 1999).
Propofol has also been shown to increase the extrasynaptic tonic conductance in hippocampal pyramidal neurons with a increased sensitivity of 7-fold over synaptic receptors, as measured by changes in the charge carried by the channel (Bai et al., 2001). Similarly, at low concentrations, etomidate enhances the tonic conductance in hippocampal neurons with a 60-fold increase of the charge transfer when compared to synaptic conductances (Cheng et al., 2006). Further, this increase in the tonic conductance has been shown to depend on α5GABAARs, which are considered the site of action for etomidate in mediating its amnestic and LTP-impairing
69 70 properties (Cheng et al., 2006). However, the sedative properties are likely mediated through an alternative receptor subunit (Belelli et al., 1997; Reynolds et al., 2003; Cheng et al., 2006;
Grasshoff et al., 2007).
In this thesis etomidate was selected for most experiments because it has a high selectivity for
GABAARs (Rudolph and Antkowiak, 2004) and possesses amnestic but not analgesic properties, which may confound behavioural tests such as fear conditioning (Davis and Cook, 1986).
Etomidate has ultra-short actions resulting from its redistribution from the brain to other tissues and its high clearance by ester hydrolysis in the liver and plasma (Gooding and Corssen, 1976).
Recently, the subunit dependency for the actions of etomidate has started to become clear through the use of point mutant mice. It has been shown that etomidate mediates its sedative and hypnotic actions through β2/3 receptor subunits (Reynolds et al., 2003; Belelli et al., 2005). For instance, mice generated to express a point mutation in the β2 subunit (β2N265S) are insensitive to the sedative effects and display reduced hypnosis to etomidate (Reynolds et al., 2003).
Correspondingly, these mice also exhibit reduced amplitudes for EEG slow sleep waves when injected with etomidate (Reynolds et al., 2003) and data suggests that β2-containing GABAARs in the ventral basal thalamus may underlie the sleep EEG profile of etomidate (Belelli et al.,
2005).
Etomidate has been shown to reduce the oscillatory activity of neurons in a β3 subunit-dependent manner. Specifically, etomidate decreases the power frequency spectra of cortical theta oscillations to 3–8 Hz in control slices and β2N265S but not in β3 point mutant mice
(β3N265M) (Drexler et al., 2005). This suggests that etomidate preferentially modifies theta oscillations primarily through the β3 compared with the β2 subunit, which is interesting because
70 71 theta oscillations are specifically implicated for their involvement in the learning and memory process (O'Keefe, 1993; Buzsaki, 2005) and β2 mRNAs are expressed at low levels in the hippocampus (McKernan et al., 1991). The β3 subunit is also more likely to partner with the α5 subunit as compared with the β2 subunit (Luddens et al., 1994; Sur et al., 1998) suggesting that at low concentrations etomidate may impair memory through an interaction with receptors containing the α5 and β3 subunits.
2.4.2 Action of anesthetics on LTP and memory
Anesthetics including enflurane, sevoflurane and isoflurane have been shown to impair learning and memory (Munte et al., 2001; Culley et al., 2003; Culley et al., 2004) and LTP (Ishizeki et al.,
2008). These volatile anesthetics have been shown to enhance GABA-evoked currents in pyramidal neurons (Nakahiro et al., 1989; Wakamori et al., 1991; Jones et al., 1992) and exhibit
GABAAR subunit selectivity and specificity (Caraiscos et al., 2004a; Sonner et al., 2005).
Isoflurane is of considerable interest because low concentrations have been shown to enhance the tonic conductance in pyramidal neurons in an α5GABAAR-dependent manner (Caraiscos et al.,
2004a).
It has been shown that the impairment of learning and memory by isoflurane is dependent on the concentration of isoflurane and the type of learning task (Dutton et al., 2001). Specifically, impairment of contextual fear conditioning occurs at 0.25 of the minimum alveolar concentration
(MAC), while impairment of tone fear conditioning is observed at 0.62 MAC (Dutton et al.,
2001). Further, α1 subunit-containing GABAARs have been implicated in mediating a portion of the contextual fear conditioning response as null mutant mice with a genetic deletion of the α1 subunit (Gabra1–/– mice) display 75% less sensitivity to the amnestic effects of isoflurane as
71 72 compared to WT mice (Sonner et al., 2005). In vitro, both isoflurane (Simon et al., 2001) and sevoflurane (Ishizeki et al., 2008) impair LTP through a GABA-dependent mechanism. While the precise mechanism and subunit dependency of volatile anesthetic action on memory and LTP has not been demonstrated, α5GABAARs remain a leading candidate, and through the use of mouse genetics these actions will start to be unraveled and become more apparent.
2.4.3 Problems associated with anesthetic-induced amnesia
Anesthetics are administered to millions of patients each year, yet we still have no true understanding of how these drugs do what they do. This remarkable lack of knowledge limits our ability to effectively design anesthetic drugs that would minimize or eliminate any unwanted side effects. The anesthetics commonly used for clinical practice are among the most potent neurodepressors used in medicine, resulting in clinically observable effects, such as amnesia, analgesia, unconsciousness, and immobility (Campagna et al., 2003). The amnesia associated with anesthetics is arguably the most important behavioural endpoint. For the patient, general anesthesia consists of amnesia, and not necessarily unconsciousness. A patient would likely choose the combination of amnesia and consciousness over the combination of recall and unconsciousness (Antognini et al., 2003). This does not diminish the necessity of anesthetic- induced analgesia and unconsciousness but merely highlights the importance of amnesia.
Anesthetics and neurodepressive drugs alike, have a particularly high affinity for GABAARs
(Rudolph and Antkowiak, 2004). Since different GABAARs predominate in different brain regions, the behavioural consequences of anesthetics are likely mediated by different populations of GABAARs. In terms of this thesis work, I was interested in determining whether anesthetic- induced amnesia was mediated by the hippocampus-dominant α5GABAARs. There is also a fine
72 73 line between the extent of anesthetic-induced amnesia during surgical procedures. It is thought that too little amnesia results in unintended patient awareness during surgical procedures and too much amnesia can result in persistent cognitive dysfunction following surgery, referred to as post-operative cognitive dysfunction (POCD). There has not been an identifiable target for the amnestic effects of anesthetics and a main goal of this thesis was to establish whether a receptor substrate or reversal agent existed that would determine the memory-impairing effects of etomidate.
2.4.4 Intraoperative awareness and post-operative cognitive
dysfunction
Unintended intraoperative awareness refers to the unexpected explicit recall of events that occurred during surgery. Unfortunately, 1 in 1000 patients who undergo general anesthesia experience this adverse outcome (Sandin et al., 2000; Sebel et al., 2004), and the incidence may be even higher among children (Lopez et al., 2007; Blusse van Oud-Alblas et al., 2009). In this situation, the patient may feel the pain or pressure of surgery, hear conversations, or feel as if they cannot breathe. The patient may be unable to communicate any distress because they have been given a paralytic and muscle relaxant. If intraoperative awareness does occur, about 42% of patients feel the pain of the operation, 94% experience panic and anxiety, and 70% experience lasting psychological symptoms (Moerman et al., 1993). These patients display an increased risk of developing post-traumatic stress disorder (PTSD), following their procedure (Osterman et al.,
2001). The increased risk of PTSD has also been studied using a rat model, where a sufficient concentration of isoflurane (> 0.67 MAC) is required to lessen the severity of stress-enhanced fear learning (Rau et al., 2009). Research has also indicated that the incidence of intraoperative awareness drastically increases when “light” anesthesia is administered during procedures such
73 74 as cardiac surgery (Sebel et al., 2004). This suggests that the depth of anesthesia possibly correlates with the behavioural endpoint. Interestingly, at the opposite end of the clinical spectrum are problems resulting from persistent memory impairment following anesthesia, which may result from too much anesthesia (Johnson et al., 2002; Hanning, 2005).
There are identifiable risk factors associated with the occurrence of post-operative cognitive dysfunction (POCD). The age of a patient is most notably the number one risk factor associated with POCD, with a higher incidence and longer persistence of POCD in patients over 65 years of age (Price et al., 2008). POCD occurs in approximately 37% of young adults (age 18 – 36 years) and 41% elderly patients (60 years or older) at the time of hospital discharge following non- cardiac surgery. Three months following surgery, 13% of elderly patients and 6% of young patients still exhibit POCD-like symptoms, with explicit memory and executive functions the most negatively affected. Other risk factors for developing POCD include lower educational level, previous cerebral vascular accident, and cardiac surgery (Hanning, 2005; Cottrell, 2008).
Despite these qualitative risk factors a molecular or receptor based construct has not been identified to determine an increased risk for POCD development. Several pathological conditions, including epilepsy and chronic alcohol abuse, alter expression of the α5GABAARs and are noted risk factors for both awareness and POCD (Glatt et al., 1992; Glatt et al., 1994;
Papadimitriou et al., 2001). Polymorphisms of the human Gabra5 gene are known to occur, although their functional significance is unknown.
The mechanistic basis of POCD needs to be addressed using animal rodent models. Although the animal based research is in its infancy, some researchers have generated data to support the human epidemiological studies. Two hours of exposure to the isoflurane (1.2 %) impaired the
74 75 performance of aged rats (18 months old) in the already acquired spatial radial arm maze task but had no effect on younger rats (6 months old). However, isoflurane decreases performance in both young and aged rats when rats are required to the learn the task following anesthesia. This effect can persist up to four weeks following anesthetic exposure (Culley et al., 2003; Culley et al.,
2004). This suggests that the acquisition of new memories following anesthesia appears to be more susceptible to anesthetic-induced impairment than previously acquired memories.
Additionally, there are notable deficits in spatial memory when mice are tested in the water maze. Following five 120 min anesthetic exposures to isoflurane (1%), anesthetized mice spent significantly less time in the target quadrant of the water maze (Bianchi et al., 2008), highlighting the deleterious effects of post-anesthetic exposure on spatial memory in animal models. To better understand the problem of intraoperative awareness and POCD it is necessary to understand the molecular mechanisms associated with the anesthetic state and this is likely to include their memory-blocking properties. Although the mechanisms underlying intraoperative awareness and POCD are complicated and are expected to involve the cerebral cortex, there are likely several subcortical structures involved, including the hippocampus (Antognini et al.,
2003).
2.5 Summary
This background section has provided a general overview of the key components that are required to fully appreciate the focus of this thesis: the hippocampus, memory, mechanisms of plasticity and GABAARs. In summary the distribution of α5GABAARs is relatively sparse and compartmentalized in the mammalian brain. Studies from our laboratory have shown that
α5GABAARs generate a tonic inhibitory conductance in CA1 hippocampal pyramidal neurons and that low concentrations of general anesthetics enhance their activity. In this thesis I
75 76
examined the effects of α5GABAARs on E-LTP and learning and memory. In CA1 neurons E-
LTP is NMDA-dependent and thought to represent a substrate for certain forms of memory.
α5GABAARs have been implicated in hippocampus-dependent memory because of their restricted distribution to the dendrites of the CA1 hippocampal region. However, it is still not certain how α5GABAARs alter CA1 neurons to regulate hippocampus-dependent memory.
Additionally, the general anesthetic etomidate was used to determine whether it enhanced the
α5GABAAR-mediated tonic conductance and whether this effect correlated with an impairment of hippocampus-dependent memory. The experiments with etomidate are also clinically relevant and have implications for the memory dysfunction associated with general anesthetics. The general hypothesis for the thesis is that α5GABAAR-activity constrains hippocampus-dependent learning and CA1 synaptic plasticity.
76 77
Chapter 3. General Methods and Materials
This general methods and materials section briefly describes the experimental animals, pharmacological agents, solutions, electrophysiological protocols and behavioural protocols used in Chapters 4, 5, and 6. Only general details, procedures and equipment set-ups are described here, as specific details regarding drug application, timing and specific protocols used are described in their respective Chapters.
3.1 Experimental animals
All experiments were approved by the Animal Care Committee of the University of Toronto.
The generation of the Gabra5–/– mouse line has been described previously (Collinson et al.,
2002). These mice do not display an overt behavioural phenotype; they breed normally and have a normal lifespan. In addition, they exhibit normal motor coordination, with no overt compensatory changes in other GABAAR subunits.
The mice were bred on a background of 50% C57Bl/6 and 50% Sv129Ev and maintained under the supervision of the Department of Comparative Medicine. The majority of mice used in the behavioural experiments were derived from homozygotic mating pairs because only male mice were used and the probability of deriving a male homozygous offspring from a heterozygous mating pair is 1 out of every 8 mice born. Thus a large colony of heterozygous mice would be needed to generate enough male homozygous mice to fulfill sample size requirements for all behavioural experiments. However, when homozygous offspring from heterozygous mating pairs
77 78 were available they were included in behavioural experiments, along with age-matched offspring from homozygous mating pairs.
Mice from heterozygous mating pairs were genotyped at the time of weaning, approximately postnatal day (P) 21. Briefly, the distal 2 mm of the tail was excised to provide the necessary
DNA for extraction and amplification using Extract-N-Amp™ Tissue PCR Kit (Sigma-Aldrich,
ON) for genotyping our mouse colony.
3.2 Electrophysiology
For the hippocampal slice experiments, tissue slicing and electrophysiological recordings were conducted in artificial cerebral spinal fluid (aCSF) of the following composition in mM: NaCl
(124), KCl (3), MgCl2 (1.3), CaCl2 (2.6), NaH2PO4 (1.25), NaHCO3 (26), D-glucose (10). The aCSF was continuously bubbled with carbogen (95% O2, 5% CO2) with the osmolarity adjusted to 300-310 mOsm.
3.2.1 Preparation of hippocampus tissue slices
Following administration of isoflurane anesthesia, mice were decapitated and their brains quickly removed and placed in ice-cold oxygenated (95% O2, 5% CO2) aCSF. Slices (350 µm thick) containing transverse sections of the hippocampus were prepared with a vibratome (VT1000E tissue slicer, Leica, Deerfield, Illinois). After a recovery period of 1 h in the oxygenated aCSF, slices were transferred to a submersion-recording chamber mounted on the stage of an upright microscope (BX51WI, Olympus, Center Valley, Pennsylvania) equipped with a water immersion objective (x40, numerical aperture 0.95, working distance 2.0 mm), Normarski optics, and
78 79 differential-interference contrast (IR-DIC) video microscopy mounted on an Olympus BX51WI microscope. The CA1 region was isolated from the CA3 region by a surgical cut, and slices were continually perfused with aCSF at a constant rate of ~2-3 ml/min.
3.2.2 Extracellular synaptic stimulation
A concentric bipolar electrode made of stainless steel (Rhodes Medical Instruments Inc.,
Summerland, California) was used to evoke synaptic responses. The stimulation electrode was positioned under 10 X magnification in the stratum radiatum approximately 50 µm from the stratum pyramidale at a depth of 100-150 µm. Baseline synaptic responses were evoked by stimulating the Schaffer collaterals with 0.1-ms pulses every 20 s (unless otherwise stated) using a isolated stimulator (Grass Technologies, W. Warwick, R.I., Model PSIU6), that yielded a half- maximal synaptic response.
3.2.3 Extracellular field recordings
Extracellular field excitatory postsynaptic potentials (fEPSPs) were recorded from CA1 pyramidal neurons using electrodes that were positioned in the stratum radiatum and contained aCSF. All responses were recorded with a Multiclamp 700A amplifier (unless otherwise stated) and headstage (Molecular Devices, Union City, CA) and low-pass filtered at 10 kHz before digitization (Digidata 1200; Molecular Devices, Union City, CA). Changes in the slope of the fEPSPs were measured as an indication of synaptic plasticity and were induced using a wide range of stimulation protocols. The specific LTP protocols are described in the Chapters in which they appear.
79 80
3.2.4 Current clamp and voltage clamp recordings in hippocampal
slices
Tight seals (> 1 GΩ) were formed on the cell body of visually identified CA1 pyramidal neurons and whole cell configuration recordings were obtained by rupturing the membrane with negative pressure. The equivalent circuit for a typical whole cell configuration is shown in Figure 3.1. For whole cell recordings, the series resistance and pipette and whole-cell capacitance were minimized electronically. To monitor series resistance, a hyperpolarizing voltage step of 10 mV or 10 pA was applied. Only cells that demonstrated stable series resistance (20% change) were used for data analysis. Current-clamp recordings were used to measure resting membrane potential, evoked inhibitory postsynaptic potentials (IPSPS), input resistance and evoked action potentials. The resting membrane potential was determined as the potential measured over a stable 5 s period. Only recordings with stable holding current and series resistance maintained below 20 MΩ were considered for analysis.
3.2.4.1 Intracellular recording solutions
The pipette solution used to record the tonic current and mIPSCs (Chapter 4, 5, 6) consisted of
(in mM); CsCl (140), HEPES (10), EGTA (10), Mg-ATP (4), CaCl2 (1), and QX-314 (5). The pH was adjusted to 7.3 with CsOH and the osmolarity was adjusted to 295–305 mOsm. The pipette solution for AMPA/NMDA currents (Chapter 6) consisted of (in mM); CsGluc (132.5), 17.5 KCl
(17.5), EGTA (0.2), HEPES (10), Mg-ATP (2), GTP (0.3), QX-314 (5) (pH 7.2-7.3 with CsOH,
Osm 290). The pipette solution for EPSP and IPSP recordings (Chapter 6) consisted of (in mM);
K-Gluc (132.5), KMeSO4 (10.3), KCl (7.2), HEPES (10), EGTA (0.2), Mg-ATP (2), GTP (0.3) with the pH adjusted to 7.3 with KOH.
80 81
Figure 3.1. The whole cell configuration equivalent circuit is shown.
The relation of the electrode to the cell membrane is illustrated for the whole-cell configuration.
Following application of negative pressure to the pipette a tight seal is formed between the glass and the cell membrane. This seal minimizes the current leak from the interior of the cell to the bath. The pipette (Rpipette) and access (Raccess) resistance are in series and constitute the series resistance. The resistance of the patch (RLeak) and cell (RM) are parallel to the series resistance in this circuit. CM = membrane capacitance, Cpipette = pipette capacitance, Probe = amplifier.
81 82
82 83
3.2.5 Recording electrodes
The recording electrodes were made from borosilicate glass capillary tubes (1.5 mm outer diameter, World Precision Instruments, Sarasota, FL) using a Narishige vertical two-stage electrode puller (Model: PP-83, Tokyo, Japan). The final resistance of electrodes was between 3-
5 MΩ for ICF-filled patch electrodes.
3.2.6 Pharmacological agents
For the experiments that required pharmacological intervention, all drugs were bath applied in aCSF (µM unless otherwise stated): bicuculline (8), APV (10 or 50), CNQX (10) were purchased from Sigma (St. Louis, MO). TTX (100 nM) was purchased from Alomne Labs Ltd. (Jerusalem,
Israel). Etomidate (1) was purchased from Bedford Laboratories (Bedford, OH). L-655,708 (10 or 20 nM), SR-95531 (5), and CGP (10) were purchased from Tocris Biosciences (Ellisville,
MO). Table 3.1 lists all pharmacological agents that were used for the electrophysiology, along with their mechanism of action.
3.2.6.1 Drug preparation
All drugs were prepared daily from frozen stock solutions. L-655,708 was diluted in HCL (1 N) to a concentration of 10 mM. It was then further diluted in H2O to make final concentration alliquots of 100 µM. The final concentration of HCL was so small that it did not significantly alter the pH. Bicuculline, APV, CNQX, and etomidate were diluted in dH2O.
83 84
Table 3.1. The primary pharmacological agents used in this thesis.
84 85
Drug Name Alternative name Pharmacological Mechanism of classification action
Etomidate N/A General anesthetic Enhances β2/3 subunit-
containing GABAA receptors
L-655,708 FG-8094 Benzodiazepine inverse Inverse agonist for α5 agonist subunit containing
GABAA receptors
Bicuculline Bicuculline methiodide Competitive GABAA receptor Inhibits all GABAA antagonist receptors
SR-95531 Gabazine Competitive GABAA receptor Inhibits synaptically
antagonist expressed GABAA receptors
(2R)-amino-5- APV Competitive NMDA receptor Inhibits all NMDA phosphonovaleric acid antagonist receptors
6-cyano-7-nitroquinoxaline- CNQX Competitive AMPA/Kainate Inhibits all 2,3-dione receptor antagonist AMPA/Kainate receptors
85 86
3.3 Behaviour
3.3.1 Morris water maze
For the water maze task, a circular pool of diameter 1.2 m was filled with tap water (25 ± 2 ºC), which was made opaque by the addition of a nontoxic white paint. A circular platform of diameter 12 cm was placed approximately 0.5 cm below the water surface so that it was not visible to the mice. Mice were pre-trained for 10 days, with 4 trials on each day. In the matching- to-place water maze, the location of the platform was changed daily. Each mouse had 60 s to search for and locate the hidden platform. If the mouse did not locate the platform within 60 s, the experimenter gently guided the mouse to the platform. Each mouse was allowed to remain on the platform for 30 s after locating it. The inter-trial interval was 30 s. Probe trials were undertaken to test for spatial learning. Four acquisition trials were conducted after a drug injection (please refer to specific methods in the respective Chapters) on the test day. The next day, a probe trial was performed to test the recall of the mice for the spatial location that had previously contained the hidden platform. During the probe trial, the hidden platform was removed from the pool, and the mouse was allowed to search for 60 s. The mouse was then rescued, and the trial ended. The percentage of time spent within the target quadrant of the pool was used to measure recall. Mice that had learned the correct location of the platform were expected to selectively search in the correct area, whereas mice that had not experienced any learning were expected to spend only 25% of the time near the correct location. The order in which the drug or vehicle was administered was randomized. Data records were made with HVS
Water 2020 software (HVS Image, Hampton, England) for off-line analysis. Briefly, a video camera captured the movement of each mouse, and the software tracked the location of the mouse by contrast comparison. The time, swim path, and latency of each mouse were recorded
86 87 during each trial, and the software later calculated the percentage of time spent in the target quadrant.
3.3.2 Fear conditioning
The conditioning chamber consisted of a Perspex acrylic arena with a light mounted in the lid
(350 × 200 × 193 mm; Technical and Scientific Equipment GmbH, Midland, Michigan, USA).
The floor consisted of stainless steel bars (4 mm diameter, 5 mm apart) connected to a computer, which controlled the duration of the test session and the timing, intensity, and duration of the audible tones and shocks. On day 1, single animals were allowed to explore the chamber for 180 s. An 800-Hz tone, created by a frequency generator, amplified to 70 dB, and lasting 20 s, was then presented. Two auditory conditioning protocols were used: (1) for cued fear conditioning, the last 2 s of each auditory tone was paired with a 0.5 mA electrical shock to the foot; (2) for trace fear learning, each auditory stimulus was followed by a quiescent period of 20 s after which the electrical shock (0.5 mA for 2 s) was applied. Each of these sequences was presented 3 times, separated by 60 s (for cued fear) or 210 s (for trace fear). On day 2, 24 h after the conditioning session, each mouse was returned to the chamber, and freezing response (defined as lack of any movement except that required for respiration) was assessed every 8 s for a total of 8 min
(contextual fear). On day 3, the conditioning chamber was modified (by covering the metal grid floor with a rubber mat and a semi-circular wall was inserted into the chamber to modify the contextual surroundings in order to the measure freezing response to the tone (either cued or trace fear learning) without any contextual influence. Freezing to tone was assessed 48 h after the conditioning period. Mice were monitored for 180 s for freezing to the modified context, to rule out contextual influences. After the monitoring period, the auditory tone was presented continuously for 300 s, and the freezing response was recorded every 8 s or continuously.
87 88
Assessment of the freezing response was performed in the same conditioning chamber in which the mice received the foot shocks.
3.3.3 Pharmacological agents
All drug sources remained the same as described in section 3.2 if they were used in the behavioural experiments. For all of the behavioural experiments drug injections were done intraperitoneal (i.p.) (mg/kg unless otherwise stated).
3.3.3.1 Drug preparation
All drugs were prepared fresh each day in a sterilized biological safety cabinet to minimize contamination of drug solutions. Etomidate was maintained in a sterile, nonpyrogenic solution with each ml containing 2 mg etomidate, propylene glycol 35% v/v. Etomidate was injected at a final volume of 2 ml/kg to yield a final dose of 4 mg/kg. L-655,708 was diluted in 100 % DMSO to a concentration of 10 mM. It was then further diluted in physiological saline to yield a final
DMSO concentration of 10 %. This dilution yields a final L-655,708 dose of 0.68 mg/kg, which was used for all of the behavioural experiments.
3.4 Statistical analysis
Statistical analyses were completed with SPSS V 13.0 (Chicago, IL) or GraphPad Prism V4.0c for MacIntosh (San Diego, CA ;www.graphpad.com). The slope of the fEPSP was calculated using the 25%–70% function in Clampfit software (Molecular Devices Corporation, Sunnyvale,
CA). All data are shown as mean ± s.e.m. Unless otherwise noted, initial statistical comparisons were completed using one- or two-way analysis of variance (ANOVA) for LTP and behavioural experiments. Secondary comparisons were conducted by unpaired student’s t-tests, the group of
88 89 interest being compared with a reference control group. Post hoc analyses were conducted using
Tukey’s HSD, Bonferroni comparisons, and unpaired student’s t-tests.
89 90
Chapter 4. α5GABAA receptors mediate the amnestic but not sedative- hypnotic effects of the general anesthetic etomidate
4.1 Introduction
General anesthetics are highly lipid-soluble, low-potency compounds that were initially thought to act by non-specifically perturbing the structure of lipid bilayers. A major advance in the 1980s was the identification of neuronal proteins as anesthetic targets (Franks and Lieb, 1988).
Behavioural and neuroimaging studies in humans have since shown that anesthetics do not non- specifically depress brain function. Rather, anesthetics cause a constellation of different behavioural endpoints, including amnesia, immobility and unconsciousness, that are mediated by different brain regions and receptor populations (Campagna et al., 2003). Recently, studies with transgenic and null mutant mice have shown that specific populations of GABAARs contribute to the sedative and immobilizing properties of certain anesthetics (for review see (Rudolph and
Antkowiak, 2004)). The molecular targets underlying the amnestic properties of anesthetics have been highly elusive but are of great clinical importance and scientific interest. Unintended intra- operative awareness during surgery occurs in 1 or 2 cases per 1000 anesthetized patients (Sebel et al., 2004). Since anesthetics are administered to over 27 million patients each year, intra- operative awareness has become a major medical concern, as highlighted by a “sentinel alert” released by the Joint Commission on Accreditation of Healthcare Organizations
(www.jcaho.org). GABAARs that contribute to the amnestic effects of general anesthetics have not previously been identified.
90 91
In vitro studies have shown that most intravenous anesthetics, including etomidate, are positive allosteric modulators of GABAAR function (Jurd et al., 2003; Reynolds et al., 2003). In the hippocampus, a brain structure that is involved in learning and memory, GABAARs generate two distinct forms of inhibition (Otis et al., 1991; Bai et al., 2001). Transient inhibitory postsynaptic currents result from the vesicular release of GABA and the activation of postsynaptic GABAARs, whereas a low-amplitude tonic inhibitory conductance is generated by low concentrations of ambient GABA (Semyanov et al., 2004). In hippocampal pyramidal neurons, GABAARs that generate the tonic and synaptic currents have distinct subunit compositions and pharmacological properties (Yeung et al., 2003; Caraiscos et al., 2004b). Our laboratory has previously shown that a tonic inhibitory conductance in hippocampal pyramidal neurons is generated primarily by
α5GABAARs (Caraiscos et al., 2004b). The α5 subunit is of particular interest in memory processes as it is predominantly expressed in the hippocampus (Sur et al., 1999). In humans and rodents, only 4% of all GABAARs in the brain but 25% of GABAARs in the hippocampus contain the α5 subunit (Sur et al., 1999). Several lines of evidence suggest that α5GABAARs play a major role in memory processes; in particular, reduced expression of the α5 subunit is associated with better performance of hippocampus-dependent learning tasks (Collinson et al.,
2002; Crestani et al., 2002; Chambers et al., 2003). Selective inverse agonists for α5GABAARs have nootropic effects in animal models (Chambers et al., 2003). Here, I test the hypothesis that etomidate, a prototypic intravenous anesthetic, increases the function of α5GABAARs and that this effect contributes to the amnestic but not the sedative-hypnotic properties of this anesthetic.
Questions:
1. Does etomidate enhance an α5GABAAR-dependent tonic conductance?
91 92
2. Are α5GABAARs a molecular target for the LTP-impairing and memory blocking
properties of etomidate?
3. Are the sedative or hypnotic properties of etomidate mediated by α5GABAARs?
4.2 Specific Methods
4.2.1 Tonic currents
The amplitude of the tonic current was measured as the difference in the holding current before and during the application of etomidate (1 µM).
4.2.2 Long-term potentiation recordings
For these experiments a TBS protocol was chosen because this induction protocol has been shown to release copious amounts of endogenous GABA (Chapman et al., 1998) and this protocol was initially used with Gabra5–/– slices (Collinson et al., 2002). The TBS was presented for a total of 1 s, and consisted of 10 stimulus trains delivered at 5 Hz, with each train consisting of 4 pulses delivered at 100 Hz.
4.2.3 Contextual fear conditioning
Vehicle (propylene glycol or sterile saline), etomidate (4 mg/kg, i.p.) or ketamine (20 mg/kg, i.p.) at concentration and dose volume equivalent to the etomidate-containing solution) was administered 30 min before the conditioning trial. On day 1, single subjects were allowed to explore the chamber for 180 s. They then received three unsignaled foot shocks (duration 2 s, intensity 1 mA) at 60 s intervals. On day 2, 24 h after the conditioning session, mice were returned to the chamber, and the freezing response was assessed immediately every 8 s for 8 min.
92 93
The freezing response was defined as the lack of any movement except that required for breathing. Assessment of the freezing response occurred in the same conditioning chamber in which the mice received the foot shocks.
4.2.4 Water maze
Mice were pre-trained for 10 d with four trials each day. In the matching-to-place water maze, the location of the platform was changed daily. Each mouse had 60 s to search for and locate the hidden platform. If the mouse did not locate the platform within 60 s, the experimenter gently guided it to the platform. Each trial ended when the mouse had sat on the platform for 30 s. The inter-trial interval was 30 s. Our laboratory initially tested the effects of etomidate on performance during the acquisition trials in WT and Gabra5–/– mice; however, the results were extremely variable, so this strategy was abandoned (data not shown). Probe trials were subsequently undertaken to test for spatial learning. During the acquisition phase of the probe trials, each mouse was randomly assigned to receive an injection of the vehicle, etomidate (4 mg/kg, i.p.) or ketamine (20 mg/kg, i.p.). Thirty minutes later, the mouse was placed in the water maze. Four acquisition trials were conducted after the injection. The next day, a probe trial was performed to test the ability of the mice to recall the spatial location that previously contained the hidden platform. During the probe trial, the hidden platform was removed from the pool and the mouse was allowed to search for 60 s. The mouse was then promptly rescued, and the trial ended. Mice that learned the correct location of the platform selectively searched in the correct area. The percentage of time spent within the correct area of the pool (four times the diameter of the platform) was used to measure recall. The correct location of the platform represented 16% of the total area of the pool. If no learning occurred, the mouse was expected to spend 16% of the time near the correct location.
93 94
4.2.5 Rotarod
Mice were tested on a rotating rod unit (rotarod) to study motor co-ordination and strength. The mice were trained to walk on a rotarod (Rotamex, Columbus Instruments, Columbus, OH) revolving at a constant speed of 12 revolutions per minute for 120 s consistently. On the test day, one pre-injection trial was performed before the animals were treated with vehicle or etomidate
(4 mg/kg). Performance was indicated by the latency to fall from the rotarod at 5 min, 30 min and 60 min after injection. The time the mouse remained on the rotarod was recorded up to a maximum of 120 s.
4.2.6 Open field
The sedative property of etomidate was tested by measuring spontaneous activity in an open field testing chamber made of Plexiglas with the dimensions 42 X 42 X 30 cm. WT and Gabra5–/– were injected with either vehicle or etomidate, 30 min before testing and then returned to their home cage. Mice were monitored for the duration of time spent walking, rearing and grooming as an index of spontaneous locomotor activity in the open field. Briefly, all mice were tested post-injection for five consecutive minutes. After placement in the open field, an event recorder was used to score the total time spent walking, rearing or grooming. Odours were masked between test subjects by wiping the floor and walls of the test chamber with a mild ethanol solution.
4.2.7 Loss of righting reflex
The loss of righting reflex (LORR) was assessed using a classical experimental protocol. The
LORR was determined in WT and Gabra5-/- mice for a wide range of etomidate doses (5–15
94 95 mg/kg). Immediately following injection, the mice were placed in a shoebox cage and observed until they stopped moving. Mice were then placed on their back and scored as anesthetized if they failed to completely right within 30 s. If the mice succeeded in righting themselves three consecutive times they were scored as awake. All mice were used only once for the LORR assay.
Data were analyzed as the percent of the population that LORR at each dose of etomidate. To determine the dose of etomidate that caused 50% of the maximal response (ED50), the dose- response plot was fit, using nonlinear regression, to the equation: Y = D + [A-D] / [1 + 10((logED50
- X) Hill slope)] where D= minimum response, A is the maximum response and X is the dose. The time to the LORR was also quantified by measuring the time from the end of the injection to when the mice first demonstrated a LORR.
4.3 Results
4.3.1 Etomidate enhanced tonic conductance in CA1 pyramidal neurons
The concentration-dependent effects of etomidate on tonic and synaptic currents were studied using whole-cell currents recorded in CA1 pyramidal neurons from hippocampus tissue slices.
We have previously shown that the application of a low concentration of etomidate (0.1 µM) to pyramidal neurons from Swiss White mice caused a reversible increase in the tonic conductance
(156% ± 24% of control, n = 8, P < 0.05) whereas etomidate (0.1 µM) failed to alter the time course or amplitude of miniature inhibitory postsynaptic currents (mIPSCs) (Cheng et al., 2006).
Etomidate, applied at a higher concentration (1 µM) further increased the tonic conductance
(314% ± 51.4% of control, n = 9, P < 0.05) and prolonged the duration of mIPSCs; however, the inhibitory net charge transfer was 60-fold greater for tonic conductance than for synaptic conductance, which indicated that higher concentrations of etomidate had a markedly greater
95 96 effect on tonic conductance than on synaptic conductance under the experimental conditions
(Cheng et al., 2006).
To determine whether the etomidate-enhanced tonic conductance was generated by
α5GABAARs, I recorded currents from hippocampal slices obtained from Gabra5–/– mice. I anticipated that low concentrations of etomidate would minimally increase the holding current in
Gabra5–/– neurons relative to WT neurons. The change in the holding current (rather than the percent change in the tonic current) was measured because the tonic current was minimal in
Gabra5–/– neurons (Caraiscos et al., 2004b). These studies were also undertaken to determine whether there was a compensatory up-regulation of other high-affinity GABAARs, such as those containing the δ subunit, which might be sensitive to low concentrations of etomidate in
Gabra5–/– neurons or differences in the etomidate sensitivity of synaptic currents recorded in
WT and Gabra5–/– neurons.
Etomidate (1 µM) increased the holding current in WT neurons (26.41 ± 2.19 pA, n = 7) but not
Gabra5–/– neurons (3.39 ± 1.62 pA, n = 7) (P < 0.05; Figure 4.1A). No differences were detected in the frequency, amplitude or time course of sIPSCs between WT and Gabra5–/– neurons, in the absence or presence of etomidate (1 µM; Table 4.1). Thus, low concentrations of etomidate that are reported to be clinically relevant (0.05 - 0.43 µM) caused a greater enhancement of the tonic current rather than the mIPSCs. Higher concentrations of etomidate caused a further increase in the holding current and prolonged the mIPSCs; however, this action could be attributed to an indiscriminate effect of etomidate on GABAARs that do not contain the
α5 subunit (Hill-Venning et al., 1997).
96 97
Figure 4.1. A low concentration of etomidate increased the holding current in CA1 pyramidal neurons from WT but not Gabra5–/– slices.
(A) Current traces illustrate the increase in an inward current following the application of etomidate in WT neurons and Gabra5–/– neurons. (B) The bar chart shows that the change in the amplitude of the holding current was greater in WT than in Gabra5–/– neurons when etomidate was applied. * denotes P < 0.05.
97 98
98 99
Table 4.1. Properties of spontaneous IPSCs recorded in CA1 pyramidal neurons of hippocampal slices.
99 100
Values are means ± sem. Parentheses enclose n values. IPSC, inhibitory postsynaptic current, aCSF, artificial cerebrospinal fluid; ETOM, etomidate.
100 101
4.3.2 Etomidate reduced LTP in WT but not Gabra5–/– brain slices
LTP is an activity-dependent strengthening of synaptic efficacy at excitatory synapses. LTP is widely regarded as a possible cellular model of learning and memory (Bliss and Collingridge,
1993; Malenka and Bear, 2004). GABAergic inhibition regulates LTP; GABAAR antagonists enhance LTP, whereas positive allosteric modulators reduce LTP and impair memory (Seabrook et al., 1997). Our aim was to determine whether etomidate differentially modulated the LTP of fEPSPs recorded at CA1 pyramidal neurons from WT and Gabra5–/– mice.
As reported previously (Collinson et al., 2002), no differences in LTP were observed between vehicle-treated WT and Gabra5–/– slices immediately following TBS (191% ± 12% vs. 208% ±
12%, n = 6 slices per condition, P > 0.05) or 60 min post-TBS (170% ± 15% vs. 175% ± 11%, n
= 6 P > 0.05, Figure 4.2A). Etomidate-treated (1 µM) slices from WT mice showed an initial increase in the slope of the fEPSPs (178% ± 16% of baseline, n = 6 slices) immediately following TBS; however, the increase was not sustained, and the slope of the fEPSPs decreased to 103% ± 13% of baseline at 60 min (n = 6, P < 0.05). Similarly, the LTP of fEPSPs was initially observed in etomidate-treated (1 µM) slices from Gabra5–/– mice, and the slope of the fEPSPs increased to 192% ± 17% of baseline (n = 6, P > 0.05; Figure 4.2B) immediately following TBS. Unlike the recordings in WT slices, enhancement of the slope of the fEPSPs in
Gabra5–/– slices was sustained at 60 min post-TBS (168% ± 14% of control, n = 6, P < 0.05,
Figure 4.2B) in the presence of etomidate (1 µM). Thus, the attenuation of LTP by etomidate was absent in slices from Gabra5–/– mice.
101 102
Figure 4.2. Etomidate reduced the LTP of fEPSPs recorded in the CA1 region of hippocampal slices prepared from WT but not Gabra5–/– mice.
(A) The superimposed traces show averaged fEPSPs recorded from WT and Gabra5–/– mice in the presence of vehicle at baseline (1) and 60 min post-TBS (2). The time-dependent change in the slope of the fEPSPs is summarized in the graphs. (B) fEPSP traces and time-dependent changes in the slope of the fEPSPs for recordings in hippocampal slices obtained from WT and
Gabra5–/– mice are shown in the presence of etomidate (1µM). Etomidate significantly reduced
LTP elicited by 1 s of TBS in pyramidal neurons from WT mice (n = 6) but not Gabra5–/– mice
(n = 6) at CA1 Schaffer collateral synapses.
102 103
103 104
4.3.3 Amnestic effects of etomidate are mediated by α5GABAARs
The electrophysiological experiments showed that low concentrations of etomidate enhanced the tonic current and reduced LTP in pyramidal neurons from WT but not Gabra5–/– mice. These findings predicted that the amnestic effect of etomidate would be attenuated in Gabra5–/– mice.
To study the effects of etomidate on hippocampus-dependent memory, two complementary behavioural assays were performed, a contextual fear conditioning paradigm and the matching- to-place version of the Morris water maze.
Contextual fear conditioning was used to measure the ability of the mice to learn and remember an association between an adverse experience and environmental cues (Fanselow, 1980). WT and Gabra5–/– mice treated with the vehicle exhibited similar freezing behaviour (83.1% ± 3.7% of time spent freezing, n = 8, vs. 86.9% ± 2.2%, n = 9, P > 0.05; Figure 4.3A). In contrast, etomidate reduced the freezing scores in WT but not Gabra5–/– mice (52.7% ± 5.0%, n = 8, vs.
78.8% ± 5.1%, n = 9, P < 0.05; Figure 4.3A). These findings indicate that etomidate impaired the acquisition of contextual fear in WT but not Gabra5–/– mice. To ensure that the reduced etomidate sensitivity exhibited by the Gabra5–/– mice was not due to a non-specific insensitivity to general anesthetics, the effect of another anesthetic that does not target GABAARs was tested.
The dissociative anesthetic ketamine causes amnesia by inhibiting the NMDA subtype of glutamate receptors. As shown previously, the vehicle had no effect in WT and Gabra5–/– mice
(76.9% ± 5.5% of time spent in freezing behaviour, n = 7, vs. 82.2% ± 4.6%, n = 7, P > 0.05), whereas ketamine (20 mg/kg i.p.) caused impairment in both WT and Gabra5–/– mice, as shown by a similar reduction in freezing scores (for WT mice, 34.8% ± 8.7% of the time, n = 8,
P < 0.05; for Gabra5–/– mice, 35.8% ± 5.7% of the time, n = 8, P < 0.05; Figure 4.3B).
104 105
The water maze was next used as an independent test of spatial learning. Etomidate impaired memory performance in WT mice, as shown by the reduced percentage of time that mice spent swimming in the area that had previously contained the hidden platform (control 21.2% ± 2.6% time spent in the correct location vs. etomidate 14.0% ± 1.6%, n = 16, P < 0.05; Figure 4.3C).
The reduction in performance was not exhibited by the etomidate-treated Gabra5–/– mice
(control 21.7% ± 2.3% vs. etomidate 19.3% ± 1.8%, n = 16, P > 0.05; Figure 4.3C). I also tested the effects of ketamine on the ability of the mice to locate the platform during a probe trial.
Ketamine (20 mg/kg i.p.) caused a similar impairment of spatial learning in WT mice (control
22.1% ± 2.7% vs. ketamine 15.4% ± 2.2%, n = 17, P < 0.05) and Gabra5–/– mice (control
21.5% ± 2.1% vs. 15.0% ± 1.8%, n = 15, P < 0.05; Figure 4.3D).
Visible platform trials were performed to test for possible genotype differences in motivational factors, perceptual and motor abilities, and non-specific effects of etomidate. The platform that was usually submerged was raised to 0.5 cm above the waterline and marked with a brightly coloured flag. The mice were injected with vehicle or etomidate (4 mg/kg) 30 min before the visible platform trial. There were no differences in the latency to locate the platform between the two genotypes, in the absence or presence of etomidate (WT control 5.4 ± 0.8 s, n = 8, vs. WT etomidate 5.5 ± 0.9 s, n = 8, P > 0.05 Gabra5–/– control 4.8 ± 0.9 s, n = 8, vs. Gabra5–/– etomidate 5.5 ± 0.8 s, n = 8, P > 0.05; Figure 4.3E). Also, no difference in mean swimming speed was detected between the genotypes during pre-training, the acquisition trial (WT control
0.23 ± 0.01 m/s, n = 16, vs. WT etomidate 0.22 ± 0.01 m/s, n = 15, P > 0.05; Gabra5–/– control
0.20 ± 0.02 m/s, n = 15, vs. Gabra5–/– etomidate 0.20 ± 0.2 m/s, n = 15, P > 0.05) (Figure 4.3F).
The lack of difference in swim speed (in the absence or presence of etomidate) for both groups confirmed the procedural ability of the mice to perform the task.
105 106
Figure 4.3. Etomidate impaired contextual fear conditioning in WT but not Gabra5–/– mice.
(A) The bar graphs show the effects of vehicle and etomidate (4 mg/kg, i.p.) on the freezing scores (mean ± s.e.m.). The scores of etomidate-treated mice were reduced in the WT but not the
Gabra5–/– mice, which indicates impairment of memory acquisition following the administration of etomidate. (B) Compared with vehicle, ketamine (20 mg/kg, i.p.) caused a similar reduction in freezing scores, which indicates a similar impairment in fear conditioning for the same context. (C) Spatial learning was impaired by etomidate in WT but not Gabra5–/– mice. The Morris water maze probe trial showed that etomidate reduced the amount of time the mice spent in the area where the platform had been located the previous day. Impaired memory retrieval was shown by WT but not Gabra5–/– mice. (D) In contrast, impairment by ketamine did not depend on the mouse genotype. (E) No differences were detected in performance on the visible platform during the probe trial (WT control 0.24 ± 0.01 m/s, n = 16 vs. WT etomidate
0.24 ± 0.01 m/s, n = 15, P > 0.05; Gabra5–/– control 0.24 ± 0.02 m/s, n = 15, vs., Gabra5–/– etomidate 0.24 ± 0.2 m/s, n = 15, P > 0.05) or (F) in swim speed between the two genotypes.
106 107
107 108
4.3.4 Sedative-hypnotic effects of etomidate are not mediated by
α5GABAARs
Performance on the rotarod measures sensorimotor control and co-ordination and is thought to depend primarily on neuronal circuits in the cerebellum and spinal cord. To ensure there were no differences in sensorimotor function in WT and Gabra5–/– mice at the dose of etomidate and time point that was selected for the behavioural experiments, I examined the impairment of walking on the rotarod. Five minutes after the injection of etomidate, both groups showed a similar reduction in the latency to fall off the rotarod (Figure 4.4A). No impairment of performance was detected 30 min after etomidate (4 mg/kg, i.p.), which was the dose selected for the fear conditioning and Morris water maze experiments.
The sedative and hypnotic properties of etomidate were studied as a positive control by measuring spontaneous motor activity and LORR, respectively. These behavioural endpoints are unlikely to be mediated by hippocampal neurons or α5GABAARs (Nelson et al., 2002). The sedative properties of etomidate were studied with an open field test which examines spontaneous locomotor activity following injection of etomidate or vehicle control. Locomotion was similar in WT and Gabra5–/– mice 30 minutes after the vehicle (Figure 4.4B). Etomidate (4 mg/kg and 10 mg/kg) caused a similar reduction in spontaneous movement at 30 min in WT and
Gabra5–/– mice as shown in Figure 4.4B.
It is widely accepted that a surrogate experimental measure for anesthetic-induced unconsciousness is the LORR (Rudolph and Antkowiak, 2004). The ability of a wide range of etomidate doses to cause the LORR was determined for WT and Gabra5–/– mice. A quantal dose-response plot for LORR showed that the dose of etomidate that caused half the maximal
108 109 response was similar in WT and Gabra5–/– mice (9.6 and 9.2 mg/kg, n-= 35 and 34 respectively,
P < 0.05, Figure 4.4C). The time to LORR was also measured following injection of the various doses of etomidate; no differences were detected between the WT and Gabra5–/– mice (Figure
4.4D). Together, the above results show that the Gabra5–/– mice do not exhibit a resistance to etomidate for the behavioural endpoints used to measure sedation and loss of consciousness.
109 110
Figure 4.4. Etomidate impairment of motor coordination, spontaneous motor activity and loss of the righting reflex (LORR) was not increased in Gabra5–/– mice.
(A) Etomidate (4 mg/kg) caused impairment of motor performance in the 12 r.p.m. rotarod test.
One pre-injection trial was performed on the test day before the mice were treated with etomidate (WT, n = 16; Gabra5–/– n = 15). The results are expressed as the latency to fall off the rotarod (mean ± s.e.m.). Both groups had impaired responses 5 min after injection (*P > 0.05, one-way ANOVA) but not at 30 or 60 min. (B) Spontaneous locomotor activity (walking) was reduced by etomidate as shown for the open field test. No difference in baseline activity was observed between WT and Gabra5–/– mice following injection of the vehicle control. Etomidate
(4 mg/kg and 10 mg/kg) caused a concentration-dependent reduction in locomotion that was similar in WT and Gabra5–/– mice. (C) Dose-response analysis for etomidate causing the LORR is shown. The data points represent the percent of the WT (blue square) or Gabra5–/– (red circle) mouse populations that exhibited LORR at the dose indicated on the abscissa.
Overlapping data points for the two genotypes are represented by a third symbol (green triangle).
Mice were tested at doses of 5, 7.5, 10, 12.5, 15 and 20 mg/kg and a total of 69 mice were studied. The fitted curves were generated using a sigmoidal equation that provided the following values: WT ED50 = 9.6 mg/kg ± 1.1, Hill slope parameter h = 5.8 ± 1.8, Gabra5–/– ED50 = 9.2 ±
1.1, h = 5.7 ± 1.5, P < 0.05). (D) The latency to LORR recorded from the time of etomidate injection. The bars show the mean ± s.e.m.. Each data point represents 6 to 8 mice and no differences were detected in between WT and Gabra5–/– mice. The time to LORR was significantly different at all doses of etomidate (ANOVA, P < 0.05) with the exception of 10 mg/kg vs. 12.5 mg/kg.
110 111
111 112
4.4 Discussion
The absence of the Gabra5–/– subunit resulted in decreased sensitivity to etomidate for impairment of learning and memory performance. The reduced sensitivity of Gabra5–/– mice to etomidate could not be attributed to differences in drug pharmacokinetics, as the brain concentrations of etomidate have been shown to be similar in WT and Gabra5–/– mice (Cheng et al., 2006). The resistance of the Gabra5–/– mice was specific for memory performance, as deletion of the α5 subunit did not influence the effect of etomidate on the LORR or spontaneous motor activity. Sensorimotor function, coordination, motor learning, perception and motivation were similar in WT and Gabra5–/– mice, in the absence and presence of etomidate, as evidenced by the rotarod and visible platform tests. These results provide evidence for a selective population of GABAARs that contribute to the amnestic but not sedative-hypnotic effects of a general anesthetic. The findings support the hypothesis that etomidate acts on different GABAAR subtypes, located in different neuronal circuits, to generate the multiple components of the anesthetic state (Eger et al., 1997).
Electrophysiological studies showed that a low concentration of etomidate primarily enhanced the tonic conductance in CA1 pyramidal neurons in slices prepared from WT but not Gabra5–/– mice, while not significantly influencing the characteristics of the mIPSCs. The free aqueous concentration of etomidate associated with amnesia in the mammalian brain has been estimated at approximately 0.43 µM (Rudolph and Antkowiak, 2004). This value was based on the concentration of etomidate that produces immobility in half of subjects (1.5 µM) and evidence suggesting that amnesia occurs at approximately 20% to 40% of the immobilizing concentration of anesthetics (Dwyer et al., 1992). The brain concentration of etomidate in mice, measured at the dose and time point selected for the Morris water maze acquisition trial and fear
112 113 conditioning, was shown to be 0.77 µM (Cheng et al., 2006). In the cultured cell model, a similar low concentration of etomidate increased the tonic inhibitory current but not synaptic currents.
This result is consistent with a previous report from our laboratory showing that low concentrations of the inhaled anesthetic isoflurane selectively enhanced tonic but not synaptic conductance in hippocampal neurons (Caraiscos et al., 2004a). Also, studies of human recombinant α5β3γ2L and α1β3γ2 GABAARs confirmed that the α5 subunit confers sensitivity to low concentrations of isoflurane and that the efficacy of isoflurane for enhancement of
GABA-evoked currents was greater for α5β3γ2L than for α1β3γ2 GABAARs. These results suggest that α5GABAARs contribute to the amnestic properties of both injectable and volatile anesthetics.
Studies of pyramidal neurons in hippocampal slices showed no differences in the induction or maintenance of LTP in WT relative to Gabra5–/– mice under control conditions, as reported previously (Collinson et al., 2002). Induction of LTP by high-frequency stimuli and maximal fEPSP strength were similar in WT and Gabra5–/– mice (Collinson et al., 2002). These authors also reported no differences in the ability of low-frequency stimuli to activate fEPSPs in CA1 or the dentate gyrus; however, the ability of paired-pulse stimuli to facilitate the amplitude of synaptic potentials was significantly increased in Gabra5–/– mice. Increased paired-pulse facilitation was specific to the CA1 region and was not observed in the dentate gyrus, where
α5GABAARs are not highly expressed. I observed that etomidate inhibited LTP in WT but not
Gabra5–/– slices, an observation that supports the hypothesis that LTP is necessary for learning and memory. The results suggest that α5GABAARs are important effector molecules in the pharmacological modulation of synaptic plasticity in CA1 pyramidal neurons. The molecular mechanisms underlying this altered excitability by α5GABAARs and modulation of LTP by other
113 114 classes of GABAergic compounds, including benzodiazepines, will be the subjects of future studies.
Genetic, pharmacological and electrophysiological studies have previously implicated
α5GABAARs in learning and memory processes. A knock-in strain of transgenic mice that expressed an α5 subunit point mutation (α5H105R) showed an unexpected reduction in
α5GABAAR expression in pyramidal neurons and improved performance for hippocampus- dependent learning tasks (Yee et al., 2004). Trace fear conditioning but not delayed conditioning or contextual conditioning was facilitated in the α5(H105R) mice. The Gabra5–/– null mutant mice used in this study have previously been reported to exhibit improved performance in the initial acquisition trials of the Morris water maze (Collinson et al., 2002). Our laboratory initially tested the effects of etomidate on performance during the acquisition trials in WT and Gabra5–/– mice. The results were highly variable, and no differences were detected between genotypes
(data not shown). The probe trial, which is often considered the true criterion for the acquisition of the Morris water maze task, showed a consistent difference between genotypes in the presence of etomidate. The normal performance of the mice in the visible platform task and their impaired performance in the hidden platform task are highly consistent with a deficit in learning and memory caused by etomidate in the WT but not Gabra5–/– mice. To complement the Morris water maze studies, I also examined fear conditioning. Our laboratory and others report no difference in the baseline performance for contextual fear conditioning (Collinson et al., 2002).
However, a reduction in contextual fear was evident following administration of etomidate in
WT mice but not Gabra5–/– mice. Contextual fear was reduced similarly in both groups by ketamine, which indicates that the resistance of Gabra5–/– mice to etomidate cannot be attributed to a general resistance to neurodepressive drugs.
114 115
The results reported here have implications for human memory processes. In humans and rodents, α5GABARs exhibit similar structural, kinetic and pharmacological properties and similar patterns of expression (Faure-Halley et al., 1993; Sur et al., 1998). Compounds that selectively reduce α5GABAAR function in animal models are currently being investigated in humans as potentially nootropic compounds (Chambers et al., 2003). Our findings suggest that drugs that selectively increase the function of α5GABAARs may also have a therapeutic role. There is a need for drugs that cause amnesia without sedation or unconsciousness. Such compounds may be of value for patients whose condition is too unstable to allow adequate doses of currently available anesthetics. Furthermore, our results raise the intriguing possibility of a genetic basis for some cases of intra-operative awareness in humans. Unpleasant recall of intra-operative events can occur despite a depth of anesthesia that is associated with lack of movement (Sandin et al., 2000). Polymorphisms occurring for the human α5 subunits are associated with a reduction in receptor function (Papadimitriou et al., 2001). Moreover, the α5 subunit is down-regulated under certain pathological conditions, including epilepsy (Houser and Esclapez, 2003). A reduction in number or function of α5GABAARs may predispose patients to intra-operative awareness. Alternatively, is it possible that anesthetic effects on α5GABAARs contribute to the post-operative cognitive and memory dysfunction that occurs in approximately 25% of elderly patients (Moller et al., 1998). Animal studies have shown that subtle long-term deficits in memory performance persisted for weeks after general anesthesia (Culley et al., 2003; Culley et al., 2004). Drugs that selectively reduce α5GABAAR function may have cognitive benefits in the memory disorders that occur after general anesthesia.
115 116
In conclusion, mice with a null mutation of the α5 subunit of the GABAAR exhibited reduced sensitivity to the amnestic but not the sedative-hypnotic effects of etomidate. Amnesia was likely caused by a direct interaction between etomidate and α5GABAARs rather than an indirect effect of α5 subunit deletion, as memory impairment by ketamine was similar in Gabra5–/– and WT mice. The results contribute growing evidence that α5GABAARs play a central role in learning and memory processes.
116 117
Chapter 5. The LTP and memory impairing effects of etomidate are prevented by pre-treatment with the α5GABAA receptor selective inverse agonist L-655,708
5.1 Introduction
The single most common fear expressed by patients who are about to undergo surgery is that they will remember traumatic surgical events (McCleane and Cooper, 1990). Unfortunately, 1 in
1000 patients who undergo general anesthesia do experience some form of awareness during surgical procedures (Sandin et al., 2000; Sebel et al., 2004), and the incidence may be even higher among children (Lopez et al., 2007; Blusse van Oud-Alblas et al., 2009). Despite the disturbing frequency of the problem, the mechanisms underlying many cases of intraoperative awareness remain elusive. At the opposite end of the clinical spectrum are problems resulting from persistent memory impairment following anesthesia. POCD occurs in both elderly patients and young adults with an incidence of approximately 26% at the time of hospital discharge and
10% at three months (Johnson et al., 2002). In older patients, the cognitive domains that are most often influenced following anesthesia and non-cardiac surgery include explicit memory and executive function (Price et al., 2008). Notably, patients who exhibit POCD at both hospital discharge and 3 months later have a higher risk of dying within one year after surgery (Steinmetz et al., 2009). These “memory disorders” associated with general anesthesia have stimulated considerable interest in understanding the molecular mechanisms underlying the memory- blocking or amnestic properties of general anesthetics (Cottrell, 2008).
117 118
Many of the behavioural end points associated with the anesthetic state are mediated, at least in part, by positive allosteric modulation of GABAA receptors (GABAARs) (Rudolph and
Antkowiak, 2004; Bonin and Orser, 2008). There is an extensive diversity of GABAARs (α1-6,
β1-3, γ1-3, δ, ε, π, θ, ρ1-3) and regional and domain-specific distributions of the different subtypes present the possibility of pharmacologically targeting specific neuronal networks and behaviours (Mohler, 2007). One population of GABAARs that has been strongly implicated in mediating the memory-blocking properties of anesthetics consists of α5GABAARs (Caraiscos et al., 2004b; Caraiscos et al., 2004a; Cheng et al., 2006). A high proportion of these receptors are expressed in CA1 and CA3 pyramidal neurons of the hippocampus, a structure that is critically involved in the encoding, consolidation and retrieval of episodic memories (Suzuki, 2006).
Electrophysiological studies have shown that α5GABAARs generate a tonic inhibitory conductance in CA1 pyramidal neurons in the hippocampus (Yeung et al., 2003; Caraiscos et al.,
2004b) that is enhanced by low, memory-blocking concentrations of anesthetics (Caraiscos et al.,
2004a; Cheng et al., 2006). More importantly, Gabra5–/– mice exhibited resistance to the amnestic properties of etomidate through mechanisms that are poorly understood (Cheng et al.,
2006).
The molecular process that is thought to be essential to the storage of information involving the hippocampus is the long-term modification of excitatory glutamatergic transmission, which is known as LTP (Bliss and Lomo, 1973; Bliss and Collingridge, 1993). It is the most widely studied cellular model for memory and is evoked by repetitive stimulation of relevant afferent pathways. Similar changes in glutamatergic synaptic strength occur in vivo during memory formation (Whitlock et al., 2006). Etomidate, studied at a concentration that occurs in vivo during memory impairment, abolished LTP induced by high frequency stimulation in
118 119 hippocampal slices from WT whereas etomidate failed to inhibit LTP in hippocampal slices from
Gabra5–/– mice (Cheng et al., 2006). We and others have shown that genetic deletion of
α5GABAARs does not alter the strength of LTP evoked by high-frequency stimulation in hippocampal slices (Collinson et al., 2002; Cheng et al., 2006). Thus, the mechanisms by which
α5GABAARs regulate plasticity and the memory-blocking property of etomidate remain to be determined. It is possible that the genetically engineered deficiency of α5GABAARs produces unrecognized compensatory changes in other GABAARs, ion channels or associated proteins
(Brickley et al., 2001), which in turn confer a resistance to etomidate. Alternatively, it is plausible that etomidate increases the activity of α5GABAARs, even under conditions where these receptors do not play a dominant physiological role, and thereby attenuates synaptic plasticity and memory. To test this hypothesis, studies were designed to determine whether pharmacologically inhibiting the activity of α5GABAARs by pre-treatment with L-655,708 altered etomidate blockade of synaptic plasticity and behavioural memory. L-655,708 (FG 8094) is a imidazobenzodiazepine inverse agonist that preferentially reduces both the function of human recombinant α5GABAARs (Casula et al., 2001) and a tonic inhibitory conductance in
CA1 pyramidal neurons (Caraiscos et al., 2004b; Glykys et al., 2008). If correct, the model could account for why subjects with a reduced complement of functional α5GABAARs who exhibit normal memory performance for hippocampus-dependent learning tasks are at risk for awareness during general anesthesia.
Questions:
1. Does selectively inhibiting α5GABAARs with L-655,708 reverse the LTP-impairing and
memory blocking properties of etomidate?
2. Does L-655,708 attenuate the enhancement of the etomidate-induced tonic current?
119 120
3. Does L-655,708 intrinsically enhance LTP or hippocampus-dependent performance?
4. Does L-655,708 influence behaviours other than learning and memory, including anxiety
and spontaneous motor movement?
5.2 Materials and Methods
5.2.1 Synaptic plasticity in hippocampus slices
The CA1 region was isolated from the CA3 region by a surgical cut. The fEPSPs, which reflect the synchronized discharge of pyramidal neurons, were evoked in CA1 stratum radiatum by stimulating the Schaffer collateral afferents. Baseline stimulation frequency was 0.05 Hz with a pulse duration of 1 ms. Stimulus intensity was adjusted to evoke a half-maximal fEPSP amplitude. Once the baseline recordings were stable for more than 20 min, LTP was induced in the slices by stimulating with a TBS protocol. The TBS protocol consisted of 10 stimulus trains at 5 Hz, with each train consisting of 4 pulses at 100 Hz. The TBS mimics the frequency of theta waves that are generated in the hippocampus of mice as they explore their environment
(O'Keefe, 1993). The slope of the fEPSP was used as an indicator of synaptic efficacy and was measured between 25 and 75 % of the maximal amplitude.
Vehicle (dimethyl sulfoxide), etomidate (1 µM), L-655,708 (20 nM), or both etomidate and L-
655,708 were perfused into the recording chamber for 15 min before to the induction of LTP.
The fEPSPs were monitored prior to and 60 min following TBS. L-655,708 is an inverse agonist, which is a compound that binds to the same receptor binding-site as the agonist but has an opposite pharmacological effect. It has a 100-fold higher functional affinity for α5GABAARs than for GABAAR that contain the α1, α2 or α3 subunits (Quirk et al., 1996; Casula et al., 2001;
Atack et al., 2006b). The concentration of L-655,708 selected for use in this study binds
120 121
preferentially to α5GABAARs in tissue slices (Atack et al., 2006b). L-655,708 is not associated with proconvulsive activity as it does not alter the dose of pentylenetetrazole required to induce seizures in a rat model (Atack et al., 2006b). The concentration of etomidate for these studies was selected because it occurs in the brains of mice injected with an amnestic dose of etomidate
(Cheng et al., 2006).
5.2.2 Voltage clamp recordings
The extracellular recording solution contained 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX)
(20 µM) and (2R)-amino-5-phosphonovaleric acid (APV) (10 µM) to block ionotropic glutamate receptors and TTX (0.3 µM) to block voltage-dependent Na+ channels. Patch pipettes had open tip resistances of 3-5 MΩ when filled with an intracellular solution that contained (in mM): CsCl
(140), HEPES (10), EGTA (10), MgATP (2), CaCl2 (1) (pH 7.3 with CsOH, 295-305 milliosmolar). Currents were sampled at 10 kHz and filtered at 2 kHz by using an eight-pole low pass Bessel filter. All cells were recorded with a holding potential of –60 mV and were recorded continuously. A stable baseline current (< 20% change) was ensured before the application of drugs. The amplitude of the tonic current under control conditions was measured as the difference in the holding current before and during the application of etomidate (1 µM), L-
655,708 (20 nM), or bicuculline (10 µM).
5.2.3 Fear-conditioned learning
In the Pavlovian fear-conditioning tasks, mice were exposed to a tone (the conditioning stimulus
[CS]) and a foot shock (the unconditioned stimulus [US]) in a novel conditioning context, with either no time delay (0 s for cued conditioning) or with an interval of 20 s between stimuli (trace
121 122 conditioning) (Quinn et al., 2002; Misane et al., 2005; Wiltgen et al., 2005). Several different associative memory tests were conducted to determine the contribution of certain brain regions for which the extent of expression of α5GABAARs differs. More specifically, the dorsal hippocampus, which has a high expression of α5GABAARs, is known to play a key role in contextual and trace fear conditioning (Wanisch et al., 2005). In contrast, the expression of
α5GABAARs is relatively low in the amygdala (Sur et al., 1999). Cued fear conditioning, which requires the basal lateral nucleus of the amygdala, served as a control (Sigurdsson et al., 2007).
Thirty minutes before being placed in the fear conditioning chamber, mice were randomly assigned to receive an intraperitoneal injection (2 ml/kg) of either vehicle (35% propylene glycol, 10% dimethyl sulfoxide), etomidate (4 mg/kg), L-655,708 (0.68 mg/kg), or the combination of etomidate and L-655,708 (administered together). For these experiments, the dose of etomidate (Cheng et al., 2006) was carefully selected to cause conscious amnesia, a state characterized by minimal sedation (which confounds the study of learning and memory) combined with loss of explicit or episodic memory (Veselis et al., 2009). Additionally, the dose of L-655,708 was selected to modify learning behaviours via preferential modulation of
α5GABAARs, as previously determined (Atack et al., 2006b; Atack et al., 2006c). On day 1, single animals were allowed to explore the chamber for 180 s. An 800-Hz tone, created by a frequency generator, amplified to 70 dB, and lasting 20 s, was then presented. For cued fear conditioning, the last 2 s of each auditory tone was paired with an electric foot shock (2 s, 1 mA); for trace fear conditioning, the auditory stimulus and foot shock (2 s, 0.5 mA) were separated by 20 s. Each of these sequences was presented 3 times, separated by 60 s (for cued fear conditioning) or 240 s (for trace fear conditioning). For contextual fear conditioning either a strong (2 s, 1 mA) or weak (2 s, 0.5 mA) foot shock was applied depending on the protocol. On
122 123 day 2, 24 h after the conditioning session, each mouse was assessed for a freezing response by placing it in the original context and scoring every 8 s for a total of 8 min to determine contextual fear. On day 3, the conditioning chamber was modified to measure the freezing response to the tone to study either cued or trace fear conditioning. Mice were monitored for 180 s for freezing to the modified context, to rule out contextual influences. After the monitoring period, the auditory tone was presented continuously for 300 s, and the freezing response was measured every 8 s.
5.2.4 Water maze learning
The water maze is a hippocampus-dependent spatial navigation task that requires the mouse to use visual cues positioned around the room to locate a hidden platform in a circular tub of opaque water (Morris et al., 1982). Mice were pre-trained for 10 days, with 4 trials on each day.
In the match-to-place paradigm, the location of the platform was changed daily and the first trial of each day was used as a comparator or reference trial to determine learning on trials 2, 3 and 4.
Each mouse had 60 s to search for and locate the hidden platform and each mouse was allowed to remain on the platform for 30 s between trials. During the acquisition phase of the probe trial, each mouse was randomly assigned to receive an intraperitoneal injection of vehicle, etomidate
(4 mg/kg), L-655,708 (0.68 mg/kg), or both etomidate and L-655,708 (administered together) 30 min before the experiment. The next day, a probe trial was performed to test the ability of the mice to recall the correct spatial location that previously contained the hidden platform. During the probe trial, the hidden platform was removed from the pool and the mouse was allowed to search for 60 s. The mouse was then promptly rescued and the trial ended. Mice that learned the correct location of the platform preferentially searched in the correct quadrant. The percentage of time spent within the correct quadrant was used to measure recall. If no learning occurred, the
123 124 mouse was expected to spend 25% of the time in the correct quadrant (i.e. ¼ of the total area).
Data records were stored with HVS Water 2020 software (VHS Image, Hampton, UK) for off- line analysis. Briefly, a video camera captured the movement of the mouse, and the software tracked the mouse by contrast comparison. The time, swim path, and latency of each mouse were recorded during each trial, and the percentage of time spent in the correct region was calculated by the software during analysis.
Visible platform trials were also performed to test for possible differences in motivational factors, perceptual and motor abilities, and any possible nonspecific effects of etomidate and L-
655,708. In these trials, the platform that was usually submerged was raised to 0.5 cm above the waterline and was marked with a brightly coloured flag. The mice were injected 30 min before the visible platform trial, similar to the treatment during the learning acquisition phase prior to the probe trial.
5.2.5 Elevated plus maze
The elevated plus maze is designed to measure the anxiety levels of the mice. The test hinges on the natural tendency of rodents to explore a novel environment and their aversion to open, elevated and brightly lit areas. The elevated plus maze consisted of four arms (5 cm × 27.5 cm) that were joined by a central area (5 cm × 5 cm). Two opposite arms were enclosed by 30 cm- high walls, and the other two arms were open. The maze was elevated to a height of 50 cm above the floor. The floor of the maze was constructed of Plexiglas to facilitate cleaning between mice.
The maze was placed in the center of the room, with illumination from fluorescent strip lights that were covered with polarizing film.
124 125
Mice were injected i.p. with either vehicle, etomidate (4 mg/kg), L-655,708 (0.68 mg/kg), or both etomidate and L-655,708 (administered together) 30 min before the experiment. Mice were placed in the central area of the maze facing an open arm and were scored for the amount of time they spent in the central area, the open arms or the closed arms. The number of entries into the open and closed arms was also monitored. All mice were allowed to explore the maze for 5 min.
The maze was wiped clean between trials with an ethanol-containing solution.
5.2.6 Statistical analysis
Statistical analyses for the electrophysiological and behavioural data were completed with
GraphPad Prism V4.0c (San Diego, CA). All pooled data are presented as mean ± standard error of the mean. Electrophysiological and behavioural statistical comparisons were completed using a one-way (i.e. drug treatment only) or two-way (drug treatment X genotype) analysis of variance (ANOVA) with two-tailed inference testing. Post hoc analyses were conducted using the Tukey-Kramer method, which accounts for both equal and unequal sample size comparisons.
For the plots of LTP, the data points (slope of the fEPSP measured between 25% and 70% of the rising phase) were binned in 1-min increments to facilitate readability. The extent of LTP was quantified for statistical comparisons by averaging the slope of the fEPSPs during the final 5 min of each experiment and normalizing to baseline values. P < 0.05 was considered as statistically significant.
125 126
5.3 Results
5.3.1 L-655,708 reverses etomidate blockade of long-term potentiation
First, to determine whether a reduction in α5GABAAR activity modifies synaptic plasticity induced by TBS, L-655,708 (20 nM) was applied at a concentration that selectively blocks the tonic inhibitory conductance in CA1 pyramidal neurons without substantially altering inhibitory synaptic transmission (Glykys et al., 2008). In vehicle-treated slices, robust LTP was induced following a 1 s presentation of TBS, such that the slope of the fEPSP was significantly increased
(P < 0.05 versus baseline; n = 8; Figure 5.1A). The application of L-655,708 failed to modify the strength of LTP (P > 0.05 versus control slices; Figure 5.1A). Next, the effect of L-655,708 on synaptic excitability was studied by comparing the slope of the input-output relationship, where current intensity was plotted against the slope of the fEPSP (Figure 5.1B). The application of L-
655,708 following TBS did not further enhance synaptic excitability. Thus, a decrease in
α5GABAAR activity, similar to a reduction in the expression of α5GABAARs, did not modify synaptic plasticity or neuronal excitability under baseline conditions.
Next, slices from WT mice were co-treated with the same concentration of L-655,708 and etomidate to determine whether inhibiting α5GABAARs can reverse etomidate-induced blockade of LTP. Etomidate inhibited LTP in WT slices stimulated with the TBS protocol (P < 0.05 versus control LTP; n = 8; Figure 5.1C) (Cheng et al., 2006). This effect of etomidate was reversed by the co-application of L-655,708 (P = 0.02 versus etomidate-treated slices; n = 8;
Figure 5.1C). Thus, inhibition of α5GABAARs reverses the LTP-blocking property of etomidate.
In addition, etomidate blocked an increase in synaptic excitability following TBS; this effect was not observed in slices treated with both etomidate and L-655,708 (Figure 5.1D).
126 127
Figure 5.1. The intrinsic activity of α5GABAARs does not enhance LTP in hippocampal slices but reverses etomidate-induced LTP impairment.
(A) L-655,708 does not potentiate LTP above control levels. This suggests minimal involvement for α5GABAARs in TBS LTP. (B) The input-output curves pre- and post-TBS in the presence and absence of L-655,708. There were no differences between vehicle-treated and L-655,708- treated slices, although excitability increased following TBS in both groups. (C) Etomidate blocked LTP and this effect was reversed by applying both L-655,708 and etomidate. (D) The input-output curves pre- and post-TBS with etomidate application in the presence and absence of
L-655,708. There were no differences between vehicle-treated and L-655,708-treated slices.
Raw traces presented above the LTP plots represent the no-drug baseline fEPSP (green), drug baseline fEPSP (black) and the drug post-TBS fEPSP (red or blue). Calibration bars: 0.5 mV, 10 ms.
127 128
128 129
5.3.2 Etomidate potentiation of the CA1 tonic conductance is blocked
by L-655,708 application
To confirm that L-655,708 reversed the enhancement of the tonic conductance by etomidate, voltage clamp experiments were performed. Etomidate application caused an significant increase in the tonic current, as evidenced by an inward shift in the holding current (IHold; n = 5, Figures
5.2A, 5.2B and 5.2C), as previously reported in cultured hippocampal cells (Cheng et al., 2006).
Next, to determine the proportion of the etomidate-potentiated tonic current that was attributed to
α5GABAARs, L-655,708 was applied. Application of L-655,708 (20 nM) resulted in a significant reduction in the IHold from the etomidate-potentiated current (n = 5, Figure 5.2) suggesting that a large proportion of the etomidate-enhanced tonic current is mediated by
α5GABAARs. Specifically, L-655,708 blocked 73 % of the total etomidate current, whereas bicuculline blocked 120% of the total current. Interestingly, etomidate and L-655,708 did not influence the kinetics of mIPSCs (Table 5.1), suggesting that L-655,708 mediates its effects on etomidate by predominantly extrasynaptic mechanisms.
129 130
Figure 5.2. L-655,708 attenuated the etomidate-induced increase in the holding current in CA1 pyramidal neurons.
(A) Current traces indicate an inward tonic current with etomidate application. The etomidate increase in the holding current is 73.18 ± 9.05 % blocked by L-655,708 application. Bicuculline was applied at the end of the recording to reveal the total tonic conductance. (B) The all point histograms for the current traces and the shifts in the holding current are shown. (C) Pooled data showing the relative changes in the holding current with etomidate, L-655,708 and bicuculline application. (D) The percentage of the block by L-655,708 and bicuculline on the total etomidate-induced tonic current is shown.
130 131
131 132
Table 5.1. Effects of etomidate and L-655,708 on spontaneous mIPSCs.
132 133
Values are means ± sem. There were no statistically significant effects between the different conditions (P > 0.05 for all comparisons). Parentheses enclose n values. mIPSC, miniature inhibitory postsynaptic current, aCSF, artificial cerebrospinal fluid; ETOM, etomidate; L6, L-
655,708.
133 134
5.3.3 L-655,708 reverses etomidate memory blockade
Contextual and cued fear conditioning protocols were used to measure the ability of mice to learn the association between a strong (2 s, 1 mA) aversive foot shock and environmental or auditory cues, respectively (Fanselow, 1980). The general procedure and the training protocols used to assess fear conditioning are shown in Figure 5.3A and Figure 5.3B, respectively. In contextual fear conditioning, mice pre-treated with L-655,708 exhibited robust freezing that was similar to that of vehicle-injected control mice (85.4 ± 3.3% versus 82.1 ± 3.1%; P < 0.001; n = 8 per group; Figure 5.3C). As such, either α5GABAARs are not important for contextual fear memory or the fear conditioning protocol produced a saturating response, such that L-655,708 could produce no further enhancement. The strong contextual fear response was considerably reduced in mice that had been injected with etomidate (50.2 ± 7.8% of time spent freezing compared to vehicle-injected controls; n = 8; Figure 5.3C). Mice treated with both etomidate and
L-655,708 (0.68 mg/kg) displayed freezing levels comparable to those observed with the vehicle control (77.2 ± 4.0%; P > 0.05; n = 8; Figure 5.3C). This later result indicates that decreasing
α5GABAAR activity completely reversed the impairment of memory by etomidate.
To address the concern that the initial experimental conditions used to study contextual fear memory produced a saturated freezing response (i.e. a ceiling effect), the level of the foot shock or US was reduced (2 s, 0.5 mA). Under these new conditions, the effects L-655,708 and etomidate were studied in WT and Gabra5–/– mice. The presentation of the weaker foot shock significantly reduced the baseline freezing in control mice when compared to mice trained with the stronger (2 s, 1 mA) foot shock (freezing with strong foot shock, 82.1 ± 3.1%; Figure 5.3C; versus weak foot shock, 53.8 ± 5.9%; Figure 5.3D; P < 0.01). Despite a weak contextual fear
134 135 conditioning protocol, there were no differences between WT and Gabra5–/– mice (53.8 ± 5.9% versus 59.9 ± 4.0%; n = 10 and n = 9, respectively; P > 0.05; Figure 5.3D). Notably, L-655,708 did not further enhance freezing in WT mice or Gabra5–/– mice (58.4 ± 7.2% versus 57.6 ±
6.4%; n = 10 and n = 9 respectively; P > 0.05; Figure 5.3D). Injections of etomidate reduced the freezing scores for WT but not Gabra5–/– mice (30.6 ± 3.3% versus 55.3 ± 6.0%; n = 11 and n =
10, respectively, P < 0.05). The ability of etomidate to reduce freezing scores was reversed by L-
655,708 (52.2 ± 5.5% versus 55.2 ± 5.6%; n = 10 and n = 9, respectively; P > 0.05; Figure 5.3D).
Consistent with the changes in synaptic plasticity, these findings indicate that α5GABAAR activity is not important for baseline contextual fear conditioning, but these receptors can be activated by etomidate to impair memory performance in this task. Finally, the effects of etomidate and L-655,708 were studied on basolateral amygdala-dependent cued fear conditioning. Neither drug had a significant effect (one-way ANOVA, P=0.52; Figure 5.3E).
This result was expected given that the expression of α5GABAARs in the amygdala is relatively low (Sur et al., 1999; Pirker et al., 2000).
5.3.4 α5GABAAR activity impairs performance for trace fear
conditioning
In trace fear conditioning, the strength of classical conditioning can be reduced by introducing a time interval (or “trace”) between the conditioned stimulus and the unconditioned stimulus. For trace fear conditioning, Gabra5−/− mice treated with vehicle outperformed their WT littermates, as indicated by significantly higher freezing scores (WT, 37.7 ± 6.8% for WT mice versus 72.8 ±
5.4% for Gabra5–/–; n = 11 and n = 10, respectively; P < 0.05, Figure 5.3F). To confirm that the difference between the genotypes was attributable to a reduction in α5GABAAR activity, mice
135 136 were injected with L-655,708, which considerably improved freezing scores for WT mice but had no effect on Gabra5–/– mice (73.6 ± 2.7% versus 67.6 ± 5.8%; n = 8 per group, respectively; P > 0.05; Figure 5.3F). Interestingly, etomidate did not significantly decrease freezing in WT and Gabra5–/– mice (33.2 ± 6.8% versus 65.1 ± 6.7%; n=10 per group, respectively) as the freezing scores were similar to those of vehicle-injected mice (P > 0.05;
Figure 5.3F). However WT mice that received both etomidate and L-655,708 displayed a high level of freezing, which was no different than Gabra5–/– mice (59.1 ± 5.4% versus 72.0 ± 4.8%; n = 10 and n = 9, respectively; P > 0.05; Figure 5.3F).
136 137
Figure 5.3. The expression and activity of α5GABAARs modify contextual and trace fear conditioning.
(A) The basic procedure used for the fear conditioning protocol is illustrated. Injection of the drug was followed 30 min later by the fear conditioning protocol, which always consisted of 3 consecutive foot shocks. During the conditioning, three foot shocks were paired with a 20 s tone
(see panel B. or the Methods for details of the fear conditioning protocols). Following the fear conditioning protocol, the mice were tested 24 hr later for freezing to context or 48 hr later for freezing to the tone. For the assessment of freezing to the tone the conditioning chamber was modified, such that the shape was circular, a rubber mat covered the shock grid and visual cues were located on the walls surrounding the chamber. (B) A schematic representation illustrating the timing for all three fear conditioning protocols is shown. In all protocols a baseline activity period of 3 min preceded the conditioning procedure. (C) L-655,708 did not enhance contextual fear conditioning when a strong foot shock was used during training. Etomidate impaired performance in contextual fear conditioning, and L-655,708 restored freezing to control values when the two drugs were co-administered. (D) When a weaker foot shock was used for the US, contextual freezing scores were not enhanced by pre-treatment with L-655,708 or in Gabra5–/– mice. Etomidate impaired contextual fear conditioning in WT but not Gabra5–/–, and L-655,708 occluded this effect. (E) Etomidate and L-655,708 did not influence performance in amygdala- dependent cued fear conditioning. (F) The performance of Gabra5–/– mice was enhanced in trace fear conditioning (a weak associative task), relative to the effect in vehicle-treated WT mice; in addition, inhibiting α5GABAARs with L-655,708 improved the performance of WT mice to the level observed in Gabra5–/– mice. Etomidate did not significantly reduce freezing scores in WT mice and Gabra5–/– mice in trace fear conditioning.
137 138
138 139
5.3.5 Etomidate impairment of spatial memory is reversed with L-
655,708
Next, the Morris water maze was used as an independent measure of hippocampus-dependent learning (Morris et al., 1982). Notably, regardless of drug treatment, the performance of the mice was similar for the acquisition trials of the water maze task (Figure 5.4A). The mean time savings (time to locate platform during trial 4 minus time required during trial 1) was used to quantify immediate memory. There were no significant differences in the mean time savings to locate the hidden platform between treatment groups (7.2 ± 2.0 s for vehicle-injected mice, 5.7 ±
2.6 s L-655,708-treated mice, 6.1 ± 3.1 s for etomidate-treated mice, 5.4 ± 3.5 s for etomidate-L-
655,708-treated mice; n = 28 per group; P < 0.05; Figure 5.4A). The results suggest that neither the up- nor down-regulation of α5GABAAR activity influenced the ability of the mice to initially learn and complete the task.
To study long-term memory performance, mice underwent a probe trial 24 h following the acquisition of the task. L-655,708 did not enhance the recall of the hidden platform location, as shown by the percentage of time spent swimming in the correct quadrant of the water maze (30.2
± 1.9% of time spent in the correct quadrant for vehicle-injected mice versus 31.5 ± 1.9 for L-
655,708-treated mice; n = 28 per group; P > 0.05; Figure 5.4B). In contrast, etomidate decreased the total amount of time spent swimming in the correct quadrant of the pool during the probe trial (21.4 ± 1.5%, n = 28, P < 0.05, Figure 5.4B). This reduction in performance was not exhibited by the mice that were injected with both etomidate and L-655,708 (31.5 ± 1.9; n = 28;
P > 0.05).
139 140
Thus the mice demonstrated equal learning during the acquisition phase of the water maze task, but etomidate impaired recall of the task 24 h later, and this effect that could be reversed by L-
655,708.
The visible platform studies revealed that there were no differences among treatment groups in the latency to locate the platform in the presence of any drug combination (4.1 ± 1.1 s for vehicle-injected mice, 4.7 ± 2.8 s for L-655,708-treated mice, 4.1 ± 2.9 s etomidate-treated groups, 3.9 ± 2.1 for etomidate-L-655,708-treated mice; n = 28; P > 0.05; Figure 5.4C). There were also no differences among the groups in terms of mean swimming speed during the acquisition trial (24.2 ± 0.60 cm/s for vehicle-injected mice, 25.2 ± 0.4 cm/s for L-655,708- treated mice, 23.4 ± 0.6 cm/s for etomidate-treated mice, 23.8 ± 0.6 cm/s for L-655,708-treated mice; n = 28; P > 0.05; Figure 5.4D). The lack of an effect on swimming speed confirmed the procedural ability of the mice to perform the tasks.
5.3.6 α5GABAARs do not contribute to the anxiety-like behaviours
The inhibition of α5GABAARs with L-655,708 has been shown to increase anxiety-like behaviours in the elevated plus maze (Navarro et al., 2002; Atack et al., 2006b). This effect could possible confound the performance of mice in the fear conditioning and water maze experiments. However, this anxiogenic effect has been attributed to inhibition of other GABAAR subtypes since α5GABAAR-selective doses were not used in other studies (Navarro et al., 2002;
Atack et al., 2006b). Nevertheless, to determine whether the activity of α5GABAARs contributed to anxiety-like behaviours, I tested mice treated with etomidate and L-655,708 in the elevated plus maze at the same doses as used for the memory assays. There were no significant differences in the time spent in the open arms (1.2 ± 0.3 s for vehicle-treated mice, 2.1 ± 1.2 s for
140 141
L-655,708-treated mice, 0.5 ± 0.4 s for etomidate-treated mice, 1.1 ± 0.4 s for etomidate-L-
655,708-treated mice; n = 8 per group; P > 0.05; Figure 5.5A) and the closed arms (87.8 ± 2.3 s for vehicle-treated mice, 87.3 ± 3.5 s for L-655,708-treated mice, 93.7 ± 1.7 s for etomidate- treated mice, 91.8 ± 1.6 s for etomidate-L-655,708-treated mice; n = 8 per group; P > 0.05;
Figure 5.4A) of the elevated plus maze. Similarly, there was no difference in the frequency of entries into the open arms (0.8 ± 0.3 for vehicle-treated mice, 1.1 ± 0.5 for L-655,708-treated mice, 0.4 ± 0.2 for etomidate-treated mice, 0.8 ± 0.3 for etomidate-L-655,708-treated mice; n = 8 per group; P > 0.05; Figure 5.5B) and closed arms (1.7 ± 0.1 s for WT mice versus 8.4 ± 4.9 s for
Gabra5–/– mice; n = 8 per group; P > 0.05; Figure 5.5C) and closed arms (94.4 ± 1.5 s for WT mice versus 82.6 ± 6.3 for Gabra5–/– mice; n = 8 per group; P > 0.05; Figure 5.5C) and the frequency of entries into the open and closed arms (data not presented; P > 0.05) were not significantly different. Furthermore, etomidate did not alter the time spent in the open arms (1.9
± 0.8 s for WT mice versus 7.1 ± 25.5 s for Gabra5–/– mice; n = 8 per group; P > 0.05; Figure
5.5D) or the closed arms (90.4 ± 2.1 s for WT mice versus 83.5 ± 9.2 s for Gabra5–/– mice; n =
8 per group; P > 0.05; Figure 5.5D), nor did it affect the frequency of entry into either type of arm (data not shown; P > 0.05).
141 142
Figure 5.4. Normal acquisition of the matching to place version of the Morris water maze but impaired recall with etomidate-treatment.
(A) Injections of either L-655,708 or etomidate do not influence the acquisition of the matching to place version of the Morris water maze. All injections were done on d 11 following 10 d of naive training in the water maze. (B) L-655,708 did not enhance free recall of the platform location, 24 h following injection, in the probe trial of the water maze. Etomidate impaired performance in the water maze, but co-application with L-655,708 returned performance to control levels. The percentage of time spent swimming in the target quadrant versus the average time spent in the other quadrants (non-target) was calculated during the probe trial. The drug treatments did not influence the swim speed (C) or the visible platform trial (D) of the mice in the water maze. ** denotes a statistically significantly difference from the control group at P <
0.05.
142 143
143 144
Figure 5.5. L-655,708 and etomidate do not contribute to anxiety-like behaviors in the elevated plus maze.
(A) L-655,708 and etomidate did not change the time spent by mice in the open or closed arms of the elevated plus maze. (B) L-655,708 and etomidate also did not change the total number of entries into the open and closed arms of the elevated plus maze. (C) There were no differences between WT and Gabra5–/– mice for the amount of time spent in the open or closed arms of the elevated plus maze. This confirms that α5GABAARs do not readily contribute to anxiety-like behaviours. (D) Etomidate did not influence the amount of time spent in either the open or closed arms of the elevated plus maze in WT or Gabra5–/– mice.
144 145
145 146
5.4 Discussion
This study supports a pharmacogenetic mechanism to account for resistance to the memory- blocking properties of etomidate. The results show that α5GABAARs do not regulate baseline plasticity evoked by high frequency stimulation in an in vitro mouse hippocampus slice model nor behavioural performance for contextual fear and spatial navigation memory; nevertheless, etomidate impairs plasticity and memory by increasing α5GABAAR activity. Most importantly, these memory-blocking effects of etomidate can be completely reversed by pre-treatment with L-
655,708.
GABAARs play a critical role in orchestrating neuronal activities by altering spike timing in neurons and synchronized rhythms in neuronal circuits. Intravenous anesthetics including etomidate, as well as most inhaled agents such as isoflurane and sevoflurane increase the activity of GABAARs (Uchida et al., 1995; Wu et al., 1996; Krasowski et al., 1998; Caraiscos et al.,
2004a). This facilitation generally causes an increase in the chloride conductance, membrane hyperpolarization and the shunting of excitatory currents, which together reduce neuronal excitability (Ben-Ari et al., 2007). The propensity for general anesthetics to block the induction of LTP by increasing GABAAR activity has been widely reported (Ishizeki et al., 2008). At the concentrations used for this study, etomidate did not block LTP via a non-selective enhancement of GABAergic neurotransmission but rather a selective potentiation of α5GABAARs. Moreover, inhibiting all GABAARs with non-selective antagonists such as picrotoxin and bicuculline is known to enhance plasticity evoked by high frequency stimulation (Wigstrom and Gustafsson,
1983) and reverse anesthetic blockade of LTP (Simon et al., 2001; Wei et al., 2002; Ishizeki et al., 2008). This study shows that pre-treatment with the α5GABAAR-preferring agent L-655,708 is sufficient to reverse etomidate blockade of LTP and memory blockade.
146 147
I found that etomidate, at the concentration studied, preferentially enhanced the tonic rather than synaptic inhibitory conductance in CA1 pyramidal neurons and this effect was attenuated by L-
655,708. Higher concentrations of etomidate, which are likely less selective, may enhance both tonic and synaptic inhibition; nevertheless, several previous comparisons of inhibitory charge mediated by the tonic and synaptic currents showed that enhancement of the tonic current by anesthetics is multiple-fold greater than synaptic inhibition (Bai et al., 1999; Mody and Pearce,
2004). Thus, I attribute the effects of etomidate on network plasticity and behaviour primarily to an enhancement of the tonic conductance.
It has been shown that α5GABAARs are predominantly expressed extrasynaptically (Brunig et al., 2002) and that these receptors generate a tonic inhibition in rodent and human hippocampus pyramidal neurons (Bai et al., 2001; Caraiscos et al., 2004b; Scimemi et al., 2005; Glykys and
Mody, 2006; Prenosil et al., 2006; Scimemi et al., 2006). However, immunocytoimaging
(Christie and De Blas, 2002) and electronmicroscopy (Serwanski et al., 2006) studies showed that the α5 subunit is also expressed at synaptic regions of hippocampal pyramidal neurons. and that these synaptic receptors do not appreciably contribute to fast IPSCS (Caraiscos et al., 2004b;
Prenosil et al., 2006; Bonin et al., 2007; Zarnowska et al., 2009). Instead synaptically expressed
α5GABAARs are thought to generate a subset of slowly activating and deactivating currents that are termed GABAA slow or IPSCslow (Pearce, 1993; Prenosil et al., 2006; Zarnowska et al.,
2009). While IPSCslow contributes to less than 1% of synaptic GABAergic inhibition, this conductance may powerfully regulate plasticity due to its position in neuronal circuitry and its temporal association with NMDA receptor activation (Kapur et al., 1997). The influence of low concentrations of etomidate on this IPSCslow current remains to be studied.
147 148
There were several unexpected results from this study. I found that the activity or expression of
α5GABAARs did not influence baseline synaptic plasticity an effect previously reported in
Gabra5–/– slices (Collinson et al., 2002; Cheng et al., 2006). Similar to these results, the strength of LTP induced by high-frequency stimulation was similar in hippocampal slices from
α5(H105R) and WT mice (Crestani et al., 2002). However, others have shown that benzodiazepine inverse agonists selective for α5GABAARs enhanced synaptic plasticity in hippocampal slices (Atack et al., 2006b; Dawson et al., 2006). Several factors could account for the differences between the previously published benzodiazepine inverse agonist data and the current results.
Notably, while many populations of GABAARs are expressed in neuronal networks and within the same cells, only some types may be active during any given condition (Glykys and Mody,
2007a). When conditions such as the intensity of the network stimulation or the amount of
GABA in the extracellular space change, different subpopulations of GABAARs are expected to become activated. Thus, the contribution of particular GABAAR populations to synaptic plasticity critically depends on the in vitro experimental protocol. Certainly caution must be exercised when making direct comparisons between studies that use different animal species, electrophysiological paradigms and even subtle differences in the behavioural protocols. For example, I found that L-655,708 failed to alter baseline synaptic plasticity induced by a 1 sec presentation of TBS, whereas others showed that L-655,708 enhanced plasticity induced by a brief priming stimulus (10 stimuli at 100 Hz) followed 30 min later by TBS (Atack et al., 2006b;
Dawson et al., 2006). Following the priming stimulus, synaptic strength was increased to 200% in both the L-655,708-treated and control slices suggesting that α5GABAAR do not play a
148 149 critical role. However, after the second phase of stimulation L-655,708-treated slices showed enhanced plasticity. Thus, the stimulation protocol can dramatically influence the subpopulation of GABAARs that influence plasticity.
Additionally, in the current experiments a reduction in α5GABAAR activity failed to enhance contextual fear conditioning or spatial navigational memory even though α5GABAARs play a key role in hippocampus-dependent memory performance (Collinson et al., 2002; Crestani et al.,
2002; Yee et al., 2004). The strength of associated learning depends on the strength of the aversive stimulus and the number of presentations (Quinn et al., 2008). To determine whether the experimental conditions contributed to the inability to demonstrate improved contextual learning
(i.e. a ceiling effect), a weaker foot shock was used in some studies. Even under these modified conditions where baseline freezing scores were reduced, L-655,708 did not strengthen contextual learning. Thus, α5GABAARs play a minimal role in processes that elicit moderate and strong contextual memory.
In contrast, the performance of Gabra5–/– mice was greater than WT control mice in trace fear conditioning. This result is consistent with studies of α5(H105R) point mutant mice, which have a partial deficit of α5GABAARs (Crestani et al., 2002). The α5H105R mice exhibit higher baseline freezing scores for trace fear conditioning compared with WT littermates. Trace fear conditioning adds complexity to the delay conditions, as the time interval requires the formation of a temporal relationship between the two stimuli. The reasons for differences in trace fear learning but not contextual fear in Gabra5–/– versus WT mice remain to be determined as the hippocampus is required for tone-shock association in rodents and humans during contextual learning (Fanselow, 1980; Clark and Squire, 1998) and spatial navigation (Morris et al., 1982).
149 150
However, these results are at odds with studies that showed a reduction in α5GABAAR activity enhances learning performance (Collinson et al., 2002; Crestani et al., 2002; Chambers et al.,
2004; Dawson et al., 2006; Gerdjikov et al., 2008) and the control of hippocampal network excitability (Crestani et al., 2002; Yee et al., 2004; Glykys and Mody, 2006). Others have shown that L-655,708 increased the performance of rats in a water maze probe trial; however, for these experiments, the probe trial was performed 15 min after completion of a series of rigorous training trials (Atack et al., 2006b). In contrast, I studied the probe trial 24 h following training and the differences in the experimental paradigms could account for the differences between studies. Clearly, αGABAARs regulate memory under baseline conditions in a subtle and intriguing manner, which is worthy of further study.
To reconcile the in vitro and behavioural data I propose the following schematic model to account for etomidate blockade of plasticity under the current experimental conditions (Figure
5.6). Since no difference in baseline plasticity was observed following high levels of stimulation,
I reasoned that α5GABAARs do not play a major role. I suggest that intense stimulation recruits both high and low affinity GABAARs such that the contribution of α5GABAARs is overshadowed or obscured by other subpopulation of GABAARs (Figure 5.6A) (Chapman et al.,
1998; Smith et al., 2001). Pharmacological modulators like etomidate that preferentially target
α5GABAARs cause a “supra-physiological” activation of these receptors causing it to exert a dominant role in inhibiting plasticity (Figure 5.6B). Finally, the “super-activated” α5GABAARs are inhibited by L-655,708, which blocks etomidate-induced LTP impairment and memory blockade. The model further implies that: 1) target receptors are expressed in neuronal circuits that normally regulate memory behaviour; 2) receptors function to regulate network activity in a constrained manner; 3) clinically-relevant concentrations of the anesthetic “super-activate”
150 151 receptors beyond normal physiological limits; and 4) L-655,708 reverses the effects of etomidate on α5GABAARs. Clearly, the current experimental results and model do not rule out actions of etomidate on other GABAA receptors under different conditions of plasticity.
There are important implications of these results regarding strategies to reverse memory impairment by general anesthetics. A critical concern for the development of reversal agents is whether the agents evoke seizures or other adverse excitatory effects. Non-selective GABAAR antagonists have little clinical utility because of proconvulant and anxiogenic properties
(Veliskova et al., 1990; Dalvi and Rodgers, 1996; Czlonkowska et al., 2000) Similarly, non- selective inverse agonists including methyl-6,7-dimethoxy-4-ethyl-beta-carboline-3-carboxylate
(DMCM) and FG 7142 are proconvulsant and anxiogenic and not suitable for clinical use. The
α5GABAAR-preferring agent L-655,708 is not proconvulsant and may offer a more favourable therapeutic profile because of the limited distribution of α5GABAARs (Atack et al., 2006b).
Another selective α5GABAAR inverse agonist, α5IA is also not proconvulsive (Dawson et al.,
2006) and has been shown in humans to reverse the memory-blocking effects of ethanol (Nutt et al., 2007). While L-655,708 or α5IA are not currently available for clinical use, these compounds serve as prototypes for drug development.
The above behavioural results provide a compelling foundation for further work, yet the studies have limitations. First, the dose of etomidate was selected to cause amnesia and not general anesthesia. At higher doses, etomidate may modify the activity of other GABAAR subtypes and other neurotransmitter systems (Hill-Venning et al., 1997; Grasshoff et al., 2007). Second, high efficacy and affinity-selective α5GABAARs compounds including L-655,708 and α5IA are anxiogenic and proconvulsant at higher doses as they reduce the activity of α1, α2 and α 3
151 152
subunit-containing GABAAR subtypes (Navarro et al., 2002; Atack et al., 2006b). Third, it remains to be determined whether α5GABAARs-preferring agents reverse memory impairment by inhaled anesthetics (Caraiscos et al., 2004a). Also, the behavioural paradigms used to study hippocampus-dependent memory do not mimic the clinical scenarios involving etomidate anesthesia. Finally, undesirable memory impairment following anesthesia may involves many cell signalling pathways and receptors.
Finally, given the striking homology between the structural and pharmacological properties of
GABAARs in non-primates and humans (NCBI Nucleotide Database Genbank, http://www.ncbi.nlm.nih.gov/Genbank/index.html), it is tempting to speculate that reduced
α5GABAAR activity contributes to some patients who experience awareness despite an apparently adequate dose of anaesthesia (McCleane and Cooper, 1990; Sandin et al., 2000; Sebel et al., 2004; Lopez et al., 2007). Intraoperative awareness, defined as the explicit recall of surgical events that occur during general anesthesia, has been associated with serious psychological outcomes, including posttraumatic stress disorders and depression (Samuelsson et al., 2007). This study raises compelling questions regarding α5GABAAR function in patients who experience inexplicable awareness during anesthesia. Animal models show that pathological conditions, including epilepsy and chronic alcohol abuse, alter expression of the α5 subunit
(Glatt et al., 1992; Glatt et al., 1994; Papadimitriou et al., 2001). Also, polymorphisms of the human Gabra5 gene occur, although their functional significance is still unknown. Studies of patients identified in the Registry for Intraoperative Awareness established by the American
Society of Anesthesiology (http://depts.washington.edu/awaredb/) might shed light on this association. Our results suggest a mouse model could be useful in developing rational prevention strategies. Finally, the implications of the study extend well beyond the purview of
152 153
anesthesiology, as they strongly implicate α5GABAARs as targets for the development of memory-modifying drugs.
153 154
Figure 5.6. A model illustrating the regulation of synaptic plasticity in the hippocampus by extrasynaptic α5GABAARs and blockade of LTP by etomidate.
(A) High-frequency stimulation leads to intense inhibitory drive and activation of postsynaptic glutamate and GABAARs. LTP is present under these conditions because of the preferentially stronger activation of glutamatergic synapses. (B) Despite strong activation of glutamate synapses, LTP is not generated during the application of a general anesthetic etomidate. This model proposes that general anesthetics selectively and robustly activate α5GABAARs, which may override glutamatergic activation and LTP because of dramatic increases in a α5GABAAR- associated shunting conductance. (C) L-655,708 inhibits α5GABAAR activity and thereby reverses etomidate blockade of LTP.
154 155
155 156
Chapter 6. Mechanism of α5 subunit-containing γ-aminobutyric acid type A receptor-mediated inhibition of learning and synaptic plasticity
6.1 Introduction
In the hippocampus, the encoding of episodic memory depends on a subtle interplay between excitatory and inhibitory neurotransmission. The neural correlate for learning and memory is thought to be the long-term alterations in the efficacy of excitatory postsynaptic neurotransmission (Malenka and Bear, 2004; Whitlock et al., 2006). At CA3–CA1 synapses, a brief period of high-frequency stimulation induces LTP of excitatory synaptic potentials, whereas low-frequency stimulation causes LTD (Collingridge et al., 1983; Dudek and Bear,
1992). In this regard, the ability of GABAAR activity to reduce synaptic plasticity in CA1 neurons has been attributed to the generation of synaptic or transient inhibitory potentials (Lu et al., 2000). The role of a recently described tonic inhibitory conductance (Bai et al., 2001;
Caraiscos et al., 2004b), generated by extrasynaptic GABAARs, in regulating memory processes and synaptic plasticity remains uncertain. Of particular interest is whether the tonic inhibition modifies the acquisition or consolidation of learning and the specific conditions that determine its impact, such as the type or intensity of the training.
Different populations of GABAARs, derived from a superfamily of 19 genes, contribute to the generation of phasic and tonic inhibitory conductances. The different subunit (α1–6, β1–3, γ1–3,
δ, ε, π, θ, ρ1–3) combinations contribute to the distinctive expression patterns and the pharmacological and biophysical characteristics of each GABAAR. For instance, a relatively high proportion of α5GABAARs populates the extrasynaptic regions of CA1 and CA3 pyramidal
156 157
neurons in the hippocampus (Sur et al., 1999; Pirker et al., 2000). The α5GABAARs generate a tonic inhibitory conductance and have higher affinity for GABA and slower desensitization than synaptic GABAARs (Burgard et al., 1996; Caraiscos et al., 2004b). When considered in light of their relatively restricted pattern of distribution and specialized function, α5GABAARs may be expected to play a critical role in modulating mnemonic processes within the hippocampus.
Correspondingly, behavioural testing in WT and Gabra5–/– mutant mice has revealed several differences in hippocampus-dependent memory tasks (Collinson et al., 2002; Cheng et al., 2006).
However, there was no difference in the LTP of excitatory synaptic transmission in hippocampal slices obtained from WT and Gabra5–/– mice following high-frequency stimulation (Collinson et al., 2002; Cheng et al., 2006). Another mutant mouse model supports these findings. A partial deficit of α5GABAARs in α5H105R point mutant mice is associated with improved trace fear- conditioning, resistance to the extinction of learning (Yee et al., 2004) and altered latent inhibition (Gerdjikov et al., 2008); however, synaptic plasticity is similar in hippocampus slices from α5H105R and WT mice (Crestani et al., 2002). Nevertheless, the memory-enhancing benzodiazepine inverse agonist L-655,708, which selectively reduces α5GABAAR function, enhances LTP (Dawson et al., 2006) and promotes spontaneous gamma oscillations in the CA3 region of the hippocampus (Glykys et al., 2008). Also, the inverse agonist a5IA improves alcohol-induced memory impairment in human volunteers (Nutt et al., 2007). Despite intense study, it remains uncertain whether or how α5GABAARs modify plasticity and hippocampus- dependent memory under physiological conditions.
Accordingly, I used stringent behavioural testing and a broad range of plasticity-inducing stimulation protocols to explore the contribution of α5GABAARs to the encoding of memory and modification of synaptic plasticity. Here, I show that inhibiting α5GABAAR activity improved
157 158 the acquisition of weakly acquired but not strongly formed fear associations. Additionally, the modification threshold for LTP was shifted to the left when α5GABAARs were genetically deleted or pharmacologically inhibited. Surprisingly, this effect only occurred within a narrow window of stimulation frequencies. Within the theta frequency, an increase in the α5GABAAR- dependent membrane conductance inhibited synaptic plasticity by shunting the EPSPs, and this attenuation of plasticity remained independent of synaptic GABAAR inhibition. Taken together, our results suggest that α5GABAARs selectively limit the ability of neurons to participate in memory formation and LTP when activated under specific conditions.
Questions:
1. Does the involvement of α5GABAARs in learning and memory depend on the type of
learning or behavioural training?
2. Do α5GABAARs actively modify LTP or LTD?
3. Is network excitability altered in Gabra5–/– mice?
4. Are there alterations in glutamate receptor signaling or expression in Gabra5–/– mice?
5. What are the relative contributions of synaptic and extrasynaptic inhibition to LTP?
6.2 Specific methods
6.2.1 Fear conditioning
In the Pavlovian fear-conditioning tasks, mice were exposed to a tone and a foot shock in a novel conditioning context, with either no time delay (0 s for cued conditioning) or with an interval of
20 s between stimuli (trace conditioning) (Misane et al., 2005; Wiltgen et al., 2005). Mice were randomly assigned to receive an i.p injection (2 ml/kg) of either vehicle (10% DMSO) or L-
158 159
655,708 (0.68 mg/kg). In different experiments, mice were injected with either vehicle or L-
655,708, 30 min before or immediately following training in the fear conditioning chamber. The dose of L-655,708 was selected to modify learning behaviours via preferential binding of
α5GABAARs, as previously determined (Atack et al., 2006b; Atack et al., 2006c). On day 1, single animals were allowed to explore the chamber for 180 s. An 800-Hz tone created by a frequency generator, was presented for 20 s at 70 dB. For cued fear conditioning, the last 2 s of each auditory tone was paired with an electric foot shock (2 s, 1 mA); for trace fear conditioning, the auditory stimulus and foot shock (2 s, 0.5 mA) were separated by 20 s. Each of these sequences was presented 3 times, separated by 60 s (for cued fear conditioning) or 240 s (for trace fear conditioning). On day 2, 24 h after the conditioning session, each mouse was assessed for a freezing response by placing it in the original context and scoring every 8 s for a total of 8 min to determine contextual fear. On day 3, the conditioning chamber was modified to measure the freezing response to the tone to study either cued or trace fear conditioning. Mice were monitored for 180 s for freezing to the modified context, to rule out contextual influences. After the monitoring period, the auditory tone was presented continuously for 300 s, and the freezing response was measured every 8 s.
6.2.2 Extracellular recordings
6.2.2.1 Long-term potentiation recordings
To examine the full extent of α5GABAARs in synaptic plasticity, the induction threshold for
LTP was investigated by varying the degree of presynaptic stimulation, which was achieved by altering the frequency of tetanic stimulation. Slices were stimulated with frequencies of 1, 5, 10,
20, 50 or 100 Hz. The number of stimulation pulses remained constant (600 pulses/stimulation frequency), with the exception of the protocol for 100 Hz, in which I stimulated each slice 3
159 160 times for a period of 1 s, separated by 20 s intervals. For recordings that used pharmacological agents, recordings were monitored for 10 min in a drug free solution (aCSF only) to ensure good slice health. Drugs, either BIC (8 µM), L-655,708 (10 nM), SR-95531 (5 µM) were applied following this period and the recording was monitored for an additional 10 min. The stimulation protocol was applied following this 20 min baseline period. Recordings were then monitored for an additional 60 min post-stimulation.
6.2.2.2 Paired-pulse facilitation
It has been suggested that the expression of LTP may include a presynaptic as well as a postsynaptic locus. I wanted to determine whether the absence of α5GABAARs might alter the expression of a presynaptically mediated form of synaptic potentiation, which would influence our results and change our interpretation. I chose to investigate paired pulse facilitation (PPF), which involves eliciting two excitatory postsynaptic potentials in succession, with the second being greater than the first. PPF is understood to be a model of presynaptic plasticity, and I wanted to determine whether some of the α5GABAAR-mediated effects were mediated presynaptically. A single pulse was delivered to the Schafer collaterals and then following an ISI of 50, 100, 150, 200, or 300 ms a second pulse was delivered; recordings were obtained from the stratum radiatum during the entire procedure.
6.2.2.3 Extrinsic excitability and coastline analysis
To determine whether deletion of the α5GABAARs increased the extrinsic excitability within the hippocampus, I used a quantification method for determining overall epileptiform activity. I recorded population spikes in the CA1 pyramidal layer before and after application of BIC (8
µM) to hippocampal slices, and analyzed the traces using the Coastline Bursting Index. The
160 161
Coastline Bursting Index (Korn et al., 1987), calculates the total linear distance of the discharged waveform of evoked population spikes and is a measure of the amount of bursting and overall excitability within the slice. Minianalysis software (version 6.0.3 from Synaptosoft) was used to analyze the coastline of the population spikes, from 20 ms after the stimulus artifact until the trace resumed returned to a stable level at 0 mV.
6.2.3 AMPA / NMDA Ratios
The ratio of the NMDA to AMPA EPSCs were determined in WT and Gabra5–/– slices to verify normal glutamatergic transmission, which could alter LTP responses. AMPA and NMDA currents were isolated by APV (10 µM) and CNQX (20 µM) respectively, in the continued presence of bicuculline (10 µM). Evoked excitatory postsynaptic currents (eEPSCs) were recorded from WT and Gabra5–/– neurons at a holding potential of –80 mV (AMPA component) and +40 mV (NMDA component). The stimulation intensity used to evoke AMPA and NMDA currents was varied and ranged from 5 to 10 V with 6 sweeps recorded at each intensity. Recording electrodes were filled with an internal solution containing (in mM); CsGluc
(132.5), 17.5 KCl (17.5), EGTA (0.2), HEPES (10), Mg-ATP (2), GTP (0.3), QX-314 (5) (pH
7.2-7.3 with CsOH, Osm 290). Only recordings with stable holding current and series resistance maintained below 20 MΩ were considered for analysis.
6.2.4 Spontaneous inhibitory recordings
Patch pipettes had open tip resistances of 3-5 MΩ when filled with an intracellular solution that contained (in mM): CsCl (140), HEPES (10), EGTA (10), MgATP (2), CaCl2 (1), QX-314 (5)
(pH 7.3 with CsOH, 295-305 milliosmolar). Currents were sampled at 10 kHz and filtered at 2 kHz by using an eight-pole low pass Bessel filter. All recordings were conducted in gap free
161 162 mode and had a stable baseline current (< 20% change) before the application of drugs. The extracellular recording solution contained CNQX (20 µM) and APV (10 µM) to block ionotropic glutamate receptors and TTX (0.3 µM) to block voltage-dependent Na+ channels. The amplitude of the tonic current under control conditions was measured as the difference in the holding current before and during the application of bicuculline (10 µM) or SR-95531 (5 µM).
6.2.5 Synaptic block of GABAAR-mediated IPSPs with SR-95531
In a separate experiment, to ensure that the GABAAR-mediated IPSPs were completely blocked by SR-95531 at the concentration used for field recordings (5 µM) and remained blocked following 10 Hz stimulation the GABAAR-mediated IPSPs were recorded in CA1 pyramidal neurons. CNQX, APV, and CGP were used to block AMPA, NMDA and GABAB responses, respectively. The GABAAR-mediated IPSPs were recorded to obtain a stable baseline response and then SR-95531 was added to block the GABAAR-mediated IPSPs. Slices were stimulated with 10 Hz to determine whether the GABAAR-mediated IPSPs remained blocked by SR-95531 following 10 Hz stimulation. Patch pipettes had open tip resistances of 3-5 MΩ when filled with an intracellular solution that contained (in mM): K-Gluc (132.5), KMeSO4 (10.3), KCl (7.2),
HEPES (10), EGTA (0.2), Mg-ATP (2), GTP (0.3) with the pH adjusted to 7.3 with KOH.
6.2.6 Current clamp recordings and input resistance measurements
A small hyerpolarizing current pulse (intensity 10 pA, duration 400 ms) was injected 500 ms before evoking EPSPs from CA1 pyramidal neurons. The hyperpolarizing pulse was injected throughout the recording to monitor the input resistance of the cells. EPSPs were evoked every
60 s to allow for recovery between sweeps. Following a stable baseline period (both current pulse and EPSP amplitude), neurons were stimulated with the 10 Hz protocol that was used for
162 163 the field recordings. EPSP amplitude, input resistance, and membrane potential were measured pre- and post-10 Hz stimulation. In order to measure the input resistance of the cell, the steady state voltage was taken and divided by the amount of injected current (10 pA). For the same data set the depolarizing envelope and membrane potential were analyzed during 10 Hz stimulation.
During 10 Hz stimulation the membrane potential was sampled 15 ms before the stimulation artifact every sec (60 total samples). The depolarizing envelope was taken as the total integrated area under the EPSPs for the entire 60 sec stimulation period. SR-95531 (5 µM), CGP (5 µM) and APV (10 µM) were added to the bath solution to inhibit synaptic GABAA, GABAB and
NMDA receptors, respectively. Patch pipettes had open tip resistances of 3-5 MΩ when filled with an intracellular solution that contained (in mM) K-Gluc (132.5), KMeSO4 (10.3), KCl
(7.2), HEPES (10), EGTA (0.2), Mg-ATP (2), GTP (0.3) with the pH adjusted to 7.3 with KOH.
6.2.7 Preparation of protein samples and western blot analysis
Whole hippocampi (n = 6 per group) from 3-month-old male mice (both WT and Gabra5–/– genotypes) were isolated rapidly and placed in ice-cold phosphate-buffered saline containing complete Mini protease inhibitors (Roche Diagnostics, Mannheim, Germany). Each group of hippocampi was then homogenized separately in 0.75 mL of radioimmunoprecipitation assay buffer (150 mmol/L NaCl, 50 mmol/L Tris pH 8.0, 5 mmol/L EDTA, 1% v/v Nonidet P-40,
0.5% w/v sodium deoxycholate and 0.1% w/v SDS) containing 5 µg/mL aprotinin, 20 µg/mL leupeptin and 1x Mini protease inhibitor cocktails (all protease inhibitors from Roche
Diagnostics, Mannheim, Germany) using 1 mL Dounce glass–glass homogenizers. The homogenates were then centrifuged at 1,000g for 5 minutes at 4°C to pellet any debris and nuclei. The supernatant was used for determining the protein concentration (BCA Protein Assay;
Pierce Biotechnology, Rockford, IL). Tissue homogenates were then resolved on discontinuous
163 164
10% sodium dodecyl sulfate polyacrylamide gel electrophoresis before transfer onto nitrocellulose membranes. Approximately 25–30 µg of protein was loaded per lane. The membranes were then washed in a mixture of Tris-buffered saline and 0.1% Tween 20 (TBS–T) and blocked for 1 h with 5% non-fat skim milk powder in TBS–T (BLOTTO) before overnight incubation at 4°C with the appropriate antibody (mouse anti-NR1 1:1,000, BD Biosciences,
Mississauga, Ontario, Canada ; rabbit anti-ΝR2Α 1:1,000, Covance, Berkeley, CA; rabbit anti-
NR2B, Santa Cruz Biotechnologies, Santa Cruz, CA; rabbit anti-GluR1 1:1,000,
Millipore/Cedarlane Laboratories, Hornby, Ontario, Canada; rabbit anti-α5 1:1,000,
Millipore/Cedarlane Laboratories, Hornby, Ontario, Canada). Membranes were then washed in
TBS–T and incubated for 1 h in HRP-conjugated anti-mouse (Santa Cruz Biotechnologies, Santa
Cruz, CA) or anti-rabbit (Invitrogen, Mississauga, Ontario, Canada) secondary antibodies
(diluted 1:5,000 in BLOTTO) at room temperature. Membranes were taken through a final series of three to five washes in TBS–T before being incubated in substrate. Size-specific bands were visualized by applying enzyme chemiluminescent substrates (SuperSignal West Pico ECL substrate; Pierce Biotechnology, Rockford State, IL) and exposing the labelled membranes on a
Kodak ImageStation 2000R for 5 to 18 minutes. Specifically labelled bands at the appropriate molecular weights were analyzed for signal intensity using Kodak Image Station 2000R software and either the automated band label function or the edge detection function to determine net band intensity. Specific band intensities from Gabra5–/– samples were normalized against control wild-type intensities. In some experiments, blots were stripped and probed with anti-β-actin antibody (1:5,000; Sigma, St. Louis, MO) to confirm equal loading of samples.
164 165
6.3 Results
6.3.1 Learning and memory co-vary with α5GABAAR activity in
hippocampus-dependent tasks
Contextual and cued fear conditioning are robust and long-lasting forms of associative learning that are dependent on hippocampus-dependent and independent processing, respectively
(Sanders et al., 2003; Fanselow and Poulos, 2005; Hamm and Weike, 2005). In these conditioning tests, mice learn to associate a distinct context or auditory tone (the CS) with an aversive stimulus such as a foot shock (the UCS). When the mice are placed back in the same training context, they exhibit a range of conditioned fear responses, including freezing. I postulated that decreasing α5GABAAR activity would enhance associative learning and memory during fear learning, similar to its effect on spatial navigational memory (Collinson et al., 2002).
I tested WT and Gabra5–/– mice in a robust contextual fear conditioning task (three 0.5 mA foot shocks of 2 s each, separated by 60 s intervals) and an equally stringent auditory cued fear conditioning task (three tone–shock pairings: 20 s, 70 dB tone paired with a 2 s, 0.5 mA foot shock, separated by 60 s intervals), with the shock delivered during the last 2 s of the tone
(Figure 6.1A). As shown previously (Cheng et al., 2006), contextual fear conditioning responses were similar in WT and Gabra5−/− mice, as indicated by high freezing scores in both groups (>
75% freezing; unpaired t test, P = 0.51). Intraperitoneal administration of L-655,708, a drug that preferentially reduces α5GABAAR activity and thus the tonic inhibitory conductance, at a dose of 0.68 mg/kg before the training, had no effect on memory performance for contextual fear conditioning (two-way ANOVA, P = 0.31; Figure 6.1B). The lack of effect of L-655,708 suggests a ceiling effect associated with the high freezing scores. I am confident that the ineffectiveness of L-655,708 is not due to a lack of functionality of the α5GABAARs in the brain
165 166 region or hippocampus sub-regions involved in contextual fear acquisition, because we have previously shown that increasing the activity of α5GABAARs pharmacologically can impair memory performance for contextual fear memory (Cheng et al., 2006). In addition, the performance of WT and Gabra5–/– mice was similar in the amygdala-dependent cued fear conditioning task (unpaired t test, P = 0.52; Figure 6.1C). This result was anticipated, as the expression of α5GABAARs in the amygdala is relatively low (Sur et al., 1999; Pirker et al.,
2000). It is possible that the role of α5GABAARs in regulating associative learning increases as the cognitive demands of the task increase. To investigate this hypothesis, I used a trace fear conditioning paradigm because rodents do not rapidly or efficiently learn this task (Misane et al.,
2005). The strength of classical conditioning can be reduced by introducing a time interval (or
“trace”) between the conditioned stimulus and the unconditioned stimulus. The procedure for trace fear conditioning was similar to that for cued fear conditioning (three tone–shock pairings:
20 s, 70 dB tone paired with a 2 s, 0.5 mA foot shock, separated by intervals of 240 s), except that an empty trace interval of 20 s was interposed between the tone and the foot shock (Figure
6.1A). Trace fear conditioning depends strongly on the dorsal hippocampus, and α5GABAAR activity is highly correlated with the performance of mice in hippocampus-dependent tasks
(Collinson et al., 2002; Collinson et al., 2006). For trace fear conditioning, Gabra5−/− mice significantly outperformed their WT littermates, as indicated by significantly higher freezing scores (two-way ANOVA, P = 0.002; Figure 6.1D). To confirm that the difference between the genotypes was attributable to a reduction in α5GABAAR activity, I inhibited the receptors pharmacologically by injecting the WT mice with L-655,708. L-655,708 dramatically improved the performance of WT mice in trace fear conditioning but had no effect on the performance of
Gabra5−/− mice (two-way ANOVA, P = 0.005; Figure 6.1D), which is consistent with the selectivity of the drug for α5GABAARs. Furthermore, L-655,708 only modified trace fear
166 167 conditioning in WT mice when it was injected before training. The administration of L-655,708 immediately after training did not enhance trace fear memory in WT mice (two-way ANOVA, P
= 0.14; Figure 6.1E), which suggests that a reduction in α5GABAAR activity does not influence the maintenance and consolidation phase of learning. Taken together, these data indicate that
α5GABAAR activity regulates the acquisition but not the consolidation of weakly acquired associative fear memory, but does not modulate strongly acquired associative fear memory.
Based on these findings, I sought to identify a correlate between associative memory tasks and synaptic plasticity.
6.3.2 α5GABAARs regulate synaptic plasticity within a narrow range of
stimulus frequencies
To identify the network substrates that underlie α5GABAAR-dependent modulation of associative learning, synaptic plasticity was studied in hippocampal slices from WT and
Gabra5–/– mice. First, I examined the plasticity of fEPSPs in response to a wide range of input stimulation frequencies (1 Hz to 100 Hz) delivered at the Schaffer commisural projections to the
CA1 pyramidal cells. To ensure that the recordings were stable, baseline responses were evoked at 0.05 Hz and were monitored in each slice for 20 min. When slices were stimulated at low frequencies (1 Hz and 5 Hz), the LTD of excitatory synaptic transmission was observed in both genotypes (Figure 6.2A). Following high-frequency stimulation (50 Hz and 100 Hz), LTP occurred in WT and Gabra5–/– slices, and the increase in amplitude of the fEPSPs was similar between the two groups (Figure 6.2B). Notably, following stimulation at submaximal frequency
(10 Hz), the plasticity in WT and Gabra5–/– was opposite in polarity: LTD was observed in WT slices, whereas LTP was observed in Gabra5–/– slices (unpaired t-test, P = 0.01; Figure 6.2C).
Accordingly, the frequency–response plot was shifted to the left for recordings from Gabra5–/–
167 168 slices, with the greatest difference observed at 10 Hz and 20 Hz (Figure 6.2D). It is noteworthy that the frequencies of stimulation where α5GABAAR activity had the greatest influence are commonly associated with large-scale theta-frequency synchronization, which is coupled with some forms of hippocampus-dependent learning (Buzsaki, 2005).
168 169
Figure 6.1. α5GABAARs physiologically regulate the acquisition of weak hippocampus- dependent associative fear memory tasks.
(A) A schematic representation showing the timing for all three fear conditioning protocols. In all protocols a baseline activity period of 3 min preceded the conditioning procedure. Three 2 s,
0.5 mA foot shocks, separated by 60 s intervals, were used for contextual fear conditioning.
Three tone–shock pairings (20 s, 70 dB tone paired with a 2 s, 0.5 mA foot shock), separated by
60 s intervals, were used for auditory cued fear conditioning. The procedure for trace fear conditioning was similar to that for cued fear conditioning (three tone–shock pairings, separated by 240 s intervals), except that an empty trace interval of 20 s was interposed between the tone and the foot shock. (B) There was no difference between WT (n = 9) and Gabra5–/– (n = 11) mice for contextual fear conditioning, which forms strong hippocampus-dependent memories; furthermore, L-655,708 had no effect on the freezing response (WT, n = 10; Gabra5–/–, n = 11).
(C) The WT (n = 9) and Gabra5–/– (n = 11) mice had similar scores during the amygdala- dependent cued fear conditioning task. (D) The performance of Gabra5–/– (n = 12) mice was enhanced in trace fear conditioning (a weak associative task), relative to the effect in naive (n =
12) and vehicle-treated (n = 13) WT mice; in addition, inhibiting α5GABAARs with L-655,708 improved the performance of WT mice (n = 13) to the level observed in Gabra5–/– mice. (E)
The performance of WT mice (n = 9) was not enhanced with L-655,708 injections immediately after training in the trace fear conditioning protocol, relative to Gabra5–/– mice (n = 8). * denotes a statistically significantly difference from the control group at P < 0.05.
169 170
170 171
Figure 6.2. α5GABAARs critically regulate the threshold for LTP within narrow range of stimulus frequencies.
(A) Persistent stimulation of the Schaffer collateral pathway induced a robust LTD of fEPSPs that was similar between the two genotypes (n = 8 slices per group). (B) Stimulation of slices with a 100 Hz tetanus induced a strong LTP response that was the same between the genotypes
(n = 8 slices per group). (C) Genetic deletion of the α5 subunit of the GABAAR lowered the threshold for LTP when stimulated at 10 Hz stimulation (n = 8 slices per group). Sample traces are shown above each figure for the times indicated by the numbers. Calibration: 0.5 mV, 10 ms.
(D) Summary of the average change from baseline for fEPSPs when stimulated with various frequencies. Persistent changes in synaptic strength were significant between the genotypes only for stimulation at 10 Hz and 20 Hz. * denotes significance at P < 0.001.
171 172
172 173
6.3.3 Synaptic transmission, excitability and glutamate receptor
expression are normal in Gabra5–/– mice
Genetic deletion of α5GABAARs could be associated with unrecognized compensatory changes, such as alterations in the expression or function of other GABAARs, ion channels or associated proteins in Gabra5–/– mice (Brickley et al., 2001). I therefore sought to determine whether there were nonspecific baseline differences in excitatory synaptic transmission between WT and
Gabra5–/– slices that might account for the observed differences in plasticity. I plotted the input–output relationship of the stimulus intensity, as the presynaptic fiber volley versus the slope of the fEPSP, and detected no difference between the genotypes (Figures 6.3A and 6.3B). I also found that PPF, which is considered to represent a form of presynaptic short-term plasticity, was similar between the genotypes (Figures 6.3C and 6.3D). Interestingly, my PPF data differ from those of Collinson et al., who reported a very modest increase in PPF in Gabra5–/– slices stimulated with an interpulse interval of 100 ms to 300 ms. Those authors attributed the increase in PPF to an increase in the amplitude of the postsynaptic potential, as enhancement of fEPSP slopes over the same paired-pulse range was not affected (Collinson et al., 2002). Notably, we and others have not detected changes in the amplitude or time course of inhibitory postsynaptic potentials (IPSPs) in CA1 pyramidal neurons from Gabra5–/– mice (Caraiscos et al., 2004b;
Cheng et al., 2006; Glykys and Mody, 2006).
I also examined the possibility that deletion of the Gabra5 gene would increase the extrinsic excitability of the slices, thus enhancing LTP. To test this possibility, I recorded population spikes from the CA1 stratum pyramidale and applied the GABAAR antagonist bicuculline (8
µM), to increase the excitability of the slices. I quantified the degree of excitation by measuring the coastline (linear distance) of the population spike before and after application of bicuculline.
173 174
As shown in Figure 6.3E, the application of bicuculline was associated with an overall increase in bursting activity. No differences were observed between the genotypes either before or after the application of bicuculline, which suggests that nonspecific enhancement of network excitability did not occur in Gabra5–/– slices. Therefore, this type of change could not account for the enhancement of LTP observed at 10 Hz (Figures 6.3E and 6.3F).
Finally, since the threshold for activity-dependent plasticity is modified by altering the effectiveness of NMDA receptors (Herron et al., 1985), I evoked glutamatergic EPSCs in WT and Gabra5–/– slices by stimulating the Schaffer-commissural pathway at 0.05 Hz. This was done to ensure that there was not an upregulation of glutamatergic currents in the Gabra5–/– slices, which may account for the LTP effect. Selective pharmacological blockers and different holding potentials were used to determine the ratio of NMDA to AMPA components of the
EPSCs. When slices were stimulated at intensities ranging from 5 to 10 V, the values were similar for WT and Gabra5–/– slices, as were the raw amplitudes of the currents (Figures 6.4A,
6.4B, 6.4C and 6.4D). Furthermore, my colleague Dr. William Ju completed a Western immunoblot assay for ionotropic glutamate receptors, which indicated no differences between genotypes in the expression of NMDA and AMPA receptor subunits (Figure 6.4E). These results indicate that up-regulation of excitatory transmission cannot account for the enhanced LTP observed in Gabra5–/– slices. Finally, to determine whether α5GABAAR modification of LTP depended on conventional NMDA receptor-dependent pathways, the competitive NMDA receptor antagonist APV (10 µM) was applied to WT and Gabra5–/– slices, which were stimulated at 10 Hz (Figure 6.5). APV completely blocked plasticity in slices from both genotypes, which indicates that α5GABAARs largely modify NMDA-dependent plasticity.
174 175
Figure 6.3. Normal synaptic transmission and excitability in Gabra5–/– slices.
α5GABAARs do not modify synaptic transmission or excitability of hippocampal slices. (A)
Representative traces demonstrating no difference between wild-type (WT) and Gabra5–/– slices after stimulation at different intensities. (B) Individual responses showed that the input–output relationships were similar for WT and Gabra5–/– slices, which suggests normal baseline presynaptic function and synaptic efficacy in the Gabra5–/– mouse model. (C) Example traces show normal paired-pulse facilitation for each genotype. (D) Grouped data show that facilitation of paired synaptic pulses (PSPs) with different inter-stimulus intervals (ISIs) was the same for
WT and Gabra5–/– slices. (E) Traces showing enhanced excitability of population spikes after bicuculline application. (F) Summarized data show changes in the coastline with application of bicuculline. There was no difference between WT and Gabra5–/– slices.
175 176
176 177
Figure 6.4. Glutamate receptor currents and expression were not altered in Gabra5–/– neurons.
(A) AMPA (–80 mV) and NMDA (+40 mV) currents were recorded in WT and Gabra5–/– neurons. (B) There was no difference in the peak NMDA to AMPA ratio in WT (n = 7) and
Gabra5–/– (n = 9) neurons. (C) AMPA currents evoked at intensities ranging from 5 to 10 V were the same in WT and Gabra5–/– neurons. (D) There was no difference in the peak amplitude of NMDA currents evoked at intensities ranging from 5 to 10 V in WT and Gabra5–/– neurons. (E) Left panel: Immunoblot assays for ionotropic glutamate receptor subtypes, including the NMDA receptor subunits NR1, NR2A and NR2B and the AMPA receptor subunit
GluR1, showed no difference in expression levels between WT and Gabra5–/– mice. Right panel: The change in the level of receptor expression was determined from the ratio of total protein expressed in Gabra5–/– tissue to total protein expressed in WT tissue. In the Gabra5–/– tissue, the total amount of protein expression did not change for glutamatergic receptor subtypes, and the ratio of total α5 was 0, as Gabra5–/– tissue does not possess the α5 protein.
177 178
178 179
Figure 6.5. LTP and LTD observed with 10 Hz stimulation is dependent on NMDA signalling pathways.
Application of 50 µM APV inhibited LTP in Gabra5–/– and inhibited LTD in WT slices when the Schaffer collateral pathway was tetanized with 10 Hz stimulation for 1 min. Black trace = normal aCSF, green trace = APV baseline, blue or red trace = post-tetanus. Calibration bar: 0.5 mV, 5 ms. fEPSP = field excitatory postsynaptic potentials.
179 180
180 181
6.3.4 Blockade of tonic but not phasic inhibitory neurotransmission
enhances submaximal LTP
Next, to determine whether a reduction in α5GABAAR activity accounted for the shift to the left of the activity-dependent plasticity curve, I studied the effects of L-655,708 in WT slices simulated at 10 Hz. L-655,708 (10 nM) altered the direction of plasticity in response to 10 Hz stimulation in WT slices such that LTD was no longer observed (two-way ANOVA, P = 0.004;
Figure 6.6A). The strength of plasticity in L-655,708-treated WT slices was similar to that observed in vehicle-treated (i.e., control) Gabra5–/– slices (Figure 6.2C). When L-655,708 was applied to Gabra5–/– slices stimulated at 10 Hz, neither the strength nor the polarity of the LTP changed (Bonferroni pair-wise comparison, P = 0.31; Figure 6.6A). I next studied whether the ability of L-655,708 to increase LTP in WT slices depended on when the drug was applied.
When L-655,708 was applied immediately after 10 Hz stimulation, no increase in synaptic strength was observed; indeed, LTD occurred, rather than LTP (Figure 6.6B). Thus, α5GABAAR activity primarily regulates the threshold for the induction rather than the maintenance or expression phases of LTP. This effect correlates with the ability of α5GABAAR activity to regulate the acquisition but not the consolidation of associative trace fear learning.
To further confirm that α5GABAARs regulate synaptic plasticity independent of other populations of GABAARs, the nonselective competitive GABAAR antagonist bicuculline (8 µM) was applied to WT slices that were stimulated at 10 Hz. Bicuculline inhibits all GABAAR subtypes, with the exception of the ρ subunit (Shimada et al., 1992), which is expressed at very low levels in the hippocampus (Enz et al., 1995; Alakuijala et al., 2005). To further confirm that
α5GABAARs predominantly regulate synaptic plasticity under certain stimulation frequencies via mechanisms independent of other populations of GABAARs, the non-selective competitive
181 182
GABAAR antagonist bicuculline (8 µM) was applied to WT slices that were stimulated at 10 Hz.
In bicuculline-treated WT slices, LTP, rather than LTD, was induced. The strength of potentiation was similar to that observed in WT slices treated with L-655,708 and in Gabra5–/– control slices (unpaired t-test, P = 0.09). Interestingly, when bicuculline was applied to Gabra5–
/– slices, the amount of LTP was not further enhanced compared with control Gabra5–/– slices
(unpaired t-test, P = 0.36; Figure 6.6C). This result indicates that α5GABAAR activity, exclusive of other GABAAR subtypes, regulates the threshold for plasticity when slices are stimulated at
10 Hz. In contrast to these findings, when slices were stimulated at high frequency (100 Hz), bicuculline enhanced plasticity in WT and Gabra5–/– slices (two-way ANOVA, P = 0.001;
Figure 6.6D) suggesting that GABAARs other than α5GABAARs regulate plasticity under conditions of high frequency stimulation.
In order to determine whether the effect of α5GABAARs on plasticity was mediated in part by synaptic inhibition I recorded mIPSCs in WT and Gabra5–/– neurons in the presence and absence of L-655,708. Under the current experimental conditions, the deletion of the α5 subunit or L-655,708 application did not affect the amplitude, rise time, decay time, or frequency of mIPSCs in CA1 pyramidal neurons. Thus, it appeared that in my preparation, α5GABAARs contributed little to phasic inhibition. The summary of the mIPSC kinetics is shown in Table 6.1.
Additionally, to further explore whether tonic GABAergic inhibition regulates plasticity, through mechanisms that are independent of synaptic inhibition, slices were treated with a selective
GABAAR antagonist for synaptic currents, SR-95531 (5 µM). Previously our laboratory showed that in CA1 pyramidal neurons SR-95531 blocks spontaneous IPSCs but not the tonic current
(Bai et al., 2001), a finding that I have replicated in this thesis (SR-95531-induced change in holding current, 1.35 ± 1.21 pA, Figure 6.7). In contrast, the concentration of L-655,708 that
182 183 blocked LTD in WT neurons induced an outward shift in the holding current, indicating a block of the tonic conductance (L-655,708-induced change in holding current, 15.63 ± 1.78 pA, Figure
6.7). Additionally, L-655,708 did not shift the holding current in Gabra5–/– neurons (L-655,708- induced change in holding current, -1.32 ± 3.76 pA, n = 5). For comparison, bicuculline, which converted the WT 10 Hz-induced LTD into LTP, was also applied and caused a large outward shift in CA1 pyramidal neurons (bicuculline-induced change in holding current, 29.35 ± 3.17 pA,
Figure 6.7). I found that selective inhibition of the IPSCs by SR-95531 failed to convert the LTD in WT slices into LTP as observed in Gabra5–/– slices when the Schaffer collaterals were stimulated with 10 Hz (Figure 6.8A). The effect of SR-95531 on 10 Hz-induced plasticity was significantly different from that of L-655,708 and bicuculline (Figures 6.6A and 6.6C, respectively). SR-95531 also failed to enhance LTP in Gabra5–/– slices (Figure 6.8A). The relative changes in plasticity in response to the various GABAA blockers, at 10 Hz stimulation, are shown in Figure 6.8B. Next, to ensure that phasic IPSPs were abolished by SR-95531 before and after 10 Hz stimulation, whole-cell current clamp recordings were obtained. The GABAAR- mediated IPSPs were studied in the absence and presence of SR-95531, with pharmacological blockers for AMPA, NMDA and GABAB receptors in the extracellular solution. SR-95531 completely blocked the IPSPs following 10 Hz stimulation (Figures 6.8C and 6.8D), which shows that that the IPSPs fail to regulate synaptic plasticity at 10 Hz.
183 184
Figure 6.6. Pharmacologic studies confirm that α5GABAARs are critical for the induction of
LTP following submaximal but not high-frequency stimulation.
Recording of fEPSPs in control conditions was followed by the application of L-655,708 for 10 min before (n = 8 per group) (A) or immediately after (n = 11 per group) (B) 10 Hz stimulation.
Application of L-655,708 before but not after 10 Hz threshold stimulation enhanced the fEPSPs of WT slices, which suggests that α5GABAARs are critical for the induction of threshold LTP but not for maintenance of the response. (C) I next measured overall involvement of GABAARs in the LTP of 10 Hz stimulation by blocking these receptors with the competitive antagonist bicuculline. Application of bicuculline to WT slices (n = 10) that had been stimulated at 10 Hz produced LTP that was indistinguishable from that observed in Gabra5–/– slices but did not further enhance the LTP in Gabra5–/–slices (n = 9). (D) Bicuculline further potentiated LTP in
WT (n = 7) and Gabra5–/– slices (n = 7) relative to drug-free conditions with 100 Hz stimulation, which suggests that GABAARs not containing the α5 subunit play an active role in
LTP when the intensity of activation is increased. Black traces = pre-tetanus baseline; blue and red traces = 60 min after tetanus in WT and Gabra5–/–slices, respectively. Calibration: 0.5 mV,
10 ms.
184 185
185 186
Table 6.1. mIPSC kinetics of WT and Gabra5–/– CA1 pyramidal neurons in the presence or absence of L-655,708.
186 187
Values are means ± sem. Parentheses enclose n values. IPSC, inhibitory postsynaptic current, aCSF, artificial cerebrospinal fluid..
187 188
Figure 6.7. The effect of SR-95531, L-655,708 and bicuculline on the tonic conductance in CA1 pyramidal neurons.
(A) SR-95531 (5 µM) blocks the transient mIPSCs but causes no outward shift in the holding current, which indicates that it does not block the extrasynaptic receptors. L-655,708 (10 nM) does not block mIPSCs but causes an outward shift in the holding current. Bicuculline (8 µM) blocks both the mIPSCs and tonic current, as indicated by an outward shift in the baseline holding current. (B) Summary bar chart indicating the relative change in holding current following application of SR-95531 (n = 7), L-655,708 (n = 5), or bicuculline (n = 8) to CA1 pyramidal neurons neurons. * denotes significance at P < .05.
188 189
189 190
Figure 6.8. Blockade of synaptic GABAA receptors does not enhance plasticity with 10 Hz stimulation.
(A) Selectively blocking synaptic GABAA receptors with SR-95531 (5 µM) does not enhance synaptic potentiation in WT (n = 8) or Gabra5–/– (n = 8) slices. (B) A summary of the change in synaptic responses after 10 Hz stimulation. The data represent the average of the last 5 min of the recording for each slice in each group. (C) The IPSP (black) was blocked with SR-95531 before
(light blue) and after (red) 10 Hz stimulation, which indicates that synaptic GABAA receptors could not account for the plasticity differences at this frequency. (D) Average data showing the raw amplitudes of the IPSP at baseline (without SR-95531), with SR-95531 and following 10 Hz stimulation for WT (n = 6) and Gabra5–/– cells (n = 6). * denotes significance at P < 0.01.
190 191
191 192
6.3.5 α5GABAARs regulate synaptic plasticity via membrane regulation
I next sought to determine whether the attenuation of plasticity via activation of α5GABAARs results from an increase in membrane conductance in the postsynaptic neuron. A “pure” shunting effect would reduce the membrane-depolarizing effect of concurrent excitatory events by increasing membrane conductance (Staley and Mody, 1992). This form of inhibition is distinct from hyperpolarization, which is a decrease in the membrane potential that results from the current flowing through this conductance. Extrasynaptic GABAARs are known to readily contribute to the shunting effect of neurons (Staley and Mody, 1992; Brickley et al., 2001;
Nusser and Mody, 2002; Wisden et al., 2002; Bonin et al., 2007); I therefore wanted to examine whether the contribution of extrasynaptic α5GABAARs to changes in membrane potential and conductance would be sufficient to account for the differences between the genotypes during 10
Hz stimulation. Current clamp techniques were used to record whole-cell EPSPs from CA1 pyramidal neurons before, during and following 10 Hz stimulation. Hyperpolarizing current pulses (intensity 10 pA, duration 400 ms) were injected at the beginning of each sweep to monitor the input resistance of the neuron. SR-95531 and CGP 55845 were added to the extracellular solution to inhibit GABAAR IPSPs and GABABR IPSPs, respectively. Under baseline condition, the amplitude of the EPSPs (WT 6.77 ± 1.87 mV, Gabra5–/– 6.93 ± 2.92 mV, L-655,708-treated WT 6.23 ± 2.34 mV), the input resistance (WT = 257.96 ± 14.17 MΩ,
Gabra5–/– = 243.75 ± 28.32 MΩ, L-655,708-treated WT = 246.67 ± 20.76 MΩ ), and the resting membrane potential (WT –59.62 ± 0.91 mV, Gabra5–/– –58.56 ± 0.81 mV, L-655,708- treated WT -58.42 ± 0.65 mV) were similar between the groups.
The mean amplitude of EPSPs was significantly reduced in WT slices 30 min following 10 Hz stimulation, but was significantly increased in both Gabra5–/– and L-655,708-treated WT slices
192 193
(Figure 6.9A1, ANOVA, P <0.05). This effect is similar to our primary effect (Figure 6.2C) and the SR-95531 effect (Figure 6.8A) for slices stimulated with 10 Hz. The summed data for the whole cell LTP are presented in Figure 6.9A2). In order to address whether α5GABAARs inhibit
LTP by decreasing the membrane resistance and possibly contribute to a shunting effect I measured the input resistance before and 30 min following 10 Hz stimulation. As mentioned above and as shown in Figures 6.9B1 and 6.9B2 there were no significant differences between any of the groups with respect to baseline input resistance. This suggests that the deletion or pharmacological inhibition of α5GABAARs does not appreciably contribute to the basal input resistance of the neuron. However, the average input resistance 30 min following 10 Hz stimulation was significantly decreased in WT neurons (182.43 ± 21.97 MΩ, Figure 6.9B1 and
6.9B2, P < 0.05 compared with baseline). In contrast, there was no change in the average input resistance 30 min following 10 Hz stimulation for both Gabra5–/– (239.19 ± 29.33 MΩ, P >0.05 compared with baseline) and L-655,708-treated WT neurons (240.34 ± 24.31 MΩ, P > 0.05 compared with baseline) (Figures 6.9B1 and 6.9B2).
193 194
Figure 6.9. Decreased membrane resistance accompanies 10 Hz induced LTD in WT neurons.
(A1) Examples of EPSP responses for single WT (blue), Gabra5–/– (red) and L-655,708-treated
WT neurons (grey). 1 = baseline EPSP, 2 = EPSP 30 min following 10 Hz (A2) Summary of the
EPSP responses for all groups. (B1) Example traces show the membrane hyperpolarizing pulse for WT, Gabra5–/– and L-655,708-treated WT neurons, before and 30 min following 10 Hz stimulation. The average input resistance was significantly decreased in WT but not Gabra5–/– or L-655,708-treated WT neurons. The individual changes in input resistance for each cell are shown in the before and after plots. Coloured symbols represent the group means. 1 = baseline
EPSP, 2 = EPSP 30 min following 10 Hz (B2) Graph showing the time course and magnitude of input resistance after 10 Hz stimulation in all groups. Note for these measurements input resistance was measured by a single -10 pA injection
194 195
195 196
Next, I wanted to determine whether α5GABAARs were also limiting the depolarization of the neuronal membrane, which would contribute to LTP impairment by limiting the ability of
NMDARs to relieve their Mg2+ block (Crunelli and Mayer, 1984) Given that the resting potential of CA1 pyramidal neurons was approximately -60 mV, which is more positive than the
calculated GABA reversal potential (-70 mV), the tonic activation of extrasynaptic GABAA receptors would be expected to hyperpolarize the membrane potential, thereby decreasing the input-output relationship in neurons (Chance et al., 2002; Mitchell and Silver, 2003). As mentioned above, the resting membrane potential was similar between all groups (Figure 6.10A).
During the 60 sec 10 Hz stimulation period, the membrane potential recorded from WT neurons did not change from baseline values (WT during 10 Hz = -60.29 ± 0.80 mV) but the membrane potential was significantly depolarized in Gabra5–/– and L-655,708-treated WT neurons
(Gabra5–/– during 10 Hz = -54.39 ± 1.26 mV, L-655,708-treated WT neurons during 10 Hz = -
53.59 ± 2.42 mV, ANOVA, P < 0.05 compared to resting values). Interestingly, following 10 Hz stimulation the membrane potential was slightly hyperpolarized in WT neurons (WT post-10 Hz
= -63.17 ± 0.85 mV, P < 0.05 compared to resting values) but returned to resting values in
Gabra5–/– and L-655,708-treated WT neurons (Gabra5–/– post-10 Hz = -58.19 ± 0.61 mV, L-
655,708-treated WT neurons post-10 Hz = -56.52 ± 1.24 mV P > 0.05 compared to resting values). Taken together, these results indicated to us that during 10 Hz stimulation the α5 subunit may act to regulate the neuronal membrane potential and decrease membrane resistance such that
LTP is inhibited. The changes in membrane potential before, during and following 10 Hz stimulation are shown in Figure 6.10A.
In order, to determine whether the regulation of the neuronal membrane by α5GABAARs influenced the initial triggering events of LTP, I analyzed the depolarizing envelope during the
196 197
10 Hz stimulation period. The depolarizing envelope can be defined as the average area under the integrated EPSPs during the simulation protocol, and is a measure of the amount of neuronal stimulation that may be required to trigger and initiate LTP. The depolarizing envelope was significantly greater in Gabra5–/– compared with WT neurons (Figures 6.10B and 6.10C, P <
0.05). L-655,708-treated WT neurons displayed a depolarizing envelope similar to Gabra5–/– neurons (Figures 6.10B and 6.10C, P > 0.05), which was significantly greater than WT control neurons (Figures 6.10B and 6.10C, P < 0.05). This suggested to us that α5GABAARs may act to decrease the overall neuronal activity during 10 Hz stimulation through the regulation of the neuronal membrane. Without this regulator (i.e. α5GABAARs) the membrane becomes depolarized during persistent (i.e. 10 Hz) stimulation, which increases overall activity, excitability and EPSP growth. The summed data for the depolarizing envelope are shown in
Figure 6.10C.
197 198
Figure 6.10. α5GABAARs regulate membrane properties, depolarizing envelope and input resistance with 10 Hz stimulation.
(A) The membrane potential during 10 Hz stimulation was significantly depolarized in Gabra5–
/– (n = 5) and L-655,708-treated WT neurons (n = 5) during 10 Hz stimulation. The average membrane potential for the 10 min period following 10 Hz stimulation was significantly hyperpolarized from baseline values in WT (n = 5). All recordings were conducted in the presence of SR-95531 and CGP 55845 to block synaptic GABAA and GABAB receptors, respectively. Statistically significant effects were compared to baseline membrane potentials for the respective treatments. (B) Example traces show that during 10 Hz stimulation the EPSPs were significantly smaller for the duration of the simulation period in WT neurons compared with Gabra5–/– and WT + L-655,708 neurons. Insets show that EPSPs are clearly larger in
Gabra5–/– and WT + L-655,708 neurons near the beginning and at the end of the stimulation period. (C) The pooled data for the depolarizing envelope are shown. The average depolarizing envelope was significantly larger in Gabra5–/– neurons and WT + L-655,708 neurons compared with WT control neurons. * denotes significance at P < 0.05.
198 199
199 200
6.4 Discussion
In this chapter, I have demonstrated a physiological role for α5GABAARs in associative learning and the regulation of long-term, activity-dependent enhancement of synaptic plasticity in CA1 hippocampal pyramidal cells. Specifically, genetic and pharmacological inhibition of
α5GABAAR activity improved weak associative learning and lowered the threshold for the induction of LTP. A narrow window of frequency input determined whether α5GABAAR activity would attenuate the potentiation of synaptic responses. The tonic inhibitory conductance generated by α5GABAARs decreased membrane resistance and prevented membrane depolarization, which coincided with impaired LTP when neurons were stimulated at 10 Hz.
Equally important, the modification of excitatory synaptic potentials by tonic inhibition was independent of phasic GABAergic transmission, as blocking IPSPs with SR-95531 failed to alter the threshold for LTP induction.
Indiscriminate reduction in GABAAR-mediated inhibition facilitates synaptic plasticity
(Wigstrom and Gustafsson, 1983) and improves memory performance; however, non-selective
GABAAR antagonists are proconvulsant and anxiogenic (Ben-Ari et al., 2007). Selective inhibition of α5GABAARs appears to facilitate learning and enhance synaptic potentiation without a profound proconvulsant effect (Atack et al., 2006b). This desirable drug profile may be attributed to the restricted pattern of α5GABAARs, which represent less than 5 % of all
GABAARs in the brain, yet are strongly expressed in the CA1 region of the hippocampus, where they contribute up to 25 % of GABAARs (Pirker et al., 2000). In the behavioural tasks, Gabra5–
/– mice outperformed WT littermates during trace but not contextual fear conditioning. I attribute this difference to the intensity of the training and a ceiling effect, rather than to the involvement of different anatomical regions in the behavioural paradigms, as pharmacologically enhancing
200 201
α5GABAAR activity attenuates memory performance in contextual fear tasks (Cheng et al.,
2006). Furthermore, our results are entirely consistent with the enhanced memory performance of α5H105R mice in trace fear conditioning (Crestani et al., 2002).
In hippocampal slices, I found that repetitive stimulation within a narrow but critical frequency range (10 Hz to 20 Hz), potentiated CA1 synapses when α5GABAAR activity was decreased.
Blocking α5GABAARs with L-655,708 abolished the LTD in WT neurons, but did not completely replicate the effect of 10 Hz stimulation on Gabra5–/– slices. However, this may be related to the concentration of L-655,708 used, which only blocked approximately 50 % of the total bicuculline tonic current. It appears that the effect of α5GABAARs on 10 Hz-induced plasticity occurs independent of synaptic inhibition because the application of L-655,708 did not significantly inhibit mIPSCs and a concentration of SR-95531, which completely inhibited the
IPSP, did not convert the LTD in WT slices to LTP. Further, blocking all GABAARs with bicuculline converted the LTD in WT slices to LTP but did not further potentiate the response of
Gabra5–/– slices when stimulated with 10 Hz. This indicates that α5GABAARs are necessary to enhance plasticity under these specific experimental conditions independent of synaptic inhibition. In contrast to these findings, bicuculline strengthened synaptic plasticity in WT and
Gabra5–/– slices when the slices were stimulated at 100 Hz. The latter finding is consistent with the recruitment of a broad range of GABAARs at higher stimulation frequencies. Others have shown an increase in inhibitory drive and GABA release during high-frequency stimulation and greater potentiation of fEPSPs with the suppression of inhibitory inputs during high- versus low- frequency stimulation (Chapman et al., 1998).
201 202
In this chapter, I also demonstrated that 10 Hz-induced LTD was accompanied by decreased input resistance, which lasted for 30 min. Previously, it has been shown that in the CA1 region, a 3 Hz pairing protocol induced LTD that is accompanied by an increase in the input resistance (Brager and Johnston, 2007). The authors attributed the increased input resistance to a physiological loss of h-channels. However, unlike the present study, the 3 Hz LTD was independent of NMDARs and instead mediated by internal stores of Ca2+. Interestingly, it has been shown that the
NMDAR can also account for changes in input resistance (Crunelli and Mayer, 1984). In the absence of Mg2+ the depolarizing response of hippocampal neurons to NMDA is accompanied by a decrease the input resistance. In the presence of Mg2+ ions, however, the response to
NMDA is always associated with an apparent increase in input resistance (Crunelli and Mayer,
1984). The Ih and NMDA changes in input resistance are converse to our current findings and suggest that these currents do not likely mediate the input resistance decrease following 10 Hz stimulation. Since L-655,708 blocked the decrease in input resistance in WT neurons, it is likely that the decrease in input resistance is mediated by α5GABAARs. If α5GABAARs were significantly activated following 10 Hz stimulation and this activation was sustained, then we would expect a decrease in the input resistance of the neuron and the accompanying membrane potential, which my data supports.
The Gabra5–/– mice lack the relatively higher-affinity GABAARs (Burgard et al., 1996; Yeung et al., 2003), which may act to lessen the “shunting” of excitatory drive within the 10–20 Hz range. However, from our data it is apparent that the effect of 10 Hz stimulation on plasticity is not a pure shunting effect, but rather a combination of shunting and membrane hyperpolarization. Together, these effects may act to decrease the EPSP following 10 Hz, possibly by limiting the relief of the Mg2+ block of NMDARs. In particular it has been shown
202 203 that different NMDAR subunits may differentially contribute to LTD and LTP (Liu et al., 2004;
Bartlett et al., 2007). For example extrasynaptic NMDA receptors are thought to preferentially contain the NR2B subunit of the NMDA receptor (Tovar and Westbrook, 1999), and these receptors have been shown to participate in LTP (Bartlett et al., 2007). The close proximity of
α5GABAARs to specific subpopulations of NMDA receptors may, through membrane regulation, limit the efficacy of NMDA receptors, which would in turn inhibit plasticity or learning.
Despite similar resting membrane potentials in WT and Gabra5–/– neurons before 10 Hz stimulation, the membrane potential of Gabra5–/– neurons was significantly depolarized and the depolarizing envelope was significantly larger compared to WT neurons. The application of L-
655,708 replicated the effect observed in Gabra5–/– neurons, as the membrane potential was depolarized and depolarizing envelope was enhanced in WT neurons. The effect on membrane potential during 10 Hz stimulation is interesting considering there are no differences in the resting membrane potential in WT, Gabra5–/– or L-655,708-treated WT neurons. This would suggest that 10 Hz stimulation is activating some process that is unaccounted for, which is significantly reduced by limiting α5GABAAR activity, such as the Ih current (Brager and
Johnston, 2007). It should also be noted that the calculated reversal potential for Cl- in the patch pipette was -70 mV, which would account for the membrane hyperpolarizing effect observed in
WT slices. However, further experiments will likely determine the contribution of the membrane potential by altering the Cl-reversal potential.
Both the direction and the degree of plasticity vary with factors that modify the synaptic “state.”
This “plasticity of plasticity” is referred to as metaplasticity, a term originally coined by
Abraham and Bear (Abraham and Bear, 1996). Plasticity refers to a change in signalling efficacy
203 204 at a given synapse whereas metaplasticity is a change in the ease with which that change in signalling efficacy can take place. Metaplasticity may serve to maintain the synaptic response within a dynamic range that is optimal for learning processes or provide a braking mechanism that prevents the development of a saturated synapse. Metaplasticity is analogous to the “sliding threshold” model of experience-dependent synaptic plasticity described by Bienenstock, Cooper, and Munro (Bienenstock et al., 1982), a model that was developed to describe visual cortex plasticity during development but was adapted to describe plasticity in other regions, including the CA1 hippocampus. Activity above the modification threshold (θM) leads to LTP, whereas low-threshold activity leads to LTD where θM can be regulated by a number of factors, including prior activity at synapses and relative contribution of NMDA subtypes (Bear, 1995; Abraham et al., 2001). My results show that α5GABAARs play a role in metaplasticity as novel regulators of the neuronal membrane potential, which may shift the ease or threshold for the induction of LTP, including the depolarizing envelope and the initial triggering events of LTP.
204 205
Chapter 7. General Discussion
7.1 Overview
The broad objective of this thesis was to understand the role of α5GABAARs in learning and memory and synaptic plasticity. Overall, the experimental data indicates that the activity of
α5GABAARs can be up- and down-regulated to bidirectionally modify learning and memory and synaptic plasticity. I found that the behavioural impairment of learning and memory correlates with an increase in the α5GABAAR-dependent tonic conductance and the impairment of LTP correlates with an α5GABAAR-dependent decrease in the membrane potential and input resistance of CA1 pyramidal neurons. The significance of these findings to our understanding of
α5GABAAR-dependent regulation of memory is considered below.
The implications of our behavioural and LTP data are extensive and bridge an existing gap in the
α5GABAAR literature. At the onset of this thesis work, published reports indicated that the genetic deletion of α5GABAARs improved the acquisition of the spatial water maze task but did not alter LTP (Collinson et al., 2002). Additional work revealed that reduced α5GABAAR expression in the α5H105R point-mutant mouse model, enhanced trace but not contextual fear conditioning (Yee et al., 2004). In particular I found the fear conditioning data striking because both contextual and trace fear conditioning are dependent on the hippocampus. This suggested to us that α5GABAARs are involved in specific forms of hippocampus-dependent learning and memory. Following these initial genetic studies, the development of a class of drugs, known collectively as inverse agonists, that possesses a high affinity for α5GABAARs and decrease the intrinsic activity of the receptor was developed (Maubach, 2003; Chambers et al., 2004; Atack et
205 206 al., 2006b; Dawson et al., 2006). Similar to the genetic studies, inverse agonists selective for
α5GABAARs improved water maze learning but they also enhanced LTP (Atack et al., 2006b).
The enhancement of LTP by inverse agonists selective for α5GABAARs is in stark contrast to the genetic reduction of α5GABAARs, as the latter does not facilitate LTP (Collinson et al., 2002).
The differences between the genetic and pharmacological evidence for involvement of
α5GABAARs in LTP, as an underlying mechanism to explain the enhanced memory performance was a primary motivating factor for this thesis. I was also interested in the potential clinical utility of modifying α5GABAAR activity to impair or improve memory.
It must be clearly appreciated that an increase in memory performance does not necessarily equal enhanced LTP. There are many studies in which changes in synaptic plasticity do not correlate with the expected changes in the behaviour of an animal (Jia et al., 1996; Gerlai et al., 1998; Jun et al., 1998; Meiri et al., 1998). For instance, mutant mice lacking the GluR2 subunit of the
AMPAR display a 2-fold enhancement of LTP in the CA1 region of hippocampal slices (Jia et al., 1996) but are impaired in the hidden platform version of the Morris water maze (Gerlai et al.,
1998). It is thought, at least for the GluR2 knockout mice that they suffer from an overall non- specific increase in excitability (particularly Ca2+ influx) that may enhance LTP and adversely alter cognitive functions. This thesis was designed in order to bridge the gap between the in vivo and in vitro influence of α5GABAARs on memory and LTP, respectively because an established link had not been convincingly demonstrated. For instance, I started with the premise: “If decreasing the overall activity of α5GABAARs enhances memory then dramatically increasing the activity of this receptor should impair learning processes.” Unfortunately, at the onset of this thesis there were no selective “agonists” or positive modulators for α5GABAAR function; however, previous observations in our laboratory demonstrated that these receptors had a
206 207 particularly high affinity for certain general anesthetics (Caraiscos et al., 2004a; Cheng et al.,
2006).
7.1.1 α5GABAAR-activity correlates with cognitive performance
In this thesis I showed that etomidate selectively potentiated the α5GABAAR-dependent tonic conductance at low concentrations and this coincided with a decrease in hippocampus-dependent memory performance. Etomidate also abolished LTP in slices prepared from WT but not
Gabra5–/– mice. In order to be certain that the memory-impairing properties of etomidate were directly mediated through α5GABAARs, I reversed the inhibition of memory and synaptic plasticity with the benzodiazepine inverse agonist L-655,708. In these experiments I found that
L-655,708 prevented LTP impairment and restored the input-output function of CA1 neurons that is decreased by etomidate application following TBS. In our experiments L-655,708 blocked the etomidate-enhanced tonic current by approximately 70 %, without influencing mIPSC events. This suggested to us that the actions of etomidate and L-655,708 on LTP were mediated by an extrasynaptic mechanism, as opposed to α5GABAARs that may be expressed synaptically
(Christie and De Blas, 2002; Serwanski et al., 2006). This data also supports our hypothesis that increases in the α5GABAAR-dependent tonic conductance impairs learning and memory.
In the behavioural experiments etomidate impaired performance in contextual fear conditioning, when mild or high foot shock intensities were used. The co-administration of L-655,708 and etomidate reversed the impairment of contextual fear conditioning regardless of shock intensity, suggesting that the level of α5GABAAR activity is modifiable allowing for the impairment or the rescue of memory. Two shock intensities were used for the contextual fear conditioning experiments because I wanted to ensure that the freezing performance was not saturated and
207 208 could undergo further enhancement. As I predicted etomidate decreased performance when a mild or high shock intensity was used, and these responses were both subsequently reversed with
L-655,708. Interestingly, L-655,708 did not enhance contextual fear responses with the mild or high foot shock intensity. This was surprising as L-655,708 possesses significant cognitive enhancing properties (Atack et al., 2006b). I found that L-655,708 enhanced performance only in the trace fear conditioning experiments and not in contextual fear conditioning or even water maze learning. This strongly suggested that the type of learning task may be important for determining the involvement or participation of α5GABAARs in cognition. This lead us to conclude that either 1) the contextual fear conditioning assay is not quite as hippocampus- specific as trace conditioning 2) L-655,708 does not possess significant intrinsic cognitive- enhancing properties or 3) targeting α5GABAARs for cognitive enhancement may be task and/or objective specific.
There is considerable evidence indicating that a high proportion of contextual fear conditioning requires the hippocampus (Kim and Fanselow, 1992; Phillips and LeDoux, 1992; Maren and
Fanselow, 1997), thus it is interesting that L-655,708 did not enhance contextual fear memory given the high proportion of α5GABAARs in this brain region (Sur et al., 1999; Pirker et al.,
2000). A typical fear conditioning protocol such as that used in our experiments (3 successive foot shocks) is highly dependent on the hippocampus (Phillips and LeDoux, 1992; Anagnostaras et al., 2001), but recent data supports the idea that strong fear association protocols (i.e. high number of intense foot shocks) are primarily mediated by the amygdala and not the hippocampus
(Quinn et al., 2008). It is likely that our contextual fear conditioning protocol is hippocampus- dependent (low number of foot shocks) although significant contributions from other brain regions cannot be ruled out (Phillips and LeDoux, 1992; Kim et al., 1993).
208 209
Further, I showed that L-655,708 significantly enhanced the performance of WT mice in trace fear conditioning. Trace fear conditioning is an unique task because it is highly complex and engages the CA1 region of the hippocampus (Moyer et al., 1996; McEchron et al., 1998;
McEchron et al., 2001) and requires NMDA-dependent signalling in this region (Misane et al.,
2005; Wanisch et al., 2005). In our experiments, Gabra5–/– mice significantly outperformed their WT littermates in the trace fear conditioning task and L-655,708 significantly enhanced the performance of WT mice. Both the trace and contextual fear conditioning protocols are dependent on the hippocampus, and it is unlikely that the differing results are related to the involvement of α5GABAARs in different hippocampal regions.
α5GABAARs are distributed throughout the septotemporal axis (Hauser et al., 2004), and similar contributions of α5GABAARs in the dorsal versus ventral hippocampus are likely. For example the acquisition and maintenance of trace fear conditioning has been shown to depend on both the dorsal (Misane et al., 2005) and ventral hippocampus (Czerniawski et al., 2008). There is also considerable evidence that indicates that dorsal hippocampal lesions abolish contextual fear conditioning (Kim et al., 1993; Maren and Fanselow, 1997; Maren, 1998), but this has not been universally demonstrated (Phillips and LeDoux, 1992; Maren et al., 1997; Wiltgen et al., 2006).
It should also be recognized that trace fear and contextual fear conditioning results do not always correlate (Crestani et al., 2002; Han et al., 2003) and the type of associative learning between the two paradigms may be fundamentally different.
It is possible that α5GABAARs are required for higher order and complex associative tasks.
Trace fear conditioning has been reported to require higher order functions such as attention and
209 210
involve the anterior cingulate cortex (Han et al., 2003). α5GABAARs appear to play a prominent role in attentional processing as genetically reducing α5GABAARs decreased prepulse inhibition, which is an attentional model for schizophrenia (Hauser et al., 2004). Conversely, the contextual fear conditioning training that was used in this thesis may represent a strong fear association and a significant proportion of the performance may be mediated by the amygdala (Quinn et al.,
2008).
It has previously been shown that the short-term memory required to complete the water maze task can remain intact, while consolidation of this information is disrupted (Devan et al., 2001).
An unexpected finding of our etomidate work was the ability of etomidate-treated mice to learn during the acquisition trials of the Morris water maze but show impairment in recalling the platform location 24 hrs after training. This is unprecedented and suggests that the effect of etomidate on spatial tasks does not influence learning, but instead the retention of information.
The acquisition of the matching to place version of the water maze may also depend on alternative brain regions such as the prefrontal cortex that are not susceptible to the cognitive- disrupting effects of etomidate (Veselis et al., 2009). Additionally, I found that inhibiting
α5GABAARs before but not following training in the trace fear conditioning protocol improved memory performance. This result indicates that α5GABAARs are important for the acquisition but not consolidation of trace fear memory. Similar results have been found in the water maze task and indicate that α5IA-II improves acquisition and retention but not consolidation of the hidden platform location (Collinson et al., 2006).
While our genetic and pharmacological data for the actions of etomidate on LTP and behaviour convincingly suggest an α5GABAAR-dependent mechanism, other reports suggest that the
210 211 potentiating effect of etomidate on the α5 subunit is minimal (Janssen et al., 2009). A recent study reported that in transfected HEK 293 cells the direct effects of etomidate are selective for
β3 subunit-containing GABAARs (Janssen et al., 2009). The potentiating effect of etomidate on
β3 subunit-containing GABAARs was more pronounced when an α2 subunit was present compared with an α5 subunit. However, these data do not suggest that etomidate does not target
α5GABAARs; it merely highlights the affinity of etomidate for additional GABAARs. In the
Janssen et al., (2009) study a relatively high concentration of etomidate was used (3 µM), which does not give insight into the actions of etomidate at significantly lower concentrations. The gating effects of etomidate could be 2-fold and at lower concentrations etomidate may directly gate or even more readily modulate the α5β3γ2 GABAARs. Nonetheless, my data shed light on the actions of etomidate and suggest that the potentiation of an α5GABAAR-dependent tonic conductance impairs hippocampal synaptic plasticity and hippocampus specific behaviours.
7.1.2 α5GABAARs do not possess intrinsic anxiety, sedative or hypnotic
properties
The cognitive-impairing effects of etomidate on water maze and fear conditioning performance were not attributed to other confounding behavioural actions of the drugs such as hypnosis or sedation. In the rotarod test for motor coordination etomidate equally impaired WT and Gabra5–
/– mice 5 min following etomidate injection but both genotypes quickly recovered 30 min after injection (same timeline for memory tests). In the open field, etomidate equally decreased the spontaneous walking time of both genotypes. The mobility of the mice is a potential concern because the main assessments of memory performance used are two motor related tasks.
211 212
The rotarod data suggests that mice injected with 4 mg/kg of etomidate will ambulate when provided with the proper motivation (i.e. placed on a rotating rod) and the open field data suggests that only exploratory behaviours are impaired. A decrease in spontaneous exploration by etomidate is not a concern for the fear conditioning studies because decreased movement in fear conditioning is indicative of enhanced performance. In our fear conditioning studies the etomidate injected mice moved more (i.e. less freezing) than control mice suggesting that the sedative properties of etomidate were not a factor. Further, it is unlikely that the effects of etomidate on fear conditioning are mediated through pain pathways because etomidate has no analgesic properties (Reeves JG, 2005), and fear memory is assessed a minimum of 24 hrs following conditioning, well after etomidate has cleared from the body (Gooding and Corssen,
1976; Davis and Cook, 1986).
It is also interesting that for WT and Gabra5–/– mice, their LORR scores were the same indicating that the deletion of α5GABAARs does not alter the hypnotic properties of etomidate. I found that the average dose for LORR for our mice was 10 mg/kg, which was the same between genotypes. Another group has reported that the ED50 for etomidate-induced amnesia was 11 mg/kg (Benkwitz et al., 2007). They further extrapolated that the EC50 aqueous concentration for etomidate-induced amnesia is 0.25 µM. The authors studied mice that were considerable younger
(3- to 4-week old animals) than those used for our etomidate experiments although most investigators perform behavioural studies with mice 3 months of age or older. It is also generally appreciated that age influences redistribution rates and drug metabolism especially with anesthetics (Hemmings et al., 2005). These factors may contribute to the change in etomidate potency with age and appear to influence the potency of intravenous anesthetics. However, the results of Benkwitz et al., (2007) suggest that subtle differences in experimental design and set-
212 213 up may account for large differences in experimental findings. Regardless, our etomidate results were remarkable and were the first demonstration that single GABAAR subunits can dissociate the behavioural endpoints of general anesthetics.
7.2 In vitro regulation of learning and memory by α5GABAARs on
hippocampal synaptic plasticity
In our behavioural experiments the contextual fear conditioning protocols induced high levels of freezing. The only intrinsic effect that I found for α5GABAARs on learning and memory was observed with trace fear conditioning, which had been previously demonstrated with a different mouse strain (Yee et al., 2004). My behavioural data seemed to suggest that a continuum may exist whereby α5GABAARs are not important for strong memory performance (such as in contextual fear conditioning) but may become increasingly important for higher order memory associations (such as trace fear conditioning). My behavioural results forced me to ask the question “how do α5GABAARs modify learning and memory?”
A close examination of previously published α5GABAAR-LTP data indicated that inverse agonists selective for α5GABAARs enhanced LTP but differences between the LTP protocols may determine whether α5GABAARs participate in the modification of LTP. In the original
Gabra5–/– study, there were no differences in LTP between WT and Gabra5–/– mice when 1 sec stimulation of TBS was used (Collinson et al., 2002). However, additional studies revealed that the benzodiazepine inverse agonist α5IA enhanced LTP (Dawson et al., 2006). In this study, plasticity was induced by a brief priming stimulus (10 stimuli at 100 Hz) followed 30 min later by theta burst stimulation. Notably, following the priming stimulus, synaptic strength was
213 214
increased to 200% in both the L-655,708-treated and control slices suggesting that α5GABAAR do not play a critical role. However, after the second phase of stimulation L-655,708-treated slices showed enhanced plasticity. Thus, the stimulation protocol dramatically influences the subpopulation of GABAARs that influence plasticity and possibly memory behaviours.
Similarly, the LTP-enhancing effects of L-655,708 have been reported to depend on the LTP stimulation protocol (Atack et al., 2006b). L-655,708 did not enhance LTP from control values when a brief 100 Hz priming pulse was delivered, but in the same slice following a 30 min stabilization period (i.e. post-100 Hz priming pulse), TBS evoked greater LTP in L-655,708- treated slices compared with vehicle-treated slices (Atack et al., 2006b). These differences suggested to me that, much like our behavioural data there is an in vitro continuum for synaptic alterations that is dependent on the protocol and conditions of activation. With this in mind I decided to investigate the BCM model of bidirectional synaptic plasticity and stimulate slices with varying degrees of presynaptic stimulation to determine whether the weight of α5GABAAR involvement in synaptic plasticity changes as a function of presynaptic stimulation.
7.2.1 α5GABAARs shift the threshold for LTP
The genetic deletion of α5GABAARs shifted the modification threshold for LTP leftward, which promoted the induction of LTP over LTD. Our most significant and consistent effect between
WT and Gabra5–/– slices occurred when the Schaffer collateral pathway was stimulated for 60 sec with 10 Hz stimulation; LTD was observed in WT slices and LTP was observed in Gabra5–
/– slices. A similar trend has been previously reported in rat slices stimulated with a 10 Hz tetanus train (Li et al., 2008). In vehicle control slices, 10 Hz stimulation resulted in an LTD of synaptic responses, which followed a similar time course to our WT slices. Interestingly, this
214 215
LTD was converted to LTP by inhibiting TRPV1 receptors with capsaicin. My results suggest that α5GABAARs limit LTP in CA1 neurons, but I cannot rule out the possibility of an interaction with another receptor population (i.e. TRPV1). My results also highlight the importance of α5GABAARs in activity-dependent plasticity and suggest that α5GABAARs are differentially activated for LTP processes.
Pharmacologically inhibiting α5GABAARs with L-655,708 abolished the LTD in WT slices. In the extracellular recordings, the fEPSPs of L-655,708-treated WT slices did not reach the degree of LTP observed in Gabra5–/– slices. It is probable that this is due to the low concentration of L-
655,708 used in our experiments (10 nM), resulting in an incomplete block of α5GABAARs
(Quirk et al., 1996; Casula et al., 2001). The L-655,708 concentration that I used only blocked 50
% of the total tonic conductance in CA1 pyramidal neurons (Figure 6.7). Strikingly, bicuculline converted the LTD in WT slices to LTP, an effect that was indistinguishable from Gabra5–/– slices, but did not further potentiate the LTP in Gabra5–/– slices. This suggested to us that blocking extrasynaptic GABAARs was important for LTP-induction with 10 Hz stimulation. I confirmed this by applying a concentration of SR-95531 that only blocks the synaptic GABAAR currents. SR-95531 did not convert the 10 Hz-induced LTD into LTP, further confirming that persistent stimulation of the Schaffer collateral pathway at 10 Hz is optimal for involvement of
α5GABAARs in plasticity. More importantly, this reinforces the concept that α5GABAARs respond differently to distinct patterns of presynaptic stimulation.
7.2.2 α5GABAARs regulate membrane properties to limit LTP
I hypothesized that α5GABAARs act to regulate neuronal membrane properties thereby preventing or limiting LTP. Given that the resting membrane potential of CA1 pyramidal
215 216 neurons was approximately –60 mV, which is more positive than the calculated GABA reversal
potential (–70 mV), the tonic activation of extrasynaptic GABAARs would be expected to hyperpolarize the membrane potential and produce a shunting inhibition. This effect would act to decrease the input-output relationship in neurons (Chance et al., 2002; Mitchell and Silver, 2003) and reduce LTP. I found that following 10 Hz stimulation, WT neurons were hyperpolarized and the input resistance was significantly decreased, as predicted. The decreased input resistance corresponded to LTD in WT neurons, while there were significant resistance changes for
Gabra5–/– and L-655,708-treated WT neurons during LTP. This effect is likely to contribute to the intrinsic excitability of neurons (Bonin et al., 2007) and acts to reduce LTP in hippocampal neurons.
It was surprising to me that following 10 Hz stimulation, there was such a large drop in the input resistance of the WT neurons. Electrophysiologists go to great lengths to ensure that the input resistance of neurons does not change during LTP recordings. A decrease in the input resistance of a neuron by greater than 25 % typically indicates poor neuronal health and electrophysiologists will typically discard these recordings. However, my data indicates that at least with 10 Hz stimulation there is a sustained drop in the input resistance of WT neurons that is not present in Gabra5–/– neurons and is prevented by pre-application with L-655,708.
Interestingly, hippocampal LTP induced by TBS decreased cellular excitability and input resistance, which has been attributed to an increase in the H-current (Fan et al., 2005). In the same study, the authors report that LTP induced with 100 Hz stimulation did not cause the input resistance of the neuron to change. Additionally, a LTD-inducing protocol increased cellular excitability and the input resistance of neurons, which was associated with a physiological loss of the H-current (Brager and Johnston, 2007). Taken together these results suggest that changes
216 217 in the input resistance of the neuron may depend on the plasticity-inducing protocol. In this thesis 10 Hz stimulation was sufficient to induce a significant and prolonged drop in the neuronal membrane resistance, which was blocked by inhibiting α5GABAARs. This suggests that 10 Hz stimulation increases an α5GABAAR-mediated conductance, which acts to decrease membrane resistance and limit LTP. Unfortunately, since input resistance measurements were not taken following stimulation at another frequency, my observation is limited to the 10 Hz stimulation protocol.
Furthermore, the slowly decreasing input resistance following 10 Hz stimulation in WT neurons is unexpected. A decrease in the neuronal input resistance due to an increase in the α5GABAAR- mediated conductance would result in an immediate drop in the input resistance. However, the drop in input resistance develops over a period of 10 min in WT neurons. Interestingly, it has been shown that increases (Brager and Johnston, 2007) and decreases (Fan et al., 2005) in the input resistance associated with changes in the H-current develop over a period of 10–20 min.
Thus, I cannot rule out the possibility that the sustained drop in the input resistance in my recordings is mediated by another conductance. However, the sustained drop in the input resistance is blocked by α5GABAARs, which indicates that the activation of α5GABAARs is necessary for the drop in input resistance to occur.
I have shown that α5GABAARs impair plasticity by limiting membrane depolarization and decreasing input resistance. In particular, the decreased input resistance following 10 Hz stimulation may correspond to a large increase in the α5GABAAR conductance although other conductances cannot be ruled out (Fan et al., 2005; Brager and Johnston, 2007). If the
α5GABAAR-mediated conductance accounts for the drop in membrane resistance then these
217 218 receptors have to become “over-activated” or selectively activated during 10 Hz stimulation. I suspect that prolonged 10 Hz stimulation creates a build-up of GABA within the extracellular space, which drives the hyperpolarizing and input resistance effect because higher affinity extrasynaptic α5GABAARs are more sensitive to a build up of GABA compared with the lower affinity synaptic GABAARs (Yeung et al., 2003). Since Gabra5–/– neurons do not have this receptor they are not susceptible to the GABA build-up and the associated membrane changes. I suggest that the extrasynaptic α5GABAARs are involved as opposed to synaptic α5GABAARs because bicuculline and L-655,708 but not SR-95531 converted the LTD to LTP in WT slices stimulated at 10 Hz.
While it is still unclear how this build up of GABA occurs, a few theoretical possibilities arise.
For example it is possible that there is a frequency-dependent increase in the rate of spontaneous quantal GABA release following action potentials (Miledi and Thies, 1971). A build up of
GABA caused by a direct activity-dependent effect is possible and has been shown to occur in cultured neurons (Jensen et al., 2000). Additionally, this basic assumption may be wrong and physiologically the system may work in the opposite direction. Instead of a build up of GABA, it is quite possible that there is an exhaustion of the inhibitory transmitter release pool. With prolonged 10 Hz stimulation the amount of GABA released may decrease to a steady state where the high affinity α5GABAARs are selectively activated. It has previously been demonstrated that there is a frequency-dependent decrease in the amplitude of the GABAergic IPSCs
(Baimoukhametova et al., 2004). In this study, the IPSCs were inhibited to 50 and 80 % of control responses during 10 and 50 Hz stimulation, respectively. Thus it is possible that 10 Hz stimulation exhausts the GABA transmitter release pool such that the residual GABA is sufficient to selectively enhance the activity of the high affinity α5GABAARs. Additionally, 10
218 219
Hz stimulation may preferentially increase α5GABAAR trafficking, which would act to increase the membrane conductance. I am left to assume that α5GABAARs are mediating this effect because there was no LTD in Gabra5–/– slices and LTD in WT slices was blocked by L-655,708 and bicuculline but not SR-95531.
Surprisingly, there are limited studies that address the effects of different LTP protocols on events that occur during tetanus. The plasticity initiating events occur within this brief time frame, yet it is rarely studied. I found that during 10 Hz stimulation the depolarization envelope was significantly larger in Gabra5–/– neurons suggesting that the effects of α5GABAARs on
LTP may be due to some disruption in the initial triggering events that are required for LTP
(Lisman and Spruston, 2005). To my knowledge, this is the first study that demonstrates that a specific receptor population limits or “shunts” the activity of the activated EPSPs during tetanus.
The term shunt is loosely used here and is only meant to reflect a decrease in the total EPSP activity during 10 Hz stimulation. Realistically, a true shunting of the EPSPs would be accompanied by a decrease in the input resistance of the neuron (Priebe and Ferster, 2002) and as
I did not monitor the input resistance during the 10 Hz stimulation period, shunting cannot be used in the electrophysiological sense. It is also possible that depolarizing the neuron during 10
Hz stimulation relieves the voltage-dependent Mg2+ block of NMDARs enhancing the EPSP and
LTP (Crunelli and Mayer, 1984). Strikingly, the membrane depolarization and the depolarizing envelope are drastically reduced by the activity of α5GABAARs suggesting that α5GABAARs may limit NMDAR function.
In this thesis I have shown that the genetic deletion or pharmacological inhibition of
α5GABAARs does not affect the basal membrane properties of hippocampal neurons. I found
219 220 that the WT, Gabra5–/–, and L-655,708-treated WT neurons have similar resting membrane potentials and input resistances. In hippocampal tissue slices, similar effects have been reported in that the contribution of α5GABAARs to the resting membrane potential and input resistance of the neuron is minimal when α5GABAARs are genetically deleted or pharmacologically inhibited
(Glykys et al., 2008). In cultured hippocampal neurons the resting membrane potential was found to be similar between WT and Gabra5–/– neurons (Bonin et al., 2007). However, in the cultured neuronal preparation the input resistance of Gabra5–/– neurons was 1.6-fold greater and the membrane potential of WT but not Gabra5–/– neurons was significantly hyperpolarized or depolarized with GABA or picrotoxin applications, respectively (Bonin et al., 2007). The differences between the cultured preparation and tissue slice preparation may account for the differences of α5GABAARs with respect to input resistance and changing membrane potential.
In this thesis I found no effect of α5GABAARs on basal membrane properties during baseline recordings and others have reported similar findings in the tissue slice preparation (Glykys et al.,
2008).
7.3 Conclusion
The primary conclusions from this thesis are that α5GABAAR-activity constrains hippocampus- dependent learning and memory and LTP when specific behavioural protocols and conditions of activation are met. In the behavioural studies, α5GABAAR-activity attenuated the performance of mice in trace fear conditioning. Conversely, α5GABAAR-activity did not play a major role in the performance of contextual or cued fear responses. The behavioural data argues in favour of a role for α5GABAARs in complex polymodal or higher order associative tasks. Similarly, the activity of α5GABAARs predominates under specific conditions of stimulation to initiate synaptic potentiation. α5GABAAR-activity limits synaptic potentiation for stimulation signals
220 221 within the 10 to 20 Hz range. The parallel that exists between the in vivo and in vitro conditions is that α5GABAARs effectively modify behaviour or plasticity when specific conditions of activation are met. It is clear that both behaviourally and electrophysiologically, α5GABAARs do not modify strong stimulation signals.
Mechanistically, I conclude that α5GABAARs oppose 10 Hz-induced LTP by damping postsynaptic changes in the properties of the membrane. Genetic or pharmacological deletion of
α5GABAARs resulted in a significant depolarization of the neuronal membrane potential during
10 Hz stimulation. This effect was also associated with an increase in the depolarizing envelope of the EPSPs during the 10 Hz stimulation, which acts to enhance LTP. This suggested that
α5GABAARs limit LTP by damping membrane depolarization and the initial triggering events for LTP. Interestingly, it is likely that the sustained changes in synaptic potentiation following 10
Hz stimulation are related to a decrease in the input resistance of the neuron. Following 10 Hz stimulation there was a large drop in the neuronal membrane resistance and this effect correlated with a significant hyperpolarization in WT neurons. The drop in input resistance and hyperpolarizing effect in WT neurons was blocked by L-655,708, suggesting that α5GABAARs are necessary to limit the initiation of synaptic potentiation. Thus, it is possible that in vivo,
α5GABAARs regulate learning and synaptic potentiation by acting to strictly regulate the membrane properties that are necessary for hippocampus-dependent memory and LTP.
Finally, the work presented here has broader implications than simply suggesting a single receptor modifies memory and weak artificial stimulation in the hippocampus. The underlying theme for this thesis is that the activity of α5GABAARs predominates under specific conditions to regulate behavioural memory and synaptic plasticity. The benzodiazepine inverse agonists
221 222
selective for α5GABAARs were initially developed as a panacea for memory enhancement, but the role of α5GABAARs in learning and synaptic potentiation is more selective than originally thought. Enhanced learning and memory correlated with reduced α5GABAAR-activity for some stages and forms of learning, but not all. Since α5GABAARs are important for subtle learning, they may be useful targets for some but not all forms of learning enhancement.
7.4 Future Directions
Future experiments will investigate a detailed mechanism for the α5GABAAR regulation of synaptic plasticity. A direct extension of the current work should ask two main questions: 1)
How are α5GABAARs selectively recruited to inhibit LTP and 2) Do α5GABAARs act as a simple membrane regulator to modify LTP? Our data suggests that α5GABAARs are selectively activated to inhibit LTP. How does this happen? I have assumed that during 10 Hz stimulation
GABA is released from presynaptic terminals, such that the high affinity receptors remain activated for a prolonged period of time. This type of question could be answered by manipulating the extracellular concentration of GABA with the use of reuptake inhibitors. For example, the anticonvulsant gabapentin is known to promote the release of GABA and prevent
GABA reuptake (Honmou et al., 1995). If the effects of 10 Hz stimulation on α5GABAARs are dependent on GABA release and extracellular concentrations then gabapentin should accentuate the LTD effect of 10 Hz stimulation. Alternatively, 10 Hz stimulation may enhance the release of
GABA by non-vesicular means. For example, asynchronous release provides a form of long lasting inhibition in the brain (Hefft and Jonas, 2005). Prolonged inhibitory release of this type would promote the rapid desensitization of synaptic GABAARs, while the activity of slowly desensitizing extrasynaptic GABAARs may be enhanced.
222 223
Future studies should also determine whether these receptors act to simply regulate membrane properties (as demonstrated in this thesis) or whether their unique pattern of cellular distribution, the quantity of inhibitory charge or other, as-yet-unidentified factors underlie the exceptional regulatory properties of α5GABAARs in network plasticity. In this thesis I have shown that
α5GABAARs contribute to LTD by limiting membrane depolarization and decreasing the membrane resistance. However, I cannot state conclusively whether membrane depolarization or decreased resistance is the main contributor. It is possible that both account for the overall effects but additional experiments are clearly required to determine the relative contribution of
α5GABAARs to limiting depolarization and decreasing resistance, which results in LTD. An obvious experiment would be to alter the GABA reversal potential to a depolarized or hyperpolarized value. If the depolarization/hyperpolarization effect was the main factor contributing to the LTD then the WT neurons should change (LTP with a depolarized potential and further LTD with a hyperpolarized potential). However, if the response remains similar to that reported in Figure 6.9 then a shunting effect is more likely. Additionally, the holding potential of pyramidal neurons could easily be adjusted, with current injections during 10 Hz stimulation to determine whether the depolarizing effect in Gabra5–/– slices significantly contributes to the LTP effect. The prediction would be that depolarizing WT neurons during 10
Hz stimulation to -54 mV should mimic the effect observed in Gabra5–/– neurons. Conversely, hyperpolarizing Gabra5–/– neurons during 10 Hz stimulation to -60 mV should mimic the effect observed in WT neurons.
Additionally, α5GABAARs may interact with other receptors to mediate the LTD effect. For example, some NMDA receptors, much like the α5GABAARs, may occur extrasynaptically
(Kohr, 2006). In particular, extrasynaptic NMDA receptors are thought to preferentially contain
223 224 the NR2B subunit of the NMDA receptor (Tovar and Westbrook, 1999), and these receptors have been shown to participate in LTP (Bartlett et al., 2007). The close proximity of
α5GABAARs to specific subpopulations of NMDA receptors may, through shunting inhibition, limit the efficacy of NMDA receptors, which would in turn inhibit plasticity or learning. It may also be of interest to determine whether other forms of tonic inhibition, such as that generated by
δGABAARs (which are strongly expressed in the dentate gyrus), serve a similar role in regulating plasticity (Nusser and Mody, 2002; Stell et al., 2003; Lindquist and Birnir, 2006). Regardless, our results raise intriguing questions about the therapeutic potential of modifying α5GABAAR activity for the improvement of weakly formed memory under task-specific conditions.
224 225
List of Appendices
Appendix 1. α5GABAARs do not regulate the cognitive impairing properties of ethanol 226 Appendix 2. Inclusion of thesis-relevant work published or submitted by candidate 252 Appendix 3. Additional publications resulting from my doctoral studies 253
225 226
Appendix 1. α5GABAARs do not regulate the cognitive impairing properties of ethanol
Introduction
GABAARs are responsible for the majority of inhibitory neurotransmission in the mammalian central nervous system, which makes these receptors prime candidates for modulation by neurodepressive agents including ethanol (Yamashita et al., 2006; Glykys et al., 2007;
Santhakumar et al., 2007), anesthetics (Caraiscos et al., 2004a; Cheng et al., 2006) and benzodiazepines (Rudolph and Mohler, 2004). Typically, GABAARs consist of α, β, and γ subunits arranged in a 2:2:1 stoichiometry, although the γ subunit may be replaced by a δ subunit
(Mohler et al., 1996). Recent evidence indicates that GABAAR populations of distinct subunit compositions mediate the different behavioural effects of ethanol such as sedation and motor incoordination (Blednov et al., 2003; McKay et al., 2004; Hanchar et al., 2005). Compelling data have shown that GABAARs containing the α1 subunit play a key role in the motor and motivational effects mediated by ethanol (Blednov et al., 2003). Specifically, deletion of α1 subunit-containing GABAARs increased sensitivity to the locomotor stimulant effects of ethanol and decreased ethanol preference drinking (Blednov et al., 2003). Alternatively, the motor slowing or sedative effects of ethanol have been linked to α5GABAARs since injections of RY
023, a benzodiazepine inverse agonist selective for α5GABAARs reversed, ethanol-induced sedation during open field testing (Cook et al., 2005).
It is still unclear whether a subunit target for the memory-impairing properties of ethanol exists.
However, inhibiting α5GABAARs with a benzodiazepine inverse agonist selective for the α5
226 227 subunit (Dawson et al., 2006) improves word-list learning in human volunteers (Nutt et al.,
2007). The relatively restricted distribution of α5GABAARs to the hippocampus makes this specific receptor subunit a promising candidate for mediating the memory-blocking effects of ethanol. Specifically, the hippocampus, a brain structure important for learning and memory, is densely populated with a high proportion of α5GABAARs (Pirker et al., 2000). Furthermore, it has been shown that Gabra5–/– mice have an enhanced learning profile in the spatially demanding water maze task (Collinson et al., 2002), while mice with a point mutation in the
Gabra5 gene (α5H105R) show a 17% reduction in α5GABAAR expression accompanied by enhanced performance in trace fear learning (Crestani et al., 2002).
The distribution of different GABAAR isoforms at the subcellular level is also an important feature that may determine the functional outcome of the effects of ethanol on GABAAR activity.
For example, α5GABAARs are expressed at extrasynaptic sites (Serwanski et al., 2006), where they generate a persistent or tonic inhibitory conductance (Caraiscos et al., 2004b; Glykys et al.,
2008) but it still is not clear whether this tonic conductance is sensitive to the effects of ethanol.
A number of studies have indicated that low concentrations of ethanol potentiate the tonic conductance generated by δ-containing GABAARs (δGABAARs) in the dentate gyrus (Glykys et al., 2007; Jia et al., 2007; Santhakumar et al., 2007). The behavioural effects of ethanol on
δGABAARs have not been investigated but null mutant mice for α4 subunit-containing
GABAARs, which concurrently express a reduction of δGABAARs in the dentate gyrus, have normal behavioural responses to ethanol (Chandra et al., 2008). δGABAARs may not account for the ethanol-impairment of learning and memory because it is not clear whether δGABAARs in the dentate gyrus actively modify the learning and memory process (Wiltgen et al., 2005). Taken together this suggests that although ethanol potentiates a δGABAAR-mediated tonic
227 228 conductance, the memory impairing effects of ethanol are likely mediated through an alternative site.
In the current chapter ethanol concentrations that are “sobriety impairing” and lethal
(Poikolainen, 1984) were used to examine effects on the tonic conductance in hippocampal pyramidal neurons. I defined sobriety impairing ethanol concentrations as the average blood- alcohol concentration for drivers involved in fatal car crashes in Ontario (http://www.safety- council.org/info/traffic/impaired/stats.html), which is ~0.16% or 30 mM. Since α5GABAARs mediate a significant tonic GABAergic conductance in hippocampal pyramidal neurons and are involved in amnesia, I hypothesized that low sobriety impairing concentrations of ethanol enhance a tonic conductance in hippocampal pyramidal neurons and this enhancement contributes to the amnestic effects of ethanol. I assessed the performance of WT and Gabra5–/– mice in conditioned fear in the presence and absence of different doses of ethanol to determine whether α5GABAARs contribute to the memory-blocking properties of ethanol. I observed that sobriety impairing concentrations of ethanol did not potentiate a tonic conductance in CA1 pyramidal neurons. Additionally, Gabra5–/– mice were not resistant to the memory impairing properties of ethanol but rather were resistant to the sedating properties of low ethanol concentrations. These results provide evidence that α5GABAAR expression is not required for the amnestic effects of ethanol but may influence the sedative properties of this drug.
Questions:
1. Does ethanol potentiate an inhibitory tonic conductance in hippocampal pyramidal
neurons?
228 229
2. Does ethanol inhibit learning and memory through an α5GABAAR-dependent
mechanism?
3. Does ethanol impair spontaneous movement in an α5GABAAR-dependent manner?
Methods
Cell cultures and whole-cell recording
Primary cultures of hippocampal neurons were prepared as described previously (MacDonald et al., 1989) from Swiss White mice on embryonic day 18. The concentration-dependent effects of ethanol on tonic currents in cultured hippocampal neurons were determined with the whole-cell patch-clamp technique (at 20–23°C). Electrodes were made from borosilicate glass pipettes and fire polished just before use. Currents were recorded with a Multiclamp 700B amplifier and headstage (Molecular Devices, Union City, CA) and low-pass filtered at 10 kHz before digitization (Digidata 1200; Molecular Devices). Series resistance and pipette and whole-cell capacitance were compensated electronically. A hyperpolarizing voltage step of 10 mV was applied periodically throughout each experiment to monitor series resistance. Only cells that demonstrated stable series resistance (< 20% change) were used for data analysis. Cells were perfused with a solution containing the following (in mM): 140 NaCl, 2.0 KCl, 1.3 CaCl2, 25
HEPES, and 28 glucose, pH 7.4. Tetrodotoxin (0.3 µM) was added to the extracellular solution to inhibit spontaneous voltage-dependent sodium channel activity. In all experiments, potassium currents were suppressed by dialyzing the cell interior with a CsCl-based internal solution, pH
7.3, that contained the following (in mM): 120 CsCl, 2.0 MgCl2, 1.0 CaCl2, 11 EGTA, 30
HEPES, 2.0 MgATP, and 2.0 tetraethylammonium. The amplitude of the tonic current under control conditions was measured as the difference in the holding current before and during the
229 230 application of bicuculline (100 µM). Ethanol (30 mM to 2 M) or the vehicle control (ECF) at equivalent concentrations was added to the extracellular solution. Solutions were applied to the cell cultures using a multibarrel fast perfusion system.
Contextual and cued-fear conditioning
In the Pavlovian fear-conditioning tasks, mice were exposed to a tone and a foot shock (the unconditioned stimulus) in a novel conditioning context with presentation of the US occurring during the last 2 s of CS presentation (Fanselow, 1980). The dorsal hippocampus, which has a high expression of α5GABAARs (Sur et al., 1999), is known to be involved in contextual fear conditioning. In contrast, the expression of α5GABAARs is low in the amygdala (Sur et al.,
1999) and cued fear conditioning, which requires the basal lateral nucleus of the amygdala, served as a control (Sigurdsson et al., 2007). On day 1, mice were injected i.p. with 0.5, 1 or 1.5 g/kg ethanol or physiological saline, 10 min before placement in the fear conditioning chamber.
Single subjects were then allowed to explore the chamber for 180 s. Following this period, a
2800 Hz tone (auditory CS) from a frequency generator, amplified to 70 dB lasting 20 s was presented 3 times, at 60 s intervals. The last 2 s of each auditory CS was paired with a 0.7 mA electrical foot shock. On day 2, 24 h after the conditioning session, mice were returned to the chamber, and freezing, defined as the lack of any movement except that necessitated for respiration, was assessed every 8 s for 8 min (total of 60 observations). On day 3, 48 h after the conditioning session, mice were placed within the conditioning chamber, which was modified by covering the metal grid floor with ceramic tiles, covering the walls with black and white stripes, and wiping the ceramic tiles with a vanilla scented cloth. Mice were then monitored for baseline freezing (every 8 s) to the “novel” context for 180 s. After this baseline period, the auditory tone
230 231 was continuously presented for 300 s and the freezing response was recorded every 8 s.
Open field test
The sedative properties of ethanol were tested by measuring spontaneous activity in an open field testing chamber made of Plexiglas with the dimensions 42 X 42 X 30 cm. WT and Gabra5–/– were injected with ethanol (0.5, 1 or 1.5 g/kg) or saline and then returned to their home cage. All mice were tested 10 min post-ethanol injection for five consecutive minutes. Mice were monitored for the duration of time spent walking, rearing and grooming as an index of spontaneous locomotor activity in the open field. After placement in the open field, a trained examiner used an event recorder to score the total time spent walking, rearing or grooming. The floor and walls of the test chamber were cleaned with a mild ethanol solution between subjects
Statistical analysis
All data were analyzed using SPSS v 11.0 and GraphPad Prism v 4.0 (GraphPad Software, San
Diego California USA, www.graphpad.com) software. The electrophysiological data were analyzed using a one-way ANOVA. The fear conditioning data were subjected to a 2 X 4 X 8
(genotype X ethanol dose X time-bin) 3-way split-plot ANOVA unless otherwise specified. The open field data were analyzed using a 2 X 4 (genotype X ethanol dose) two-way ANOVA. Post- hoc analysis for any main effects and interactions consisted of Tukey’s-HSD for fear conditioning and Bonferroni’s post hoc-testing for open field testing and electrophysiology data with the significance level set at 0.05.
231 232
Results Ethanol does not potentiate a GABAergic tonic conductance in cultured hippocampal pyramidal neurons
Sobriety-impairing concentrations of ethanol (within the range of 30-100 mM) have been shown to enhance tonically active δGABAARs in various neuronal structures including the dentate gyrus (Wei et al., 2004), cerebellum (Yamashita et al., 2006) and the thalamus (Jia et al., 2007). I wanted to determine whether such low concentrations of ethanol enhance a tonic GABAergic conductance in CA1 pyramidal neurons. Applications of bicuculline (100 µM) revealed a large tonic conductance in pyramidal neurons from Swiss white mice (88.05 ± 12.45 pA, n = 7, P <
0.05 compared with control). Interestingly, the application of low concentrations of ethanol resulted in no significant change in the holding current (30 mM, 11.49 ± 4.91 pA, n = 5, and 100 mM, 0.60 ± 4.23 pA, n = 5; P > 0.05 compared with baseline). However, concentrations of ethanol that are considered not to be clinically relevant (> 100 mM) enhanced the tonic current in pyramidal neurons (300 mM, -18.09 ± 4.02 pA, n = 4, and 1 mM, -116.43 ± 23.80 pA, n = 2; P <
0.05 compared with baseline). The change in the holding current for bicuculline and the different concentrations of ethanol is summarized in Appendix Figure 1.1. Thus, I concluded that the tonic conductance in cultured hippocampal neurons, which is predominantly mediated by
α5GABAARs, is not potentiated by low, clinically relevant concentrations of ethanol.
Effect of ethanol on cued and contextual fear conditioning Next, to determine whether ethanol-impairment of memory performance was influenced by
α5GABAARs, Gabra5–/– mice were subjected to fear conditioning paradigms following the
232 233 administration of different ethanol doses. I hypothesized that Gabra5–/– mice would demonstrate resistance to the memory impairing properties of ethanol because it has previously been shown that α5GABAAR-selective benzodiazepine inverse agonists reverse ethanol- mediated memory impairment in humans (Nutt et al., 2007), as well as other ethanol-mediated behaviours in rats (McKay et al., 2004).
Day 1: Training Performance The assessment of freezing to the tone-shock pairings on the first day of testing was analyzed to examine 1) whether the genetic deletion of α5GABAARs or the administration of ethanol influenced the acquisition of the fear conditioning task (i.e. learning during conditioning) and 2) the development of conditioning during successive presentations of the CS-US pairings. The percentage of time spent freezing on successive presentations to the tone was analyzed using a 2
X 4 X 3 (genotype X dose X shock trials) split-plot ANOVA.
Baseline freezing performance did not differ between the genotypes [F(1,56) = 0.175, P > 0.05,
Appendix Figure 1.2] and was not influenced by the dose of ethanol injected [F(3,56) = 1.51, P >
0.05, Appendix Figures 1.2A and 1.2B]. Furthermore, the interaction between genotype and dose of ethanol did not reach statistical significance on the conditioning day [F(3,56) = 2.02, P >
0.05], (Appendix Figure 1.2C). The amount of freezing increased as a function of shock-trials in all groups [F(2,112) = 39.52, P < 0.001], (Appendix Figures 1.2A and 1.2B). There were no interactions between genotype and shock-trials [F(2,112) = 0.188, P > 0.05], the dose of ethanol and shock-trials [F(6,112) = 0.501, P > 0.05] or genotype, dose of ethanol and shock-trials
[F(6,112) = 0.826, P > 0.05], suggesting that genotype nor the dose of ethanol did not influence the acquisition of the fear conditioning task.
233 234
Day 2: Contextual Fear Conditioning
α5GABAARs are highly expressed within the hippocampus particularly within the dendritic layer of CA1 (Sur et al., 1999), which is an important brain region for the acquisition and expression of conditioned fear to contextual cues (Anagnostaras et al., 2001; Maren, 2001). Thus, I assessed contextual fear memory upon re-exposing the mice to the conditioning chamber 24 hrs following conditioning. All contextual freezing scores were combined into 1 min-bins in order to analyze the freezing data over the entire session. There was a general decrease in the level of contextual freezing over time in all groups, which plateaued around 4-5 min (Appendix Figures 1.3A and
1.3B). There was no genotypic difference in freezing scores between WT (Appendix Figure
1.3A) and Gabra5–/– (Appendix Figure 1.3B) [F(1,56) = 0.11, P > 0.05]. A 2 X 4 X 8 (genotype
X dose X 1 min-bins) split-plot ANOVA revealed a significant effect for time-bins [F(7,392) =
6.193, P < 0.001] and the dose of ethanol injected [F(3,56) = 45.487, P < 0.0001]. The moderate
(1 g/kg) and high (1.5 g/kg) but not the low (0.5 g/kg) dose of ethanol equally impaired contextual freezing in both genotypes, indicating that ethanol impairs contextual fear memory through a mechanism other than α5GABAARs (Appendix Figure 1.3C). The interaction between the time-bins and dose of ethanol injected was approaching statistical significance [F(21,392) =
1.58, P = 0.052], but no other main effects or interactions reached statistical significance [All F values < 1.52].
234 235
Appendix Figure 1.1. Low clinically-relevant concentrations of ethanol do not potentiate a tonic inhibitory conductance in CA1 pyramidal neurons.
(A) Current trace illustrates that low concentrations of ethanol do not potentiate a tonic conductance in cultured hippocampal pyramidal neurons. (B) The change in holding current is presented for each concentration of ethanol. Only high concentrations of ethanol potentiated the tonic current (> 300 mM) but these concentrations are not clinically relevant. The number of cells for each concentration is represented above each bar and data are presented as mean ± s.e.m.
235 236
236 237
Appendix Figure 1.2. Ethanol does not impair the acquisition of contextual fear conditioning.
There were no differences for the percentage of time spent freezing during baseline (3 min time bins) and freezing to the first, second and third foot shock (1 min time bins) for the dose of ethanol injected for WT (A) and Gabra5–/– mice (B). (C) Post-shock data were further pooled and are represented as a bar chart for a direct comparison between the genotypes at the different doses of ethanol. Data are represented as mean ± s.e.m.
237 238
238 239
Appendix Figure 1.3. Gabra5–/– mice are not resistant to the memory impairing properties of ethanol in contextual fear conditioning.
The freezing scores of WT and Gabra5–/– mice are shown for each min (1–8) during the contextual monitoring session. (A) The high dose of ethanol (1.5 g/kg) impaired the overall freezing performance of WT mice for the entire monitoring session, when compared to saline injected controls. The moderate dose of ethanol (1 g/kg) impaired performance of WT mice as the monitoring session progressed, with high scores during the first min (~ 80%) and low scores during the last min (~35%; P < 0.05). (B) Similar trends were observed for Gabra5–/– mice, with the moderate and high doses impairing performance. (C) Contextual freezing data were pooled and are represented as a bar chart for a direct comparison between the genotypes at the different doses of ethanol. Ethanol equally impaired freezing responses at the moderate and high doses suggesting that ethanol impairs memory independent of α5GABAARs. * denotes significantly different from saline-injected control mice. Data are represented as mean ± s.e.m.
239 240
240 241
Day 3: Auditory-Cued Fear Conditioning I also assessed the conditioned fear response to an auditory-cued fear stimulus. This task is highly-dependent upon the amygdala (Maren, 2001) and I wanted to determine the effects of ethanol on a brain region with a low expression of α5GABAARs (Sur et al., 1999). All fear conditioning data were combined into 1 min bins, in order to assess freezing over the entire duration of testing (ie. baseline freezing and tone freezing) (Appendix Figure 1.4). During the baseline assessment of the freezing response to the novel context, there were no differences between WT (Appendix Figure 1.4A) and Gabra5–/– mice (Appendix Figure 1.4B) [F(1,56) =
2.48, P = 0.121] and dose of ethanol injected [F(3,56) = 0.826, P > 0.05]. Freezing was greatest in all groups during the 3 min time bin coinciding with the initiation of the tone CS [F(7,392) =
164.207, P = 0.0001].
There were no genotypic differences during the tone presentation [F(1,56) = 0.095, P > 0.05] but there was a significant effect for the dose of ethanol injected [F(3,56) = 43.084, P < 0.0001] with
1.5 g/kg ethanol reducing the overall expression of cued fear in both genotypes similarly, compared with all other groups (P < 0.001) (Appendix Figure 1.4C). In ethanol-treated groups there was a significant decrease in the overall freezing during the 5 minute exposure to the tone
[F(4,224) = 9.950, P < 0.0001]. There also was a time bin X dose of ethanol interaction
[F(12,224) = 2.245, P < 0.01], which was the result of a sharper decline in the freezing response over time in the mice injected with 1.5 g/kg of ethanol (P < 0.001). No additional interactions reached statistical significance [All F’s < 0.953]. The average cued freezing scores over the 5 min of testing with the tone present are displayed in Appendix Figure 1.4C.
241 242
Effects of ethanol on Gabra5–/– mice in the open field
The contribution of α5GABAARs to spontaneous locomotion remains uncertain. A previous group has reported that a high dose of the benzodiazepine-inverse agonist RY 024 partially reverses ethanol-induced sedation and ambulation in the open field but the effects and selectivity of inverse agonists remain unclear (McKay et al., 2004). In vitro results suggest that RY 024 has
~ 67-fold greater selectivity for α5GABAARs, but the specificity of RY 024 in vivo is not known.
Therefore, I chose to investigate whether the sedative or locomotor-impairing properties of ethanol are mediated through α5GABAARs by using Gabra5–/– mice.
Ethanol dose-dependently decreased overall spontaneous locomotion and activity in both genotypes (Appendix Figures 1.5A and 1.5B) [F(3,56) = 61.23, P < 0.0001]. Further analysis revealed an interaction between the genotype and the dose of ethanol injected [F(3,56) = 2.966, P
= 0.0396], indicating that the two genotypes responded differently to the different ethanol doses.
Interestingly, WT mice displayed a greater sensitivity to ethanol at 0.5 g/kg (P < 0.05) and 1 g/kg
(P < 0.01) when total walk time was analyzed compared to Gabra5–/– mice. There were no genotypic differences at the 1.5 g/kg (P > 0.05) dose of ethanol (Appendix Figure 1.5A).
However, ethanol reduced grooming [F(3,56) = 26.79, P = 0.0004] and rearing [F(3,56) = 20.14,
P = 0.003] similarly in both genotypes in a dose dependent manner (Appendix Figures 1.5B and
1.5C). There were no genotypic differences or an interaction between the genotype and dose of ethanol injected for the rearing and grooming data [All F’s < 1.61].
242 243
Appendix Figure 1.4. Gabra5–/– mice are not resistant to the memory impairing properties of ethanol in cued fear conditioning.
The freezing scores of WT and Gabra5–/– mice are shown for each min (1–8) during the monitoring session. Baseline freezing to the modified context was monitored for minutes 1–3 and the tone was subsequently presented continuously for minutes 4-8 of the fear conditioning protocol. (A) In WT mice, low baseline freezing scores were observed during the first 3 min. A dramatic increase in the freezing scores coincided with the onset of the auditory tone. The high dose of ethanol (1.5 g/kg) impaired the overall freezing performance of WT mice when the tone was presented. (B) A similar trend was observed for Gabra5–/– mice, with the high dose of ethanol impairing performance. (C) Cued freezing data were pooled and are represented as a bar chart for a direct comparison between the genotypes at the different doses of ethanol. Ethanol equally impaired freezing responses at the high doses suggesting that ethanol impairs memory independent of α5GABAARs. * denotes significantly different from saline-injected control mice.
Data are represented as mean ± s.e.m.
243 244
244 245
Appendix Figure 1.5. Gabra5–/– mice are resistant to the sedative properties of ethanol when injected with a low and moderate dose.
(A) The spontaneous walking of WT mice was significantly impaired when they were injected with ethanol at a dose of 0.5 and 1 g/kg in the open field. These same doses had no significant effect on Gabra5–/– mice. Ethanol injected at the 1.5 g/kg dose significantly impaired spontaneous walking in both genotypes. (B) There was a general trend for all ethanol doses to decrease the time spent grooming in the open field and Gabra5–/– were not resistant to this trend. (C) All doses of ethanol decreased the total time spent rearing compared to saline injected mice. Gabra5–/– mice were not resistant to this impairment. * denotes significantly different from saline-injected control mice. Data are represented as mean ± s.e.m.
245 246
246 247
Discussion
Electrophysiological and behavioural strategies were used to investigate whether ethanol enhances a tonic conductance in hippocampal pyramidal neurons, and whether such an effect could account for ethanol’s memory-impairing properties. At sobriety impairing concentrations, ethanol did not significantly enhance the tonic conductance in cultured hippocampal pyramidal neurons. In the behavioural tasks, WT and Gabra5–/– mice were equally impaired in contextual fear conditioning by ethanol at doses of 1 g/kg and 1.5 g/kg. Similarly for cued fear conditioning,
WT and Gabra5–/– mice injected with 1.5 g/kg of ethanol were equally impaired. Interestingly, the total time spent walking in the open field was reduced in WT mice injected with 0.5 g/kg and
1 g/kg of ethanol whereas Gabra5–/– mice were unaffected. Taken together our results indicate that ethanol impairs fear memory through mechanisms that are independent of α5GABAARs, but may mediate its sedative properties at low doses via an α5GABAAR-dependent mechanism.
Previous in vitro experiments have shown that δGABAARs are sensitive to low concentrations of ethanol (Santhakumar et al., 2007). Application of ethanol within the 10-30 mM range enhanced the δGABAAR tonic conductance in brain areas, including the cerebellum (Hanchar et al., 2005), dentate gyrus (Wei et al., 2004) and thalamus (Jia et al., 2007), where the δ subunit mediates a large tonic conductance. α5GABAARs mediate a large proportion of the tonic conductance in
CA1 and CA3 pyramidal neurons (Caraiscos et al., 2004b) with a smaller residual current that is attributed to δGABAARs in these neurons (Glykys and Mody, 2006). It is thus surprising that I found low concentrations of ethanol did not enhance the tonic current and even inhibited it, as seen with the 30 mM ethanol concentration. Given that ethanol did not potentiate the tonic
247 248
current in CA1 pyramidal neurons, it appears that α5GABAARs do not readily contribute to an ethanol-potentiated tonic current in CA1 hippocampal pyramidal neurons.
In the behavioural studies, I found that acute injections of ethanol impaired contextual fear memory at moderate (1 g/kg) and high (1.5 g/kg), but not low (0.5 g/kg) ethanol doses. However, the impaired fear memory was not the result of an impaired learning process during training, as all mice had comparable freezing to shock scores on the training day (Appendix Figures 1.2A and 1.2B). The impairment of hippocampus-dependent contextual fear memory by ethanol was not dependent on the presence of α5GABAARs since Gabra5−/− mice were not resistant to the performance impairing effects of ethanol on memory (Appendix Figure 1.3). This was surprising because it has recently been shown that an inverse agonist selective for α5GABAARs (α5IA) attenuated the adverse effect of ethanol on a hippocampus-dependent memory task in human subjects (Nutt et al., 2007). The discrepancy between these two studies could simply be related to different actions of ethanol in human and rodent models, or the result from a non-specific action of the benzodiazepine inverse agonist. α5IA becomes increasingly non-selective for
α5GABAARs when greater concentrations are used (Dawson et al., 2006) Because the concentration at which α5IA is selective for α5GABAARs is unknown in humans, α5IA may prevent ethanol-induced memory impairment in humans through non-selective actions. Ethanol has several different effects on memory: it robustly inhibits spatial memory (hidden platform water maze) (Berry and Matthews, 2004) but may spare (Melia et al., 1996; White et al., 1998;
Berry and Matthews, 2004) and even facilitate (Matthews et al., 1999) non-spatial memory
(visible platform water maze). In our study, ethanol impaired non-spatial fear learning without increasing α5GABAAR activity, but it remains uncertain whether α5GABAARs contribute to the impairment of spatial memory by ethanol.
248 249
My results, along with those of others (McKay et al., 2004) suggest that α5GABAARs are important for ethanol-induced motor-impairment. RY 024, a selective α5GABAAR inverse agonist, reverses the motor impairing and sedative effects of ethanol in the absence of intrinsic effects (McKay et al., 2004). The decreased freezing scores observed across the different doses of ethanol cannot be solely attributed to the motor-impairing properties of ethanol because low doses that did not impair freezing during fear conditioning (Appendix Figures 1.3 and 1.4), impaired exploration time in the open field (Appendix Figure 1.5). This is in contrast to evidence that the sedative effects of the general anesthetic etomidate are not mediated through
α5GABAARs (Cheng et al., 2006). We have previously shown that while the general anesthetic etomidate enhances a α5GABAAR tonic conductance and impairs hippocampus-dependent memory and LTP through an α5GABAAR-dependent mechanism, the α5 subunit is not sensitive to the sedative or hypnotic properties of etomidate (Cheng et al., 2006). These results are in direct contrast to our current findings, as ethanol does not enhance a CA1 tonic conductance or impair hippocampus-dependent memory in an α5 subunit-dependent manner. This is striking because the sedative actions of different GABAergic compounds are thought to require
α1GABAARs (Rudolph et al., 1999).
It has been shown that mice can develop tolerance to the motor-depressant actions of diazepam, and this depends on the chronic activation of two competitive mechanisms orchestrated by α1 and α5GABAARs (van Rijnsoever et al., 2004). It appears as though the phasic contribution of
α1GABAARs in the forebrain areas mediating motor control alters the weight of tonic inhibition mediated by α5GABAARs in the hippocampus. The change in the α1GABAARs phasic transmission is reflected in diazepam tolerant mice by a 15% diminution in binding labelling for
249 250
α5GABAARs in the dentate gyrus. This in turn renders the mice insensitive to the motor- depressant effects of diazepam and suggests that α5GABAARs can interact with another
GABAAR subunit to alter motor behaviours. Additionally, Savic and colleagues (Savic et al.,
2008) have shown that newly developed benzodiazepine site agonists, devoid of activity at
α1GABAARs but with partial or full selectivity at α5GABAARs contributes to the sedative profiles of these drugs. Furthermore, there is a relatively high expression of α5GABAARs in the ventral horn of the spinal cord (Ruano et al., 2000), which could contribute to the motor- impairing effects of ethanol. This seems more likely rather than an interaction with another
GABAAR subunit such as the α1 subunit because our ethanol treatment was acute and the effects of ethanol seemed to impair mobility as opposed to sedation (Appendix Figures 1.5B and 1.5C).
Finally, our data suggests that the role of α5GABAARs in ethanol pharmacology appears to be strongest for ethanol-induced sedation and not for non-spatial hippocampus-dependent memory, however the possibility exists that α5GABAARs are simply mediating an anxiolytic effect
(Navarro et al., 2002).
Despite the role of α5GABAARs in ethanol-induced sedation, these receptors have emerged as targets for the rewarding-reinforcing properties of ethanol (June et al., 2003; Cook et al., 2005).
The benzodiazepine-like inverse agonist RY 023 (RY) exhibits both a high affinity (Ki of ~ 2.7 nM) and high selectivity (~ 75-fold) at recombinant GABAARs composed of α5β2γ2 subunits
(Liu et al., 1996) and suppresses ethanol-maintained responding in a lever-pressing task (June et al., 2001) and ethanol-induced motor-impairment (Cook et al., 2005). Further, it has been confirmed that the benzodiazepine site mediates the ethanol effect as the non-specific competitive β-carboline benzodiazepine antagonist ZK 93426 reversed the RY-induced suppression of ethanol-maintained responding (Cook et al., 2005). However, studies using
250 251
Gabra5–/– mice indicate that α5GABAARs are not essential for the rewarding properties mediated by ethanol as Gabra5–/– mice did not differ from WT littermates in operant responding for 10 % ethanol containing solutions (Stephens et al., 2005). Still, in the same study, the inverse agonist α5IA-II decreased ethanol-maintained responding in WT mice at 27-fold lower concentrations than Gabra5–/– mice. Taken together these results suggest that benzodiazepine inverse agonists selective for α5GABAA receptors are instrumental for suppressing ethanol- mediated responding, although the role of these same receptors in signalling ethanol reward is less convincing.
Initially, the effects of ethanol on cognition and behaviour were believed to result from a non- specific global depression of neuronal activity (White et al., 2000). It was thought that the effects of ethanol were mediated by a non-specific perturbation of the neuronal lipid bilayer. Substantial evidence now indicates that ethanol, like general anesthetics, does not cause a non-specific neuronal depression but interacts with membrane bound proteins including ligand-gated ion channels (Crews et al., 1996). Ethanol is considered to exert its neurodepressive effects primarily through the GABAergic system (Boehm et al., 2004), although it also acts via ligand-gated
NMDA (Ron, 2004) and glycine receptors (Mihic, 1999). Given the high distribution of
α5GABAARs in the hippocampus (Pirker et al., 2000), their role in cognition (Collinson et al.,
2002; Collinson et al., 2006) and anesthetic-induced memory impairment (Cheng et al., 2006), these receptors can be viewed as prime targets for the cognitive disrupting effects of ethanol. It is therefore surprising that the results of this study suggest that α5GABAARs are not important for the cognitive-impairing properties of ethanol.
251 252
Appendix 2. Inclusion of thesis-relevant work published or submitted by candidate
1. Cheng, VY*., Martin, LJ*., Elliot, EM., Kim, J., Mount, HT., Taverna, FA., Roder, JC.,
MacDonald, JF., Bhambri, A., Collinson, N., Wafford, KA., Orser, BA. 2006. α5GABAA receptors mediate the amnestic but not sedative-hypnotic effects of the general anesthetic, etomidate. J Neurosci 26(14):3713-3720. *co-first author [Division of labour: LJM contributed
70% of experiments and analyses, and 40% of the writing].
2. Martin, LJ., Jackson, MF., Ju. W., Macdonald, JF., Oh, GHT., Roder, JC., Orser, BA. 2009.
α5 subunit-containing γ-aminobutyric acid type A receptors regulate synaptic plasticity and anesthetic memory blockade. J Neurosci (In revision) [Division of labour: LJM contributed
100% of the experiments and analyses, and 75% of the writing].
3. Martin, LJ., Oh, GHT., Orser, BA. 2009. Etomidate targets α5 γ-aminobutyric acid subtype
A receptors to regulate synaptic plasticity and memory blockade Anesthesiology (In Press)
[Division of labour: LJM contributed 100% of experiments and analyses, and 90% of the writing].
4. Martin, LJ., Bonin, RP., Kim, J., Mount, HTJ., Orser, BA. 2009. The cognitive impairing effects of ethanol are not mediated through tonically conducting α5GABAA receptors. Eur
Neuropsychopharm. (submitted) [Division of labour: LJM contributed 90% of the experiments and analyses, and 90% of the writing].
252 253
Appendix 3. Additional publications resulting from my doctoral studies
1. Bonin, RP., Martin, LJ., MacDonald, JF., Orser. BA. 2007. α5GABAA receptors regulate the intrinsic excitability of mouse hippocampal pyramidal neurons. J Neurophys 98(4):2244-54.
[Division of labour: LJM contributed 10% of the experiments]
2. Sun, HS., Jackson, MF., Martin, LJ., Jansen, K., Teves, L., Cui, H., Mori, Y., Jones, M.,
Forder, JP., Golde, TE., Orser, BA., MacDonald, JF., Tymianski, M. 2009. Suppression of hippocampal TRPM7 enhances neuronal survival, function and learning following cerebral ischemia. Nat Neurosci (Accepted) [Division of labour: LJM contributed 20% of experiments and analyses, and 20% of writing]
3. Vargas-Caballero, M., Martin, LJ., Orser, BA., Paulsen, O. Feedforward inhibition limiting an NMDAR-mediated response is reduced by blocking α5GABAA receptors. 2009. J
Neurophys (submitted) [Division of labour: LJM contributed 30% of experiments and analyses, and 20% of writing].
4. Saab, BJ., MacLean, AJB., Kanisek, M., Martin, LJ., Zurek, A., Roder, JC., Orser, BA. Pre- treatment with the α5GABAA receptor inverse agonist, L-655,708 prevents persistent memory impairment following isoflurane. 2009. Anesthesiology (submitted) [Division of labour: LJM contributed 20% of experiments and 20% of writing]
253 254
References
Abraham WC (2008) Metaplasticity: tuning synapses and networks for plasticity. Nat Rev Neurosci.
Abraham WC, Bear MF (1996) Metaplasticity: the plasticity of synaptic plasticity. Trends Neurosci 19:126-130.
Abraham WC, Mason-Parker SE, Bear MF, Webb S, Tate WP (2001) Heterosynaptic metaplasticity in the hippocampus in vivo: a BCM-like modifiable threshold for LTP. Proc Natl Acad Sci U S A 98:10924-10929.
Adams B, Fitch T, Chaney S, Gerlai R (2002) Altered performance characteristics in cognitive tasks: comparison of the albino ICR and CD1 mouse strains. Behav Brain Res 133:351- 361.
Akbarian S, Huntsman MM, Kim JJ, Tafazzoli A, Potkin SG, Bunney WE, Jr., Jones EG (1995) GABAA receptor subunit gene expression in human prefrontal cortex: comparison of schizophrenics and controls. Cereb Cortex 5:550-560.
Alakuijala A, Palgi M, Wegelius K, Schmidt M, Enz R, Paulin L, Saarma M, Pasternack M (2005) GABA receptor ρ subunit expression in the developing rat brain. Brain Res Dev Brain Res 154:15-23.
Alford S, Frenguelli BG, Schofield JG, Collingridge GL (1993) Characterization of Ca2+ signals induced in hippocampal CA1 neurones by the synaptic activation of NMDA receptors. J Physiol 469:693-716.
Alger BE, Nicoll RA (1979) GABA-mediated biphasic inhibitory responses in hippocampus. Nature 281:315-317.
Ali AB, Thomson AM (2008) Synaptic α5 subunit-containing GABAA receptors mediate IPSPs elicited by dendrite-preferring cells in rat neocortex. Cereb Cortex 18:1260-1271.
Amzica F, Steriade M (1999) Spontaneous and artificial activation of neocortical seizures. J Neurophysiol 82:3123-3138.
Anagnostaras SG, Gale GD, Fanselow MS (2001) Hippocampus and contextual fear conditioning: recent controversies and advances. Hippocampus 11:8-17.
Andersen P (2007) The hippocampus book. New York: Oxford University Press.
Andersen P, Sundberg SH, Sveen O, Wigstrom H (1977) Specific long-lasting potentiation of synaptic transmission in hippocampal slices. Nature 266:736-737.
Antognini J, Carstens E, Raines DE (2003) Neural mechanisms of anesthesia. Totowa, N.J.: Humana.
254 255
Asai Y, Takano A, Ito H, Okubo Y, Matsuura M, Otsuka A, Takahashi H, Ando T, Ito S, Arakawa R, Asai K, Suhara T (2008) GABAA/Benzodiazepine receptor binding in patients with schizophrenia using [11C]Ro15-4513, a radioligand with relatively high affinity for α5 subunit. Schizophr Res 99:333-340.
Atack JR, Bayley PJ, Fletcher SR, McKernan RM, Wafford KA, Dawson GR (2006a) The proconvulsant effects of the GABAA α5 subtype-selective compound RY-080 may not be α5-mediated. Eur J Pharmacol 548:77-82.
Atack JR, Bayley PJ, Seabrook GR, Wafford KA, McKernan RM, Dawson GR (2006b) L- 655,708 enhances cognition in rats but is not proconvulsant at a dose selective for α5- containing GABAA receptors. Neuropharmacology 51:1023-1029.
Atack JR, Pike A, Clarke A, Cook SM, Sohal B, McKernan RM, Dawson GR (2006c) Rat pharmacokinetics and pharmacodynamics of a sustained release formulation of the GABAA α5-selective compound L-655,708. Drug Metab Dispos 34:887-893.
Attwell D, Barbour B, Szatkowski M (1993) Nonvesicular release of neurotransmitter. Neuron 11:401-407.
Bai D, Pennefather PS, MacDonald JF, Orser BA (1999) The general anesthetic propofol slows deactivation and desensitization of GABAA receptors. J Neurosci 19:10635-10646.
Bai D, Zhu G, Pennefather P, Jackson MF, MacDonald JF, Orser BA (2001) Distinct functional and pharmacological properties of tonic and quantal inhibitory postsynaptic currents mediated by γ-aminobutyric acidA receptors in hippocampal neurons. Mol Pharmacol 59:814-824.
Baimoukhametova DV, Hewitt SA, Sank CA, Bains JS (2004) Dopamine modulates use- dependent plasticity of inhibitory synapses. J Neurosci 24:5162-5171.
Ballard TM, Knoflach F, Prinssen E, Borroni E, Vivian JA, Basile J, Gasser R, Moreau JL, Wettstein JG, Buettelmann B, Knust H, Thomas AW, Trube G, Hernandez MC (2009) RO4938581, a novel cognitive enhancer acting at GABAA α5 subunit-containing receptors. Psychopharmacology (Berl) 202:207-223.
Barnes CA (1979) Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. J Comp Physiol Psychol 93:74-104.
Bartlett TE, Bannister NJ, Collett VJ, Dargan SL, Massey PV, Bortolotto ZA, Fitzjohn SM, Bashir ZI, Collingridge GL, Lodge D (2007) Differential roles of NR2A and NR2B- containing NMDA receptors in LTP and LTD in the CA1 region of two-week old rat hippocampus. Neuropharmacology 52:60-70.
Bartos M, Vida I, Frotscher M, Geiger JR, Jonas P (2001) Rapid signaling at inhibitory synapses in a dentate gyrus interneuron network. J Neurosci 21:2687-2698.
Bashir ZI, Tam B, Collingridge GL (1990) Activation of the glycine site in the NMDA receptor is necessary for the induction of LTP. Neurosci Lett 108:261-266.
255 256
Baumann SW, Baur R, Sigel E (2003) Individual properties of the two functional agonist sites in GABAA receptors. J Neurosci 23:11158-11166.
Bear MF (1995) Mechanism for a sliding synaptic modification threshold. Neuron 15:1-4.
Bear MF (2003) Bidirectional synaptic plasticity: from theory to reality. Philos Trans R Soc Lond B Biol Sci 358:649-655.
Bear MF, Kirkwood A (1993) Neocortical long-term potentiation. Curr Opin Neurobiol 3:197- 202.
Bear MF, Malenka RC (1994) Synaptic plasticity: LTP and LTD. Curr Opin Neurobiol 4:389- 399.
Beattie EC, Carroll RC, Yu X, Morishita W, Yasuda H, von Zastrow M, Malenka RC (2000) Regulation of AMPA receptor endocytosis by a signaling mechanism shared with LTD. Nat Neurosci 3:1291-1300.
Belelli D, Lambert JJ, Peters JA, Wafford K, Whiting PJ (1997) The interaction of the general anesthetic etomidate with the γ-aminobutyric acid type A receptor is influenced by a single amino acid. Proc Natl Acad Sci U S A 94:11031-11036.
Belelli D, Peden DR, Rosahl TW, Wafford KA, Lambert JJ (2005) Extrasynaptic GABAA receptors of thalamocortical neurons: a molecular target for hypnotics. J Neurosci 25:11513-11520.
Ben-Ari Y (2006) Seizures beget seizures: the quest for GABA as a key player. Crit Rev Neurobiol 18:135-144.
Ben-Ari Y, Gaiarsa JL, Tyzio R, Khazipov R (2007) GABA: a pioneer transmitter that excites immature neurons and generates primitive oscillations. Physiol Rev 87:1215-1284.
Benkwitz C, Liao M, Laster MJ, Sonner JM, Eger EI, 2nd, Pearce RA (2007) Determination of the EC50 amnesic concentration of etomidate and its diffusion profile in brain tissue: implications for in vitro studies. Anesthesiology 106:114-123.
Bernston G, Sarter M, Ruland S, Hart S, Ronis V (1996) Benzodiazepine receptor agonists and inverse agonists yield concordant rather than opposing effects on startle responses. Journal of Psychopharmacology 10:309-312.
Berry RB, Matthews DB (2004) Acute ethanol administration selectively impairs spatial memory in C57BL/6J mice. Alcohol 32:9-18.
Bianchi SL, Tran T, Liu C, Lin S, Li Y, Keller JM, Eckenhoff RG, Eckenhoff MF (2008) Brain and behavior changes in 12-month-old Tg2576 and nontransgenic mice exposed to anesthetics. Neurobiol Aging 29:1002-1010.
Bienenstock EL, Cooper LN, Munro PW (1982) Theory for the development of neuron selectivity: orientation specificity and binocular interaction in visual cortex. J Neurosci 2:32-48.
256 257
Birnir B, Korpi ER (2007) The impact of sub-cellular location and intracellular neuronal proteins on properties of GABAA receptors. Curr Pharm Des 13:3169-3177.
Bito H, Deisseroth K, Tsien RW (1996) CREB phosphorylation and dephosphorylation: a Ca2+- and stimulus duration-dependent switch for hippocampal gene expression. Cell 87:1203- 1214.
Blatt GJ, Fitzgerald CM, Guptill JT, Booker AB, Kemper TL, Bauman ML (2001) Density and distribution of hippocampal neurotransmitter receptors in autism: an autoradiographic study. J Autism Dev Disord 31:537-543.
Blednov YA, Walker D, Alva H, Creech K, Findlay G, Harris RA (2003) GABAA receptor α1 and β2 subunit null mutant mice: behavioral responses to ethanol. J Pharmacol Exp Ther 305:854-863.
Bliss TV, Lomo T (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J Physiol 232:331- 356.
Bliss TV, Collingridge GL (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361:31-39.
Bliss TV, Gardner-Medwin AR, Lomo T (1973) Synaptic plasticity in the hippocampal formation. In: Macromolecules and behavior (Ansell GB, Bradley PB, eds). Baltimore: University Park Press.
Bliss TVP, Collingridge GL, Morris RGM (2004) Long-term potentiation : enhancing neuroscience for 30 years. Oxford: Oxford University Press.
Blusse van Oud-Alblas HJ, van Dijk M, Liu C, Tibboel D, Klein J, Weber F (2009) Intraoperative awareness during paediatric anaesthesia. Br J Anaesth 102:104-110.
Boehm SL, 2nd, Ponomarev I, Jennings AW, Whiting PJ, Rosahl TW, Garrett EM, Blednov YA, Harris RA (2004) γ-Aminobutyric acid A receptor subunit mutant mice: new perspectives on alcohol actions. Biochem Pharmacol 68:1581-1602.
Bonin RP, Orser BA (2008) GABAA receptor subtypes underlying general anesthesia. Pharmacol Biochem Behav 90:105-112.
Bonin RP, Martin LJ, MacDonald JF, Orser BA (2007) α5GABAA receptors regulate the intrinsic excitability of mouse hippocampal pyramidal neurons. J Neurophysiol 98:2244- 2254.
Borden LA, Smith KE, Vaysse PJ, Gustafson EL, Weinshank RL, Branchek TA (1995) Re- evaluation of GABA transport in neuronal and glial cell cultures: correlation of pharmacology and mRNA localization. Receptors Channels 3:129-146.
257 258
Braff DL, Geyer MA, Swerdlow NR (2001) Human studies of prepulse inhibition of startle: normal subjects, patient groups, and pharmacological studies. Psychopharmacology (Berl) 156:234-258.
Brager DH, Johnston D (2007) Plasticity of intrinsic excitability during long-term depression is mediated through mGluR-dependent changes in Ih in hippocampal CA1 pyramidal neurons. J Neurosci 27:13926-13937.
Brickley SG, Cull-Candy SG, Farrant M (1996) Development of a tonic form of synaptic inhibition in rat cerebellar granule cells resulting from persistent activation of GABAA receptors. J Physiol 497 ( Pt 3):753-759.
Brickley SG, Revilla V, Cull-Candy SG, Wisden W, Farrant M (2001) Adaptive regulation of neuronal excitability by a voltage-independent potassium conductance. Nature 409:88-92.
Bright DP, Aller MI, Brickley SG (2007) Synaptic release generates a tonic GABAA receptor- mediated conductance that modulates burst precision in thalamic relay neurons. J Neurosci 27:2560-2569.
Brioni JD, McGaugh JL (1988) Post-training administration of GABAergic antagonists enhances retention of aversively motivated tasks. Psychopharmacology (Berl) 96:505-510.
Brunig I, Scotti E, Sidler C, Fritschy JM (2002) Intact sorting, targeting, and clustering of γ- aminobutyric acid A receptor subtypes in hippocampal neurons in vitro. J Comp Neurol 443:43-55.
Bu DF, Tobin AJ (1994) The exon-intron organization of the genes (GAD1 and GAD2) encoding two human glutamate decarboxylases (GAD67 and GAD65) suggests that they derive from a common ancestral GAD. Genomics 21:222-228.
Bu DF, Erlander MG, Hitz BC, Tillakaratne NJ, Kaufman DL, Wagner-McPherson CB, Evans GA, Tobin AJ (1992) Two human glutamate decarboxylases, 65-kDa GAD and 67-kDa GAD, are each encoded by a single gene. Proc Natl Acad Sci U S A 89:2115-2119.
Burgard EC, Tietz EI, Neelands TR, Macdonald RL (1996) Properties of recombinant γ- aminobutyric acid A receptor isoforms containing the α5 subunit subtype. Mol Pharmacol 50:119-127.
Buzsaki G (2005) Theta rhythm of navigation: link between path integration and landmark navigation, episodic and semantic memory. Hippocampus 15:827-840.
Campagna JA, Miller KW, Forman SA (2003) Mechanisms of actions of inhaled anesthetics. N Engl J Med 348:2110-2124.
Caraiscos VB, Newell JG, You-Ten KE, Elliott EM, Rosahl TW, Wafford KA, MacDonald JF, Orser BA (2004a) Selective Enhancement of Tonic GABAergic Inhibition in Murine Hippocampal Neurons by Low Concentrations of the Volatile Anesthetic Isoflurane. J Neurosci 24:8454-8458.
258 259
Caraiscos VB, Elliott EM, You-Ten KE, Cheng VY, Belelli D, Newell JG, Jackson MF, Lambert JJ, Rosahl TW, Wafford KA, MacDonald JF, Orser BA (2004b) Tonic inhibition in mouse hippocampal CA1 pyramidal neurons is mediated by α5 subunit-containing γ- aminobutyric acid type A receptors. Proc Natl Acad Sci U S A 101:3662-3667.
Caron MJ, Mottron L, Rainville C, Chouinard S (2004) Do high functioning persons with autism present superior spatial abilities? Neuropsychologia 42:467-481.
Carpenter MB, Sutin J (1983) Human neuroanatomy, 8th Edition. Baltimore: Williams & Wilkins.
Castellani GC, Quinlan EM, Cooper LN, Shouval HZ (2001) A biophysical model of bidirectional synaptic plasticity: dependence on AMPA and NMDA receptors. Proc Natl Acad Sci U S A 98:12772-12777.
Casula MA, Bromidge FA, Pillai GV, Wingrove PB, Martin K, Maubach K, Seabrook GR, Whiting PJ, Hadingham KL (2001) Identification of amino acid residues responsible for the β1 α5 subunit binding selectivity of L-655,708, a benzodiazepine binding site ligand at the GABAA receptor. Journal of Neurochemistry 77:445-451.
Cavus I, Teyler TJ (1998) NMDA receptor-independent LTP in basal versus apical dendrites of CA1 pyramidal cells in rat hippocampal slice. Hippocampus 8:373-379.
Chadderton P, Margrie TW, Hausser M (2004) Integration of quanta in cerebellar granule cells during sensory processing. Nature 428:856-860.
Chambers MS, Atack JR, Broughton HB, Collinson N, Cook S, Dawson GR, Hobbs SC, Marshall G, Maubach KA, Pillai GV, Reeve AJ, MacLeod AM (2003) Identification of a Novel, Selective GABAA α5 Receptor Inverse Agonist Which Enhances Cognition. J Med Chem 46:2227-2240.
Chambers MS, Atack JR, Carling RW, Collinson N, Cook SM, Dawson GR, Ferris P, Hobbs SC, O'Connor D, Marshall G, Rycroft W, MacLeod AM (2004) An Orally Bioavailable, Functionally Selective Inverse Agonist at the Benzodiazepine Site of GABAA α5 Receptors with Cognition Enhancing Properties J Med Chem 47:5829-5832.
Chance FS, Abbott LF, Reyes AD (2002) Gain modulation from background synaptic input. Neuron 35:773-782.
Chandra D, Werner DF, Liang J, Suryanarayanan A, Harrison NL, Spigelman I, Olsen RW, Homanics GE (2008) Normal acute behavioral responses to moderate/high dose ethanol in GABAA receptor α4 subunit knockout mice. Alcohol Clin Exp Res 32:10-18.
Chandra D, Jia F, Liang J, Peng Z, Suryanarayanan A, Werner DF, Spigelman I, Houser CR, Olsen RW, Harrison NL, Homanics GE (2006) GABAA receptor α4 subunits mediate extrasynaptic inhibition in thalamus and dentate gyrus and the action of gaboxadol. Proc Natl Acad Sci U S A 103:15230-15235.
259 260
Chapman CA, Perez Y, Lacaille JC (1998) Effects of GABAA inhibition on the expression of long-term potentiation in CA1 pyramidal cells are dependent on tetanization parameters. Hippocampus 8:289-298.
Chen C, Tonegawa S (1997) Molecular genetic analysis of synaptic plasticity, activity-dependent neural development, learning, and memory in the mammalian brain. Annu Rev Neurosci 20:157-184.
Cheng VY, Martin LJ, Elliott EM, Kim JH, Mount HTJ, Taverna FA, Roder JC, MacDonald JF, Bhambri A, Collinson N, Wafford KA, Orser BA (2006) α5GABAA Receptors Mediate the Amnestic But Not Sedative-Hypnotic Effects of the General Anesthetic Etomidate. J Neurosci 26:3713-3720.
Chowdhury N, Quinn JJ, Fanselow MS (2005) Dorsal hippocampus involvement in trace fear conditioning with long, but not short, trace intervals in mice. Behav Neurosci 119:1396- 1402.
Christie SB, De Blas AL (2002) α5 Subunit-containing GABAA receptors form clusters at GABAergic synapses in hippocampal cultures. Neuroreport 13:2355-2358.
Chupakhin VI, Palyulin VA, Zefirov NS (2006) Modeling the open and closed forms of GABAA receptor: analysis of ligand-receptor interactions for the GABA-binding site. Dokl Biochem Biophys 408:169-174.
Clark RE, Squire LR (1998) Classical conditioning and brain systems: the role of awareness. Science 280:77-81.
Coan EJ, Collingridge GL (1985) Magnesium ions block an N-methyl-D-aspartate receptor- mediated component of synaptic transmission in rat hippocampus. Neurosci Lett 53:21- 26.
Coan EJ, Saywood W, Collingridge GL (1987) MK-801 blocks NMDA receptor-mediated synaptic transmission and long term potentiation in rat hippocampal slices. Neurosci Lett 80:111-114.
Cobb SR, Buhl EH, Halasy K, Paulsen O, Somogyi P (1995) Synchronization of neuronal activity in hippocampus by individual GABAergic interneurons. Nature 378:75-78.
Collingridge GL, Kehl SJ, McLennan H (1983) Excitatory amino acids in synaptic transmission in the Schaffer collateral-commissural pathway of the rat hippocampus. J Physiol 334:33- 46.
Collingridge GL, Olsen RW, Peters J, Spedding M (2009) A nomenclature for ligand-gated ion channels. Neuropharmacology 56:2-5.
Collinson N, Atack JR, Laughton P, Dawson GR, Stephens DN (2006) An inverse agonist selective for α5 subunit-containing GABAA receptors improves encoding and recall but not consolidation in the Morris water maze. Psychopharmacology 188:619-628.
260 261
Collinson N, Kuenzi FM, Jarolimek W, Maubach KA, Cothliff R, Sur C, Smith A, Otu FM, Howell O, Atack JR, McKernan RM, Seabrook GR, Dawson GR, Whiting PJ, Rosahl TW (2002) Enhanced learning and memory and altered GABAergic synaptic transmission in mice lacking the α5 subunit of the GABAA receptor. J Neurosci 22:5572- 5580.
Conti F, Minelli A, Melone M (2004) GABA transporters in the mammalian cerebral cortex: localization, development and pathological implications. Brain Res Brain Res Rev 45:196-212.
Cook JB, Foster KL, Eiler WJ, 2nd, McKay PF, Woods J, 2nd, Harvey SC, Garcia M, Grey C, McCane S, Mason D, Cummings R, Li X, Cook JM, June HL (2005) Selective GABAA α5 benzodiazepine inverse agonist antagonizes the neurobehavioral actions of alcohol. Alcohol Clin Exp Res 29:1390-1401.
Cormier RJ, Greenwood AC, Connor JA (2001) Bidirectional synaptic plasticity correlated with the magnitude of dendritic calcium transients above a threshold. J Neurophysiol 85:399- 406.
Cossart R, Bernard C, Ben-Ari Y (2005) Multiple facets of GABAergic neurons and synapses: multiple fates of GABA signalling in epilepsies. Trends Neurosci 28:108-115.
Costa E (1998) From GABAA receptor diversity emerges a unified vision of GABAergic inhibition. Annu Rev Pharmacol Toxicol 38:321-350.
Cottrell JE (2008) We care, therefore we are: anesthesia-related morbidity and mortality: the 46th Rovenstine Lecture. Anesthesiology 109:377-388.
Crestani F, Keist R, Fritschy JM, Benke D, Vogt K, Prut L, Bluthmann H, Mohler H, Rudolph U (2002) Trace fear conditioning involves hippocampal α5 GABAA receptors. Proc Natl Acad Sci U S A 99:8980-8985.
Crews FT, Morrow AL, Criswell H, Breese G (1996) Effects of ethanol on ion channels. Int Rev Neurobiol 39:283-367.
Crunelli V, Mayer ML (1984) Mg2+ dependence of membrane resistance increases evoked by NMDA in hippocampal neurones. Brain Res 311:392-396.
Culley DJ, Baxter M, Yukhananov R, Crosby G (2003) The memory effects of general anesthesia persist for weeks in young and aged rats. Anesth Analg 96:1004-1009, table of contents.
Culley DJ, Baxter MG, Yukhananov R, Crosby G (2004) Long-term impairment of acquisition of a spatial memory task following isoflurane-nitrous oxide anesthesia in rats. Anesthesiology 100:309-314.
Czepita D, Reid SN, Daw NW (1994) Effect of longer periods of dark rearing on NMDA receptors in cat visual cortex. J Neurophysiol 72:1220-1226.
261 262
Czerniawski J, Yoon T, Otto T (2008) Dissociating space and trace in dorsal and ventral hippocampus. Hippocampus.
Czlonkowska AI, Krzascik P, Sienkiewicz-Jarosz H, Siemiatkowski M, Szyndler J, Bidzinski A, Plaznik A (2000) The effects of neurosteroids on picrotoxin-, bicuculline- and NMDA- induced seizures, and a hypnotic effect of ethanol. Pharmacol Biochem Behav 67:345- 353.
Dalvi A, Rodgers RJ (1996) GABAergic influences on plus-maze behaviour in mice. Psychopharmacology (Berl) 128:380-397.
Davies CH, Collingridge GL (1993) The physiological regulation of synaptic inhibition by GABAB autoreceptors in rat hippocampus. J Physiol 472:245-265.
Davies CH, Starkey SJ, Pozza MF, Collingridge GL (1991) GABA autoreceptors regulate the induction of LTP. Nature 349:609-611.
Davis PJ, Cook DR (1986) Clinical pharmacokinetics of the newer intravenous anaesthetic agents. Clin Pharmacokinet 11:18-35.
Dawson GR, Maubach KA, Collinson N, Cobain M, Everitt BJ, MacLeod AM, Choudhury HI, McDonald LM, Pillai G, Rycroft W, Smith AJ, Sternfeld F, Tattersall FD, Wafford KA, Reynolds DS, Seabrook GR, Atack JR (2006) An inverse agonist selective for α5 subunit-containing GABAA receptors enhances cognition. J Pharmacol Exp Ther 316:1335-1345.
DeLong GR, Ritch CR, Burch S (2002) Fluoxetine response in children with autistic spectrum disorders: correlation with familial major affective disorder and intellectual achievement. Dev Med Child Neurol 44:652-659.
Delong R (2007) GABAA receptor α5 subunit as a candidate gene for autism and bipolar disorder: a proposed endophenotype with parent-of-origin and gain-of-function features,with or without oculocutaneous albinism. Autism 11:135-147.
Destexhe A, Contreras D, Steriade M (1998) Mechanisms underlying the synchronizing action of corticothalamic feedback through inhibition of thalamic relay cells. J Neurophysiol 79:999-1016.
Devan BD, Goad EH, Petri HL, Antoniadis EA, Hong NS, Ko CH, Leblanc L, Lebovic SS, Lo Q, Ralph MR, McDonald RJ (2001) Circadian phase-shifted rats show normal acquisition but impaired long-term retention of place information in the water task. Neurobiol Learn Mem 75:51-62.
Dias R, Sheppard WF, Fradley RL, Garrett EM, Stanley JL, Tye SJ, Goodacre S, Lincoln RJ, Cook SM, Conley R, Hallett D, Humphries AC, Thompson SA, Wafford KA, Street LJ, Castro JL, Whiting PJ, Rosahl TW, Atack JR, McKernan RM, Dawson GR, Reynolds DS (2005) Evidence for a significant role of α3-containing GABAA receptors in mediating the anxiolytic effects of benzodiazepines. J Neurosci 25:10682-10688.
262 263
Dingledine R, Korn SJ (1985) γ-aminobutyric acid uptake and the termination of inhibitory synaptic potentials in the rat hippocampal slice. J Physiol 366:387-409.
Dor-Shav NK, Horowitz Z (1984) Intelligence and personality variables of parents of autistic children. J Genet Psychol 144:39-50.
Dorow R, Horowski R, Paschelke G, Amin M (1983) Severe anxiety induced by FG 7142, a beta-carboline ligand for benzodiazepine receptors. Lancet 2:98-99.
Drexler B, Roether CL, Jurd R, Rudolph U, Antkowiak B (2005) Opposing actions of etomidate on cortical theta oscillations are mediated by different γ-aminobutyric acid type A receptor subtypes. Anesthesiology 102:346-352.
Dudek SM, Bear MF (1992) Homosynaptic long-term depression in area CA1 of hippocampus and effects of N-methyl-D-aspartate receptor blockade. Proc Natl Acad Sci U S A 89:4363-4367.
Dudek SM, Bear MF (1993) Bidirectional long-term modification of synaptic effectiveness in the adult and immature hippocampus. J Neurosci 13:2910-2918.
Dunning DD, Hoover CL, Soltesz I, Smith MA, O'Dowd DK (1999) GABAA receptor-mediated miniature postsynaptic currents and α-subunit expression in developing cortical neurons. J Neurophysiol 82:3286-3297.
Dutton RC, Maurer AJ, Sonner JM, Fanselow MS, Laster MJ, Eger EI, 2nd (2001) The concentration of isoflurane required to suppress learning depends on the type of learning. Anesthesiology 94:514-519.
Dwyer R, Bennett HL, Eger EI, 2nd, Heilbron D (1992) Effects of isoflurane and nitrous oxide in subanesthetic concentrations on memory and responsiveness in volunteers. Anesthesiology 77:888-898.
Eger EI, 2nd, Koblin DD, Harris RA, Kendig JJ, Pohorille A, Halsey MJ, Trudell JR (1997) Hypothesis: inhaled anesthetics produce immobility and amnesia by different mechanisms at different sites. Anesth Analg 84:915-918.
Eichenbaum H (2000) A cortical-hippocampal system for declarative memory. Nat Rev Neurosci 1:41-50.
Eichenbaum H (2004) Hippocampus: cognitive processes and neural representations that underlie declarative memory. Neuron 44:109-120.
Enz R, Brandstatter JH, Hartveit E, Wassle H, Bormann J (1995) Expression of GABA receptor ρ1 and ρ2 subunits in the retina and brain of the rat. Eur J Neurosci 7:1495-1501.
Erlander MG, Tillakaratne NJ, Feldblum S, Patel N, Tobin AJ (1991) Two genes encode distinct glutamate decarboxylases. Neuron 7:91-100.
Essrich C, Lorez M, Benson JA, Fritschy JM, Luscher B (1998) Postsynaptic clustering of major GABAA receptor subtypes requires the γ2 subunit and gephyrin. Nat Neurosci 1:563-571.
263 264
Eugene E, Depienne C, Baulac S, Baulac M, Fritschy JM, Le Guern E, Miles R, Poncer JC (2007) GABAA receptor γ2 subunit mutations linked to human epileptic syndromes differentially affect phasic and tonic inhibition. J Neurosci 27:14108-14116.
Fan Y, Fricker D, Brager DH, Chen X, Lu HC, Chitwood RA, Johnston D (2005) Activity- dependent decrease of excitability in rat hippocampal neurons through increases in Ih. Nat Neurosci 8:1542-1551.
Fanselow MS (1980) Conditioned and unconditional components of post-shock freezing. Pavlov J Biol Sci 15:177-182.
Fanselow MS, Helmstetter FJ (1988) Conditional analgesia, defensive freezing, and benzodiazepines. Behav Neurosci 102:233-243.
Fanselow MS, Poulos AM (2005) The neuroscience of mammalian associative learning Annual Review of Psychology 56:207-234.
Farrant M, Nusser Z (2005) Variations on an inhibitory theme: phasic and tonic activation of GABAA receptors. Nat Rev Neurosci 6:215-229.
Faure-Halley C, Graham D, Arbilla S, Langer SZ (1993) Expression and properties of recombinant α1 β2 γ2 and α5 β2 γ2 forms of the rat GABAA receptor. Eur J Pharmacol 246:283-287.
Fox CJ, Russell KI, Wang YT, Christie BR (2006) Contribution of NR2A and NR2B NMDA subunits to bidirectional synaptic plasticity in the hippocampus in vivo. Hippocampus 16:907-915.
Fradley RL, Guscott MR, Bull S, Hallett DJ, Goodacre SC, Wafford KA, Garrett EM, Newman RJ, O'Meara GF, Whiting PJ, Rosahl TW, Dawson GR, Reynolds DS, Atack JR (2007) Differential contribution of GABAA receptor subtypes to the anticonvulsant efficacy of benzodiazepine site ligands. J Psychopharmacol 21:384-391.
Franks NP, Lieb WR (1988) Volatile general anaesthetics activate a novel neuronal K+ current. Nature 333:662-664.
Frey U, Huang YY, Kandel ER (1993) Effects of cAMP simulate a late stage of LTP in hippocampal CA1 neurons. Science 260:1661-1664.
Fritschy JM, Mohler H (1995) GABAA-receptor heterogeneity in the adult rat brain: differential regional and cellular distribution of seven major subunits. J Comp Neurol 359:154-194.
Fritschy JM, Kiener T, Bouilleret V, Loup F (1999) GABAergic neurons and GABAA-receptors in temporal lobe epilepsy. Neurochem Int 34:435-445.
Furukawa H, Singh SK, Mancusso R, Gouaux E (2005) Subunit arrangement and function in NMDA receptors. Nature 438:185-192.
Gafford GM, Parsons RG, Helmstetter FJ (2005) Effects of post-training hippocampal injections of midazolam on fear conditioning. Learn Mem 12:573-578.
264 265
Ganter P, Szucs P, Paulsen O, Somogyi P (2004) Properties of horizontal axo-axonic cells in stratum oriens of the hippocampal CA1 area of rats in vitro. Hippocampus 14:232-243.
Gasnier B (2000) The loading of neurotransmitters into synaptic vesicles. Biochimie 82:327-337.
Gaspary HL, Wang W, Richerson GB (1998) Carrier-mediated GABA release activates GABA receptors on hippocampal neurons. J Neurophysiol 80:270-281.
Gazzaniga MS (2004) The cognitive neurosciences, 3rd Edition. Cambridge, Mass.: MIT Press.
Gean PW, Lin JH (1993) D-2-amino-5-phosphonovaleate blocks induction of long-term depression of the NMDA receptor-mediated synaptic component in rat hippocampus. Neurosci Lett 158:170-172.
Geiger JR, Melcher T, Koh DS, Sakmann B, Seeburg PH, Jonas P, Monyer H (1995) Relative abundance of subunit mRNAs determines gating and Ca2+ permeability of AMPA receptors in principal neurons and interneurons in rat CNS. Neuron 15:193-204.
Gerdjikov TV, Rudolph U, Keist R, Mohler H, Feldon J, Yee BK (2008) Hippocampal α5 subunit-containing GABAA receptors are involved in the development of the latent inhibition effect. Neurobiol Learn Mem 89:87-94.
Gerlai R, Henderson JT, Roder JC, Jia Z (1998) Multiple behavioral anomalies in GluR2 mutant mice exhibiting enhanced LTP. Behav Brain Res 95:37-45.
Geula C, Nagykery N, Nicholas A, Wu CK (2008) Cholinergic neuronal and axonal abnormalities are present early in aging and in Alzheimer disease. J Neuropathol Exp Neurol 67:309-318.
Glatt K, Sinnett D, Lalande M (1994) The human gamma-aminobutyric acid receptor subunit beta 3 and alpha 5 gene cluster in chromosome 15q11-q13 is rich in highly polymorphic (CA)n repeats. Genomics 19:157-160.
Glatt KA, Sinnett D, Lalande M (1992) Dinucleotide repeat polymorphism at the GABAA receptor alpha 5 (GABRA5) locus at chromosome 15q11-q13. Hum Mol Genet 1:348.
Gloor P (1997) The temporal lobe and limbic system. New York: Oxford University Press.
Glykys J, Mody I (2006) Hippocampal network hyperactivity after selective reduction of tonic inhibition in GABAA receptor α5 subunit-deficient mice. J Neurophysiol 95:2796-2807.
Glykys J, Mody I (2007a) Activation of GABAA Receptors: Views from Outside the Synaptic Cleft. Neuron 56:763-770.
Glykys J, Mody I (2007b) The main source of ambient GABA responsible for tonic inhibition in the mouse hippocampus. J Physiol 582:1163-1178.
Glykys J, Mann EO, Mody I (2008) Which GABAA receptor subunits are necessary for tonic inhibition in the hippocampus? J Neurosci 28:1421-1426.
265 266
Glykys J, Peng Z, Chandra D, Homanics GE, Houser CR, Mody I (2007) A new naturally occurring GABAA receptor subunit partnership with high sensitivity to ethanol. Nat Neurosci 10:40-48.
Gooding JM, Corssen G (1976) Etomidate: an ultrashort-acting nonbarbiturate agent for anesthesia induction. Anesth Analg 55:286-289.
Grasshoff C, Jurd R, Rudolph U, Antkowiak B (2007) Modulation of presynaptic β3-containing GABAA receptors limits the immobilizing actions of GABAergic anesthetics. Mol Pharmacol 72:780-787.
Gray JA (1982) The neuropsychology of anxiety : an enquiry into the functions of the septo- hippocampal system. New York: University Press.
Grover LM, Teyler TJ (1990) Two components of long-term potentiation induced by different patterns of afferent activation. Nature 347:477-479.
Gulledge AT, Stuart GJ (2003) Excitatory actions of GABA in the cortex. Neuron 37:299-309.
Guptill JT, Booker AB, Gibbs TT, Kemper TL, Bauman ML, Blatt GJ (2007) [3H]- flunitrazepam-labeled benzodiazepine binding sites in the hippocampal formation in autism: a multiple concentration autoradiographic study. J Autism Dev Disord 37:911- 920.
Gurley D, Amin J, Ross PC, Weiss DS, White G (1995) Point mutations in the M2 region of the α, β, or γ subunit of the GABAA channel that abolish block by picrotoxin. Receptors Channels 3:13-20.
Haley JE, Schaible E, Pavlidis P, Murdock A, Madison DV (1996) Basal and apical synapses of CA1 pyramidal cells employ different LTP induction mechanisms. Learn Mem 3:289- 295.
Hamm AO, Weike AI (2005) The neuropsychology of fear learning and fear regulation. Int J Psychophysiol 57:5-14.
Han CJ, O'Tuathaigh CM, van Trigt L, Quinn JJ, Fanselow MS, Mongeau R, Koch C, Anderson DJ (2003) Trace but not delay fear conditioning requires attention and the anterior cingulate cortex. Proc Natl Acad Sci U S A 100:13087-13092.
Hanchar HJ, Dodson PD, Olsen RW, Otis TS, Wallner M (2005) Alcohol-induced motor impairment caused by increased extrasynaptic GABAA receptor activity. Nat Neurosci 8:339-345.
Hanning CD (2005) Postoperative cognitive dysfunction. Br J Anaesth 95:82-87.
Hardingham GE, Chawla S, Johnson CM, Bading H (1997) Distinct functions of nuclear and cytoplasmic calcium in the control of gene expression. Nature 385:260-265.
Harris D, Clayton T, Cook J, Sahbaie P, Halliwell RF, Furtmuller R, Huck S, Sieghart W, DeLorey TM (2008) Selective influence on contextual memory: physiochemical
266 267
properties associated with selectivity of benzodiazepine ligands at GABAA receptors containing the α5 subunit. J Med Chem 51:3788-3803.
Harrison NL (2007) Mechanisms of sleep induction by GABAA receptor agonists. J Clin Psychiatry 68 Suppl 5:6-12.
Harvey J, Collingridge GL (1992) Thapsigargin blocks the induction of long-term potentiation in rat hippocampal slices. Neurosci Lett 139:197-200.
Hauser J, Rudolph U, Keist R, Mohler H, Feldon J, Yee BK (2004) Hippocampal α5 subunit- containing GABAA receptors modulate the expression of prepulse inhibition. Mol Psychiatry 10:201-207.
Hebb DO (1949) The organization of behavior; a neuropsychological theory. New York: Wiley.
Hefft S, Jonas P (2005) Asynchronous GABA release generates long-lasting inhibition at a hippocampal interneuron-principal neuron synapse. Nat Neurosci 8:1319-1328.
Hell JW, Maycox PR, Jahn R (1990) Energy dependence and functional reconstitution of the γ- aminobutyric acid carrier from synaptic vesicles. J Biol Chem 265:2111-2117.
Hemmings HC, Jr., Akabas MH, Goldstein PA, Trudell JR, Orser BA, Harrison NL (2005) Emerging molecular mechanisms of general anesthetic action. Trends Pharmacol Sci 26:503-510.
Herron CE, Lester RA, Coan EJ, Collingridge GL (1985) Intracellular demonstration of an N- methyl-D-aspartate receptor mediated component of synaptic transmission in the rat hippocampus. Neurosci Lett 60:19-23.
Hertz L, Dringen R, Schousboe A, Robinson SR (1999) Astrocytes: glutamate producers for neurons. J Neurosci Res 57:417-428.
Hevers W, Lüddens H (1998 ) The diversity of GABAA receptors. Pharmacological and electrophysiological properties of GABAA channel subtypes. Mol Neurobiol 18:35-86.
Heynen AJ, Abraham WC, Bear MF (1996) Bidirectional modification of CA1 synapses in the adult hippocampus in vivo. Nature 381:163-166.
Hill-Venning C, Belelli D, Peters JA, Lambert JJ (1997) Subunit-dependent interaction of the general anaesthetic etomidate with the γ-aminobutyric acid type A receptor. Br J Pharmacol 120:749-756.
Hille B (2001) Ion channels of excitable membranes, 3rd Edition. Sunderland, Mass.: Sinauer.
Honmou O, Oyelese AA, Kocsis JD (1995) The anticonvulsant gabapentin enhances promoted release of GABA in hippocampus: a field potential analysis. Brain Res 692:273-277.
Horsthemke B, Wagstaff J (2008) Mechanisms of imprinting of the Prader-Willi/Angelman region. Am J Med Genet A 146A:2041-2052.
267 268
Houser CR, Esclapez M (2003) Downregulation of the α5 subunit of the GABAA receptor in the pilocarpine model of temporal lobe epilepsy. Hippocampus 13:633-645.
Howland JG, Harrison RA, Hannesson DK, Phillips AG (2008) Ventral hippocampal involvement in temporal order, but not recognition, memory for spatial information. Hippocampus 18:251-257.
Hu Y, Lund IV, Gravielle MC, Farb DH, Brooks-Kayal AR, Russek SJ (2008) Surface expression of GABAA receptors is transcriptionally controlled by the interplay of cAMP- response element-binding protein and its binding partner inducible cAMP early repressor. J Biol Chem 283:9328-9340.
Huang Y, Lu W, Ali DW, Pelkey KA, Pitcher GM, Lu YM, Aoto H, Roder JC, Sasaki T, Salter MW, MacDonald JF (2001) CAKbeta/Pyk2 kinase is a signaling link for induction of long-term potentiation in CA1 hippocampus. Neuron 29:485-496.
Huang YY, Li XC, Kandel ER (1994) cAMP contributes to mossy fiber LTP by initiating both a covalently mediated early phase and macromolecular synthesis-dependent late phase. Cell 79:69-79.
Impagnatiello F, Guidotti AR, Pesold C, Dwivedi Y, Caruncho H, Pisu MG, Uzunov DP, Smalheiser NR, Davis JM, Pandey GN, Pappas GD, Tueting P, Sharma RP, Costa E (1998) A decrease of reelin expression as a putative vulnerability factor in schizophrenia. Proc Natl Acad Sci U S A 95:15718-15723.
Ino T, Kaneko T, Mizuno N (1998) Direct projections from the entorhinal cortical layers to the dentate gyrus, hippocampus, and subicular complex in the cat. Neurosci Res 32:241-265.
Ishikawa M, Mizukami K, Iwakiri M, Hidaka S, Asada T (2004) Immunohistochemical and immunoblot study of GABAA α1 and β2/3 subunits in the prefrontal cortex of subjects with schizophrenia and bipolar disorder. Neurosci Res 50:77-84.
Ishizeki J, Nishikawa K, Kubo K, Saito S, Goto F (2008) Amnestic concentrations of sevoflurane inhibit synaptic plasticity of hippocampal CA1 neurons through γ-aminobutyric acid- mediated mechanisms. Anesthesiology 108:447-456.
Ito I, Sugiyama H (1991) Roles of glutamate receptors in long-term potentiation at hippocampal mossy fiber synapses. Neuroreport 2:333-336.
Ivanov A, Tyzio R, Zilberter Y, Ben-Ari Y (2008) (R)-roscovitine, a cyclin-dependent kinase inhibitor, enhances tonic GABA inhibition in rat hippocampus. Neuroscience 156:277- 288.
Izumi Y, Clifford DB, Zorumski CF (1991) 2-Amino-3-phosphonopropionate blocks the induction and maintenance of long-term potentiation in rat hippocampal slices. Neurosci Lett 122:187-190.
Jacob TC, Moss SJ, Jurd R (2008) GABAA receptor trafficking and its role in the dynamic modulation of neuronal inhibition. Nat Rev Neurosci 9:331-343.
268 269
Janssen MJ, Ade KK, Fu Z, Vicini S (2009) Dopamine modulation of GABA tonic conductance in striatal output neurons. J Neurosci 29:5116-5126.
Jenkins A, Andreasen A, Trudell JR, Harrison NL (2002) Tryptophan scanning mutagenesis in TM4 of the GABAA receptor α1 subunit: implications for modulation by inhaled anesthetics and ion channel structure. Neuropharmacology 43:669-678.
Jenkins A, Greenblatt EP, Faulkner HJ, Bertaccini E, Light A, Lin A, Andreasen A, Viner A, Trudell JR, Harrison NL (2001) Evidence for a common binding cavity for three general anesthetics within the GABAA receptor. J Neurosci 21:RC136.
Jensen K, Lambert JD, Jensen MS (2000) Tetanus-induced asynchronous GABA release in cultured hippocampal neurons. Brain Res 880:198-201.
Jia F, Pignataro L, Harrison NL (2007) GABAA receptors in the thalamus: α4 subunit expression and alcohol sensitivity. Alcohol 41:177-185.
Jia F, Yue M, Chandra D, Keramidas A, Goldstein PA, Homanics GE, Harrison NL (2008) Taurine is a potent activator of extrasynaptic GABAA receptors in the thalamus. J Neurosci 28:106-115.
Jia Z, Agopyan N, Miu P, Xiong Z, Henderson J, Gerlai R, Taverna FA, Velumian A, MacDonald J, Carlen P, Abramow-Newerly W, Roder J (1996) Enhanced LTP in mice deficient in the AMPA receptor GluR2. Neuron 17:945-956.
Jin H, Wu H, Osterhaus G, Wei J, Davis K, Sha D, Floor E, Hsu CC, Kopke RD, Wu JY (2003) Demonstration of functional coupling between γ-aminobutyric acid (GABA) synthesis and vesicular GABA transport into synaptic vesicles. Proc Natl Acad Sci U S A 100:4293-4298.
Johnson T, Monk T, Rasmussen LS, Abildstrom H, Houx P, Korttila K, Kuipers HM, Hanning CD, Siersma VD, Kristensen D, Canet J, Ibanaz MT, Moller JT (2002) Postoperative cognitive dysfunction in middle-aged patients. Anesthesiology 96:1351-1357.
Jones MV, Brooks PA, Harrison NL (1992) Enhancement of gamma-aminobutyric acid-activated Cl- currents in cultured rat hippocampal neurones by three volatile anaesthetics. J Physiol 449:279-293.
Jun K, Choi G, Yang SG, Choi KY, Kim H, Chan GC, Storm DR, Albert C, Mayr GW, Lee CJ, Shin HS (1998) Enhanced hippocampal CA1 LTP but normal spatial learning in inositol 1,4,5-trisphosphate 3-kinase(A)-deficient mice. Learn Mem 5:317-330.
June HL, Harvey SC, Foster KL, McKay PF, Cummings R, Garcia M, Mason D, Grey C, McCane S, Williams LS, Johnson TB, He X, Rock S, Cook JM (2001) GABAA Receptors Containing α5 Subunits in the CA1 and CA3 Hippocampal Fields Regulate Ethanol-Motivated Behaviors: An Extended Ethanol Reward Circuitry. J Neurosci 21:2166-2177.
269 270
June HL, Foster KL, McKay PF, Seyoum R, Woods JE, Harvey SC, Eiler WJ, Grey C, Carroll MR, McCane S, Jones CM, Yin W, Mason D, Cummings R, Garcia M, Ma C, Sarma PV, Cook JM, Skolnick P (2003) The reinforcing properties of alcohol are mediated by GABAA1 receptors in the ventral pallidum. Neuropsychopharmacology 28:2124-2137.
Jurd R, Arras M, Lambert S, Drexler B, Siegwart R, Crestani F, Zaugg M, Vogt KE, Ledermann B, Antkowiak B, Rudolph U (2003) General anesthetic actions in vivo strongly attenuated by a point mutation in the GABAA receptor β3 subunit. Faseb J 17:250-252.
Kaibara T, Leung LS (1993) Basal versus apical dendritic long-term potentiation of commissural afferents to hippocampal CA1: a current-source density study. J Neurosci 13:2391-2404.
Kandel ER, Schwartz JH (2000) Principles of neural science, 4th Edition. New York: McGraw- Hill.
Kapur A, Pearce RA, Lytton WW, Haberly LB (1997) GABAA-mediated IPSCs in piriform cortex have fast and slow components with different properties and locations on pyramidal cells. J Neurophysiol 78:2531-2545.
Kaufman DL, Houser CR, Tobin AJ (1991) Two forms of the γ-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions. J Neurochem 56:720-723.
Kawakami R, Shinohara Y, Kato Y, Sugiyama H, Shigemoto R, Ito I (2003) Asymmetrical allocation of NMDA receptor ε2 subunits in hippocampal circuitry. Science 300:990-994.
Kemp A, Manahan-Vaughan D (2004) Hippocampal long-term depression and long-term potentiation encode different aspects of novelty acquisition. Proc Natl Acad Sci U S A 101:8192-8197.
Kenakin T (2004) Efficacy as a vector: the relative prevalence and paucity of inverse agonism. Mol Pharmacol 65:2-11.
Ketterman JK, Li YV (2008) Presynaptic evidence for zinc release at the mossy fiber synapse of rat hippocampus. J Neurosci Res 86:422-434.
Kim JJ, Fanselow MS (1992) Modality-specific retrograde amnesia of fear. Science 256:675- 677.
Kim JJ, Rison RA, Fanselow MS (1993) Effects of amygdala, hippocampus, and periaqueductal gray lesions on short- and long-term contextual fear. Behav Neurosci 107:1093-1098.
Kirkwood A, Rioult MC, Bear MF (1996) Experience-dependent modification of synaptic plasticity in visual cortex. Nature 381:526-528.
Kirkwood A, Dudek SM, Gold JT, Aizenman CD, Bear MF (1993) Common forms of synaptic plasticity in the hippocampus and neocortex in vitro. Science 260:1518-1521.
270 271
Kneussel M, Brandstatter JH, Gasnier B, Feng G, Sanes JR, Betz H (2001) Gephyrin- independent clustering of postsynaptic GABAA receptor subtypes. Mol Cell Neurosci 17:973-982.
Kohr G (2006) NMDA receptor function: subunit composition versus spatial distribution. Cell Tissue Res 326:439-446.
Komaki A, Shahidi S, Lashgari R, Haghparast A, Malakouti SM, Noorbakhsh SM (2007) Effects of GABAergic inhibition on neocortical long-term potentiation in the chronically prepared rat. Neurosci Lett 422:181-186.
Korn SJ, Giacchino JL, Chamberlin NL, Dingledine R (1987) Epileptiform burst activity induced by potassium in the hippocampus and its regulation by GABA-mediated inhibition. J Neurophysiol 57:325-340.
Korpi ER, Grunder G, Luddens H (2002) Drug interactions at GABAA receptors No one is listening to you no one likes you in this house. Prog Neurobiol 67:113-159.
Kozlov AS, Angulo MC, Audinat E, Charpak S (2006) Target cell-specific modulation of neuronal activity by astrocytes. Proc Natl Acad Sci U S A 103:10058-10063.
Krasowski MD, Koltchine VV, Rick CE, Ye Q, Finn SE, Harrison NL (1998) Propofol and other intravenous anesthetics have sites of action on the γ-aminobutyric acid type A receptor distinct from that for isoflurane. Mol Pharmacol 53:530-538.
Krook-Magnuson E, Huntsman MM (2007) The transience of interneuron circuit diversity just "sped" up. Proc Natl Acad Sci U S A 104:16723-16724.
Kumari V, Fannon D, Sumich AL, Sharma T (2007) Startle gating in antipsychotic-naive first episode schizophrenia patients: one ear is better than two. Psychiatry Res 151:21-28.
Laurie DJ, Seeburg PH (1994) Regional and developmental heterogeneity in splicing of the rat brain NMDAR1 mRNA. J Neurosci 14:3180-3194.
Laurie DJ, Wisden W, Seeburg PH (1992) The distribution of thirteen GABAA receptor subunit mRNAs in the rat brain. III. Embryonic and postnatal development. J Neurosci 12:4151- 4172.
Lee HJ, Choi JS, Brown TH, Kim JJ (2001) Amygdalar nmda receptors are critical for the expression of multiple conditioned fear responses. J Neurosci 21:4116-4124.
Lee HK, Barbarosie M, Kameyama K, Bear MF, Huganir RL (2000) Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature 405:955-959.
Li HB, Mao RR, Zhang JC, Yang Y, Cao J, Xu L (2008) Antistress effect of TRPV1 channel on synaptic plasticity and spatial memory. Biol Psychiatry 64:286-292.
Lim C, Blume HW, Madsen JR, Saper CB (1997) Connections of the hippocampal formation in humans: I. The mossy fiber pathway. J Comp Neurol 385:325-351.
271 272
Lindquist CE, Birnir B (2006) Graded response to GABA by native extrasynaptic GABA receptors. J Neurochem 97:1349-1356.
Lisman J, Spruston N (2005) Postsynaptic depolarization requirements for LTP and LTD: a critique of spike timing-dependent plasticity. Nat Neurosci 8:839-841.
Liu L, Wong TP, Pozza MF, Lingenhoehl K, Wang Y, Sheng M, Auberson YP, Wang YT (2004) Role of NMDA receptor subtypes in governing the direction of hippocampal synaptic plasticity. Science 304:1021-1024.
Liu QS, Pu L, Poo MM (2005) Repeated cocaine exposure in vivo facilitates LTP induction in midbrain dopamine neurons. Nature 437:1027-1031.
Liu QY, Schaffner AE, Chang YH, Maric D, Barker JL (2000) Persistent activation of GABAA receptor/Cl- channels by astrocyte-derived GABA in cultured embryonic rat hippocampal neurons. J Neurophysiol 84:1392-1403.
Liu R, Zhang P, McKernan R, Wafford K, Cook JM (1995) Synthesis of novel imidazobenzodiazepines selective for the α5β3γ2 (Bz5) GABAA benzodiazepine receptor subtype. Med Chem Res 5:700-709.
Liu R, Hu RJ, Zhang P, Skolnick P, Cook JM (1996) Synthesis and pharmacological properties of novel 8-substituted imidazobenzodiazepines: high-affinity, selective probes for α5- containing GABAA receptors. J Med Chem 39:1928-1934.
Loebrich S, Bahring R, Katsuno T, Tsukita S, Kneussel M (2006) Activated radixin is essential for GABAA receptor α5 subunit anchoring at the actin cytoskeleton. Embo J 25:987-999.
Lopez U, Habre W, Laurencon M, Haller G, Van der Linden M, Iselin-Chaves IA (2007) Intra- operative awareness in children: the value of an interview adapted to their cognitive abilities. Anaesthesia 62:778-789.
Low K, Crestani F, Keist R, Benke D, Brunig I, Benson JA, Fritschy JM, Rulicke T, Bluethmann H, Mohler H, Rudolph U (2000) Molecular and neuronal substrate for the selective attenuation of anxiety. Science 290:131-134.
Lu W, Man H, Ju W, Trimble WS, MacDonald JF, Wang YT (2001) Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. Neuron 29:243-254.
Lu YM, Roder JC, Davidow J, Salter MW (1998) Src activation in the induction of long-term potentiation in CA1 hippocampal neurons. Science 279:1363-1367.
Lu YM, Mansuy IM, Kandel ER, Roder J (2000) Calcineurin-mediated LTD of GABAergic inhibition underlies the increased excitability of CA1 neurons associated with LTP. Neuron 26:197-205.
Luddens H, Seeburg PH, Korpi ER (1994) Impact of β and γ variants on ligand-binding properties of gamma-aminobutyric acid type A receptors. Mol Pharmacol 45:810-814.
272 273
Lynch G, Larson J, Kelso S, Barrionuevo G, Schottler F (1983) Intracellular injections of EGTA block induction of hippocampal long-term potentiation. Nature 305:719-721.
Maccaferri G, Lacaille JC (2003) Interneuron Diversity series: Hippocampal interneuron classifications--making things as simple as possible, not simpler. Trends Neurosci 26:564-571.
MacDonald JF, Mody I, Salter MW (1989) Regulation of N-methyl-D-aspartate receptors revealed by intracellular dialysis of murine neurones in culture. J Physiol 414:17-34.
MacDonald JF, Jackson MF, Beazely MA (2006) Hippocampal long-term synaptic plasticity and signal amplification of NMDA receptors. Crit Rev Neurobiol 18:71-84.
Madison DV, Schuman EM (1991) LTP, post or pre? A look at the evidence for the locus of long-term potentiation. New Biol 3:549-557.
Maffei A, Nataraj K, Nelson SB, Turrigiano GG (2006) Potentiation of cortical inhibition by visual deprivation. Nature 443:81-84.
Maguire J, Mody I (2008) GABAAR plasticity during pregnancy: relevance to postpartum depression. Neuron 59:207-213.
Malenka RC (1994) Synaptic plasticity in the hippocampus: LTP and LTD. Cell 78:535-538.
Malenka RC, Bear MF (2004) LTP and LTD: an embarrassment of riches. Neuron 44:5-21.
Malinow R (1994) LTP: desperately seeking resolution. Science 266:1195-1196.
Manahan-Vaughan D, Braunewell KH (1999) Novelty acquisition is associated with induction of hippocampal long-term depression. Proc Natl Acad Sci U S A 96:8739-8744.
Manahan-Vaughan D, Kulla A, Frey JU (2000) Requirement of translation but not transcription for the maintenance of long-term depression in the CA1 region of freely moving rats. J Neurosci 20:8572-8576.
Mann EO, Paulsen O (2007) Role of GABAergic inhibition in hippocampal network oscillations. Trends Neurosci 30:343-349.
Maren S (1998) Overtraining does not mitigate contextual fear conditioning deficits produced by neurotoxic lesions of the basolateral amygdala. J Neurosci 18:3088-3097.
Maren S (2001) Neurobiology of Pavlovian fear conditioning. Annu Rev Neurosci 24:897-931.
Maren S, Fanselow MS (1997) Electrolytic lesions of the fimbria/fornix, dorsal hippocampus, or entorhinal cortex produce anterograde deficits in contextual fear conditioning in rats. Neurobiol Learn Mem 67:142-149.
Maren S, Aharonov G, Fanselow MS (1997) Neurotoxic lesions of the dorsal hippocampus and Pavlovian fear conditioning in rats. Behav Brain Res 88:261-274.
273 274
Martin DL, Barke KE (1998) Are GAD65 and GAD67 associated with specific pools of GABA in brain? Perspect Dev Neurobiol 5:119-129.
Massey PV, Bashir ZI (2007) Long-term depression: multiple forms and implications for brain function. Trends Neurosci 30:176-184.
Matthews DB, Ilgen M, White AM, Best PJ (1999) Acute ethanol administration impairs spatial performance while facilitating nonspatial performance in rats. Neurobiol Learn Mem 72:169-179.
Maubach K (2003) GABAA receptor subtype selective cognition enhancers. Curr Drug Targets CNS Neurol Disord 2:233-239.
McCauley JL, Olson LM, Delahanty R, Amin T, Nurmi EL, Organ EL, Jacobs MM, Folstein SE, Haines JL, Sutcliffe JS (2004) A linkage disequilibrium map of the 1-Mb 15q12 GABAA receptor subunit cluster and association to autism. Am J Med Genet B Neuropsychiatr Genet 131B:51-59.
McCleane GJ, Cooper R (1990) The nature of pre-operative anxiety. Anaesthesia 45:153-155.
McEchron MD, Weible AP, Disterhoft JF (2001) Aging and learning-specific changes in single- neuron activity in CA1 hippocampus during rabbit trace eyeblink conditioning. J Neurophysiol 86:1839-1857.
McEchron MD, Bouwmeester H, Tseng W, Weiss C, Disterhoft JF (1998) Hippocampectomy disrupts auditory trace fear conditioning and contextual fear conditioning in the rat. Hippocampus 8:638-646.
McKay PF, Foster KL, Mason D, Cummings R, Garcia M, Williams LS, Grey C, McCane S, He X, Cook JM, June HL (2004) A high affinity ligand for GABAA-receptor containing α5 subunit antagonizes ethanol's neurobehavioral effects in Long-Evans rats. Psychopharmacology (Berl) 172:455-462.
McKernan RM, Quirk K, Prince R, Cox PA, Gillard NP, Ragan CI, Whiting P (1991) GABAA receptor subtypes immunopurified from rat brain with α subunit-specific antibodies have unique pharmacological properties. Neuron 7:667-676.
McNaughton BL, Douglas RM, Goddard GV (1978) Synaptic enhancement in fascia dentata: cooperativity among coactive afferents. Brain Res 157:277-293.
Mehta AK, Ticku MK (1999) An update on GABAA receptors. Brain Res Brain Res Rev 29:196- 217.
Meiri N, Sun MK, Segal Z, Alkon DL (1998) Memory and long-term potentiation (LTP) dissociated: normal spatial memory despite CA1 LTP elimination with Kv1.4 antisense. Proc Natl Acad Sci U S A 95:15037-15042.
274 275
Melia KR, Ryabinin AE, Corodimas KP, Wilson MC, Ledoux JE (1996) Hippocampal- dependent learning and experience-dependent activation of the hippocampus are preferentially disrupted by ethanol. Neuroscience 74:313-322.
Mercer A, Trigg HL, Thomson AM (2007) Characterization of neurons in the CA2 subfield of the adult rat hippocampus. J Neurosci 27:7329-7338.
Mihic SJ (1999) Acute effects of ethanol on GABAA and glycine receptor function. Neurochem Int 35:115-123.
Miledi R, Thies R (1971) Tetanic and post-tetanic rise in frequency of miniature end-plate potentials in low-calcium solutions. J Physiol 212:245-257.
Misane I, Tovote P, Meyer M, Spiess J, Ogren SO, Stiedl O (2005) Time-dependent involvement of the dorsal hippocampus in trace fear conditioning in mice. Hippocampus 15:418-426.
Mitchell CP, Grayson DR, Goldman MB (2005) Neonatal lesions of the ventral hippocampal formation alter GABAA receptor subunit mRNA expression in adult rat frontal pole. Biol Psychiatry 57:49-55.
Mitchell SJ, Silver RA (2003) Shunting inhibition modulates neuronal gain during synaptic excitation. Neuron 38:433-445.
Mody I (2001) Distinguishing between GABAA receptors responsible for tonic and phasic conductances. Neurochem Res 26:907-913.
Mody I, Pearce RA (2004) Diversity of inhibitory neurotransmission through GABAA receptors. Trends Neurosci 27:569-575.
Moerman N, Bonke B, Oosting J (1993) Awareness and recall during general anesthesia. Facts and feelings. Anesthesiology 79:454-464.
Mohler H (2007) Molecular regulation of cognitive functions and developmental plasticity: impact of GABAA receptors. J Neurochem 102:1-12.
Mohler H, Fritschy JM, Luscher B, Rudolph U, Benson J, Benke D (1996) The GABAA receptors. From subunits to diverse functions. Ion Channels 4:89-113.
Moller JT, Cluitmans P, Rasmussen LS, Houx P, Rasmussen H, Canet J, Rabbitt P, Jolles J, Larsen K, Hanning CD, Langeron O, Johnson T, Lauven PM, Kristensen PA, Biedler A, van Beem H, Fraidakis O, Silverstein JH, Beneken JE, Gravenstein JS (1998) Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. ISPOCD investigators. International Study of Post-Operative Cognitive Dysfunction. Lancet 351:857-861.
Montgomery JM, Madison DV (2002) State-dependent heterogeneity in synaptic depression between pyramidal cell pairs. Neuron 33:765-777.
275 276
Monyer H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH (1994) Developmental and regional expression in the rat brain and functional properties of four NMDA receptors. Neuron 12:529-540.
Morishita W, Lu W, Smith GB, Nicoll RA, Bear MF, Malenka RC (2006) Activation of NR2B- containing NMDA receptors is not required for NMDA receptor-dependent long-term depression. Neuropharmacology.
Morris RG, Garrud P, Rawlins JN, O'Keefe J (1982) Place navigation impaired in rats with hippocampal lesions. Nature 297:681-683.
Mosbacher J, Schoepfer R, Monyer H, Burnashev N, Seeburg PH, Ruppersberg JP (1994) A molecular determinant for submillisecond desensitization in glutamate receptors. Science 266:1059-1062.
Moss SJ, Smart TG (2001) Constructing inhibitory synapses. Nat Rev Neurosci 2:240-250.
Moyer JR, Jr., Thompson LT, Disterhoft JF (1996) Trace eyeblink conditioning increases CA1 excitability in a transient and learning-specific manner. J Neurosci 16:5536-5546.
Mozrzymas JW, Zarnowska ED, Pytel M, Mercik K (2003) Modulation of GABAA receptors by hydrogen ions reveals synaptic GABA transient and a crucial role of the desensitization process. J Neurosci 23:7981-7992.
Mulkey RM, Malenka RC (1992) Mechanisms underlying induction of homosynaptic long-term depression in area CA1 of the hippocampus. Neuron 9:967-975.
Munte S, Munte TF, Kuche H, Osthaus A, Herzog T, Heine J, Leuwer M, Piepenbrock S (2001) General anesthesia for interventional neuroradiology: propofol versus isoflurane. J Clin Anesth 13:186-192.
Nakahiro M, Yeh JZ, Brunner E, Narahashi T (1989) General anesthetics modulate GABA receptor channel complex in rat dorsal root ganglion neurons. Faseb J 3:1850-1854.
Navarro JF, Buron E, Martin-Lopez M (2002) Anxiogenic-like activity of L-655,708, a selective ligand for the benzodiazepine site of GABAA receptors which contain the α5 subunit, in the elevated plus-maze test. Prog Neuropsychopharmacol Biol Psychiatry 26:1389-1392.
Nelson N (1998) The family of Na+/Cl- neurotransmitter transporters. J Neurochem 71:1785- 1803.
Newell JG, Czajkowski C (2003) The GABAA receptor α1 subunit Pro174-Asp191 segment is involved in GABA binding and channel gating. J Biol Chem 278:13166-13172.
Nicoll RA, Schmitz D (2005) Synaptic plasticity at hippocampal mossy fibre synapses. Nat Rev Neurosci 6:863-876.
Nicoll RA, Tomita S, Bredt DS (2006) Auxiliary subunits assist AMPA-type glutamate receptors. Science 311:1253-1256.
276 277
Nishimura T, Schwarzer C, Gasser E, Kato N, Vezzani A, Sperk G (2005) Altered expression of GABAA and GABAB receptor subunit mRNAs in the hippocampus after kindling and electrically induced status epilepticus. Neuroscience 134:691-704.
Nusser Z, Mody I (2002) Selective modulation of tonic and phasic inhibitions in dentate gyrus granule cells. J Neurophysiol 87:2624-2628.
Nusser Z, Sieghart W, Somogyi P (1998) Segregation of different GABAA receptors to synaptic and extrasynaptic membranes of cerebellar granule cells. J Neurosci 18:1693-1703.
Nutt DJ, Little HJ, Taylor SC, Minchin MC (1984) Investigating benzodiazepine receptor function in vivo using an intravenous infusion of DMCM. Eur J Pharmacol 103:359-362.
Nutt DJ, Besson M, Wilson SJ, Dawson GR, Lingford-Hughes AR (2007) Blockade of alcohol's amnestic activity in humans by an α5 subtype benzodiazepine receptor inverse agonist. Neuropharmacology 53:810-820.
Nyiri G, Freund TF, Somogyi P (2001) Input-dependent synaptic targeting of α2-subunit- containing GABAA receptors in synapses of hippocampal pyramidal cells of the rat. Eur J Neurosci 13:428-442.
O'Dell TJ, Kandel ER, Grant SG (1991) Long-term potentiation in the hippocampus is blocked by tyrosine kinase inhibitors. Nature 353:558-560.
O'Keefe J (1993) Hippocampus, theta, and spatial memory. Curr Opin Neurobiol 3:917-924.
Obenaus A, Mody I, Baimbridge KG (1989) Dantrolene-Na (Dantrium) blocks induction of long-term potentiation in hippocampal slices. Neurosci Lett 98:172-178.
Ohnuma T, Augood SJ, Arai H, McKenna PJ, Emson PC (1999) Measurement of GABAergic parameters in the prefrontal cortex in schizophrenia: focus on GABA content, GABAA receptor α1 subunit messenger RNA and human GABA transporter-1 (HGAT-1) messenger RNA expression. Neuroscience 93:441-448.
Olsen RW, Sieghart W (2008) International Union of Pharmacology. LXX. Subtypes of γ- aminobutyric acidA receptors: classification on the basis of subunit composition, pharmacology, and function. Update. Pharmacol Rev 60:243-260.
Olton DS, Samuelsson RJ (1976) Remembrance of places passed: Spatial memory in rats. J Exp Psych 2:97-116.
Orser BA, Wang LY, Pennefather PS, MacDonald JF (1994) Propofol modulates activation and desensitization of GABAA receptors in cultured murine hippocampal neurons. J Neurosci 14:7747-7760.
Osterman JE, Hopper J, Heran WJ, Keane TM, van der Kolk BA (2001) Awareness under anesthesia and the development of posttraumatic stress disorder. Gen Hosp Psychiatry 23:198-204.
277 278
Otis TS, Staley KJ, Mody I (1991) Perpetual inhibitory activity in mammalian brain slices generated by spontaneous GABA release. Brain Res 545:142-150.
Papadimitriou GN, Dikeos DG, Karadima G, Avramopoulos D, Daskalopoulou EG, Stefanis CN (2001) GABAA receptor β3 and α5 subunit gene cluster on chromosome 15q11-q13 and bipolar disorder: a genetic association study. Am J Med Genet 105:317-320.
Parra P, Gulyas AI, Miles R (1998) How many subtypes of inhibitory cells in the hippocampus? Neuron 20:983-993.
Pearce RA (1993) Physiological evidence for two distinct GABAA responses in rat hippocampus. Neuron 10:189-200.
Peng Z, Huang CS, Stell BM, Mody I, Houser CR (2004) Altered expression of the delta subunit of the GABAA receptor in a mouse model of temporal lobe epilepsy. J Neurosci 24:8629-8639.
Perriol MP, Dujardin K, Derambure P, Marcq A, Bourriez JL, Laureau E, Pasquier F, Defebvre L, Destee A (2005) Disturbance of sensory filtering in dementia with Lewy bodies: comparison with Parkinson's disease dementia and Alzheimer's disease. J Neurol Neurosurg Psychiatry 76:106-108.
Petersen EN (1983) DMCM: a potent convulsive benzodiazepine receptor ligand. Eur J Pharmacol 94:117-124.
Phillips RG, LeDoux JE (1992) Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behav Neurosci 106:274-285.
Philpot BD, Cho KK, Bear MF (2007) Obligatory Role of NR2A for Metaplasticity in Visual Cortex. Neuron 53:495-502.
Pirker S, Schwarzer C, Wieselthaler A, Sieghart W, Sperk G (2000) GABAA receptors: immunocytochemical distribution of 13 subunits in the adult rat brain. Neuroscience 101:815-850.
Poikolainen K (1984) Estimated lethal ethanol concentrations in relation to age, aspiration, and drugs. Alcohol Clin Exp Res 8:223-225.
Poulter MO, Brown LA, Tynan S, Willick G, William R, McIntyre DC (1999) Differential expression of α1, α2, α3, and α5 GABAA receptor subunits in seizure-prone and seizure- resistant rat models of temporal lobe epilepsy. J Neurosci 19:4654-4661.
Prenosil GA, Schneider Gasser EM, Rudolph U, Keist R, Fritschy J-M, Vogt KE (2006) Specific Subtypes of GABAA Receptors Mediate Phasic and Tonic Forms of Inhibition in Hippocampal Pyramidal Neurons. J Neurophysiol 96:846-857.
Price CC, Garvan CW, Monk TG (2008) Type and severity of cognitive decline in older adults after noncardiac surgery. Anesthesiology 108:8-17.
278 279
Priebe NJ, Ferster D (2002) A new mechanism for neuronal gain control (or how the gain in brains has mainly been explained). Neuron 35:602-604.
Pritchett DB, Seeburg PH (1990) γ-aminobutyric acid A receptor α5-subunit creates novel type II benzodiazepine receptor pharmacology. J Neurochem 54:1802-1804.
Quinn JJ, Oommen SS, Morrison GE, Fanselow MS (2002) Post-training excitotoxic lesions of the dorsal hippocampus attenuate forward trace, backward trace, and delay fear conditioning in a temporally specific manner. Hippocampus 12:495-504.
Quinn JJ, Loya F, Ma QD, Fanselow MS (2005) Dorsal hippocampus NMDA receptors differentially mediate trace and contextual fear conditioning. Hippocampus 15:665-674.
Quinn JJ, Wied HM, Ma QD, Tinsley MR, Fanselow MS (2008) Dorsal hippocampus involvement in delay fear conditioning depends upon the strength of the tone-footshock association. Hippocampus 18:640-654.
Quirk K, Blurton P, Fletcher S, Leeson P, Tang F, Mellilo D, Ragan CI, McKernan RM (1996) 3 [ H]L-655,708, a novel ligand selective for the benzodiazepine site of GABAA receptors which contain the α5 subunit. Neuropharmacology 35:1331-1335.
Radian R, Ottersen OP, Storm-Mathisen J, Castel M, Kanner BI (1990) Immunocytochemical localization of the GABA transporter in rat brain. J Neurosci 10:1319-1330.
Rau V, Oh I, Laster M, Eger EI, 2nd, Fanselow MS (2009) Isoflurane suppresses stress-enhanced fear learning in a rodent model of post-traumatic stress disorder. Anesthesiology 110:487-495.
Reeves JG GP, Lubarsky DA, McEvoy MD (2005) Intravenous nonopioid anesthetics. In: Miller's anesthesia, 6th Edition (Miller RD, ed), pp 317–378. Philadelphia, Pa.: Elsevier Churchill Livingstone.
Reynolds DS, Rosahl TW, Cirone J, O'Meara GF, Haythornthwaite A, Newman RJ, Myers J, Sur C, Howell O, Rutter AR, Atack J, Macaulay AJ, Hadingham KL, Hutson PH, Belelli D, Lambert JJ, Dawson GR, McKernan R, Whiting PJ, Wafford KA (2003) Sedation and anesthesia mediated by distinct GABAA receptor isoforms. J Neurosci 23:8608-8617.
Richerson GB, Wu Y (2003) Dynamic equilibrium of neurotransmitter transporters: not just for reuptake anymore. J Neurophysiol 90:1363-1374.
Rivera C, Voipio J, Payne JA, Ruusuvuori E, Lahtinen H, Lamsa K, Pirvola U, Saarma M, Kaila K (1999) The K+/Cl- co-transporter KCC2 renders GABA hyperpolarizing during neuronal maturation. Nature 397:251-255.
Ron D (2004) Signaling cascades regulating NMDA receptor sensitivity to ethanol. Neuroscientist 10:325-336.
279 280
Ruano D, Letang V, Biton B, Avenet P, Benavides J, Scatton B, Vitorica J (2000) Subunit composition of rat ventral spinal cord GABAA receptors, assessed by single cell RT- multiplex PCR. Neuroreport 11:3169-3173.
Rudolph U, Mohler H (2004) Analysis of GABAA receptor function and dissection of the pharmacology of benzodiazepines and general anesthetics through mouse genetics. Annu Rev Pharmacol Toxicol 44:475-498.
Rudolph U, Antkowiak B (2004) Molecular and neuronal substrates for general anaesthetics. Nat Rev Neurosci 5:709-720.
Rudolph U, Crestani F, Benke D, Brunig I, Benson JA, Fritschy JM, Martin JR, Bluethmann H, Mohler H (1999) Benzodiazepine actions mediated by specific γ-aminobutyric acidA receptor subtypes. Nature 401:796-800.
Ruppersberg JP, Mosbacher J, Gunther W, Schoepfer R, Fakler B (1993) Studying block in cloned N-methyl-D-aspartate (NMDA) receptors. Biochem Pharmacol 46:1877-1885.
Samuelsson P, Brudin L, Sandin RH (2007) Late psychological symptoms after awareness among consecutively included surgical patients. Anesthesiology 106:26-32.
Sanders MJ, Wiltgen BJ, Fanselow MS (2003) The place of the hippocampus in fear conditioning. Eur J Pharmacol 463:217-223.
Sandin RH, Enlund G, Samuelsson P, Lennmarken C (2000) Awareness during anaesthesia: a prospective case study. Lancet 355:707-711.
Santhakumar V, Wallner M, Otis TS (2007) Ethanol acts directly on extrasynaptic subtypes of GABAA receptors to increase tonic inhibition. Alcohol 41:211-221.
Sarter M, Bruno JP, Berntson GG (2001) Psychotogenic properties of benzodiazepine receptor inverse agonists. Psychopharmacology (Berl) 156:1-13.
Savic MM, Huang S, Furtmuller R, Clayton T, Huck S, Obradovic DI, Ugresic ND, Sieghart W, Bokonjic DR, Cook JM (2008) Are GABAA receptors containing α5 subunits contributing to the sedative properties of benzodiazepine site agonists? Neuropsychopharmacology 33:332-339.
Scanziani M (2000) GABA spillover activates postsynaptic GABAB receptors to control rhythmic hippocampal activity. Neuron 25:673-681.
Schoups A, Vogels R, Qian N, Orban G (2001) Practising orientation identification improves orientation coding in V1 neurons. Nature 412:549-553.
Scimemi A, Schorge S, Kullmann DM, Walker MC (2006) Epileptogenesis is associated with enhanced glutamatergic transmission in the perforant path. J Neurophysiol 95:1213-1220.
Scimemi A, Semyanov A, Sperk G, Kullmann DM, Walker MC (2005) Multiple and Plastic Receptors Mediate Tonic GABAA Receptor Currents in the Hippocampus. J Neurosci 25:10016-10024.
280 281
Scoville WB, Milner B (1957) Loss of recent memory after bilateral hippocampal lesions. J Neurol Neurosurg Psychiatry 20:11-21.
Seabrook GR, Easter A, Dawson GR, Bowery BJ (1997) Modulation of long-term potentiation in CA1 region of mouse hippocampal brain slices by GABAA receptor benzodiazepine site ligands. Neuropharmacology 36:823-830.
Sebel PS, Bowdle TA, Ghoneim MM, Rampil IJ, Padilla RE, Gan TJ, Domino KB (2004) The incidence of awareness during anesthesia: a multicenter United States study. Anesth Analg 99:833-839.
Selig DK, Hjelmstad GO, Herron C, Nicoll RA, Malenka RC (1995) Independent mechanisms for long-term depression of AMPA and NMDA responses. Neuron 15:417-426.
Semyanov A, Walker MC, Kullmann DM (2003) GABA uptake regulates cortical excitability via cell type-specific tonic inhibition. Nat Neurosci 6:484-490.
Semyanov A, Walker MC, Kullmann DM, Silver RA (2004) Tonically active GABAA receptors: modulating gain and maintaining the tone. Trends Neurosci 27:262-269.
Serwanski DR, Miralles CP, Christie SB, Mehta AK, Li X, De Blas AL (2006) Synaptic and nonsynaptic localization of GABAA receptors containing the α5 subunit in the rat brain. The Journal of Comparative Neurology 499:458-470.
Shimada S, Cutting G, Uhl GR (1992) γ-Aminobutyric acid A or C receptor? γ-Aminobutyric acid ρ1 receptor RNA induces bicuculline-, barbiturate-, and benzodiazepine-insensitive gamma-aminobutyric acid responses in Xenopus oocytes. Mol Pharmacol 41:683-687.
Sieghart W (1989) Multiplicity of GABAA--benzodiazepine receptors. Trends Pharmacol Sci 10:407-411.
Sieghart W, Sperk G (2002) Subunit composition, distribution and function of GABAA receptor subtypes. Curr Top Med Chem 2:795-816.
Sigel E (2002) Mapping of the benzodiazepine recognition site on GABAA receptors. Curr Top Med Chem 2:833-839.
Sigurdsson T, Doyere V, Cain CK, LeDoux JE (2007) Long-term potentiation in the amygdala: a cellular mechanism of fear learning and memory. Neuropharmacology 52:215-227.
Silva AJ, Stevens CF, Tonegawa S, Wang Y (1992a) Deficient hippocampal long-term potentiation in alpha-calcium-calmodulin kinase II mutant mice. Science 257:201-206.
Silva AJ, Paylor R, Wehner JM, Tonegawa S (1992b) Impaired spatial learning in alpha- calcium-calmodulin kinase II mutant mice. Science 257:206-211.
Silva AJ, Kogan JH, Frankland PW, Kida S (1998) CREB and memory. Annu Rev Neurosci 21:127-148.
281 282
Simon W, Hapfelmeier G, Kochs E, Zieglgansberger W, Rammes G (2001) Isoflurane blocks synaptic plasticity in the mouse hippocampus. Anesthesiology 94:1058-1065.
Sine SM, Engel AG (2006) Recent advances in Cys-loop receptor structure and function. Nature 440:448-455.
Sipila ST, Huttu K, Soltesz I, Voipio J, Kaila K (2005) Depolarizing GABA acts on intrinsically bursting pyramidal neurons to drive giant depolarizing potentials in the immature hippocampus. J Neurosci 25:5280-5289.
Smith AJ, Alder L, Silk J, Adkins C, Fletcher AE, Scales T, Kerby J, Marshall G, Wafford KA, McKernan RM, Atack JR (2001) Effect of α subunit on allosteric modulation of ion channel function in stably expressed human recombinant γ-aminobutyric acid subtype A receptors determined using (36)Cl ion flux. Mol Pharmacol 59:1108-1118.
Smith SS, Shen H, Gong QH, Zhou X (2007) Neurosteroid regulation of GABAA receptors: Focus on the α4 and δ subunits. Pharmacol Ther 116:58-76.
Sonner JM, Cascio M, Xing Y, Fanselow MS, Kralic JE, Morrow AL, Korpi ER, Hardy S, Sloat B, Eger EI, 2nd, Homanics GE (2005) α1 subunit-containing GABA type A receptors in forebrain contribute to the effect of inhaled anesthetics on conditioned fear. Mol Pharmacol 68:61-68.
Sperk G, Schwarzer C, Tsunashima K, Fuchs K, Sieghart W (1997) GABAA receptor subunits in the rat hippocampus I: immunocytochemical distribution of 13 subunits. Neuroscience 80:987-1000.
Staley KJ, Mody I (1992) Shunting of excitatory input to dentate gyrus granule cells by a depolarizing GABAA receptor-mediated postsynaptic conductance. J Neurophysiol 68:197-212.
Standaert DG, Landwehrmeyer GB, Kerner JA, Penney JB, Jr., Young AB (1996) Expression of NMDAR2D glutamate receptor subunit mRNA in neurochemically identified interneurons in the rat neostriatum, neocortex and hippocampus. Brain Res Mol Brain Res 42:89-102.
Steinmetz J, Christensen KB, Lund T, Lohse N, Rasmussen LS (2009) Long-term consequences of postoperative cognitive dysfunction. Anesthesiology 110:548-555.
Stell BM, Brickley SG, Tang CY, Farrant M, Mody I (2003) Neuroactive steroids reduce neuronal excitability by selectively enhancing tonic inhibition mediated by δ subunit- containing GABAA receptors. Proc Natl Acad Sci U S A 100:14439-14444.
Stephens DN, Pistovcakova J, Worthing L, Atack JR, Dawson GR (2005) Role of GABAA α5- containing receptors in ethanol reward: the effects of targeted gene deletion, and a selective inverse agonist. Eur J Pharmacol 526:240-250.
Steriade M, Amzica F (1999) Intracellular study of excitability in the seizure-prone neocortex in vivo. J Neurophysiol 82:3108-3122.
282 283
Sur C, Quirk K, Dewar D, Atack J, McKernan R (1998) Rat and human hippocampal α5 subunit- containing γ-aminobutyric AcidA receptors have α5 β3 γ2 pharmacological characteristics. Mol Pharmacol 54:928-933.
Sur C, Fresu L, Howell O, McKernan RM, Atack JR (1999) Autoradiographic localization of α5 subunit-containing GABAA receptors in rat brain. Brain Res 822:265-270.
Suzuki WA (1999) The long and the short of it: memory signals in the medial temporal lobe. Neuron 24:295-298.
Suzuki WA (2006) Encoding new episodes and making them stick. Neuron 50:19-21.
Sweatt JD (2003) Mechanisms of memory. San Diego, Calif.: Academic Press.
Swerdlow NR, Weber M, Qu Y, Light GA, Braff DL (2008) Realistic expectations of prepulse inhibition in translational models for schizophrenia research. Psychopharmacology (Berl) 199:331-388.
Swope SL, Moss SJ, Raymond LA, Huganir RL (1999) Regulation of ligand-gated ion channels by protein phosphorylation. Adv Second Messenger Phosphoprotein Res 33:49-78.
Takahashi H, Iwase M, Ishii R, Ohi K, Fukumoto M, Azechi M, Ikezawa K, Kurimoto R, Canuet L, Nakahachi T, Iike N, Tagami S, Morihara T, Okochi M, Tanaka T, Kazui H, Yoshida T, Tanimukai H, Yasuda Y, Kudo T, Hashimoto R, Takeda M (2008) Impaired prepulse inhibition and habituation of acoustic startle response in Japanese patients with schizophrenia. Neurosci Res 62:187-194.
Takeuchi A, Takeuchi N (1971) Anion interaction at the inhibitory post-synaptic membrane of the crayfish neuromuscular junction. J Physiol 212:337-351.
Tao YH, Yuan Z, Tang XQ, Xu HB, Yang XL (2006) Inhibition of GABA shunt enzymes' activity by 4-hydroxybenzaldehyde derivatives. Bioorg Med Chem Lett 16:592-595.
Tardin C, Cognet L, Bats C, Lounis B, Choquet D (2003) Direct imaging of lateral movements of AMPA receptors inside synapses. Embo J 22:4656-4665.
Thomson AM, Bannister AP, Hughes DI, Pawelzik H (2000) Differential sensitivity to Zolpidem of IPSPs activated by morphologically identified CA1 interneurons in slices of rat hippocampus. Eur J Neurosci 12:425-436.
Toth K, McBain CJ (1998) Afferent-specific innervation of two distinct AMPA receptor subtypes on single hippocampal interneurons. Nat Neurosci 1:572-578.
Tovar KR, Westbrook GL (1999) The incorporation of NMDA receptors with a distinct subunit composition at nascent hippocampal synapses in vitro. J Neurosci 19:4180-4188.
Towers SK, Gloveli T, Traub RD, Driver JE, Engel D, Fradley R, Rosahl TW, Maubach K, Buhl TLEH, Whittington MA (2004) α5 subunit-containing GABAA receptors affect the dynamic range of mouse hippocampal kainate-induced gamma frequency oscillations in vitro. J Physiol (Lond) 559:721-728.
283 284
Triller A, Choquet D (2005) Surface trafficking of receptors between synaptic and extrasynaptic membranes: and yet they do move! Trends Neurosci 28:133-139.
Tsien JZ, Huerta PT, Tonegawa S (1996) The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell 87:1327-1338.
Tsumoto T (1993) Long-term depression in cerebral cortex: a possible substrate of "forgetting" that should not be forgotten. Neurosci Res 16:263-270.
Uchida I, Kamatchi G, Burt D, Yang J (1995) Etomidate potentiation of GABAA receptor gated current depends on the subunit composition. Neurosci Lett 185:203-206. van der Linden JA, Joels M, Karst H, Juta AJ, Wadman WJ (1993) Bicuculline increases the intracellular calcium response of CA1 hippocampal neurons to synaptic stimulation. Neurosci Lett 155:230-233. van Rijnsoever C, Tauber M, Choulli MK, Keist R, Rudolph U, Mohler H, Fritschy JM, Crestani F (2004) Requirement of α5-GABAA receptors for the development of tolerance to the sedative action of diazepam in mice. J Neurosci 24:6785-6790.
Veliskova J, Velisek L, Mares P, Rokyta R (1990) Ketamine suppresses both bicuculline- and picrotoxin-induced generalized tonic-clonic seizures during ontogenesis. Pharmacol Biochem Behav 37:667-674.
Veselis RA, Pryor KO, Reinsel RA, Li Y, Mehta M, Johnson R, Jr. (2009) Propofol and midazolam inhibit conscious memory processes very soon after encoding: an event- related potential study of familiarity and recollection in volunteers. Anesthesiology 110:295-312.
Vilardaga JP, Steinmeyer R, Harms GS, Lohse MJ (2005) Molecular basis of inverse agonism in a G protein-coupled receptor. Nat Chem Biol 1:25-28.
Wagner DA, Czajkowski C (2001) Structure and dynamics of the GABA binding pocket: A narrowing cleft that constricts during activation. J Neurosci 21:67-74.
Wakamori M, Ikemoto Y, Akaike N (1991) Effects of two volatile anesthetics and a volatile convulsant on the excitatory and inhibitory amino acid responses in dissociated CNS neurons of the rat. J Neurophysiol 66:2014-2021.
Walker MC, Semyanov A (2008) Regulation of excitability by extrasynaptic GABAA receptors. Results Probl Cell Differ 44:29-48.
Wang DD, Kriegstein AR (2009) Defining the Role of GABA in Cortical Development. J Physiol.
Wanisch K, Tang J, Mederer A, Wotjak CT (2005) Trace fear conditioning depends on NMDA receptor activation and protein synthesis within the dorsal hippocampus of mice. Behav Brain Res 157:63-69.
284 285
Wei H, Xiong W, Yang S, Zhou Q, Liang C, Zeng BX, Xu L (2002) Propofol facilitates the development of long-term depression (LTD) and impairs the maintenance of long-term potentiation (LTP) in the CA1 region of the hippocampus of anesthetized rats. Neurosci Lett 324:181-184.
Wei W, Faria LC, Mody I (2004) Low Ethanol Concentrations Selectively Augment the Tonic Inhibition Mediated by δ Subunit-Containing GABAA Receptors in Hippocampal Neurons. J Neurosci 24:8379-8382.
Wei W, Zhang N, Peng Z, Houser CR, Mody I (2003) Perisynaptic localization of δ subunit- containing GABAA receptors and their activation by GABA spillover in the mouse dentate gyrus. J Neurosci 23:10650-10661.
White AM, Matthews DB, Best PJ (2000) Ethanol, memory, and hippocampal function: a review of recent findings. Hippocampus 10:88-93.
White AM, Elek TM, Beltz TL, Best PJ (1998) Spatial performance is more sensitive to ethanol than nonspatial performance regardless of cue proximity. Alcohol Clin Exp Res 22:2102- 2107.
Whiting PJ (2003) GABAA receptor subtypes in the brain: a paradigm for CNS drug discovery? Drug Discov Today 8:445-450.
Whitlock JR, Heynen AJ, Shuler MG, Bear MF (2006) Learning induces long-term potentiation in the hippocampus. Science 313:1093-1097.
Whittington MA, Traub RD (2003) Interneuron diversity series: inhibitory interneurons and network oscillations in vitro. Trends Neurosci 26:676-682.
Whittington MA, Jefferys JG, Traub RD (1996) Effects of intravenous anaesthetic agents on fast inhibitory oscillations in the rat hippocampus in vitro. Br J Pharmacol 118:1977-1986.
Wigstrom H, Gustafsson B (1983) Facilitated induction of hippocampal long-lasting potentiation during blockade of inhibition. Nature 301:603-604.
Wikstrom MA, Matthews P, Roberts D, Collingridge GL, Bortolotto ZA (2003) Parallel kinase cascades are involved in the induction of LTP at hippocampal CA1 synapses. Neuropharmacology 45:828-836.
Wilson CJ (1987) Morphology and synaptic connections of crossed corticostriatal neurons in the rat. J Comp Neurol 263:567-580.
Wiltgen BJ, Sanders MJ, Ferguson C, Homanics GE, Fanselow MS (2005) Trace fear conditioning is enhanced in mice lacking the δ subunit of the GABAA receptor. Learn Mem 12:327-333.
Wiltgen BJ, Sanders MJ, Anagnostaras SG, Sage JR, Fanselow MS (2006) Context fear learning in the absence of the hippocampus. J Neurosci 26:5484-5491.
285 286
Wisden W, Cope D, Klausberger T, Hauer B, Sinkkonen ST, Tretter V, Lujan R, Jones A, Korpi ER, Mody I, Sieghart W, Somogyi P (2002) Ectopic expression of the GABAA receptor α6 subunit in hippocampal pyramidal neurons produces extrasynaptic receptors and an increased tonic inhibition. Neuropharmacology 43:530-549.
Wu J, Harata N, Akaike N (1996) Potentiation by sevoflurane of the γ-aminobutyric acid- induced chloride current in acutely dissociated CA1 pyramidal neurones from rat hippocampus. Br J Pharmacol 119:1013-1021.
Yamashita M, Marszalec W, Yeh JZ, Narahashi T (2006) Effects of ethanol on tonic GABA currents in cerebellar granule cells and mammalian cells recombinantly expressing GABAA receptors. J Pharmacol Exp Ther 319:431-438.
Yee BK, Hauser J, Dolgov VV, Keist R, Mohler H, Rudolph U, Feldon J (2004) GABA receptors containing the α5 subunit mediate the trace effect in aversive and appetitive conditioning and extinction of conditioned fear. Eur J Neurosci 20:1928-1936.
Yeung JY, Canning KJ, Zhu G, Pennefather P, MacDonald JF, Orser BA (2003) Tonically activated GABAA receptors in hippocampal neurons are high-affinity, low-conductance sensors for extracellular GABA. Mol Pharmacol 63:2-8.
York A, von Fraunhofer N, Turk J, Sedgwick P (1999) Fragile-X syndrome, Down's syndrome and autism: awareness and knowledge amongst special educators. J Intellect Disabil Res 43 ( Pt 4):314-324.
Zarnowska ED, Keist R, Rudolph U, Pearce RA (2009) GABAA Receptor α5 Subunits Contribute to GABAA, slow Synaptic Inhibition in Mouse Hippocampus. J Neurophysiol 101:1179-1191.
Zhang N, Wei W, Mody I, Houser CR (2007) Altered localization of GABAA receptor subunits on dentate granule cell dendrites influences tonic and phasic inhibition in a mouse model of epilepsy. J Neurosci 27:7520-7531.
286