-
PROPOFOL AND BENZODIAZEPINE MODULATION OF GABA,R FUNCTION
Laura Catherine McAdam
A thesis submitted in conformity with the requirements for the
degree of Master of Science Graduate Department of Physiology
University of Toronto
G Copyright by Laura Catherine McAdam (1 997)
-
National Library m*m of Canada Bibliothèque nationale du Canada
Acquisitions and Acquisitions et Bibliographic Senrices services
bibliographiques 395 Wellington Street 395, rue Wellington Ottawa
ON KIA ON4 ûttawa ON K 1 A ON4 Canada Canada
Your Me Votre reIsrmœ
Our Ne Noire reftirencB
The author has granted a non- L'auteur a accordé une Licence non
exclusive licence allowing the exclusive permettant à la National
Library of Canada to Bibliothèque nationale du Canada de reproduce,
han, distribute or sell reproduire, prêter, distribuer ou copies of
tbis thesis in microform, vendre des copies de cette thèse sous
paper or electronic formats. la forme de microfiche/film, de
reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la
propriété du copyright in this thesis. Neither the droit d'auteur
qui protège cette thèse. thesis nor substantial extracts fkom it Ni
la thèse ni des extraits substantiels may be printed or otherwise
de celle-ci ne doivent être imprimés reproduced without the
author's ou autrement reproduits sans son permission.
autorisation.
-
A. ABSTRACT
PROPOFOL AND BENZODIAZEPM MODULATION OF GABAAR FUNCTION
Master of Science, 1997
Laura Catherine McAdam
Department of Physiology, University of Toronto
The general anaesthetic, propo fol, has multiple effects on the
GABA,R: it potentiates
GABA-evoked responses; it activates the receptor, and it alters
the kinetics of receptor
desensitization. In contrast, sedative benzodiazepines only
potentiate GABA-evoked responses.
It is not known if the multiple effects of propofol on the
GABA,R are mediated by distinct
binding sites or which effect accounts for propofol's
anaesthetic properties. The whole-ceIl patch
clamp method was used to study the effects of benzodiazepines
and propofol on GABA,R's
present in cultured, murine hippocarnpal neurons.
The concentration of propofol that activated the GABA,R (EC,, =
24.3 k 3.4 PM) was
significantly greaeater than that which decreased receptor
desensitization (EC,, = 1.15 + 0.33 PM). Furthemore. midazolam
potentiated propofol-induced currents but did not alter
propofol-
induced changes in receptor desensitization, suggesting that
receptor activation and
desensitization were mediated by distinct mechmisms.
Low concentrations of propofol (1 PM) and midazolam (0.5 plu)
interacted in a
superadditive rnanner to enhance GABA (0.3 PM)-evoked responses
and isobolographic analysis
suggested a synergistic interaction. Ln contrast, higher
concentrations of these dmgs produced an
additive interaction. As low concentrations of propofol and
midazolam interact synergistically to
produce hypnosis and higher concentrations do not interact
synergistically to induce anaesthesia,
Our results suggest that hypnosis and anaesthesia may be
mediated by distinct changes in
GABAAR hnction.
-
B. ACKNOWLEDGMENTS
1 wish to express my sincere gratitude to my supervisors, Dr.
B.A. Oner and Dr. J.F.
MacDonald for their guidance and encouragement over the past two
years. 1 would also like to
thank Dr. Orser for introducing me to the clinical aspects of
research and for the opportunities to
visit Sunnybrook Health Science Centre. I also appreciate the
advice provided by my cornmittee
member, Dr. Wojtowicz. 1 would like to thank my labrnates for
providing a stimulating, helphl
and fnendly working atmosphere. I also gratefully acknowledge
Lidia Brandes and Ella
Czerwinska for overseeing the ce11 cultures. Finally, and most
importantly. 1 would like to thank
my parents and sisters for al1 of their love and support
throughout the paçt Z years.
This research was supported by an International Anesthetists
Research Society Frontiers
in Anesthesia Research Grant Award to Dr. Orser.
-
C . TABLE OF CONTENTS
A. ABSTRACT
B. ACLN0 WLEDGiMENTS
C. TABLE OF CONTENTS
D. LIST OF FIGURES
E. LIST OF TABLES
F. ABBREVIATION N D E X
t NTRODUCTION
.LVAESTHESIA
NON-SPECIFIC AND SPECEIC THEORJES OF ANAESTHETIC
ACTION
N I B I T O R Y NEUROTRANSMISSION AND GABA, RECEPTORS
GABA
GN3A,R PROPERTIES
KMETIC MODEL OF GABA,R FUNCTION
DESENSITIZATION
S U B W T S OF THE GABA,R
GABAAR AND ANAESTHESIA
PROPOFOL EFFECTS ON GABAAR
BENZODIAZEPNE ACTION AT THE GABA,R
-. I I
-. . 111
i v
... V l l l
i x
X
-
1 -4.1. BENZODIAZEPINE B W W G IS NFLüENCED BY SUBUNIT
COMPOSITION OF THE GABA,+R
1.5. DRUG INTERACTIONS BETWEEN PROPOFOL AND
BENZODIAZEPMS
1 S. 1 DETERMNTNG DRUG NTERACTIONS
7 - . OBJECTIVES AND WORKING HYPOTHESIS
3.1. SECTION 1
3 9 - . - . SECTION 3
LMETHODS
CELL CULTURES
RECORDNG PIPETTES
PIPETTE SOLUTION
WHOLE-CELL VOLTAGE-CLAMP RECORDINGS
DRL'G Ai' AGONIST PERFUSION SYSTEM
WHOLE-CELL CURRENT AYALYSIS
DRUG INTERACTION ANALYSIS
DRUGS AND OTHER CHEMICALS
STATISTICS
-
4. RESC'LTS
4.1. SECTION 1
il. 1.1. MIDAZOLAM MCREASED THE AFFNTY OF THE GABA,R FOR
GABA -4ND PROPOFOL 37
3.1.2. MIDAZOLALM DOES NOT INFLUENCE DESENSITIZATION OF THE
GABA,R 47
3.1.3. PROPOFOL DECREASED RECEPTOR DESENSITIZATION 50
4.1 .4. PROPOFOL-NUCED .MODULATION OF GABA,R 55
4.2. SECTION 2 64
3 . 1 . EFFECTS OF MIDAZOLAM AND PROPOFOL ON CC'RREXTS
EVOKED BY SUBSATURATNG CONCENTRATIONS OF GABA 64
1.2.2. SYNERGISTIC NI'ERACTION BETWEEN PROPOFOL AiÿD
MIDAZOLAM IN THE PRESENCE OF LOW CONCENTR4TIONS OF
GABA 77
5. DISCUSSION 88
5.1. SECTION 1 88
5.1.1. MIDAZOLAM POTENTUTION OF GABA-EVOKED CURRENTS 88
5.1.2. ,MIDAZOLAM POTENTlATION OF PROPOFOL-EVOKED
Cb'RRENTS 89
5.1.3. PROPOFOL BLOCKA.DE OF THE GABA,,R 92
-
5 4 MID.LVOLAM DiD NOT I'CTLLT'YCE DESENSITIZ.4TIO-Y OF THE
G.ABX,R 93
5.1 3. PROPOFOL MODLZXïION OF GrZBA,R DESENSITIZATIOX 94
- 9 3 .-. SECTION 2 97
5 . 1 . PH.4K'.'ACOLOGICAL SYNERGISM BETWEEX PROPOFOL .%.ND
'tlIDAZOL-X.kl 97
5 . 2 CLiS1C.U SYXERGISM BETWEE'V PROPOFOL .k\D .MID=VOL.I.f
98
6 . CONCLCSIONS
-7
. . REFEREXCES
-
D. LIST OF FIGURES
GN3Aergic synapses and the topology of a G.4BA,R subunit
?/lultibarrel fast-perfusion system
Desensitization of GABA induced currents
Isobolographic anal ysis
Midazolam increased the affinity of the GABA,R for GXBA
.Midazolam increased the affinity of the GABA,R for propofol
Midazolam potentiated currents evoked by propofol
The W relationship of propofol-induced currents recorded in the
absence and
presence of midazolarn
Midazolam did not influence GABA,R desensitization of the GABA,R
for
currents evoked by GABA
.Midamlm did not aiter desensitization of propo fol-evoked
currents
Propo fo 1 decreased GN3A ,R desensitization in a
concentration-dependent
mann er
Midazo lam did not e ffect propo fol-induced modulation of GAB
A,R
desensitization
Midazolam did not influence GABA,R desensitization at lower
concentrations
of D ~ O D O ~ O ~
-
Propofol and midazolam (or flurazepam) produced an additive
enhancement of
GABA (3 p.M)-evoked currents From hippocampal neurons 65
Widazolam and propofol did not alter the reversa1 potential of
the GABA-evoked
currents 68
Propofol and midazolam produced an additive enhancernent of GABA
(3 pu)-
evoked currents when the concentration of propofol was reduced 7
1
Propofol and midazolam produced a superadditive enhancement of
G.4E5.4 (3
uM)-evoked currents when the concentration of GABA was reduced
73
Propofol and midazolam produced an additive enhancement of
G.îB.4 ( 3 ~ ~ 3 4 ) -
evoked currents in spinal cord neurons
'vlidazolam potentiation of GABA ( 1 ,uM)-evoked response
Propofol potentiation of GABA ( 1 FM)-evoked response
Propofol enhancement of currents evoked by co-application of
midazolam and
G.%l3-4
Synergisric interactions between propofol and midazolarn
E. LIST OF TABLES
1. Propofoi increased the rate of GAB.4,R desensitization and
deactivation at 56
saturating concentrations of GABA
7 - - Widazolam increased the rate of GAB.A4R deactivation at
two concentrations of 63 propofol in the presence of saturating
concentrations of G M . 4
-
F. .U3BREVIATION INDEX
. M P A
ATP
BDZ
BZ 1
BZ2
CDPX
CNS
DMSO
DRC
E G O
EGTA
FLU
FMZ
G-proteins
GABA
GABA,R
GABA,R
GAi3bR
GAD
HEPES
a-amino-3-hydroxyl-5-methyl-4-isoxazole propionic acid
adenosine-5-triphosphate
benzodiazepine
benzodiazepine type 1
benzodiazepine type 2
chlordiazepoxide
central nenious system
dimethyt sulfoxide
dose response curve
concentration that produces 50 % of the observed response
ehylene glycol-0,0'-bis(2-aminoethy1)-N,N,N,N',N7-te~~cetc
acid
flurazepam
flumazenil
guanine nucleotide binding protein
7-aminobutync acid
a subtype of y-arninobutyric acid receptor
a subtype of y-arninobutyric acid receptor
a subtype of 7-aminobutyric acid receptor
glutamic acid decarboxylase
25 N-2-hydroxy-ethylpiperazine-N'-2-ethanesulphoic acid
-
IC,,
IPSC
KA
MDZ
MEM
nAChR
NMDA
'CMDAR
PA
PRO
P s
TE,4
TM
TTX
concentration that inhibits the response by 50 %
inhibitory post-synaptic current
kainate
midazo larn
minimum essential media
nicotinic acety lcholine recep tor
N-methyl-D-aspartate
NMDA receptor
p icoamperes
propo fol
picosiemens
tetraethylarnrnonium
transmem brane domain
tetrodotoxin
-
1. INTRODUCTION
1.1. ANAESTHESIA
haesthesia is a complex phenornenon defined as a behavioural
state associated with the
loss of awareness and absence of pain (Tanelian. 1993). The
mechanisms of dmg action that
produce general anaesthesia are not well understood despite the
use of these drugs for over 200
years (Harris et al.. 1995). It is commonly thought that general
anaesthetics alter synaptic
transmission rather than inhibit the propagation of impulses
along the length of the nerve fiber
(Franks and Lieb, 1994). Aithough some anaesthetics decrease the
synthesis, uptake and release
of neurotransmitters, evidence for major presynaptic target
sites is minimal (Tanelian et al.,
1993). Anaesthetics are thought to pnmady influence receptors
present in the post-synaptic
membrane (Franks and Lieb. 1994). The y-aminobutyric acid type,,
receptor (GABA,R) is
thought to play an important role in mediating the behavioural
effects of intravenous
anaesthetics including propofol. barbiturates, several
neuroactive steroids and sedative
benzodiazepines (Franks and Lieb. 1994; Tanelian et al.. 1993;
Lin et al.. 1993). The focus of
this thesis is the interactions of propofol and benzodiazepines
on GABA,\R hnction.
1.1.1. NON-SPECIFIC AND SPECIFIC THEORTES OF ANAESTHETIC
ACTION
Two general theories of anaesthetic action have evolved; the
unitary theory, whereby
anaesthetic dmgs mediate their effect by a common mechanism, and
a theory in which different
dmgs act by specific sites of action.
The great diversity in structure of anaesthetic drugs suggests a
nonspecific mechanism of
action. At the tum of the century, Meyer and Overton noted that
the potency of an anaesthetic
was correlated with its lipid solubility (Overton, 1901). This
relationship can be applied to a wide
-
variety of anaesthetic compounds and is true, whether the
potency is measured in whole animais
or at the cellular or subce1luIa.r level, over a wide
concentration range (Urbane, 1985; Miller
198 1 ; Moore et al., 1964).
Unitary models of anaesthetic action suggest that anaesthetics
dissolve into the cell
membrane and perturb the structure and dynamic properties of the
lipid bilayer (Franks and Lieb,
1994; Tanelian et al., 1993). Thus, lipids are the primary
target and membrane proteins are
influenced secondarily. At clinically relevant concentrations,
anaesthetics alter lipid bilayer
fluidity. However, this effect was mimicked by small increases
in temperature (+l O C ). These
changes of temperature have no anaesthetic effect suggesting
that the action of anaesthetics on
lipid bilayers does not underlie the drugs clinical effects
(Franks and Lieb, 1994). Further, the
unitary theory of anaesthetic action does not account for
cimg-specific effects observed at the
molecular, cellular and whole animal level (Tanelian et al.,
1993; Franks and Lieb, 1994).
Therefore, it has been increasingly difficult to support a
unitary theory of general anesthesia.
Ueda and Karnaya (1973) provided the first convincing evidence
that anaesthetics
modulate protein function by demonstrating the inhibitory effect
of volatile agents on the enzyme
luciferase. Subsequently it was demonstrated, in a photoaffinity
labeling study, that the binding
of halothane to rat brain synaptosomes was saturable (K, = 490
FM). These data suggested
anaesthetics bind to a specific number of sites in the brain
(El-Maghrabi et al., 1992). These and
other studies resulted in a shift of interest £?om non-specific
mechanisms of anaesthetics to the
study anaesthetic interactions with specific receptors (for
review Franks and Lieb, 1994).
-
1.2. INHIBITORY NEUROTRANSMISSION AND GABA, RECEPTORS
1.2.1. GABA
Approximately 50 years ago, GABA and its synthesizing enzyme,
glutamic acid
decarboxylase (GAD) were discovered in the gray matter of the
mammalian central nervous
system (CNS) (Tanelian et. al., 1993). Not long afker, GABA was
shown to be an inhibitory
neurotransmitter (Bazemore et al., 1957; Tanelian et al., 1993).
GABA is present in high
concentrations in brain and spinal cor& and trace amounts
have been round in peripheral nervous
tissues (Miyata and Otsuka, 1972; Rabow et al., 1995).
GABA is synthesized fhm glutamate by GAD and stored for release
in the presynaptic
neuron (Figure A). During an action potential, a cascade of
presynaptic events leads to an
increase in intracellular calcium. The increase in cytosolic Ca-
stimulates the release of GABA
from the presynaptic terminal. Based on the low variance at the
peak of synaptic responses and
their smdl amplitudes, receptor saturation at the synapse has
been postulated (Mody et al., 1994).
Further, it is thought that there are a smdl number of GABA
receptors on the post synaptic
membrane activated during an inhibitory post synaptic potential
(IPSC). Jones and Westbrook
(1995) compared the on rates of brief pulses of GABA to those of
IPSCs. They estirnated the
peak concentration of GABA in the synaptic cleft was 526 FM.
GABA acts on pre- and postsynaptic GABA, recepton (GABABRs),
postsynaptic
GABAARs and postsynaptic GABA, receptors (GABA&). GABA,Rs
act via guanine-
nucleotide-binding proteins (G-protehs) to inhibit Ca* channels
or activate K' channels. The
GABA,R is selectively activated by baclofen, is insensitive to
GABA analogs isoguavane or
THIP and is not inhibited by the GABAAR antagonist bicuculline
(Bomann, 1992; Curtis et al.,
-
Figure A
GABAergic synapses and the topology of a GABAAR subunit
A . Diagram of a GABAergic synapse, showing GABA synthesis and
release from a presynaptic
terminal and GABA actions on the pre- and postsynaptic neuron.
GABA is synthesized from
glutamate by the enzyme GAD and stored in the presynaptic neuron
for release. Once *
released. G.4BA c m bind to post synaptic GABAARs and GABA,Rs.
GABA also binds to
presynaptic GABA,Rs. GABA is removed From the synaptic clef? by
diffusion and by Na--
coupled active transport into pre- and postsynaptic neurons and
glial cells. Once GABA is in
the cell it is rnetabolized in the mitochondria to
glutamate.
B. Genenc GABA,R protein subunit sequence and topological
stmcture. The Y- and C-
terminal of the polypeptide is suggested to be extracellular.
There are 4 TM domains (TM 1 -
TM). Between the third and fourth transrnernbrane domain there
is a hydrophilic putative
cytoplasrnic region of highly variable sequence involved in
intracellular regulatory
mechanisms such as phosphorylation.
-
GLIAL CELL
GAD
-\c . 7, - - f , ' v - - -+ -,
b 4- '- Na Ca"
P p . -. - - . .- - - - - . - - GABA \. A
A v r
POSTSYNAPTIC
- GABA, . .
r C 1'
.,+. -'
TMtTM2 TM3 TM4 Lipid Bilayer
-
1970). Presynaptic GABABRs have been implicated in the regdation
of Ca" dependent
transmitter release (Bowery et ai., 198 1 and 1980) while
activation of postsynaptic GABABRs
increase K* conductance (Dutar and Nicol, 1988; Karlsson et al.,
1988). Recently two isoforms
of the GABA,R have been cloned (Kaupmann et al., 1997). GABABRs
have been suggested to
be important targets for therapy. Therefore, the cloning of the
GABA,R will allow for the study
of their molecular and functional diversity and lead to a
greater understanding of their clinical
importance.
GABA, recepton (GAB&Rs) are associated wiîh a Cl' channel
but are
pharmacologically distinct fiom the GABAAR (Bormann and
Feigenspan, 1 995). They are
insensitive to bicuculline and baclofen and selectively bind
Cis-4-aminopent-2-enoic acid
(Krogsgaard-Larsn et al., 1994). The p subunit present in the
GAB&R is thought to confer
baclofen resistance. The only place in the CNS where GAB&Rs
have been found to be
expressed is in the retina (Krogsgaard-Lmn et al., 1994; B o m m
and Feigenspan, 1995).
The major postsynaptic action of GABA is activation of GABAARs.
Activation of the
GABAAR results in an increase in chloride conductance which
hyperpolarizes the neuron. This
creates a curent shunt that reduces the effects of any other
currents on membrane voltage
thereby decreasing the response to excitatory input (Mody et
al., 1994). Synaptic inhibition is
terminated when GABA dimises out of the cleft or is actively
transport& into neurons or glial
cells by a sodium-coupled active transporter (for review see
Kanner, 1997; Tanelian et al., 1993).
GABA is mainly metabolized by GABA-transaminase, an enzyme which
is pnmarily localized
in mitochondria (Tanelian et al., 1993).
-
1.2.2 GABAAR PROPERTIES
The GABAAR belongs to the superfamily of ligand-gated channels
(Stephenson, 1995).
The GABAAR channel is formed kom five glycoprotien subunits that
CO-assemble to form a
functional Cl' channel. GABA bùids to the GABAAR to regulate
channel gating (opening and
closing of the channel pore). The GABA dose response c w e (DRC)
is sigrnoidal and has a Hill
coefficient of approximately 2 suggesting that at Ieast two
molecules of GABA bind in a
cooperative manner to the GABAAR for full activation of the
receptor (Macdonald and Olsen,
1994).
There are multiple high amnity (nM) and low aanity (PM) binding
sites on the
GABAAR for GABA. It is not known if the low and high &nity
sites are distinct or
interconvertable. However, the low afiinity binding site for
GABA is thought to be associated
with the conformational change that results in an open channel
(Macdonald and Olsen, 1994).
Single channel analysis has revealed that the GABAAR channel
opens to multiple conductance
levels. The main conductance level, responsible for 95% of the
current, has a single channel
conductance of is 27-30 pS. The two less fiequent conductance
levels have a single channel
conductance of 17-19 pS and 11-12 pS.
The GABAAR has agonist recognition sites for GABA, barbiturates,
propofol and
anaesthetic steroids, as well as antagonist recognition sites
for bicuculline and flumazenil
(Stephenson, 1995; Tanelian et al., 1993). Further, the GABAAR
complex is modulated by
several second messenger systems. It is subject to
phosphorylation by several protein kinase
systems including protein kinase C, protein kinase A, and
tyrosine kinases (Tanelian et al., 1993;
Moss et al., 1995).
-
1 e2.3. KINETIC MODEL OF GABAAR FCTNCTION
B inding Gating
To explain the complex gating properties of the GABAAR, various
kinetic modeis have
been described (Orser et al., 1994; Macdonald and Olsen, 1994;
Jones and Westbrook, 1995 and
1996; Lavoie and Twyman, 1996). In the model presented above two
stages regulate channel
isomerization fkom a closed to open state; ligand binding and
channel gating. Receptor
activation is regulated by the sequential binding of two agonist
molecules (C to CG and C?,)
followed by an isomenzation step to the open state (O). The
rates of channel opening and
closing are P and a and the rates of association and
dissociation of agonist molecules are 2k,[,,,
kZpl and k-,p,, 2k-2,,,, respectively. The rates of entry into,
and recovery Ekom a nonconducting
desensitized @) state are indicated by &, and K-,,
respectively. Orser et al. (1994) suggested
that there was a high probability the desensitized receptor will
enter the open state prior to
closing. However, other models suggest the desensitized state
proceeds f?om a closed
configuration (Macdonald and Olsen, 1994). in addition, Orser et
al. (1994) proposed a third
binding site for agonist which governed the transition to the
desensitized state, however this
model is simplified and does not include this binding state.
-
Drugs other than GABA have been shown to alter the transition
rates between the closed,
open and desensitized states (Macdonald and Olsen, 1994).
Benzodiazepines, propofol and
barbiturates increase the affinity of the GABA,R for GABA and
enhance the rate of GABA
binding (Macdonald and Olsen, 1994, ûrser et al., 1994 and
Lavoie and Twyman, 1997). In
addition propofol and barbiturates alter the rates governing
receptor desensitization (Oeer et al.,
2 994; Macdonald and Olsen, 2 994).
1.2-4- DESENSITIZATION
Ligand-gated channeis open in response to the binding of
neurotransmitter but also close
or "desensitize" with the agonist still bound (Jones and
Westbrook, 1995). Desensitization has
been viewed ~aditionally as a negative-feedback rnechanisrn to
prevent the undesirable
consequences of excess receptor activation such as
excitotoxicity (Jones and Westbrook, 1996).
However, recent expenments suggest desensitization of GABAARs
prolong, rather than curtail
synaptic currents (Jones and Westbrook, 1995 and 1996).
Traditionally, it was thought that the decay of the post
synaptic current (PSC) was
govemed by agonist clearance and reuptake of transrnitter.
However, several kinetic
mechanisms contribute to the decay of the IPSC, including
unbinding, desensitization and
deactivation. Jones and Westbrook (1995 and 1996) suggested
receptor desensitization might
conaibute to the decay of the slow component of the IPSC. In
expenments with pairs of closely
timed responses, the second response was strongly depressed as
predicted by fast desensitization
of GABAAR. The second response returned to the full amplitude
with a time course similar to
the fast decay time course of the initial response. They
interpreted these data to indicate the
decay of the IPSC was related to the movement through fast
desensitized states (Jones and
Westbrook, 1996). Further, they suggested the peak open
probability and duration of the fast
-
decay component of the IPSCs were lirnited by fast
desensitization, whereas the slow component
was produced by channel reopening after exit fiom desensitized
states. Therefore, h g s that
increase receptor desensitization maintain the channel within a
set of states that had a high
probability of reopening before the agonist could dissociate.
These dmgs would pmlong the
decay curent of the P S C (Otis and Mody, 1992; Mody et ai.,
1994). Additionally, receptor
desensitization could limit the fiquency at which GABA recepton
produced full amplitude
responses to GABA (Lavoie and Twyman, 1996). The slow recovery
from desensitization might
also result in the attenuation of successive IPSCs during high
fkequency release. Attenuation of
IPSCs might also develop in the presence of low background
levels of GABA.
1.2.5. SUBUNITS OF GABAAR
To date 18 subunits of the GABAAR have been identified and are
classified based
according to the conservation of their amino acid sequence.
There are six GABA,R subunit
types and various isofoms: al-6, B 1-4, y 1-4, pl-2, a and E
(Sieghart, 1995; Davies et al.,
1997). Additionally, some of the GABAAR gene products undergo
alternative splicing such as
the human P3 (Kirkness and Fraser, 1993), the rat a6 (Korpi et
al., 1994) and the y2 subunit
which exists in two foms, y2 short (y2S) and y2 long (y2L). The
y2L splice variant has an 8
arnino acid insert (Whiting et al., 1990; Kofûji et al.,
1991).
Eac h GABAAR subunit has an extended extracellular, hydrop hilic
N-teminal domain of
the order of 220 arnino acids and has four putative
trausmembrane domains (TM 1 -TM4), which
form a wall of the chloiide ion channel (Figure A). The
positively charged residues in the TM2
region are thought to form one side of the selectivity filter
which permits negatively charged
chlonde ions to pass through the channel central pore (Tanelian
et al., 1993). The TMI-TM3
-
regions are adjacent to the N-terminal domain, whereas TM4 is at
the C-terminal end of the
protein. This transmernbrane topology predicts that the N- and
C- terminal regions are
extracellular however, this remains unproved. Separating TM3 and
TM4, there is a putative
hydrophilic cytoplasmic region that is possibly involved in
intracellular regulatory rnechanisms.
It contains several putative sequences for phosphorylation by
various protein kinases. It is here
that the y2L splice variant has an 8 amino acid insert which
contains a putative sequence for PKC
phosphorylation (Khan et al., 1 994).
GABAAR subunits are unevenly distributed throughout the brain
and different neuronal
populations possess different compliment of subunits (Wisden et
al, 1992). In situ hybridization
for rnRNA of the various subunits revealed that the pl subunit
was largely confined to the
hippocampus (Widsen et al., 1992). Ln contrast the P2 and P3
subunits were widely disîributed
within the CNS. The al, a2, a4, a5, pl , P2, y2 mRNA were
strongly expressed in the
hippocampus, whereas there was lower expression of the P3 and yl
subunits. The a3, y3, 6 and
a 6 subunits were minimally or not expressed in the
hippocampus.
Over 10,000 possible pentameric subunit combinations of the
GABAAR potentially exist
however, some combinations are more likely to occur. To detemùne
which subunits co-existed in
the same receptor cornplex, McKemnan and Whiting (1996)
summarized studies where
GABA,Rs were immunoprecipitated using different combinations of
antisera. They proposed
almost half of al1 GABAARs in the brain contained the alp2y2
subunits. Two other
triheteromeric combinations constituted a M e r 35% of the total
GABAARs present in the
brain; a2P3y2 and a3py2/y3. The following combinations are found
in the hippocarnpus;
alp2y2 present in intemeurons, a2P2/3y2 and aSP3y2/y3 present in
pyramidal cells and u4P6
-
present in the dentate gyms. In spinal cord motor neurons,
a2PU3-12 is the most cornmon
subunit combination.
-
1.2.6. GABAAR AND ANAESTHESIA
It is now recognized that anaesthetics influence GABAAR fûnction
by several different
mechanisms. However, it is not known which mechanisrn is the
most important for sedation,
hypnosis and anaesthesia (Tanelian et al., 1993). Agents such as
propofol, barbiturates and
neurosteroids directly activate the GABAAR., enhance GAB A
binding to the receptor and
influence desensitization (Orser et al., 1994; Rabow et al.,
1995; Harris et al., 1995). In contrast
sedative benzodiazepines only potentiate agonist evoked
responses (Choi et al., 1 98 1 ).
It is thought that there are a small number of postsynaptic
recepton which are saturated
during synaptic transmission (Mody et al., 1994; Jones and
Westbrook, 1995). This limits the
number of mechanisms by which anaesthetics might exert a
positive allosteric influence upon
GABAergic transmission (Mody et al., 1994). For example, in the
presence of saturating
concentrations of GABA it is unlikeiy that anaesthetics increase
the amplitude of the IPSC since
anaesthetics increase the amnity rather than efficacy of GABA
for the GABAAR. However,
anaesthetics could proiong the decay and thereby modify the
temporal summation of inhibitory
tone (Lambert et ai., 1997). Benzodiazepines, barbiturates and
propofol have been show to
prolong the decay of the P S C (Orser et al., 1994; Otis and
Mody, 1992; Poncer et al., 1996).
This could be explained by an increase in the affinity of the
receptor for GABA and a decrease in
the rate of deactivation since GABA would remain bound longer.
Further, prolongation of the
decay could be due to an increased channel open time which would
increase the probability of
the channel being open during the decay.
1.3. PROPOFOL EFF'ECTS ON GABAAR
Propofol(2,6-di-isopropylphenol) is an aikylphenol recently
introduced for clinical use as
a general anesthetic. It is an intravenous agent which rapidly
and reliably causes loss of
-
consciousness. Propofol is an important anaesthetic because it
does not cause use-dependent
tolerance, as seen with the barbiturates and benzodiazepines
(Fassoulaki et al., 1994). Use-
dependent tolerance refers to an increase in dose requixements
of drug during anaesthesia to
maintain a constant same level of unconsciousness.
Propofol is a highly lipophilic molecule (octanoVwater partition
coefficient = 4,300)
(Tonner et al., 1992; Veintemilla et al., 1992). In principle,
propofol could disturb GABAAR
channel function indirectly by pertmbing the plasma membrane.
However, when propofol was
applied to the extracellular domain of the GABAAR it directly
activated channel opening,
whereas propofol added to the pipette solution did not enhance a
GABA evoked current. Thus,
propofol exhibited a clear membrane asymmetry suggesting it did
not difhise into the membrane
to activate the GABAAR (Hales and Lambert, 1991). Additionally,
bicuculline inhibited receptor
activation by propofol M e r suggesting that propofol is acting
at a specific site on the GABAAR
(Hara et al., 1993).
Propofol has been reported to enhance inhibitory synaptic
transmission and potentiate
GABA-induced depolarizations in slices fkom the rat olfactory
cortex (Collins, 1988). In
addition, propofol reversibly, and dose-dependently potentiated
GABA-evoked responses
recorded from acutely dissociated CA1 pyramidal neurons fiom the
rat hippocampus (Hales and
Lambert, 199 1). Potentiation of the GABA-evoked responses by
propofol was associated with a
parailel shift to the lefi of the GABA DRC, indicating an
apparent increase in the f i n i t y of the
receptor for GABA (Orser et al, 1994; Hara et al., 1994; Adodra
and Hales, 1995). Propofol has
also been shown to reduce the extent of desensitization during
prolonged drug applications
(Orser et al., 1994). Single charnel analysis, indicated that
low concentrations of propofol greatly
-
increased the frequency and open probability of GABAAR channels
but had little effect on the
open duration (Orser et al., 1994).
Recombinant GABAAR receptors have shown that propofol cm
potentiate GABA-evoked
currents in al1 combinations of subunits tested (for review see
Harris et al., 1995). The extent of
propofol potentiation of recombinant receptors did not differ
between receptor combinations that
contained the al, a4 or a6 isoform. In contrast, pentobarbitone
produced a much greater
potentiation of currents fiom receptors composed of the a4 and
a6 subunit compared to al
containing receptors, suggesting pentobarbitone and propofol
modulatory sites are distinct
(Wafford et al., 1996).
Propofol, ai concentrations generally greater than those
required for allostenc modulation
of the GABAAR directly activate the GABAAR (Haies and Lambert,
1991 ; Hara et al., 1993;
Orser et al., 1994; Adodra and Hales, 1995; Sanna et al.,
1995qb). Propofol-induced currents are
potentiated by diazepam, and antagonized by bicuculline and zinc
(Haies and Lambert, 199 1 ;
Hara et al 1993; Adodra and Hales, 1995). The modulation of
propofol-evoked currents is
similar to the modulation of GABA-evoked cments which are
potentiated by benzodiazepines
and antagonized by bicuculline and zinc. The direct actions of
propofol and pentobarbital require
the p subunit, unlike the potentiating action of propofol which
does not require a specific subunit
(Sanna et al., 1995qb). Correlation analysis showed no
relationship between the direct and
potentiating actions of propofol. Therefore, Sanna et al.
(1995b) postulated two sites of action
for propofol: one on the P subunit responsible for direct
activation, and a second site that could
not be placed on any single subunit, responsible for
potentiating GABA-evoked currents.
-
The P subunit is essential for direct-activation of the GABA,R
by propofol. However, the
extent of receptor activation is modulated by the a subunits.
Recepton composed of aJy2s
subunits were maximally activated by propofol when the
combination contained a6, activation
was reduced when the al subunit was substituted for a6. Further,
when the a 4 subunit was
present the receptor was not activated by propofol (Waf5ord et
al., 1996). Hence, the P subunit is
essential however, the a subunit can modulaie the extent of
propofol activation of the GABAAR.
1.4. BENZODIAZEPINE ACTION AT THE GABAAR
There are three types of benzodiazepine (BDZ) ligands: 1)
agonists which increase Cl'
conductance, 2) inverse agonisa which decrease CI- conductance
and 3) antagonists which have
little intrinsic activity but block agonist and inverse-agonist
effects (Haefely, 1990 and 199 1 ;
Tanelian et al., 1993; Macdonald and Olsen, 1994).
Potentiation of GABAAR-meàiated synaptic inhibition likely
underlies the therapeutic
efficacy of benzodiazepines as anxiolytics, anticonvulsants and
sedatives. Benzodiazepines bind
to an integral site on the GABAAR and increase the apparent
affinity of the receptor for GABA
(Pritchett et al., 1989; Choi et al., 1981; Study and Barker,
1981). Radiolabelling studies
demonstrated hinctional coupling between the benzodiazepine
receptor and the GABAAR; both
GABA and benzodiazepine agonists reciprocally increase the
binding affinity for the other agent
(Costa and Guidotti, 1979; Braestrup and Neilson, 1 98 1 ).
Fluctuation studies and single channel
analysis suggested that diazepam increased GABAAR currents by
increasing opening frequency
without altering channel conductance, open duration or bursting
properties (Study and Barker,
1981; Twyman et al., 1989). Application of the benzodiazepine
agonist, rnidazolarn, (< 0.3 PM)
increased the frequency at which single charnels open, whereas
higher concentrations decreased
-
the fkequency of channel opening (Rogers et al., 1994). Detailed
kinetic analysis indicated the
increase in fkequency of opening reflected an increase in the
binding affinity of the receptor for
the first of the two GABA molecules (Rogers et al., 1994; Lavoie
and Twyman et al, 1996).
1.4.1. BENZODIAZEPINE BINDING IS INFLUENCED BY SUBUNIT
COMPOSITION
OF THE GABAAR
Historically, benzodiazepine agonists were classified into two
groups depending on their
affinity for the GABAAR; benzodiazepine Type 1 and 2 agonists
(BZ1 and BZ2) bind with hi&
affinity to BZ1 receptors and BZ2 recepton, respectively. More
recently it has been shown that
BZ1 agonists have a high affinity for GABAARs containing the a 1
subunit, whereas BZ2
agonists have a high affinity for receptors containing a2, a 3
or a 5 subunits.
In order for benzodiazepines to modulate GABA-evoked currents,
the y subunit must be
present in the GABAAR complex (Rovira and Ben-An, 1993). The a
subunit influences the
efficacy, as well as the amnity of benzodiazepine ligands
(Macdonald and OIsen, 1994). Site-
directed mutagenesis has confirmed the sequence specificity for
berwdiazepine ligand binding
on the a subunit (Pritchett and Seeburg, 1991). Cornparison of
sequences for a 1, a 2 and a3
subunits revealed differences in the N-texminal putative
extracellular domain. Using chimenc
cDNAs which mixed the N-terminal domains of the a 1 and a3
subunits, lead to the identification
of a single residue that detemiined the affuiity of GABAARs for
benzodiazepine agonist. There
was a Gly in the al subunit at amino acid 201, whereas in the
a2, a3 and a 5 subunits the amino
acid in homologous position was Glu. Therefore, for BZ1 agonists
to have a high afEnity for
GABAARs the a subunit must have a glycine at arnino acid
201.
-
Recepton containing the a6 subunit have a unique pharmacology
where they do not bind
flunitrazeparn. Comparisons of the al and a 6 subunit revealed a
single residue that conferred
the ability of the receptor with the isoform to bind
'H-fl~nitraze~am. The kg-100 was mutated
to His-100 which was found in a 1, a2, a 3 and a5, it restored
the flunitrazepam binding for
recombinant receptors containing the a6 subunit (Wieland et al.,
1992). Therefore, rnolecular
studies have revealed that the difference in agonist affinity
depends on the a subunit present in
the GABAAR cornplex.
1.5. DRUG INTERACTIONS BETWEEN PROPOFOL AND BENZODIAZEPINES
Propofol and midazolam are comrnonly used clinically in
combination. Propofol and
midazolam have been shown to act synergistically for the
induction of hypnosis (Short and Chui,
1991). The ED,,'s for propofol and midazolarn to induce hypnosis
were reduced by 44% when
used in combination compared to each agent acting individually.
Additionally, midazolam
reduced the ED,, for anesthesia of propofol by 52% (Short and
Chui, 1991). However, a recent
paper by Oxom et al. (1997) has s h o w that midazolarn, given
immediately prior to propofol did
not influence the dose of propofol required for the induction of
hypnosis or maintenance of
anaesthesia.
The mechanism of synergism which underlies the sedative and
hypnotic effects of
midazolam and propofol is not certain but may be due to
interactions occurring at the GABA,R.
The interactions between midazolam and propofol at the GABAAR
are studied in this thesis.
1.5.1. DETERMIMNG DRUG INTERACTIONS
Synergism occurs when the effect of the combination of drugs is
greater than that
expected Eom the additive effect of each drug alone (Berenbaurn,
1989). Drugs do not have to
-
act at the sarne site to have a synergistic action; they can
indirectly modulate each other by
increasing or decreasing the dlinity or efficacy of the other h
g . Alternatively, they may act in
concert by activating the same recepton (Wessinger, 1986).
Several methods have been used to
detemine if drug interactions are additive, superadditive or
subadditive, including the Fixed-
Dose Model and Isobolographic Analysis.
The Fixed-Dose Model predicts that the combination of drug 1 and
dmg 2 will produce
an effect equai to the sum of each effect alone. The response
fkom the combination of drugs is
compared to the theoretical additive response. If response of
the drugs in combination is
significantly lower than the theoretical response, the drugs
interact in a superadditive rnanner
(Wessinger, 1986). The Fixed-Dose Model does not take into
consideration that the DRCs for
dmgs are not linear (Tallarida et al., 1989). Hence, for a
specific effect size the
potencykoncentration of a drug has to be determined. Further, to
adequately explore dmg
interactions, a wide variety of doses and dose combinations need
to be tested (Wessinger, 1986).
A method that takes both of these concepts into consideration is
Isobolographic Anaiysis
(Wessinger, 1986). For Isobolographic analysis, the potency of a
cimg, or mixture of two dmgs,
that produces a specified "effect level" is measured as a dose
or concentration. The relative
potencies of drug 1 and 2 are plotted on the isobolograph plot
and the diagonal line connecting
these endpoints is termed the isobol of additivity (Wessinger,
1986). If the two dmgs do not
interact and the effects are additive, the concentrations of the
two drugs to reach the effect level
will lie on the isobol of additivity (Berenbaum, 1989;
Tallarida, 1992). When agents in
combination are more effective than expected fkom thek dose
response curves (Le. they interact
synergistically), smaller amounts of drugs are needed to produce
the effect under consideration
-
and would lie below the isobol of additivity. This mode1 of
studying h g interactions will
determine if the dmgs interact in a synergistic, additive of
subadditive manner.
-
2. OBJECTIVES AND WORKRYG HYPOTHESES
2.1.1. SECTION 1
Anaesthetics, such as propofol and barbiturates, have been shown
to have multiple
actions at the GABA,R: they directly activate the receptor; they
potentiate GABA-evoked
responses, and they influence receptor desensitization. Previous
studies suggest that GABA and
anaesthetics activate the GABAAR by binding to distinct sites.
In addition, distinct sites likely
modulate the anaesthetic-induced potentiation and the direct
activation of the GABAAR.
However, it is not known if a distinct site mediates
anaesthetic-induced changes in GABAAR
desensitization.
Our hypothesis is that the effect of propofol on GABAAR
desensitization is mediated by a
site distinct fiom the two binding sites involved in
potentiation of GABA-evoked responses and
direct activation. Here, we will determine if the abilities of
propofol to directly activate the
GABAAR and decrease receptor desensitization are differentially
modulated by benzodiazepines.
In Section 1, we will address the following questions:
1. Does midazolam increase the &ty of the GABAAR for
GABA?
2. Does midazolam increase the f i t y of the GABAAR for
propofol?
3. Does midazolam alter desensitization of the GABAAR?
4. What are the characteristics of the propofol modulation of
GABAAR desensitization?
5 . Does midazolam alter propofol modulation of GABAAR
desensitization?
-
2.2. SECTION 2
A variety of agents are used for CO-induction of anaesthesia.
This practice is based on the
principle of synergism, whereby dnigs in combination produce a
greater-than-additive effect.
Propofol and midazolam are used clinically in combination, and
it has been suggested that they
act synergistically for the induction of hypnosis (Short and C h
i , 1991). Our hypothesis is that
the synergistic effects observed in the clinicd studies are due
to the interactions between
midazolam and propofol at the GABAAR
Ln Section 2, we will address the following questions:
1. Do midazolam and propofol potentiate GABA-evoked responses in
an additive,
superadditive, or antagonistic manner?
2. 1s the interaction between propofol and midazolam dependent
on the concentrations of
GAB A and propofol?
3. Does ce11 type influence the interaction between midazolarn
and propofol?
-
3. METHODS
3.1. CELL CULTURES
Cultures of ernbryonic hippocarnpal or spinal cord neurons were
prepared fkom Swiss
white mice as previously described (MacDonald et al., 1989).
Bnefly, feu1 pups ( 1 7 days in
utero) were removed from mice sacrificed by cervicai
dislocation. The hippocampi were
dissected fiom each f e w , then placed in an ice-cooled petri
dish. Neurons were dissociated by
mechanical trituration using two Pasteur pipettes (tip diameter
150-200 mm) and plated at a
density of 1x106 cellu'ml on 35-mm culture dishes. The culture
dishes had been coated with
collagen (from cdf skin) or poly-D-lysine (Sigma Chemical Co.,
St. Louis, MO). For the fint 1 O
days in vitro, cells were maintained in Minimal Essential Media
(MEM) (Life Technologies,
Grand Island, NY) supplemented with glucose ( h a 1
concentration 33.6 mM), NaHCO, (final
concentration 31.56 mM), 10% horse se-, 10% fetal bovine serum,
and 1% insulin (Life
Technologies, Grand Island, NY). The neurons were cultured at
36.S°C in a 5% CO2 I 95% air
environment. Once the background cells had grown to confluence
(4 to 7 days), 0.1 mL of
FUDR mixture (4 mg 5-fluorodeoxyuridine and 10 mg uridine in 20
rnL MEM) was added to
each dish to arrest ce11 division. The supplemented media was
changed every 3-4 days. A f er 10
days in culture, the media was changed to a new media containing
MEM supplemented with 10%
hone senim and 1 % insulin. There was a heterogeneous population
of neurons in culture as there
were neurons kom d l of the hippocampal regions.
The sarne procedure was followed for spinal cord neurons with
the following additions.
The spinal cord was dissected nom fetal pups after 15 days in
utero. Once the spinal cord was
dissected it was subjected to a trypsin digestion for 30 min
then dissociated by mechanical
trituration. The rest of the procedure was not aitered.
-
Rior to recording, neurons were rinsed with an extracellular
recording solution
containing (in mM): 140 NaCl, 1.3 CaCI,, 5.4 KCl, 25
N-2-hydroxy-ethylpiperazine-N'-2-
ethanesulphonic acid (HEPES), 33 glucose, 1 MgCl, and 0.0003
tetrodotoxin (TTX) (pH 7.4,
325-335 mOsm). TTX was added to inhibit voltage-activated Na*
channels. Cells were studied
1 2- 1 6 days after dissociation.
3.2. RECORDING PIPETTES
Whole-ce11 patch-pipettes (3- 10 MR) were consûucted fiom
borosilicate glass capillary
nibing containing an inner filament (outer diameter 1.5 mm,
TW150F-4, World Precision
Instruments Inc., Sarasota, FL). Electrodes were pulled using a
two-stage vertical puller
(Narishige PP-83, Narishige Scientific Lab., Setagaya-Ku, Tokyo)
and tips were occasionally
firepolished with a microforge (Narishige MF-83, Narishige
Scientific Lab., Setagaya-Ku,
Tokyo) .
Pipettes were filled with the pipette solution (see below) and
mounted in a teflon holder
that contained a silvedsilver chloride-coated silver wue
(Ag/AgCl). ui this system, the Ag/AgCI
acted as a reversible electrode that converted ionic CI- cwent
in the pipette solution into electron
current in the wire. A ground electrode (AglAgCI pellet) was
placed in the bath solution and the
offset potential was nulled pnor to the seal formation.
3.3. PIPETTE SOLUTION
Patch-pipettes were filled with a pipette solution containing
(in mM): 140 choline-CI, 10
HEPES, 11 EGTA, 1 CaCl,, 10 tetraethylarnrnonium chlonde
(TEA-CI), 2 MgCl, and 4 Mg-ATP
(pH 7.4, 290-300 mOsm). Choline was used as an irnpemeant cation
which has been used in
previous studies to record GABA-evoked currents (Orser et al.,
1994).
-
3.4. WOLE-CELL VOLTAGE-CLAMP RJECORDINGS
Whole-ce11 voltagetlamp recording techniques measure ion flow
across the ce11
membrane and indirectly provide information regarding receptor
activity. The voltage-clamp
system ernploys a negative feedback amplifier which compares a
signal received fiom the
recording electrode to a reference or "command potential". The
potential difference between the
command potential and the reference potential is arnplified and
transmitted to the recording
electrode.
Whole-ce11 currents were recorded using the Axopatch 1D
amplifier (Axon Instruments
Inc., Foster City, CA). The recording-pipette was filled with
pipette solution and secured into the
pipette holder. The holder was inserted into the head stage of
the amplifier and the pipette tip
was positioned close to the ce11 using a course rnanipulator
then a fine hydraulic
micromanipulator (Narishige Scientific Lab., Setagaya-Ku,
Tokyo). The junction potential
between the bath and patch-pipette solutions was nullified by
adjusting the input offset of the
amplifier. A +20 mV test pulse was used to measure the electrode
resistance and to track the
development of the seal between the pipette and the ce11
membrane. Following seal formation,
the command potential was set and the membrane patch disrupted
by applying negative pressure
and a brief current pulse.
Neurons were voltage clamped at -60 mV. Once patched, we waited
10 minutes for the
cell to stabilize and to reduce the effect of rundown. Responses
were recorded and analyzed on a
cornputer using the pClamp program (Axon Instruments Inc.,
Foster City, CA). Ali experiments
were conducted at room temperature (20-25°C).
-
3.5. DRUG AND AGONIST PERFUSION SYSTEM
Drugs and agonists were applied to the neurons using a
multi-banel pefision system
(Figure B; Johnson and Ascher, 1987). A senes of 2 or 3 square
glass capillary tubes (400 x 400
Fm, Longreach Scientific Resources, Orr's Island, ME) were
horizontally aligned and glued
together with cyanoacrylate glue. The array of barrels was then
placed in a holder. Lateral
movement of the pefision barrels was regulated using a stepping
motor (Vexta, Oriental Motor
Co.) which was attached to a Leitz manipulator (Germany). The
manipulator was controlled by a
custom-made, computerized switching system. The solution bathing
the patched neuron was
changed by laterally displacing the barrels. The speed of the
fast perfusion system was
previously detemined f?om the rate of omet of Mg+-induced
inhibition of NMDA currents
recorded at hyperpolerizing potentials and was estimated to be
less than 50 msec. n i e barrels
were comected by silastic tubing to separate soiution reservoirs
and the flow rate (approxirnately
0.5 mllrnin) was regulated by changing the height of the
reservoirs above the bath. Agonists
were applied only once every few minutes to allow receptors to
recover f?om desensitization.
Drugs and agonists were applied for 15 or 18 seconds at a time
to obtain the peak and relative
steady-state current.
3.6. WOLE-CELL CURRENT ANALYSIS
Whole-ce11 data were analyzed using the CLAMPFIT program of
pClamp (Axon
Instruments Inc., Foster City, CA). For a11 experiments the
cells used were Eom at les t 3
different dissections.
For dose-response analysis, the current amplitudes were plotted
versus the agonist
concentrations and fit using a standard logistic equation (I = h
a x (l+CEC,,)", where 1 is the
-
Figure B
iMulti-barre1 fast-perfusion system
Three square g l a s capillary tubes (400 x 400 pm) were aligned
and glued together as shown.
The barrels were connected by silastic tubing to separate
solution reservoirs. Perfusion solutions
were exchanged by laterally displacing the barrels a distance of
one barre1 diameter. Solutions
perfused cultured munne hippocampal neurons. Currents were
recorded using the whole-ceIl
patch clamp technique. The holding potential was -60 mV.
-
current amplitude and C is the concentration of the agonist)
(Graph Pad Prism, San Diego, Ca).
The concentration of agonist that produced 50% of the maximal
response (EC,,) and the Hill
coefficient (n) were determined fkom die equation.
The extent of receptor desensitization was quantified by
detexmining the steady-state to
peak current ratio (IssAp; see Figure C). An increase in the
Iss/Ip ratio correlated with a decrease
in receptor desensitization. The IssAp ratio was plotted versus
the concentration of agonist. This
relationship was best fit using a nonlinear regression for
GABA-evoked currents (equation
above) and a linear regression for propofol-evoked currents. The
rate of receptor desensitization
was estimated from the rate of current decay during prolonged
agonist application. The decay of
the current was best fit using a biexponential equation I(t) =
A, exp(-Vrf) + A, exp(-t/r3 + C
where I(t) is the current amplitude at any given time t, C is
the baseline current, r, and r, are the
fast and slow time constants of current decay, respectively . A,
and 4 are the estimated fast and
slow intercepts of the cornponents at time zero,
respectively.
Deactivation refers to the current decay during agonist washout
(Figure C). The rate of
deactivation was best fit using the above equation. The
biexponential equation above was
modified, the baseline current, C, was zero.
3.7. DRUG INTERACTION ANALYSIS
Fixed-Dose Ratio mode1 and Isobolic Analysis were used to
detemine if the interactions
of rnidazolam and propofol, with respect to potentiation of
GABA-evoked currents, were
synergistic, additive or subadditive. Using the Fixed-Dose Ratio
method, low, clinically relevant
doses of propofol, midazolarn and GABA were chosen. The
concentrations were previously
noted by Reynolds et al. (1996) to have synergistic interactions
for the enhancement of GABA-
-
Figure C
Desensitization and deactivation of GABA induced currents.
The bar on top of the current represents the duration of drug
application. Desensitization is the
process by which receptors are inactivated in the prolonged
presence of agonist. The extent of
receptor desensitization was calculated from the steady-state to
peak current ratio (Iss/lp).
Additionaily. the extent of desensitization was quantified by
fitting the decay with a
biexponential equation. AAer agonist application. the current
retumed to a base line current.
This is t e n e d deactivation and was quantified by fit~ing the
decay with a biexponential equation.
-
evoked currents in oocytes expressing human GABAARs. Since the
degree of potentiation of
GAE3A-evoked currents dependents on the concentration of
agonist, doses of propofol and
GABA were varied in different experiments (Harris et al.,
1995).
Potentiation of GABA-evoked currents were quantified by
estimating the increase in the
peak response induced by midazolam or propofol compared to the
response evoked by GABA.
The theoretical additive response was estimated by adding the
enhancement produced by each
dmg alone to the amplitude of the response evoked by GABA. These
values were compared to
the amplitude of the currents evoked by the two dmgs in
combination. In addition, charge
transfer was measured (using pClamp Software) by taking the
integral of the area under the
current response. The charge transfer was nomalized to the
control GABA response. The
tieoretical additive value was estimated by adding the relative
enhancement by each dmg alone
and these values were compared to the measured charge transfer
of GABA-evoked currents in
combination with the two drugs.
Isobolographic analysis was used to determine if the dmg
interactions were synergistic.
Testing a pair of drugs for synergism first required the
determination of the potency of each dmg
alone then exarnining the potency of the two dnigs in
combination. The potency of a dmg or
mixture, was defined as the dose that produced a chosen effect
level. Here, we chose an effect
level where the dmg potentiated the GABA-evoked response by a
factor of 3, that is the current
was 3 times the control or 200% greater than the control.
The doses of propofol and midazolam that enhanced the GABA
response by a factor of 3
were determined from individual DRCs (2,. and +, respectively;
Figure D). The average dose of
midazolam and propofol that produced the 3 times effect level
was plotted on the X-and Y-axis
of the isobologram, respectively. These points were joined to
fom the isobol of ad&iviy.
-
Figure D
Isobolographic analysis
The Y awis represents the concentration of dnig 1 and the X axis
represents the concentration of
drug 2. 2,. is the concentration of dnig 1 that potentiated the
control response by a factor of 3.
Z,. is the concentration of dmg 2 that potentiated the control
response by a factor of 3. The line
connecting 2,. and z2. contains coordinates which are dose pairs
that produce an additive
interaction. The line is called the isobol of additivity. P
represents a dose pair that interacts in a
synergistic manner. Q represents a dose pair that interacts in a
subadditive manner.
-
synergistic
subadditive
-
Subsequently, we selected a dose of midazolam that enhanced the
GABA-evoked curent by less
than the 3 times effect level. in the presence of this fixed
concentration of midazolarn, the
concentration of propofol that produced the 3 times effect was
less than the propofol
concentrations estimated ftom the isobol of additivity (Figure
D).
The measured concentration of propofol that produced the 3
tirnes effect level in the
presence of rnidazolarn was then compared to the theoretical
additive concentration. To
detemine the theoretical additive concentration of propofol, the
following equations were used
as descnbed by Tallarinda (1992). The doses of propofol and
rnidazolam that each produced the
3 times effect level were denoted as z,* and q*, respectively
(Figure D). The relative potency
(R) was defined as R = z,*/q*. The confidence of this ratio, or
the variance of R, was calculated
according to :
C = V(Z, *)/(q*)' + (2, * )2~(q*) / (~2*)4,
where V was the variance of the average dose. The theoretical
additive concentration of propofol
was calculated by:
z,ad,j = z1* - E-i
where q W ~ S the fixed concentration of midazolarn. The
variance of this value was caIculated
b y:
V(z,,&J = V(z,*) + (ZJZC - 2qV(z,*)/q*.
If the theoretical additive concentration (z,& was
significantly less than the actual
concentration, the dmg interaction was not additive. If the
theoretical value was significantly
greater than the actual concentration, the dmg interaction was
synergistic.
-
3.8. DRUGS AND OTHER CHEMICALS
Propofol was prepared fiom ~ipnvan' (Zeneca Pharma, Missassauga,
ON). Each mL of
Diprivan contained (in mg): 10 2,6 di-isopropylphenol, 100
soybean oil, 12 egg lecithin and 22.5
glycerol. Stocks of propofol (1 0 mM and 1 rnM) were prepared
every 3 days and stored at 4°C.
Midazolarn was prepared fiom Versedm (Hohann-LaRoche Ltd.,
Missassauga, ON). Each mL
of Versed contained (in mg): 1 midazolam, 0.1 disodium edetate,
10.45 beruyl alcohol and 8
NaCl. In addition, rnidazolam was generously donated by
Hofiann-LaRoche. There was no
difference in potentiation of GABA-evoked currents when
midazolam was prepared from v e n d
compared to midazolam which was dissolved in dimethylsulfoxide
(DMSO). EGTA, HEPES
and KCl were purchased from Fluka Biochemika (Bucks,
Switzerland), DMSO from Fischer
Scientific (Fairlawn, NJ), CNQX nom Tocns Cookson (St. Louis,
MO), TTX fiom Precision
Biochernicals, h c . (Vancouver, Canada) and NaCl, CaC12,
glucose and NaOH fiom BDH inc.
(Toronto, ON). Unless specified otherwise, al1 other compounds
were purchased fkorn the Sigma
Chernical Co. (St. Louis, MO).
3.9. STATISTICS
Al1 data are presented as mean t S.E.M unless otherwise
indicated. An analysis of
variance (ANOVA) was used to compare multiple groups of data.
Al1 groups of data were tested
to ensure that the requirernents for a paramenic test were met.
If the data were not normally
distributed, a non-parametric test was substinited for the
parametric test. For experirnents
including only two cornparison a Student's t-test was used The
EC, values and Hill slopes of
the DRC were compared using a paired t-test.
-
4. RESULTS
4.1. SECTION 1
Sedative benzodiazepines, such as diazepam, are known to enhance
GABA-evoked
currents by increasing the affinity of the GABAAR for GABA
(Study and Barker, 1981).
However, benzodiazepines have no inainsic GABA-mimetic
properties (Choi et al., 1 98 1 ).
Propofol activates the GABAAR possibly by binding to a site
distinct from the GABA binding
domain (Sanna et al., 1995a,b). The purpose of these
experirnents was to determine if midazolarn
infiuenced propofol-evoked currents or altered the f i n i t y
of the GABA,R for propofol. In
addition, we investigated the effects of midazolarn on
GABA-evoked responses.
4.1.1. MIDAZOLAM INCREASED THE AFFINITY OF THE GABAAR FOR GABA
AND
PROPOFOL
In the absence of GABA or propofol, rnidazolarn (0.5 PM) did not
induce a GABA,R-
mediated current. However, midazolarn (0.5 PM) potentiated
cuments evoked by subsaturating
concentrations of GABA (0.6-20 PM; Figure 1). The amplitudes of
currents evoked by near-
saturating concentrations of GABA (1 00-600 PM) were not
enhanced. Midazolam (0.5 pM)
shifted the GABA dose-response curve (DRC) to the lefl and
significantly decreased the EC,,
from 9.49 + 1.85 PM to 6-92 i 1.32 p M @ < 0.05, paired
t-test, n = 11, Figure lb). However, midazolarn did not change the
Hill coefficient (1 -54 + 0.19 to 1.65 t 0.25, p > 0.05, paired
t-test).
As previously reported (Orser et al., 1994), propofol (1-600 PM)
induced GABAAR-
mediated inward currents in al1 hippocampal neurons tested
(Figure 2). The thresho Id
concentration for propofol-activated cments was approximately 1
pM and maximal currents
were observed at 600 pM propofol. Higher concentrations of
propofol (approximately 600-1 000
-
Figure 1
Midazolam increased the aflinity of the GABAAR for GABA.
A. Responses evoked by 0.6, 3. 6. 10. 100, and 600 pM GABA, in
the absence and
presence of 0.5 FM rnidazolam. Midazolam increased the amplitude
of currents
evoked by subsaturating concentrations of GABA. However,
rnidazolarn did not
change the amplitude of the maximal response. The bars indicate
the duration of drug
application.
B. The dose response curves for peak currents activated by GABA
in the absence or
presence of 0.5 u M midazolam (n = I I ) are shown. iMidazolam
shifted the GABA
DRC to the Ieft and shified the EC,, from 9.49 f 1.85 p M to
6.92 21.32 pM (p <
0.05, paired t-test) but did not affect the Hill coefficient (1
-54 r 0.19 to 1.65 k0.25. p
> 0.05, paired t-test).
-
I 4 sec
GABA
GABA+MDZ
1 10
GABA (PM)
-
Figure 2
Midazolam increased the affinity of the GABAAR for propofol.
A. Responses evoked by 1. 6. 10. 60. 100, and 600 pM propofol.
in the absence and
presence of 0.5 pM midazolam. Midazolam increased the amplitude
of currents
evoked by subsaturating concentrations of propofol. However.
midazolam did not
change the amplitude of the maximal response. Note that is 3 of
6 cells. midazolam
increased the peak of the after-response.
B. Dose response cuves for peak currents activated by propofol,
in the absence or
presence of 0.5 p M midazolam are shown (n = 6). Midazolam
shifted the propofol
DRC to the left and decreased the EC,, from 24.3 & 3.4 pM to
16.1 + 2.3 @M (p < 0.05. paired t-test) but did not affect the
Hill coefficient (1 2 4 + 0.1 1 to 1.35 +O. 16. p
> 0.05, paired t-test). Further, midazolam did not change the
amplitude of the
maximal response.
-
3 sec
i PRO
A P R O t MD2
10 1 O0 Propofol (PM)
-
42
CiM) activate progressively smaller currents (Orser et al.,
1994). Therefore, the highest
concentration of propofol used in these experiments was 600 pM.
Upon cessation of the
application of a high concentration of propofol (600pM), a
current surge, also refexred to as an
"after-response", was observed. in contrast, there was no
decrease in the peak response or after-
response observed at high concentrations GABA.
Propofol-activated currents were inhibited by
bicuculline suggesting they were mediated by the GABAAR. The
rate of onset and offset of
currents evoked by propofol (1-100 FM) were slower than the
rates for currents activated by
GABA (Orser et al., 1994).
Midazolam (0.5 pM) potentiated currents evoked by subsaturating
concentrations of
propofol (1-100 PM, Figure 2a) but did not alter the maximal
amplitude of responses evoked by
600 p M propofol. As illustrated in Figure 2b, midazolam shified
the propofol DRC to the left
and decreased EC,, fiom 24.3 f 3.4 pM to 16.4 f 2.3 p M @ <
0.05, paired t-test, n = 6) but
caused no change in the Hill coefficient (1.24 t 0.1 1 to 1.35 f
0.16, p > 0.05, paired t-test). In 3
of the 6 cells tested, midazolam increased the amplitude of the
propofol-induced "after
response".
Flumazenil, a selective antagonist for the benzodiazepine site,
inhibited the effect of
midazolam on propofol-evoked currents (Figure 3) suggesting
evoked currents by acting at the benzodiazepine site.
The Current Voltage (W) relationship of propofol (20
midazolam potentiated propo fol-
LM)-evoked currents, recorded in
the absence and presence of midazolam (0.5 PM), are sumrnarized
in Figure 4. Responses were
noxmalized to the peak currents measured at a holding potential
of -60 mV. The W
relationships were nonlinear and indicated currents reversed
polarity at approxirnately -6 mV.
-
Figure 3
Midazolam potentiated currents evoked by propofol.
In the absence of propofol (PRO). rnidazolam (MDZ) did not evok
ent. However.
PRO (20 FM) directly-acrivated a current. When ;MD2 was
CO-applied with PRO the
current was enhanced. When flumazenil (FMZ) was CO-applied with
PRO and MDZ. the
potentiation was depressed.
-
--, --y.-- -,-- --- MDZ 0.5pM
200 pA 1 300 msec
-
Figure 4
The IN relationship of propofol-induced currents recorded in the
absence and
presence of midazolam.
Propofol (20 piM) was applied to neurons voltage clamped at
holding potentials ranging
from -60 mV to +20 mV in the absence and presence of midazolam
(0.5 FM). -411
currents were normalized to the current evoked at -60 mV (n =
4). The W relationship
was nonlinear and currents reversed polarity at approximately -6
mV.
-
The reversa1 potential was close to the calculated Nernst
equilibnum potential for chloride ions
(+1 mV) suggesting the currents evoked by propofol are mediated
by chloride ions (Hales and
Lambert, 1991 ; Hales et ai., 1993).
4.1.2. MIDAZOLAM D1D NOT INFLUENCE DESENSITIZA TION OF THE
GABA$
Previous reports suggest that benzodiazepines including
diazeparn and chlordiazepoxide,
increase desensitization of the GABAAR (Frosch et al., 1992;
Mierlak and Farb, 1988).
However, the increase in GABAAR desensitization by
benzodiazepines could possibly result
from an increase in agonist binding rather than a direct effect
on channel gating. Therefore, the
purpose of these experiments was to detemiine if benzodiazepines
have an intrinsic effect on
GABAAR desensitization.
Currents recorded in the presence of a prolonged agonist
application peaked then decayed
to an apparent steady-state. In order to examine the effects of
midazolam on desensitization of
the GABAAR we calculated the Iss/Ip ratio for GABA (0.1-600
PM)-evoked cwents in the
absence and presence of midazolam (0.5 PM, Figure 5). Midazolam
increased the peak current
evoked by subsaturating concentrations of GABA (0.1 - 60 PM) and
decreased the IssAp ratio as
the concentration of GABA increased. Midazolam did not
significantly change the IssAp ratio at
any concentration of GABA tested @ > 0.05, ANOVA, n = 11).
The Iss/Ip ratios for GABA (0.6-
600 FM) and GABA + MDZ (0.5 PM) were plotted versus the
concentration of GABA (Figure
5b). The EC,, and Hill coefficient for this relationship were
not different for currents recorded in
the absence or presence of midazolam (4.15 f 0.69 to 5.27 t 1.1
3 FM, and - 1.36 f 0.25 to - 1.34 k
0.18, respectively, p > 0.05, paired t-test). Therefore,
although midazolam increased the
amplitude of the currents, it did not significantly alter the
extent of GABAAR desensitization.
-
Figure 5
Midazolam did not influence desensituation of the GABA,%R for
currents evoked by
GABA.
A. Responses evoked by 3 or 10 pM GABA, in the absence and
presence of 0.5 piM
midazolam. peaked and then desensitized.
B. The dose response curved for the Iss/Ip ratios of GABA-evoked
currents, in the
absence and presence of midazolarn, are shown. Midazolam did not
change the Iss/Ip
value at any concentration of GABA (p > 0.05. ANOVA). In
Addition. there was no
difference in the absence or presence of midazolarn for the EC,,
values (4.15 2 0.69
uiM versus 5.27 2 1.13 FM . p > 0.05, paired t-test) or the
Hill coefficients (-1.36 IT 0.25 versus -1.34 + 0.18. p > 0.05,
paired t-test) in the absence and presence of
midazolam, respectively.
-
The effects of midazolam on desensitization of the GABAAR by
propofol-evoked cments
were also examined. The IssAp ratio for propofol (1-600
PM)-evoked currents recorded in the
absence or presence of midazolam (0.5 pM) are shown in Figure 6.
The Iss/Ip ratios for propofol
(1-600 FM) and propofol + MDZ (0.5 PM) responses were plotted
venus agonist concentration
(Figure 6b). Midazolam had no significant effect on the IssAp
ratios at al1 concentrations of
propofoi, (p > 0.05, ANOVA, n = 6). The Iss/Ip relationship
was linear for the concentrations of
propofol exarnined. The Iss/Ip concentration relationship was
fit with a linear regression; the
dopes were not different in the absence and presence of
midazolarn (-0.285 t 0.008 to -0.275 i
0.006, p > 0.05, paired t-test), nor were the correlation
coefficients (? values) different (0.95 +_
0.02 to 0.93 k 0.02, p > 0.05, paired t-test). Note that the
Iss/Ip relationship is strikingly different
than the sigrnoidal relationship observed with GABA-evoked
currents. This could be due to a
decrease in receptor desensitization in the presence of propofol
compared to GABA. in
summary, midazolarn enhanced currents activated by subsaturating
concentrations of propofol,
but did not alter the IssAp ratios significantly. These data
suggest, midazolarn did not directly
alter receptor desensitization.
4.1.3. PROPOFOL DECREASED GABAAR DESENSITIZATION
Propofol decreases receptor desensitization and increases the
steady-state current
observed during prolonged applications of agonist. However, the
effects of propofol on
GABAAR desensitization had not been fûlly characterized ( b e r
et al., 1994). Here we
investigated the effect of various concentrations of propofol
(0.0 1 - 100 PM) on currents evoked
by GABA (600 FM) (Figure 7). A saturating concentration of GABA
(600 PM) induced currents
that peaked rapidly, then decayed to an apparent steady-state.
When propofol (0.01-100 pM)
-
Figure 6
Midazolam did not alter desensitization of propofol-evoked
currents.
A. Responses evoked by 6 or 100 pM propofol. in the absence and
presence of 0.5 pM
midazolarn peaked. and then desensitized.
B. The dose response curves for the Iss/Ip ratios of
propofol-evoked currents in the
absence and presence of midazolam are show. Midazolam did not
change the Iss/Ip
value at any concentration of propofol (p > 0.05. ANOVA). The
propofol DRCs for
the W I p ratios were linear in the absence and presence of
midazolarn. There was no
difference in the slopes in the absence and presence of
midazolarn (-0.285 k 0.008
venus -0.775 + 0.006, p > 0.05, paired t-test, n = 6) or the
$ values (0.95 - 0.02 versus 0.93 f 0.02, p > 0.05, paired
t-test, n = 6).
-
Figure 7
Propofol decreased GABA,R desensitization in a concentration
dependent manner
A. Responses to 600 pM GABA and increasing concentrations of
propofol (1-20 PM)
are shown.
B. The Iss/Ip ratio decreased as the concentration of propofol
increased from 0.01-20
PM, suggesting that propofol decreased GABAAR desensitization.
The Iss/Ip ratios
were normalized to the Iss/Ip ratio at 10 pM propofol. The
nurnber of cells tested at
each concentration of propofol is noted below the mean. The
EC5,] of the propofol
DRC was 1.15+ 0.33 pM and the Hill coefficient was 1.14 2 0.13.
When the
concentration of propofol was increased from 70 piM to 100 PM,
there was an
decreased steady-state current. Thus, the IssiIp DRC for
propofol was biphasic. The
decrease in GAEIA,R current observed at higher concentrations of
propofol was
possibly due to channel blockade.
-
rl GABA+PRO 1 pM \
' - GABA +PRO 6 p M GABA +PRO 10 p M
1 / GABA + PRO 20 pM
4 sec
GABA 0.0 1 600 pM
o. 1 I Propofol (PM)
-
was CO-applied in the presence of GABA, the amplitude of the
steady-state current increased
suggesting a decrease in receptor desensitization. The threshold
concentration of propo fol which
induced a change in the steady-state current was approxirnately
0.1 FM (Figure 7). The
concentration of propofol was plotted versus the I s d p ratio.
The Iss/Ip values were normalized
to the maximal IssAp value. In the presence of 0.01 to 20 pM
propofol the I s d p ratio increased
in a dose-dependent manner. The relationship was biphasic as the
Iss/Ip ratio decreased with
concentrations of propofol greater than 20 @A. The EC, for the
propofol DRC representing the
decrease in GABAAR desensitization was 1.15 t 0.33 pM and the
Hill coefficient was 1.14 f
O. 132 (Figure 7b).
We also investigated the effects of propofol on the rate of
desensitization. The decay
phase of the response was fitted using a biexponential equation.
For cu~~en t s illustrated in Figure
7a propofol did not influence the fast time constant (TJ but
increased the slow time constant (rJ
of the decay (Table la).
In addition, the deactivation of currents was slowed in the
presence on propofol (Figure
7a). The decay of the response, observed following temination of
agonist application, was fit
with a biexponential equation. As summarized in Table lb,
propofol increased the fast and slow
time constant of deactivation in a dose dependent manner.
4.1.4. PROPOFOL-INDUCED MODULA TION OF GABAJ2 DESENSITIZA TION
WAS
NOT INFLUENCED B Y MIDAZOLAM
In the previous section, we demonstrated that rnidazolam
potentiated propofol-evoked
currents but did not modulate GABAAR desensitization. We then
wished to determine if
midazolarn influenceci propofol induced changes in GABAAR
desensitization in both
-
Table 1
A
Desensitization Lt (sec) %O,V (sec) . GABA 600 PM 1 .O3 4.30
1 GABA 600 uM + PRO 1 DM 1.04 5.49 1 GABA 600 PM + PRO 6 pM 1
.O7 8.36 GABA 600 pM + PRO 10 pM 1.04 9.30 GABA 600 uM + PRO 20 MM
1 .O8 8.56 1
Deactivation T h , (sec) rdow (sec) GABA 600 pM 0.232 2.456 GABA
600 PM + PRO 1 MM 0.383 2.593
1 GABA 600 uM + PRO 6 uM 0.52 1 2.785 1 GABA 600 p M + PRO 10
ph4 0.735 2.991 GkBA 600 PM + PRO 20 p M 0.766 3.273
Propofol decreased the rate of GABAAR deseasitization and siowed
receptor deactivation in the presence of saturating concentrations
of G M A .
A. The rate of receptor desensitization was determined for
cuments in Figure 7. The fast and slow components were calculated
using a biexponential fit. As the concentration of propofol
increased there was linle change in the fast time constant however,
there was dose-dependent enhancement of the slow time constant.
B. The rate of deactivation for currents in Figure 7. There was
a dose dependent increase in both the fast and slow time constants
for deactivation.
-
hippocampal and spinal cord neurons. Further, the effect of
midazolam was examined at 3
di fferent concentrations of propofol.
ï h e IssAp ratios were calculated for currents evoked by 600 pM
GABA, 600 p M GABA
and 0.5 p M midazolam (GABA + MDZ), 600 FM GABA and 10 p M
propofol (GABA + PRO),
and GABA, propofol and midazolarn (GABA + PRO + MDZ) in
hippocampal neurons. As
previously described, midazolam did not alter the steady-state
current evoked by 600 p M GABA
(Figure 8a). The IsslIp ratios for currents evoked by GABA and
GABA + MDZ were 0.14 k
0.01 and 0.13 t 0.01, respectively @ > 0.05, ANOVA, n = 8,
Figure 8b). When 10 pM propofol
was CO-applied with GABA, the IssAp ratio increased nom 0.14 t
0.01 to 0.20 i 0.02 @ c 0.05,
ANOVA). However, when midazolam was CO-applied with propofol and
GABA, the IsslIp ratio
was not significantly different fiom that observed in the
presence of propofol and GABA (0.19 f
0.01 versus 0.20 + 0.02, p > 0.05, ANOVA). Hence, midazolam
did not alter the effects of 10
pM propofol on GABAAR desensitization (600 pM GABA).
in spinal cord neurons, the I s d p ratios for currents
activated by 600 p M GABA, GABA
+ 0.5 p M MDZ, GABA + 10 pM PRO or GABA + PRO + MDZ were 0.16 f
0.02, 0.16 + 0.02, 0.21 f 0.02 and 0.21 + 0.02, respectively (n =
7; Figure 8 c,d). The IsslIp values by GABA and GABA + MDZ evoked
responses were significantly different fiom those evoked by GABA
+
PRO and GABA + PRO + MDZ @ c 0.05, ANOVA). Notably, the
conclusions were not
different between hippocampal and spinal cord neurons.
To ensure that the lack of effect of midazolam on
propofol-modulated currents was not
due to the relatively high concentrations of propofol, propofol
was decreased fiom 10 FM to 1 or
6 p M and tested in hippocampal neurons. In the presence of
lower concentrations of propofol(1
-
Figure 8
Midazolam did not effect propofol-induced modulation of GABAAR
desensitization.
A,C. Responses evoked in hippocampal and spinal cord neurons by
600 FM GABA in
the presence of 0.5 pM midazolarn, 10 pM propofol, or midazolam
plus propofol are
shown.
B,D. The bar graphs illustrate mean Iss/Ip ratios calculated for
currents evoked by
GABA. GABA A MDZ. GABA + PRO and GABA + PRO + MD2 in hippocampal
(n =
S ) and Spinal cord neurons (n = 5) . The W I p ratios for
currents evoked by GABA and
GABA - MDZ were different from those of GABA + PRO and GABA -i
PRO - MDZ in both hippocampal neurons and spinal cord neurons
(p
-
or 6 PM), midazolam did not influence the extent of GABAAR
desensitization (Figure 9).
Therefore, midazolarn did not modulate GABAAR desensitization,
nor did it effect the ability of
propo fol to decrease GABAAR desensitization.
Despite the lack of effect that rnidazolarn has on the extent of
receptor desensitization,
midazolarn slowed the rate of current deactivation in the
presence of propofol. The rate of
deactivation for the currents in Figure 9 were fit with a
biexponential equation. Midazolam
increased the fast and slow time constants for both
concentrations of propofol (Table 2).
-
Figure 9
Midazolam did not influence GABAAR desensitization at lower
concentrations of
propofol.
A. Responses evoked by 600 FM GABA, in the presence o f propofol
(1 FM), or
propofol plus rnidazolarn (0.5 FM), were recorded from
hippocampal neurons.
B. Responses evoked by 600 pM GABA. in the presence of propofol
(6 FM). or
propofol plus midazolarn were recorded From hippocampal neurons.
also indicate that
rnidazolarn does not modulate receptor desensitization
regardless of the dose of
propo fol.
-
PRO 1 pM PRO 6 pM
GABA
GABA+PRO
GABA+PRO+MDZ
-
Table 2
Deac tivation Sy,t (sec) 5 , h W (sec) GABA 600 pM 0.232
2.456
- - -- - -
GABA 600 PM + PRO 1 FM 0.383 2.593 1 GABA 600 MM + PRO 1 IJM +
MDZ 0.5 uM 0.475 2.773 1 1 GABA 600 p M + PRO 6 p M ( GABA 600 pM +
PRO 6 pM + MDZ 0.5 MM 0.797 4.337 1
The rate of current deactivation was decreased by midazolam
The deactivation rates f?om currents in Figure 9 were fit with a
bi-exponential equation. The fast and slow time constant were
increased in the presence of propofol and they were further
increased in the presence of midazolarn and propofol.
-
4.2. SECTION 2
4.2.1. EFFECTS OF MIDAZOLAM AND PROPOFOL ON CURRENT EVOKED
BY
SUBSA TURA TING CONCENTRA TIONS OF GABA
The purpose of these experiments was to determine if propofol
and rnidazolarn enhanced
currents evoked by subsaturating concentrations of GABA in an
additive, superadditive or
subadditive manner. Two models were employed; a Fixed-Dose Model
and Isobolographic
Analysis.
Using the Fixed-Dose Model, we examined the eff i t of each dmg
aione and the
combination of propofol and midamlam. The calculated additive
response was then compared to
the effect of the dmgs in combination. To determine if the
interactions were influenced by ce11
type we recorded responses from hippocampal and spinal cord
neurons for GABA (3 PM),
propofol (10 PM) and rnidazolarn (0.5 FM). Expenments were also
conducted using different
concentrations of propofol (10 and 1 PM) and GABA (3 and 0.3 pM)
in hippocampal neurons.
These concentrations were previously reported by Reynolds et al.
(1996) to produce a
superadditive enhancement of GABA-evoked responses recorded fiom
oocytes expressing
hurnan a 1 B2y2L and a2P2y2L subunits.
In cultured hippocampal neurons, PRO (10 FM), and MDZ (0.5 PM),
enhanced GABA-
evoked currents (3 PM) recorded fiom hippocampal neurons (Figure
Na). The currents
potentiated by MD2 andor PRO were less than the maximal current
generated by 600 pM
GABA. GABA-activated currents were enhanced by MDZ, PRO and PRO
+ MDZ by a factor of
1-45 + 0.21, 3.44 2 0.41 and 4.15 f 0.54, (n = 8; Figure lob).
There were no significant differences between currents enhanced by
PRO + MDZ and the theoretical additive response @ >
-
Figure 1 0
Propofol and midazolam (or flurazepam) produced an additive
enbancement of
GABA (3 PM)-evoked currents from hippocarnpal neurons.
A,D. Responses evoked by combinations of 3 pM GABA, 10 pM
propofol, and 0.5 pM
midazolarn (or flurazepam) were srnaller than the responses
evoked by 600 pM GABA in
hippocampal neurons.
B,E. Currents were normalized to the peak response evoked by 3
piM GABA. There was
no difference between the measured response to PRO + MDZ (n = 8)
(or F L ü (n = 7))
and the theoretical additive values (p > 0.05, Student's
t-test).
C,F. The charge transfer was calcuiated for al1 combinations of
dmgs and normalized to
the value calculated for 3 pM GABA. There was no difference
between the measured
PRO + MDZ (or FLU) response and the theoretical additive value
for charge transfer (p >
0.05. Student's t-test).
-
0.05, S tudents t-test). Similar results were obtained when
another benzodiazepine, flurazepam
(FLU, 0.5 FM), was substituted for midazolam. GABA-activated
currents were enhanced by
FLU, PRO and PRO + FLU by a factor of 1.26 k 0.06,2.49 f 0.36,
and 3.24 f 0.56, respectively
(n = 7; Figure 10 d,e). This suggested that the additive
interaction was not influenced by the
benzodiazepine being examined.
In addition to examining the effects of propofol and midazolam
on peak currents, we also
measured the total current recorded during the application of
agonist. Charge transfer was
measured by integrating the area under the current (see
methods). Charge transfer was enhanced
by MDZ, PRO or MD2 + PRO by a factor of 1.83 t 0.26, 3.99 +
1.08, and 5.04 + 1.75 (n = 8; Figure 10c). The enhancement of
charge transfer by FLU, PRO or PRO + FLU was 1.30 t 0.07,
2.20 + 0.31 and 2.83 k 0.39, respectively (n = 7; Figure [Of).
Again, the theoretical additive enhancement of charge transfer was
not significantly different than the enhancement for MD2
(or FLU) + PRO modulated currents @ > 0.05, Students t-test).
Therefore, the interaction
between propofol and midazolam or flurazepam, at these
concentrations, was additive.
The W relationship for GABA-evoke