Available via license: CC BY-NC-SA 4.0
Content may be subject to copyright.
Systems/Circuits
Location of the Mesopontine Neurons Responsible for
Maintenance of Anesthetic Loss of Consciousness
XAnne Minert, Shai-Lee Yatziv, and XMarshall Devor
Department of Cell and Developmental Biology, Institute of Life Sciences, and Center for Research on Pain, The Hebrew University of Jerusalem, Jerusalem
91904, Israel
The transition from wakefulness to general anesthesia is widely attributed to suppressive actions of anesthetic molecules distributed by
the systemic circulation to the cerebral cortex (for amnesia and loss of consciousness) and to the spinal cord (for atonia and antinocice-
ption). An alternative hypothesis proposes that anesthetics act on one or more brainstem or diencephalic nuclei, with suppression of
cortex and spinal cord mediated by dedicated axonal pathways. Previously, we documented induction of an anesthesia-like state in rats
by microinjection of small amounts of GABA
A
-receptor agonists into an upper brainstem region named the mesopontine tegmental
anesthesia area (MPTA). Correspondingly, lesioning this area rendered animals resistant to systemically delivered anesthetics. Here,
using rats of both sexes, we applied a modified microinjection method that permitted localization of the anesthetic-sensitive neurons with
much improved spatial resolution. Microinjected at the MPTA hotspot identified, exposure of 1900 or fewer neurons to muscimol was
sufficient to sustain whole-body general anesthesia; microinjection as little as 0.5 mm off-target did not. The GABAergic anesthetics
pentobarbital and propofol were also effective. The GABA-sensitive cell cluster is centered on a tegmental (reticular) field traversed by
fibers of the superior cerebellar peduncle. It has no specific nuclear designation and has not previously been implicated in brain-state
transitions.
Key words: anesthesia; arousal; MPTA; reticular formation; syncope; wet blanket hypothesis
Introduction
Transient loss of consciousness (LOC, “reversible coma”) and the
other components of the anesthetic state, immobility, analgesia
and amnesia, are classically attributed to generalized pharmaco-
logical suppression of CNS function. Once ascribed to altered
physical properties of lipid membranes, suppression is now
known to be mediated by ubiquitous inhibitory ligand-gated ion
channels, most prominently the GABA
A
-receptor (GABA
A
-R;
Campagna et al., 2003;Franks, 2008). A refinement of this “gen-
eralized suppression” or “wet blanket” hypothesis adds that each
component of anesthesia is realized in a distinct CNS structure.
Thus, circulating anesthetic molecules suppress the neocortex
and hippocampus to bring about LOC and amnesia and the
brainstem and spinal cord to cause atonia and analgesia (Lukatch
and MacIver, 1996;Antognini et al., 2003;Grasshoff et al., 2005;
Hentschke et al., 2005;Bonhomme et al., 2012;Raz et al., 2014).
An alternative hypothesis holds that anesthetic molecules act
at a focus from which they engage dedicated axonal pathways.
The axons, in turn, project to the far-flung effector structures,
from cortex to cord, inducing anesthesia secondarily. This “ded-
icated pathways” hypothesis originated with Moruzzi and Ma-
Received Feb. 26, 2017; revised Aug. 6, 2017; accepted Aug. 9, 2017.
Authorcontributions:A.M.andM.D.designedresearch;A.M.,S.-L.Y.,andM.D.performedresearch;A.M.,S.-L.Y.,
and M.D. analyzed data; A.M. and M.D. wrote the paper.
This work was supported by the Israel Health Ministry, the Seymour and Cecile Alpert Chair in Pain Research, the
Willem Bean Legacy Fund, and the Hebrew University Center for Research on Pain. The funders played no role in the
planning or execution of the experiments or the drafting of this manuscript. We thank Yelena Fishman, Adi Gold-
enberg,TamirAvigdor, and Mark Baron for their contributions and Clif Saper and Jun Lu for helpful comments on the
study and for hosting the senior author at BIDMC where the initial experiments were performed. The bonus time
method as a means of improving spatial resolution was originally suggested by Clif Saper.
The authors declare no competing financial interests.
Correspondenceshouldbe addressed to: Prof. M. Devor, Department of Cell and Developmental Biology, Institute
of Life Sciences, The Hebrew University of Jerusalem, Jerusalem 91904, Israel. E-mail: marshlu@mail.huji.ac.il.
DOI:10.1523/JNEUROSCI.0544-17.2017
Copyright © 2017 the authors 0270-6474/17/379320-12$15.00/0
Significance Statement
General anesthesia permits pain-free surgery. Furthermore, because anesthetic agents have the unique ability to reversibly switch
the brain from wakefulness to a state of unconsciousness, knowing how and where they work is a potential route to unraveling the
neural mechanisms that underlie awareness itself. Using a novel method, we have located a small, and apparently one of a kind,
cluster of neurons in the mesopontine tegmentum that are capable of effecting brain-state switching when exposed to GABA
A
-
receptor agonists. This action appears to be mediated by a network of dedicated axonal pathways that project directly and/or
indirectly to nearby arousal nuclei of the brainstem and to more distant targets in the forebrain and spinal cord.
9320 •The Journal of Neuroscience, September 20, 2017 •37(38):9320 –9331
goun, who suggested that anesthetics suppress the mesopontine
ascending reticular activating system (aRAS) which, in turn, leads
to cortical suppression and LOC (Bremer, 1936;Moruzzi and
Magoun, 1949;French et al., 1953;Magni et al., 1959). Contem-
porary models parse the aRAS into numerous brainstem and
diencephalic sleep-promoting and sleep-suppressing “arousal
nuclei” with associated ascending and descending effector path-
ways. Anesthesia is induced when agents substitute for an endog-
enous neurotransmitter(s) in one or more nodes in this switching
circuitry, “hijacking” network function. Proposals about where
exactly this might occur include the locus coeruleus, tuberomam-
millary nucleus, medial thalamus, lateral hypothalamus, and pre-
optic area (Correa-Sales et al., 1992;Alkire et al., 2008;Franks,
2008;Lu et al., 2008;Brown et al., 2010;Moore et al., 2012;
Zecharia et al., 2012;Scharf and Kelz, 2013;Anaclet et al., 2014;
Baker et al., 2014;Weber et al., 2015).
A significant advance toward defining the locus of anesthetic
action was the discovery that an anesthesia-like state is induced
rapidly and reversibly by microinjecting GABA
A
-R agonists into
a circumscribed, bilaterally symmetrical region in the upper
brainstem, the mesopontine tegmental anesthesia area (MPTA)
(Devor and Zalkind, 2001;Voss et al., 2005; see video at http://
links.lww.com/AA/B466). This state closely resembles systemic
anesthesia by behavioral, electrographic, and functional criteria
and can be sustained by repeated microinjections (Sukhotinsky et
al., 2007;Abulafia et al., 2009;Namjoshi et al., 2009;Devor et al.,
2016;Sukhotinsky et al., 2016). Anesthesia comes on much too
rapidly to permit drug diffusion to distant brain regions (seconds
to a few minutes) and effective doses are far too small to survive
dilution during vascular redistribution. These observations are
therefore inconsistent with generalized CNS suppression, but
they are compatible with the dedicated pathways hypothesis. In
addition, we found that lesioning the MPTA renders animals
resistant to anesthetic induction by the systemic route (Minert
and Devor, 2016). The MPTA cluster may thus be singular. No
comparable anesthetic-sensitive structure has been found despite
systematic searching and, if another exists, engaging it is apparently
insufficient because anesthesia cannot be induced in MPTA-
lesioned animals at clinically relevant doses. The intrinsic prop-
erties of MPTA neurons and their connectivity makes them both
sufficient and necessary for GABAergic anesthetic induction
(Sukhotinsky et al., 2016).
A limitation of our prior studies, however, is limited spatial
resolution. MPTA microinjections through indwelling catheters
likely exposed additional structures, including adjacent arousal
nuclei, to the agonists. This is the reason that we chose to call the
anesthetic-sensitive region an “area” without stipulating a specific
nuclear designation. More precise localization of the relevant neu-
rons is an important prelude to characterizing them in more detail
and determining the mechanism whereby they switch brain and spi-
nal cord state between wakefulness and anesthesia (Devor et al.,
2016). With this aim in mind, we have implemented a modified drug
delivery protocol with enhanced spatial resolution.
Materials and Methods
Overview of the “bonus time” method
Briefly, rats were anesthetized using the short-acting agent propofol and
mounted in a stereotaxic instrument. Then, minute volumes of GABA
A
-R
agonists were microinjected unilaterally into the MPTA via a fine glass
micropipette. We monitored the degree to which localized deposition of
drug delayed the time to emergence from anesthesia (“anesthesia time“)
beyond what would have been expected without drug microinjection.
Although maintenance of anesthesia is not identical to anesthetic induc-
tion, it is a reasonable surrogate in light of the fact that induction per se
has already been documented after microinjection into the MPTA area
during the awake state. Prolongation of anesthesia time is referred to here
as a positive “bonus” and the approach is called “the bonus time”
method.
Animals and surgery
Experiments were performed in two series using similar, but nonidenti-
cal methods. The first series, performed in the Department of Neurology,
Beth Israel Deaconess Medical Center (BIDMC)/ Harvard University
(Boston, MA) used adult male Sprague Dawley rats (270–330 g). In the
subsequent series, performed at the Institute of Life Sciences, the Hebrew
University of Jerusalem Israel (HUJI), we used adult female Wistar-
derived Sabra strain rats (250–350 g). At both venues, animals were
maintained under specific pathogen-free conditions, 1–3 per cage, room
temperature 21–23°C, 12 h:12 h day:night cycle, with lights on at 07:00.
Food and water were available ad libitum. Experimental protocols were
approved by the Institutional Animal Care and Use Committees of
BIDMC and HUJI and, at both venues, they were conducted in accor-
dance with the United States Public Health Service’s Policy on Humane
Care and Use of Laboratory Animals.
At BIDMC, a preliminary surgical procedure was performed to pro-
vide intravenous access for the main experiment. Under chloral hydrate
anesthesia (350 mg/kg, i.p.), a sterile chronic indwelling catheter was
introduced into the vena cava via a femoral vein tributary and firmly
anchored to local fascia. The intravenous part of the catheter was made of
soft silicone tubing. This was glued to a polyethylene tube (Intramedic
PE50) that was tunneled under the skin and exteriorized at the back of the
neck. Rats were given an analgesic postoperatively (Flunixin, 1 mg/kg, s.c.)
and were allowed to recover for 4–14 d, at which time they underwent a
terminal experiment in which drugs were microinjected into the MPTA.
During the recovery time, catheters were flushed intermittently with
heparin solution to help keep them patent. In experiments performed at
HUJI, the previously implanted catheter for intravenous access was sub-
stituted by placing, on the day of the experiment itself, an indwelling
neoflon catheter (26 G; Becton Dickinson) into the tail vein under brief
isoflurane sedation.
At both venues, experiments began by anesthetizing the rat with a
bolus dose of propofol (13 mg/kg, i.v., at BIDMC, propofol 1%; Gensia
Sicor Pharma; at HUJI, Propofol-Lipuro 1%; B. Braun Medical). It was
then mounted in a stereotaxic apparatus with head level between bregma
and lambda and a small craniotomy was made over the MPTA target
using a dental burr. At BIDMC, frontoparietal screw electrodes were
placed bilaterally for recording EEG and a pair of needle electrodes were
placed in the biceps femoris muscle to register the hindlimb EMG. The
empirical endpoint used to mark the emergence from anesthesia was the
sudden appearance of repetitive “pacing” movements of one or both
hindlimbs. As soon as this was observed, animals were given a fixed
supplemental dose of 10 mg/kg intravenous propofol to reinitiate anes-
thesia. This dose was determined in preliminary titration experiments to
yield ⬃15 min of anesthesia before pacing reappeared. Pacing behavior
served as a convenient and unambiguous marker of emergence. However,
we also established that its onset was always accompanied by rhythmic
bursts in the biceps femoris EMG (Fig. 1C), nocifensive response to
hindpaw, and/or earlobe pinch. Slightly later theta waves (4– 8 Hz) and
sometimes higher-frequency EEG components (15–30 Hz) appeared, add-
ing to and later replacing the prominent propofol-induced delta-band
peak (0.5–4.0 Hz). In addition, in calibration trials in which the animal
was released from the stereotaxic and the supplemental propofol post-
poned, we observed prompt return of righting. Having determined in
this way that pacing reliably marks emergence from anesthesia, in subse-
quent experiments the behavioral endpoint was used as the primary
marker of emergence. In particular, pinch stimuli were avoided until
after pacing began because they are likely to accelerate emergence in an
uncontrolled manner and introduce an element of experimenter bias.
After one or more supplemental intravenous injections of propofol at
the fixed bolus dose, we began MPTA microinjections. The duration of
anesthesia provided by the fixed bolus dose before commencement of
drug microinjections was remarkably stable from trial to trial (Fig. 2) and
served as a basis for determining whether drugs microinjected into the
Minert et al. •Transient Loss of Consciousness J. Neurosci., September 20, 2017 •37(38):9320 –9331 • 9321
MPTA significantly extended the expected du-
ration of anesthesia, providing a positive bonus
time.
MPTA microinjection
Procedure. First, a micropipette pulled from
fine borosilicate glass (inner diameter 0.3 or 0.5
mm (catalog #B100-30-7.5HP or #BF100-50-
15, respectively; Sutter Instruments) with tip
broken to a diameter ⬃20–30
m was passed
through the burr hole and dura and lowered to
the level of the MPTA. In some experiments,
we deliberately targeted nearby off-target loca-
tions. We began with relatively large microin-
jections (200 and 500 nL) delivered to initial
coordinates located within the boundaries of
the MPTA as defined using the indwelling can-
nula method (⫺7.6 to ⫺8.8 mm caudal to
bregma, 1.3 mm lateral to the midline, and 6.3–
7.3 mm below the dura). Our strategy was to
roam from there, progressively reducing the
volume of drug microinjected, seeking a site(s)
that provided a significant extension of anes-
thesia time. Once such a “hotspot” was found,
additional small volume microinjections were
placed rostrocaudal, mediolateral, and dorso-
ventral to effective site. Other trials targeted
specific candidate nuclei including the pedun-
culopontine tegmental nucleus (PPTg), the
parabrachial complex (PB), and the oral pon-
tine reticular nucleus (PnO). All microinjec-
tions in this study were unilateral, favoring the
left side, although to ensure bilateral symme-
try, trials were also made on the right side.
Before insertion, microinjection pipettes
were filled to several millimeters above the
taper with one of the following test agents:
muscimol (2 mMin saline; Sigma-Aldrich, Re-
hovot, Israel), pentobarbital (200 mg/ml; Pen-
tal, CTS, Kiryat Malachi, Israel), propofol
(Propofol-Lipuro 1%; B. Braun Medical, or
97%, 2,6-diisopropylphenol; Sigma-Aldrich),
Figure 2. Anesthesia time was independent of the number of antecedent propofol boluses given. A, In rats given 2 (n⫽60), 3
(n⫽50), 4 (n⫽20), or 5 (n⫽7) consecutive bolus doses of propofol mean anesthesia time was the same on the first bolus as on
the n
th
.B, Lines represent anesthesia times of 60 individual rats upon repeated bolus dosing (all at HUJI). To ease separation, lines
representing the seven rats given five consecutive doses are colored. Error bars in Aindicate ⫾SD.
Figure1. Microinjections of muscimol into the MPTA. A,Sagittal section showingthe dye spread associatedwith a microinjectionof 200 nLof muscimol into theMPTA. B, Coronalsection showing
the dye spread associated with on-target muscimol microinjections of 20 nL, first on the left, then 20 min later on the right (rat #340). These trials yielded bonus times of 10 and 21 min, respectively.
The boundaries of the MPTA as defined by Devor and Zalkind (2001) are overlaid on the right-sided microinjection (1000 ⫻1500
m, solid outline). The common core of the MPTA is shown within
the 500 ⫻750
m dashed frame. C, The sudden appearance of repetitive hindlimb “pacing” movements was accompanied by rhythmic bursts in the biceps femoris electromyogram. Both were
silenced by the standard intravenous bolus of propofol.
9322 •J. Neurosci., September 20, 2017 •37(38):9320 –9331 Minert et al. •Transient Loss of Consciousness
lidocaine (5% in PBS, prepared fresh from powder; Sigma-Aldrich), te-
trodotoxin (TTX, 1 mMin artificial CSF; Alomone Laboratories), or
saline (0.9% NaCl). In most experiments, the solution microinjected
included pontamine sky blue dye (1 or 2%; Gurr/BDH Chemicals) to
mark the microinjection site.
For microinjection, the butt of the pipette was attached by a length
of polyethylene tubing to a valve apparatus that delivered pulses of
positive air pressure. The pressure, duration, and duty cycle of the
pulses were adjustable and were varied with tip diameter and solution
viscosity to provide controlled extrusion of the test agent into the
brain. Typically, we used pressure of 1 kg/cm
2
, 20 ms pulses, 1–10
pulses/s. Between pulses, the back pressure on the pipette fell rapidly
to zero. Calibrated volumes of test solutions, 500-10 nL, were micro-
injected by monitoring the change in the position of the meniscus
inside the pipette using an ocular rule under optical magnification
(0.5 mm bore: 200 nL/mm, 0.3 mm bore: 70 nL/mm). The volume per
millimeter of internal pipette bore was calibrated previously using a 1
l Hamilton syringe.
Micropipettes were lowered to the target position and microinjection
was initiated 1–3 min before the anticipated time of emergence of the
animal from effects of the most recent intravenous bolus dose of propo-
fol. Microinjection took ⬃20 –60 s. The micropipette was left in place for
3–5 min after microinjection, but was withdrawn immediately if pacing
began. Time was measured from the completion of the microinjection
until hindlimb pacing began. If this equaled the anticipated emergence
time (i.e., the time anticipated for the beginning of pacing), we registered
a “bonus” of zero for that trial. If the time to pacing was greater than
anticipated, then the difference was registered as a positive bonus, indi-
cating prolongation of the duration of anesthesia. As soon as the bonus
time was noted, an additional bolus dose of intravenous propofol was
given. Animals were not allowed to awaken from anesthesia while re-
strained in the stereotaxic.
Evaluation of bonus time. In the first experiments, we established the
anticipated time of emergence in each rat individually by averaging sev-
eral cycles of fixed-dose intravenous propofol injections. The time from
bolus injection until the beginning of pacing, “anesthesia time,” proved
to be stable with no consistent change over repeated cycles of bolus
dosing within and across rats (Fig. 2). This justified a modification of the
procedure such that, for each rat, we used the running average of (pre-
microinjection) anesthesia times obtained in all previous experimental
trials. This value converged rapidly. Over 130 injection cycles (at HUJI),
it averaged 11.0 ⫾2.0 min (⫾SD, coefficient of variation ⫽0.18; 42 rats
till #359). This value remained stable over subsequent experiments
with no detectable drift. Curiously, in experiments performed at
BIDMC, anesthesia time after the identical bolus dose of propofol was
significantly longer [19.0 ⫾6.3 min, coefficient of variation ⫽0.33).
The difference is p⬍0.001, 95% confidence interval (CI) ⫽7.1–9.5].
Although this could be due to a variety of venue-specific factors, a
likely factor is the use of male Sprague Dawley rats at BIDMC and
female Sabra rats at HUJI. Clinical studies report significantly longer
emergence times from propofol-induced anesthesia in men than in
women when measurements begin at an equivalent anesthetic depth
based on bispectral index values (Hoymork and Raeder, 2005;
Choong et al., 2013).
In some experiments, only a single microinjection was given. In others,
we gave more either at an adjacent location or when attempting to rep-
licate a successful microinjection using a smaller volume (maximum of
four microinjections in a single rat; Table 1). Follow-up microinjections
were also given when, during the roaming process, muscimol yielded a
short bonus time, when saline was delivered as a control, or when a novel
Table 1. Summary of muscimol microinjection trials showing venue, volume delivered, and bonus time obtained (in minutes, 91 trials in 53 rats)
BIDMC rat #
500 nL or 200 nL
bonus time
HUJI rat #
200 nL
bonus time
100 nL or 50 nL
bonus time
20 nL
bonus time
10 nL
bonus time
Significant NS Significant NS Significant NS Significant NS Significant NS
465(R) 321 1, 2
5 13 358 10
6 35 359 21(R)
7 57(R) 322 26 5 35
839 323 14
9 26 324 36 24
10 34 325 10
11 22 327 50 23(R)
12 50 329 33 3(R) 23
13 45 370 9 2(R)
14 10 371 1, 2(R)
15 5 372 5(R), 6
29 28 330 32, 34 2(R)
30 45 331 22, 46(R)
31 45 332 15, 18(R)
32 28 338 3, 4(R), 6, 6(R)
33 36 339 1, 4(R)
36 28 351 2(R) 3
37 28 340 10, 21(R)
341 18, 30(R)
342 8
500 nL or 200 nL
bonus time
343
345
15, 19, 22(R)
11(R), 20 18
HUJI rat # Significant NS 346 16(R), 19(R) 5
369 15 347 30 4(R)
314 21 4 348 18, 22(R)
316 15 3 349 3(R) 31
317 2, 2, 3 350 3(R) 1(R)
318 58, 8 6 352 4(R) 3
Note that some rats had up to four microinjection trials. Scores marked (R) were from right-sided microinjections; italics indicate use of the volume set in italics. Criterion for statistical significance is described in the Materials and Methods.
NS, Not significant.
Minert et al. •Transient Loss of Consciousness J. Neurosci., September 20, 2017 •37(38):9320 –9331 • 9323
drug was involved for which we wanted to verify that the microinjection
had in fact been on target by follow-up delivery of muscimol or pento-
barbital to the same locus.
Histology
Definition of the microinjection site. At the end of each experiment, as
soon as pacing commenced, a final intravenous injection of propofol was
given and the animal was rapidly killed by transcardial perfusion with
0.9% saline followed by 10% neutral 0.1 MPO
4
buffered formalin (pH
7.3; Sigma-Aldrich), both at 37°C. The brain was dissected out shortly
after perfusion and postfixed at 4°C for at least 24 h in the same fixative.
It was then transferred to 20% sucrose in PBS containing 0.02% sodium
azide (PBS-azide) at 4°C. At least 24 h later, tissue was cut on a freezing
microtome into 50- or 100-
m-thick coronal sections. We collected all
sections in the region of microinjection beginning from the level at which
the first pontamine dye spot was seen in the block face ⬃200–300
m
before dye actually appeared in sections and continuing until after dye
was no longer present. Every third section through each dye spot was
immediately mounted onto a glass slide in serial order. Intervening sec-
tions were saved in PBS-azide (4°C). The number and relative location of
individual dye spots seen on the block face as cutting proceeded always
matched the microinjection protocol. The slides were air dried over-
night, cleared with xylene, and coverslipped using Entellan (Merck).
Tissue sections containing dye spots were optically projected onto the
corresponding coronal level in the rat brain atlas of Paxinos and Watson
(1998) and the magnification was varied until a best fit was obtained to
the dorsal contour of the colliculi and the ventral contour of the pons.
The maximal visible extent of the contour of the dye spot was then
transferred by hand onto the atlas section by an individual blinded to
microinjected drug, volume, and the experimental outcome. The maxi-
mal cross-section of each microinjection was measured planimetrically
based on a section in which the micropipette tract emerged from the dye
spot (Fig. 1A,B).
Estimation of drug spread. We estimated the extent of drug spread
beyond the sphere defined by the volume microinjected based on the
maximal dye-spread cross-section measured. The calculation assumed
that spread was approximately isotropic and formed a sphere around the
tip of the injection micropipette. This assumption was based on the fact
that dye spots were round in both frontal and sagittal sections (Fig.
1A,B). The radius (r) of the droplet formed by drug solutions microin-
jected, before spread in the tissue, was calculated from the volume in-
jected. For the smallest volumes used (e.g., 10 nL), rwas taken as 134
m,
calculated by solving for rin the formula for the volume of a sphere V⫽
4/3
r
3
. The radius of the sphere formed by dye spread after microinjec-
tion was based on the cross-sectional area of dye-spread measured, cal-
culated by solving for rin the formula for a circle, A⫽
r
2
. The
corresponding volume of spread was then calculated using 4/3
r
3
.
Neuronal density. Estimating the number of MPTA neurons exposed
to microinjected anesthetics requires a measure of neuronal density.
Four naive rats were perfused as above and brains were cut in serial 50
m frozen sections and immunolabeled for NeuN, a selective nuclear
marker of viable neurons. Briefly, sections were incubated overnight in
polyclonal mouse anti-NeuN antibody (Millipore catalog #MAB377,
1:25,000, room temperature), followed by 1.5 h in secondary antibody
(biotinylated goat anti-mouse IgG; Vector Laboratories, 1:1000; at 37°C).
Bound antibody was visualized using the ABC reaction and diaminoben-
zidine (Vectastain Elite ABC kit; Vector Laboratories).
One section from the mesopontine border (level ⫺8.0 or ⫺8.3 mm
relative to bregma; Paxinos and Watson, 1998) was selected for counting.
The field of view was overlaid with a 1000 ⫻1500
m counting frame
representing the caudal MPTA (Fig. 1B) and all immunolabeled neurons
within this region of interest (ROI) were counted bilaterally at ⫻250
magnification using the Neurolucida system (version 10.51; MBF Biosci-
ence) and their locations were plotted. Plotting began in an upper corner
of the counting frame and proceeded systematically in a zig-zag manner
one field of view at a time until the entire frame was covered. Cells on the
edge of the ROI had to have ⬎50% of their area within the frame to be
included in the count. The two counts (ROIs in the left and right MPTA)
for each rat were averaged and these values were averaged over the four
rats used. The average cell count per counting frame divided by the area
of the counting frame provided neuronal density (neurons per square
millimeter). This was converted to neurons per unit volume taking into
consideration the section thickness of 50
m. Therefore, the crude den-
sity in units of neurons per cubic millimeter was 20⫻the measured
number of neurons per square millimeter. This crude density measure is
an overestimate, however, because some neurons are split by the mi-
crotome blade and counted twice. To correct for this factor, we used
Abercrombie’s method (Koningsmark, 1970), which estimates the de-
gree of overcounting based on section thickness (50
m) and mean
neuronal diameter (11.8
m) based on a measured mean cell area of
109
m
2
(see Results). The overcount was rectified by multiplying by the
calculated correction factor ⫽0.83.
Neuronal size and shape. Neurons within a smaller ROI that included a
region we call the common core of the MPTA (see Results) were evalu-
ated for soma size and shape. This ROI covered 25% of the area of the
caudal MPTA counting frame (500 ⫻750
m) and was separated from
its upper, medial, and lateral borders by 250
m(Fig. 1B, dashed out-
line). Using the same sections as for neuronal counting, the perimeter of
each neuron for which at least 50% of the soma area fell within the
smaller ROI was traced with an on-screen cursor. From these measure-
ments, Neurolucida calculated the somatic area and aspect ratio as fol-
lows: (min diameter) ⫼(max diameter). As aspect ratio approaches 1,
the soma becomes less flat and more symmetric (e.g., circular or square).
We also calculated the form factor as follows: [(4
⫻area) ⫼(perimeter
2
)], a
parameter reflecting the tortuosity of the perimeter, and roundness as
follows: [(4/
⫻area) ⫼(maximum diameter)
2
], a parameter reflecting
the compactness of the soma, with a circle having a roundness of 1.
Statistical evaluation
Expected duration of anesthesia after repeated fixed-dose bolus injec-
tions of propofol was based on the observed average anesthesia time ⫾
SD as described above. In light of the stability of anesthesia time after
bolus injections of propofol (Fig. 2), it was possible to evaluate for indi-
vidual microinjection trials whether anesthesia time was increased sig-
nificantly; that is, whether the bonus time obtained was significantly
greater than zero. In doing this, we applied a highly conservative crite-
rion. Specifically, bonus time in a particular trial needed to be at least 3
SDs ⬎0 to be declared statistically significant (Z-score ⫽3, 1-tailed p⫽
0.001). Therefore, the minimally significant bonus time was 18.8 min
(BIDMC) and 6.1 min (HUJI). Differences between group means were
evaluated using 2-tailed Student’s ttests after confirming normality of
the underlying data distributions. Significance of linear regressions (R
2
)
was tested using Pearson tests. The 95% CIs were calculated using Med-
calc (https://www.medcalc.org/calc/comparison_of_means.php). All
means are given ⫾SD except where use of SEM is noted.
Results
Locating the MPTA
A total of 91 unilateral muscimol microinjections, made in 53
rats, were targeted to the MPTA area and adjacent areas of inter-
est (Table 1). We began with a few 500 nL microinjections to
match the ones given in our prior studies using unilateral micro-
injections via chronically implanted cannulae (Devor and Zal-
kind, 2001;Devor et al., 2016). Then, we gradually reduced the
volume to as little as 10 nL at each step, attempting to refine the
boundaries of the effective area (Fig. 1A,B). The main result was
identification of a small subregion within the originally de-
fined MPTA area at which very small microinjections reliably
extended anesthesia time, whereas closely adjacent microin-
jections did not.
Figure 3 shows the location of the injection centers of all 91
trials, where filled symbols indicate loci that yielded a significant
bonus time and open symbols no bonus. The scatter of effective
loci using volumes of 500–50 nL was strikingly larger than that
using smaller volumes (20–10 nL). This is because, given the
greater spread of drug associated with larger volumes, the center
9324 •J. Neurosci., September 20, 2017 •37(38):9320 –9331 Minert et al. •Transient Loss of Consciousness
of effective microinjections could be farther from the effective cell
cluster (hotspot). The loci of unsuccessful microinjections sur-
rounded the central effective core in all three dimensions. The
spatial spread of the pontamine sky blue dye used to track the
microinjected muscimol was typically spherical with a tear-
shaped extension running vertically along the track of the mi-
cropipette used for injection (Figs. 1A,B,4). Plots of all 17 small
volume (10 or 20 nL) microinjections that were effective (i.e.,
which yielded a significant bonus) clustered in a bilaterally sym-
metrical region just ventral and lateral to the periaqueductal gray,
at the far lateral edge of the decussation of the superior cerebellar
peduncle (SCP) (Fig. 5A). The dye spread area of each of these
microinjections is indicated in light pink in the figure, along with
the core area common to all, which is dark red (left and right
sides). Dye spread associated with the 14 microinjections cen-
tered on this plane that failed to yield a significant bonus sur-
round this area (volumes 10 –200 nL; light blue diagonal lines). In
only two of these ineffective trials was there likely to have been
overlap with the common core area (rat #15, 200 nL centered
⫺8.0 mm from bregma, and rat #322, 200 nL centered ⫺8.3 mm
from bregma, both bonus times ⫽5 min; Fig. 5A, left).
In contrast, in our original study in which relatively large
(ⱖ500 nL) microinjections were made through previously im-
planted guide cannulae, we logged failure to induce anesthesia in
nearly half of trials (46%) in which targeting appeared to be ad-
equate. A number of technical causes for this were suggested
(Devor and Zalkind, 2001). Here, we logged only two such fail-
ures (5.4% of trials). We attribute the improvement to the better
control of drug delivery afforded by the bonus time method and
to the use of smaller, more focal microinjections. For example, it
has been reported that muscimol delivered to the PnO promotes
arousal (Xi et al., 2001;Watson et al., 2011). Large microinjec-
tions that exposed much of the PnO together with the MPTA
might have attenuated effects obtained here using more focal
MPTA microinjections.
Figure 5Bshows all 35 effective microinjections (23 rats, in-
cluding the 50 –500 nL microinjections). In 33 of the cases (94%),
the dye spread overlapped the common core area on the left
and, in the remaining two, it approached very closely and in
fact did include the mirror common core (right side). Herein-
after, the term “common core” will refer to the dark red region
on the right side in Figure 5A. A number of effective microin-
jections using larger volumes had their centers rostral to the
⫺8.0 plane or caudal to the ⫺9.16 plane (n⫽10 and 8, re-
spectively; Fig. 3). Extrapolating based on dye spread radius
(Myers, 1966;Myers and Hoch, 1978) in each case, the mus-
cimol would likely have reached the common core. Overall,
bonus times for the n⫽37 muscimol microinjections that
included the common core averaged 25.1 ⫾14.0 min, whereas
bonus times of the n⫽12 surrounding microinjections that
did not include the common core averaged 3.6 ⫾2.4 min
(Table 1). The difference is statistically significant (p⬍
0.0001; 95% CI ⫽29.7 ⫺13.3).
Microinjections that fell within the PPTg, the PB, and the
PnO, but in which the dye spread did not include the common
core of the MPTA, failed to yield a significant bonus (PPTg: 4 rats,
4 trials, bonus ⫽3.0 ⫾1.8; PB: 3 rats, 6 trials, bonus ⫽4.2 ⫾3.1
min; PnO: 3 rats, 5 trials, bonus ⫽3.2 ⫾1.0 min; Figs. 3,4). A few
microinjections, including these nuclei, did produce a significant
bonus, but, in each one, the marker dye and presumably the
muscimol, reached the MPTA common core (Fig. 3). Saline
microinjections centered on the common core never yielded a
significant bonus (10 trials in 5 rats, bonus ⫽2.6 ⫾1.1 min,
horizontal dark blue lines; Fig. 5C).
The laterodorsal tegmental nucleus (LDTg) is another nearby
nucleus implicated in sleep–wake regulation. No microinjections
were restricted to the LDTg, but 13 fell in the MPTA without
impinging on the LDTg and all of these yielded significant bonus
scores. Together, these observations indicate that the functional
hotspot of the MPTA region is not the PPTg, PB, PnO, or LDTg.
In addition, except for the LTDg, for which data are not available,
we can rule out the others as able to support anesthesia when ex-
posed to muscimol.
Figure3. Small volume muscimol microinjections(10 –20 nL) locate the hotspot for anesthetic induction withinthe MPTA area. Symbols mark centers of the microinjections listedin Table 1, with
the volume microinjected indicated for each symbol. Filled symbols indicate trials that yielded a significant bonus time; open symbols indicate bonus times that were not significant. Numbers above
the coronal sections indicate the section’s location caudal to bregma (in millimeters; outlines from Paxinos and Watson, 1998). Microinjections that fell on planes ⫺7.0 to ⫺8.0 are gathered on the
most rostral section outline (⫺7.8 mm), those on planes ⫺8.0 to ⫺9.16 are gathered on the middle section (⫺8.8 mm), and those that fell on planes ⫺9.3 to ⫺10.8 are gathered on the caudal
section (⫺9.8 mm). Many effective microinjections overlap, especially in the dark (red) clusters.
Minert et al. •Transient Loss of Consciousness J. Neurosci., September 20, 2017 •37(38):9320 –9331 • 9325
Exposure of MPTA neurons to
GABAergic anesthetic agents
The MPTA area was originally discovered
using systematic intracerebral microin-
jections of the barbiturate anesthetic pen-
tobarbital (Devor and Zalkind, 2001).
However, because muscimol is a direct
GABA
A
-R agonist, in contrast to pento-
barbital, which acts at an allosteric modu-
latory site on the receptor (Lo¨scher and
Rogawski, 2012), and because it has a
much higher affinity for the GABA
A
-R
than pentobarbital (K
D
⫽2–10 nM), mus-
cimol was used here as a tool to refine the
functional boundaries of the effective MPTA
cluster (Beaumont et al., 1978;Davies et al.,
1998;Johnston, 2014). Muscimol, how-
ever, does not cross the blood– brain bar-
rier and is hence ineffective as a general
anesthetic agent upon systemic adminis-
tration. We therefore investigated whether
neurons in the MPTA common core also
respond to pentobarbital (Fig. 6B). Micro-
injected into the common core area, a sig-
nificant bonus was obtained on each of the 7
trials using pentobarbital (n⫽6 rats, mean
bonus ⫽19.3 ⫾8.4 min). Thus, accurately
targeted pentobarbital sustained anesthesia
for ⬃20 min at ⬃1/4000
th
of the systemic
dose (⬃15 mg i.v. vs 4
g microinjected).
In addition, we tested microinjected
propofol, a second GABAergic anesthetic
currently in wide clinical use. Propofol
1% was effective in 7 microinjection trials
(4 rats, mean bonus ⫽12.3 ⫾3.1 min)
and probably also in 2 additional trials in a
fifth rat (total 9 trials in 5 rats, 10.9 ⫾3.9
min all at HUJI; Fig. 6B). Both trials in the
fifth rat yielded bonus times of 6.0 min,
very close to our 6.1 min, p⬍0.001 criterion. The short bonus
times obtained despite use of a saturating concentration of
propofol was probably due to rapid washout by the circulation
and perhaps also to a particularly low affinity to the GABA
A
-R
isoform(s) expressed by the relevant MPTA neurons. With this in
mind, we also tested 97% propofol, which would have washed out
more slowly. Indeed, this significantly increased bonus times
(22.7 ⫾14.2 min, 3 trials in 3 rats compared with 1% propofol,
9 trials in 5 rats, ttest, p⫽0.035; Fig. 6B). Rapid washout
could explain why Voss et al. (2005) obtained only a partial
effect with propofol using the indwelling cannula method (an-
algesia without LOC). Other contributing factors might have
included suboptimal targeting or co-recruitment of nearby
pro-arousal neurons by the relatively large volumes they mi-
croinjected (1000 nL).
Critical mass of neurons and neuronal characteristics
How many MPTA neurons need to be exposed to muscimol to
extend anesthesia time significantly? An upper-limit estimate of
this number can be derived from the number of neurons exposed
by 10 nL microinjections, the smallest volume used in this study.
The radius of dye marks associated with such microinjections
(mean ⫽395
m, based on 6 on-target microinjections) repre-
sents a volume of 0.26 mm
3
(260 nL). Because the sphere radius
of 10 nL ⫽134
m, the linear spread was ⬃⫻3 in each dimension
(volume spread is ⫻26). Average neuronal density within the
MPTA counting frame based on NeuN-labeled histological sec-
tions was ⬃435 neurons/mm
2
(Fig. 7B). This converts to ⬃7200
neurons/mm
3
after correcting for double counting of split cells
using Abercrombie’s formula (Koningsmark, 1970). Muscimol
and the marker dye are both small, water-soluble molecules
(m.w. ⫽114 vs 992). Assuming that the spread of muscimol and
the marker was approximately the same (0.26 mm
3
), up to
⬃1900 neurons would have been exposed to the drug by a 10 nL
microinjection. The actual number is probably less (see
Discussion).
The spherical volume associated with the MPTA common
core is considerably smaller than the dye-spread volume from 10
nL microinjections. It has a radius of 0.22 mm and a spherical
volume of 0.04 mm
3
(40 vs 260 nL; Fig. 5B, right). This volume
contains only ⬃300 neurons. It is unclear whether it would suf-
fice to expose the common core neurons alone to muscimol to
obtain a significant bonus and clinical anesthesia. Microinjec-
tions smaller than 10 nL that do not spread beyond the common
core will be required to answer this question. However, it appears
that inclusion of the common core neurons is important, if not
absolutely necessary. Specifically, there were three microinjection
trials (Fig. 5A, left, blue diagonal lines) in which muscimol reached
Figure 4. Ineffective microinjections of muscimol surround the cluster of effective microinjections in all three planes. The full
extent of dye spread is shown in red (dark fill) for the 17 small volume microinjections (10 or 20 nL) that yielded a significant bonus
time. Dye spread of all of the ineffective microinjections represented at these levels (⫺7.6 to ⫺9.8 mm, n⫽28 in 18 rats),
including large and small volume trials (10 –500 nL), are shown in light blue fill.
9326 •J. Neurosci., September 20, 2017 •37(38):9320 –9331 Minert et al. •Transient Loss of Consciousness
neurons in the effective zone (pink), but missed the common core
(Fig. 5A, left, red). These trials failed to induce a significant bonus.
Conversely, two microinjections that induced a significant bonus
(Fig. 5B, left) hit the pink area, but missed the common core.
Given the ambiguity as to whether common core neurons are
sufficient and/or essential, we examined the relation between the
volume of muscimol microinjected and the bonus time obtained.
For successful microinjection trials, bonus values varied con-
siderably, averaging in the range of 20–30 min. This was much
longer than the anesthesia time provided by our standard intra-
venous bolus dose of propofol (11 min). However, interestingly,
bonus values did not vary systematically with the volume micro-
injected and thus with the overall muscimol dose or the number
of MPTA area neurons exposed to the drug (Fig. 6B). This was so
even for individual rats that received mi-
croinjections of several volumes. For ex-
ample, in 3 rats that received 2 muscimol
microinjections at the same location, 200
nL (45.6 ng) and later 50 nL (11.4 ng),
bonus scores obtained were 26 and 35
min, 36 and 24 min, and 33 and 23 min,
respectively (Table 1). Likewise, bonus
time values did not differ despite numer-
ous differences at the two venues where
the experiments were performed (Fig.
6A). This suggests that as long as the neu-
rons in the common core (and perhaps a
limited area beyond) are exposed to mus-
cimol, there is little further effect of expos-
ing a much larger number of neighboring
neurons to the drug.
Neurons residing in the common core
area vary in size and shape. Although most
had a somatic area ⬍220
m
2
(90%), there
was a significant population of much larger
size extending up to 1080
m
2
(Fig. 7A,B).
Interestingly, projection neurons, MPTA
neurons retrogradely labeled from a variety
of projection targets (Reiner et al., 2008),
are considerably larger than the average
(188 vs 109
m
2
). MPTA common core
neurons were also heterogeneous in aspect ratio, form factor, and
roundness (Fig. 7C–E). In addition to their size, the larger cells
tended to be flatter (smaller aspect ratio and form factor) and had
a more tortuous circumference (lower roundness) than the
smaller cells. A considerable amount is known about the afferent
and efferent connectivity of neurons in the MPTA area and on
other cellular characteristics (Sukhotinsky et al., 2016), but fur-
ther analysis is needed with respect to the MPTA’s common core.
Silencing MPTA neurons by exposing them to
local anesthetics
The GABAergic agents tested generally suppress activity in neu-
rons that bear GABA
A
-Rs. In previous microinjection studies
Figure 5. Location of the MPTA common core. A–C, Maximal cross-sectional area of effective (solid pink) and noneffective (hatched blue) microinjections are collected onto a single rostrocaudal
plane (⫺8.8 mm). The track of the injection pipette in these trials actually lay over the range of ⫺8.0 to ⫺9.16 mm caudal to bregma. A, Dye spread for the 17 small effective muscimol
microinjections largely overlap (10 or 20 nL, 9 rats, 10 on the left and 7 on the right). The region in common to all (intersection), on the right and on the left, is shown in solid red (dark). The term
“common core” refers to this area on the right. Noneffective microinjections (10 –200 nL) surround the effective ones (14 microinjections in 12 rats). B,AsinA, but showing areas of both large and
small effective muscimol microinjections (24 left, 11 right in 23 rats). Noneffective trials are not shown. The solid red area on the right is common to all 11 microinjections on that side. The (smaller)
solid red area on the left is common to 22 of the 24 left-sided microinjections. Dye spread areas of the two microinjections not included fell closely nearby (black outlines, rats #322 and #15, 50 and
200 nL, respectively). C, Dye spread areas of 10 saline (control) microinjections (5 rats; 50 –200 nL; horizontal dark blue lines) overlapped the common core zone bilaterally (5 left, 5 right). None of
these yielded a significant bonus time.
Figure 6. A, Average of the significant bonus times was the same (p⫽0.24) at the two experimental venues (left:
BIDMC, male Sprague Dawley rats; right: HUJI, female Wistar-derived Sabra rats). B, Anesthesia bonus time did not
correlate significantly with the volume (and thus dose) of muscimol microinjected (open circles; 39 trials in which there was
a significant bonus; R
2
⫽0.086, p⫽0.07). The regression line for this relation is shown. Also shown are bonus times
obtained for on-target microinjections of pentobarbital (7 trials in 6 rats, filled diamonds) and propofol (12 trials in 9 rats,
filled triangles). The regression of all bonus times on volume microinjected also failed to reach statistical significance
(R
2
⫽0.011, p⫽0.43). All data shown were from experiments performed at HUJI. Horizontal dashed lines indicate the
degree of variability of anesthesia times obtained using the standard bolus dose of propofol in trials in which there was no
drug microinjection (⫹1SD⫽2.0 min; ⫹3 SDs ⫽6.1 min).
Minert et al. •Transient Loss of Consciousness J. Neurosci., September 20, 2017 •37(38):9320 –9331 • 9327
using the indwelling cannula method, we showed that delivery of
lidocaine, a pan-sodium channel blocker that is expected to si-
lence MPTA neurons nonselectively, does not induce anesthesia.
This unexpected result is important because it constrains the
mechanism by which the MPTA might influence the brain-state
switching network (Devor and Zalkind, 2001;Devor et al., 2016).
For this reason, we attempted to replicate this experiment using
the new bonus time method. Indeed, on-target microinjection of
200 nL of 5% lidocaine yielded a nonsignificant bonus in each of
9 trials (9 rats, all at BIDMC, bonus ⫽2.7 ⫾5.8 min). Finally,
because lidocaine action is pH sensitive and we could not mea-
sure the in situ pH in the MPTA, we repeated this experiment
with 1 mMTTX. Again, no significant bonus was obtained (4
trials in 2 rats, all at HUJI, 3 ⫻50 nL, 1 ⫻100 nL, all on-target,
mean bonus time ⫽1.0 ⫾1.8 min).
Discussion
The original borders of the MPTA were set by framing the centers
of effective microinjections (Devor and Zalkind, 2001). Agent
spread was not taken into account. The bonus time method, with
its improved spatial resolution, permitted a significant upgrade
in our ability to define the actual location of the MPTA neurons
that mediate general anesthesia upon exposure to GABAergic
anesthetics. Relatively large (ⱖ200 nL) muscimol microinjec-
tions at a distance from the MPTA common core yielded signif-
icant bonus times, but the effect was lost using smaller volumes.
A progressive search using ever smaller volumes yielded a dorsal
region in the caudal half of the (original) MPTA frame, where as
little as 10 nL of muscimol reliably prolonged anesthesia time.
Deviation by as little as 0.5 mm in any direction eliminated the
effect. In a recent complementary experiment, we showed that
bilateral MPTA lesions markedly reduced sensitivity to sys-
temic GABAergic anesthetics (Minert and Devor, 2016). Impor-
tantly, the territory common to all effective lesions was centered
on the common core hotspot for anesthetic induction identified
in the present study.
Size and location of the MPTA
Size
From knowledge of neuronal density and the extent of agent
spread, we estimated that unilateral exposure of ⬃1900 MPTA
neurons to muscimol is sufficient to maintain anesthesia for 10s
of minutes. This number is probably an overestimate. The
marker dye, although somewhat larger than muscimol, had 1–2 h
to diffuse, the typical time from microinjection to perfusion. The
drug itself, in contrast, had to act within ⬃2–3 min to prevent
pacing and termination of the trial. This brief diffusion interval
would probably have permitted the drug to access ⬍1900 neu-
rons. If exposure of common core neurons is sufficient, then the
minimal number could be as few as ⬃300 neurons. Concerning
the overall size of the MPTA, we adopt the conservative position
that GABA-receptive MPTA neurons involved in brain-state
switching might be present wherever small-volume trials were
successful (Fig. 8, right, pink zone). We refer to this as the “max-
imal MPTA” (mMPTA). With an equivalent radius of 0.75 mm
(volume ⫽1.77 mm
3
), it contains ⬃11,700 neurons
Nuclear designation
In the authoritative rat brainstem atlases of Paxinos and Watson
(Paxinos and Watson, 1998;Paxinos et al., 1999), the MPTA
common core occupies a tegmental (reticular) field traversed by
fibers of the SCP that has no nuclear designation. Common core
neurons lie interstitial to the SCP. The larger mMPTA is likewise
undesignated; the region ventral to the SCP is simply labeled
“subpeduncular tegmentum.” The mMPTA merges ventrally
with the PnO and is laterally and caudally adjacent to the PPTg,
Figure 7. Sizes and shapes of neurons populating the MPTA common core are heterogeneous. A total of 1198 neurons were measured bilaterally in four naive rats. A, Histogram showing somatic
area.B, Photomicrograph showing a typical field ofNeuN-immunolabeled neurons, includesone neuron that isexceptionally larger (arrow).C–E, Histograms showingthe distribution of threeshape
parameters in the population of neurons shown in A. These were as follows: aspect ratio (a measure of flatness and radial symmetry), form factor (a measure of the complexity of the somatic
perimeter), and roundness.
9328 •J. Neurosci., September 20, 2017 •37(38):9320 –9331 Minert et al. •Transient Loss of Consciousness
caudally adjacent to the PB, and medially adjacent to the LDTg.
These nuclei are implicated in sleep and arousal (Fuller et al.,
2011;Watson et al., 2011;Roeder et al., 2016), but all were ruled
out as the location of the anesthesia-promoting MPTA neurons.
The mMPTA approximately overlaps the rodent homolog of the
“mesencephalic locomotor region” (Sherman et al., 2015), an
area in which Shik et al. (1969) documented (in cats) that electri-
cal stimulation triggers locomotion. This is also the region where
direct application of teflurane or cyclopropane were shown to
induce sedation and sleep (in cats; Folkman et al., 1968). It con-
tains neurons that send ascending and descending axons to
nearby arousal nuclei and to more distant forebrain, hindbrain,
and spinal cord structures associated with arousal, memory,
movement, and pain. These pathways are presumably the effec-
tors of the canonical components of anesthesia. A subset of neu-
rons in the MPTA are known to express GABA
A
-R subunits.
These are presumably the neurons that are sensitive to GABAer-
gic anesthetics (Sukhotinsky et al., 2003,2007,2016).
Indwelling cannula versus bonus time methods
Both the original indwelling cannula method and the bonus time
method permit focal deposition of GABA
A
-R agonists in the
brain. A major difference is that, using the former, anesthetic
induction begins from the awake state; in the latter, the animal is
already anesthetized (with propofol). Effective drugs prolong an-
esthesia time, postponing emergence, but they do not actually
cause induction. This difference is potentially significant because
emergence from anesthesia might be an active process, not simple
reversal of induction as the anesthetic agent dissipates (Sleigh et
al., 2001;Kelz et al., 2008). A variety of actions promote arousal
and emergence from anesthesia. Noxious peripheral input (e.g.
by tail pinch) is a prime example. Reanimation can also be evoked
by activation of central pain pathways (Schiff et al., 2007;
Brischoux et al., 2009;Anaclet et al., 2014;Roeder et al., 2016;
Taylor et al., 2016;Morales and Margolis, 2017). It is not clear
whether arousal by pain-provoking stimuli is mechanistically
linked to mesopontine sleep–wake and anesthesia circuitry or is
more general in nature. Either way, MPTA neurons appear to be
key players in induction as well as emergence as between the two
experimental methods both have been documented. Whether the
same individual MPTA neurons contribute to both remains un-
certain. Also uncertain is the degree to which our findings apply
to anesthesia using non-GABAergic agents and to what extent
they apply to clinical anesthesia in humans.
Finally, we confirmed our earlier observation that nonselec-
tive suppression of the MPTA using lidocaine (and now TTX) do
neither. Our explanation of this apparent paradox is in a pro-
posed model of anesthetic induction by MPTA neurons (Minert
and Devor, 2016;Devor et al., 2016).
Network actions
Observations, mostly based on noninvasive imaging in humans,
have highlighted large-scale network connectivity and top-down
processing as central factors in conscious awareness, with
thalamocortical interactions taking center stage (Brown et al.,
2010;Bonhomme et al., 2012;Vijayan et al., 2013;Hudetz and
Mashour, 2016;Tononi et al., 2016;Mashour and Hudetz, 2017).
Degradation and recovery of effective connectivity are striking
correlates of anesthetic induction and emergence. These obser-
vations, however, do not bear directly on the question of where in
the brain anesthetics act. The temporal resolution of the BOLD
signal is too low to discriminate generalized suppression by cir-
culating drugs from activation of dedicated axonal pathways. In
principle, both are capable of mediating the observed changes.
Magnetoencephalography might have the required temporal
resolution.
GABA
A
-Rs are ubiquitous in the CNS, including in the cere-
bral cortex, and they are surely engaged by circulating agonists.
Our microinjection results, however, suggest that their engage-
ment is not necessary for anesthetic LOC. It is probably also not
sufficient because, after MPTA lesions with the thalamus and
cortex intact, clinically relevant drug doses no longer induce an-
esthesia (Minert and Devor, 2016). We cannot exclude the pos-
sibility that certain parameters thought to be bona fide neural
correlates-of-consciousness are not in fact causal, as they require
direct agonist binding to cortical receptors. These could include
anesthetic-induced reduction in cortical blood flow and metab-
olism, aspects of cortical connectivity, and some features of al-
tered EEG. We note, for example, that bolus doses of systemic
propofol induce burst suppression, but agonists microinjected
into the MPTA, even at very high concentrations, do not (Devor
et al., 2016). Assuming that in clinical (systemic) anesthesia
MPTA-driven changes are engaged in addition to generalized
suppression-driven changes, it will be important moving forward
to unravel which changes are due to which process. This is true
for both the cortical correlates of LOC and for the sensory and
motor changes that accompany anesthesia, typically attributed to
the brainstem and spinal cord. Such dissociation is straightfor-
ward in rats using MPTA microinjection (Abulafia et al., 2009),
but in humans, microinjection is clearly impracticable. There
Figure 8. Graphical illustration of the MPTA common core and mMPTA. Left, Noneffective
microinjectionsfrom the right side ofFigure 5Aare mirrortransposed andplotted together with
noneffective microinjections actually made on the left side in Figure 5A(light blue shading).
These contours surround the transposed MPTA common core (Fig. 5A, right). Note the non-
shaded halo surrounding the common core that is devoid of failed microinjections. Right, Dye
spread areas of noneffective microinjections of all volumes (light blue shading) have been
mirror transposed and combined with all of the effective small volume microinjections
(10 –20 nL; Fig. 5A, both sides, pink fill). The union of these pink contours is the mMPTA. The
successfulmicroinjections (pink) are surrounded on allsides by failedmicroinjections (blue).For
clarity, the two negative trials noted in the Results that included the common core, whereas
indicated on the right side in this illustration, are absent on the left.
Minert et al. •Transient Loss of Consciousness J. Neurosci., September 20, 2017 •37(38):9320 –9331 • 9329
might be a pharmacological approach, however, if MPTA neu-
rons express GABA
A
-R isoforms or other propitious receptors
that can be activated using target-selective anesthetics.
Conclusions
Our results add to the growing understanding of brain-state
switching in general and the neural mechanisms underlying an-
esthesia in particular. Most notably, they support the dedicated
pathways hypothesis of general anesthesia by suggesting that sys-
temic drug administration induces immobility, analgesia, amne-
sia, and LOC by a primary action in the brainstem, followed by
recruitment of dedicated ascending and descending axonal path-
ways that secondarily modulate function in far-flung effector lo-
cations in the CNS. We have located the relevant GABA-sensitive
neurons to a very small, focal population within the larger region
originally designated as MPTA. This nucleus appears to consti-
tute a key node in a network that effects brain-state switching.
The improved localization of the functionally important
MPTA cluster will facilitate a more detailed characterization of
the relevant neurons, including their cellular properties and
functional connectivity. Of particular importance is determining
whether the same neurons participate in all of the functional
endpoints of anesthesia or if separate subpopulations mediate
atonia, analgesia, amnesia, and LOC (Reiner et al., 2007). Heter-
ogeneity might make it possible to recruit specific neuronal sub-
populations selectively permitting, for example, surgical quality
pain control without sedation or selective promotion of sleep (or
alertness). The improved localization will also facilitate determi-
nation of whether the flip-flop circuit associated with GABAergic
anesthesia participates in other instances of transient LOC such
as fainting due to hypotension (syncope), blood gas imbalance,
fear, or natural sleep (Meiri et al., 2016). Finally, the identifica-
tion of a localized cluster of neurons that are intimately related to
brain-state switching may provide a viable experimental lead into
the circuitry that realizes conscious experience itself.
References
Abulafia R, Zalkind V, Devor M (2009) Cerebral activity during the
anesthesia-like state induced by mesopontine microinjection of pento-
barbital. J Neurosci 29:7053–7064. CrossRef Medline
Alkire MT, Hudetz AG, Tononi G (2008) Consciousness and anesthesia.
Science 322:876– 880. CrossRef Medline
Anaclet C, Ferrari L, Arrigoni E, Bass CE, Saper CB, Lu J, Fuller PM (2014)
The GABAergic parafacial zone is a medullary slow wave sleep-promoting
center. Nat Neurosci 17:1217–1224. CrossRef Medline
Antognini JF, Carstens E, Raines DE, eds (2003) Neural mechanisms of an-
esthesia. Totowa, NJ: Human Press.
Baker R, Gent TC, Yang Q, Parker S, Vyssotski AL, Wisden W, Brickley SG,
Franks NP (2014) Altered activity in the central medial thalamus pre-
cedes changes in the neocortex during transitions into both sleep and
propofol anesthesia. J Neurosci 34:13326–13335. CrossRef Medline
Beaumont K, Chilton WS, Yamamura HI, Enna SJ (1978) Muscimol bind-
ing in rat brain: association with synaptic GABA receptors. Brain Res
148:153–162. CrossRef Medline
Bonhomme V, Boveroux P, Brichant JF, Laureys S, Boly M (2012) Neural
correlates of consciousness during general anesthesia using functional
magnetic resonance imaging (fMRI). Arch Ital Biol 150:155–163. CrossRef
Medline
Bremer F (1936) Nouvelles recherches sur le mecanisme du someil. CR Se-
ances Soc Biol Fil (Paris) 122:460– 464.
Brischoux F, Chakraborty S, Brierley DI, Ungless MA (2009) Phasic excita-
tion of dopamine neurons in ventral VTA by noxious stimuli. Proc Natl
Acad Sci U S A 106:4894 –4899. CrossRef Medline
Brown EN, Lydic R, Schiff ND (2010) General anesthesia, sleep, and coma.
N Engl J Med 363:2638–2650. CrossRef Medline
Campagna JA, Miller KW, Forman SA (2003) Mechanisms of actions of
inhaled anesthetics. N Engl J Med 348:2110–2124. CrossRef Medline
Choong E, Loryan I, Lindqvist M, Nordling Å, el Bouazzaoui S, van Schaik
RH, Johansson I, Jakobsson J, Ingelman-Sundberg M (2013) Sex differ-
ence in formation of propofol metabolites: a replication study. Basic Clin
Pharmacol Toxicol 113:126–131. CrossRef Medline
Correa-Sales C, Rabin BC, Maze M (1992) A hypnotic response to dexme-
detomidine, an alpha 2 agonist, is mediated in the locus coeruleus in rats.
Anesthesiology 76:948–952. CrossRef Medline
Davies M, Thuynsma RP, Dunn SM (1998) Effects of propofol and pento-
barbital on ligand binding to GABAA receptors suggest a similar mecha-
nism of action. Can J Physiol Pharmacol 76:46–52. Medline
Devor M, Zalkind V (2001) Reversible analgesia, atonia, and loss of con-
sciousness on bilateral intracerebral microinjection of pentobarbital. Pain
94:101–112. CrossRef Medline
Devor M, Zalkind V, Fishman Y, Minert A (2016) Model of anaesthetic
induction by unilateral intracerebral microinjection of GABAergic ago-
nists. Eur J Neurosci 43:846– 858. CrossRef Medline
Folkman J, Mark VH, Ervin F, Suematsu K, Hagiwara R (1968) Intracere-
bral gas anesthesia by diffusion through silicone rubber. Anesthesiology
29:419– 425. CrossRef Medline
Franks NP (2008) General anaesthesia: from molecular targets to neuronal
pathways of sleep and arousal. Nat Rev Neurosci 9:370–386. CrossRef
Medline
French JD, Verzeano M, Magoun HW (1953) A neural basis of the anes-
thetic state. AMA Arch Neurol Psychiatry 69:519 –529. CrossRef Medline
Fuller PM, Fuller P, Sherman D, Pedersen NP, Saper CB, Lu J (2011) Reas-
sessment of the structural basis of the ascending arousal system. J Comp
Neurol 519:933–956. CrossRef Medline
Grasshoff C, Rudolph U, Antkowiak B (2005) Molecular and systemic
mechanisms of general anaesthesia: the ‘multi-site and multiple mecha-
nisms’ concept. Curr Opin Anaesthesiol 18:386–391. CrossRef Medline
Hentschke H, Schwarz C, Antkowiak B (2005) Neocortex is the major target
of sedative concentrations of volatile anaesthetics: strong depression of
firing rates and increase of GABAA receptor-mediated inhibition. Eur
J Neurosci 21:93–102. CrossRef Medline
Hoymork SC, Raeder J (2005) Why do women wake up faster than men
from propofol anaesthesia? Br J Anaesth 95:627–633. CrossRef Medline
Hudetz AG, Mashour GA (2016) Disconnecting consciousness: is there a
common anesthetic endpoint? Anesth Analg 123:1228–1240. CrossRef
Medline
Johnston GA (2014) Muscimol as an ionotropic GABA receptor agonist.
Neurochem Res 39:1942–1947. CrossRef Medline
Kelz MB, Sun Y, Chen J, Cheng Meng Q, Moore JT, Veasey SC, Dixon S,
Thornton M, Funato H, Yanagisawa M (2008) An essential role for
orexins in emergence from general anesthesia. Proc Natl Acad Sci U S A
105:1309–1314. CrossRef Medline
Koningsmark BW (1970) Methods for the counting of neurons. In: Con-
temporary research methods in neuroanatomy (Nauta WJH, Ebbesson
SOE, eds), pp 315–340. New York: Springer.
Lo¨ scher W, Rogawski MA (2012) How theories evolved concerning the
mechanism of action of barbiturates. Epilepsia 53:12–25. CrossRef Medline
Lu J, Nelson LE, Franks N, Maze M, Chamberlin NL, Saper CB (2008) Role
of endogenous sleep-wake and analgesic systems in anesthesia. J Comp
Neurol 508:648– 662. CrossRef Medline
Lukatch HS, MacIver MB (1996) Synaptic mechanisms of thiopental-
induced alterations in synchronized cortical activity. Anesthesiology 84:
1425–1434. CrossRef Medline
Magni F, Moruzzi G, Rossi G, Zanchetti A (1959) EEG arousal following
inactivation of the lower brain stem by selective injection of barbiturate
into the vertebral circulation. Arch Ital Biol 97:33–41.
Mashour GA, Hudetz AG (2017) Bottom-up and top-down mechanisms of
general anesthetics modulate different dimensions of consciousness.
Front Neural Circuits 11:44. CrossRef Medline
Meiri G, Lanir S, Minert A, Devor M (2016) Transient loss of consciousness
during hypercapnia and hypoxia: Involvement of pathways associated
with general anesthesia. Exp Neurol 284:67–78. CrossRef Medline
Minert A, Devor M (2016) Brainstem node for loss of consciousness due to
GABA receptor-active anesthetics. Exp Neurol 275:38– 45. CrossRef
Medline
Moore JT, Chen J, Han B, Meng QC, Veasey SC, Beck SG, Kelz MB (2012)
Direct activation of sleep-promoting VLPO neurons by volatile anesthet-
ics contributes to anesthetic hypnosis. Curr Biol 22:2008 –2016. CrossRef
Medline
9330 •J. Neurosci., September 20, 2017 •37(38):9320 –9331 Minert et al. •Transient Loss of Consciousness
Morales M, Margolis EB (2017) Ventral tegmental area: cellular heteroge-
neity, connectivity and behaviour. Nat Rev Neurosci 18:73–85. CrossRef
Medline
Moruzzi G, Magoun HW (1949) Brain stem reticular formation and activa-
tion of the EEG. Electroencephalogr Clin Neurophysiol 1:455–473.
CrossRef Medline
Myers R (1966) Injection of solutions into cerebral tissue: relation between
volume and diffusion. Physiol Behav 1:171–174. CrossRef
Myers RD, Hoch DB (1978) 14C-dopamine microinjected into the brain-
stem of the rats: dispersion kinetics, site content and functional dose.
Brain Res Bull 3:601–609. CrossRef Medline
Namjoshi DR, McErlane SA, Taepavarapruk N, Soja PJ (2009) Network ac-
tions of pentobarbital in the rat mesopontine tegmentum on sensory
inflow through the spinothalamic tract. J Neurophysiol 102:700–713.
CrossRef Medline
Paxinos G, Watson C (1998) The rat brain in stereotaxic coordinates, Ed 4.
San Diego: Academic.
Paxinos G, Carrive P, Wang H, Wang PY (1999) Cytoarchitectural atlas of
the rat brainstem. San Diego: Academic.
Raz A, Grady SM, Krause BM, Uhlrich DJ, Manning KA, Banks MI (2014)
Preferential effect of isoflurane on top-down vs. bottom-up pathways in
sensory cortex. Front Syst Neurosci 8:191. CrossRef Medline
Reiner K, Sukhotinsky I, Devor M (2007) Mesopontine tegmental anesthe-
sia area projects independently to the rostromedial medulla and to the
spinal cord. Neuroscience 146:1355–1370. CrossRef Medline
Reiner K, Sukhotinsky I, Devor M (2008) Bulbospinal neurons implicated
in mesopontine-induced anesthesia are substantially collateralized.
J Comp Neurol 508:418– 436. CrossRef Medline
Roeder Z, Chen Q, Davis S, Carlson JD, Tupone D, Heinricher MM (2016)
Parabrachial complex links pain transmission to descending pain modu-
lation. Pain 157:2697–2708. CrossRef Medline
Scharf MT, Kelz MB (2013) Sleep and anesthesia interactions: a pharmaco-
logical appraisal. Curr Anesthesiol Rep 3:1–9. CrossRef Medline
Schiff ND, Giacino JT, Kalmar K, Victor JD, Baker K, Gerber M, Fritz B,
Eisenberg B, Biondi T, O’Connor J, Kobylarz EJ, Farris S, Machado A,
McCagg C, Plum F, Fins JJ, Rezai AR (2007) Behavioural improvements
with thalamic stimulation after severe traumatic brain injury. Nature 448:
600– 603. CrossRef Medline
Sherman D, Fuller PM, Marcus J, Yu J, Zhang P, Chamberlin NL, Saper CB,
Lu J (2015) Anatomical location of the mesencephalic locomotor region
and its possible role in locomotion, posture, cataplexy and parkinsonism.
Front Neurol 6:140. CrossRef Medline
Shik ML, Severin FV, Orlovsky GN (1969) Control of walking and running
by means of electrical stimulation of the mesencephalon. Electroencepha-
logr Clin Neurophysiol 26:549. Medline
Sleigh JW, Steyn-Ross DA, Steyn-Ross ML, Williams ML, Smith P (2001)
Comparison of changes in electroencephalographic measures during in-
duction of general anaesthesia: influence of the gamma frequency band
and electromyogram signal. Br J Anaesth 86:50–58. CrossRef Medline
Sukhotinsky I, Zalkind V, Devor M (2003) Anesthetic effect of barbiturates
microinjected into the brainstem: neuroanatomy. In: Proccedings of the
10th world congress on pain (Dostrovsky J, Carr D, Koltzenburg M, eds),
pp 305–313. Seattle: IASP.
Sukhotinsky I, Zalkind V, Lu J, Hopkins DA, Saper CB, Devor M (2007)
Neural pathways associated with loss of consciousness caused by intrace-
rebral microinjection of GABA A-active anesthetics. Eur J Neurosci 25:
1417–1436. CrossRef Medline
Sukhotinsky I, Minert A, Soja P, Devor M (2016) Mesopontine switch for
the induction of general anesthesia by dedicated neural pathways. Anesth
Analg 123:1274–1285. CrossRef Medline
Taylor NE, Van Dort CJ, Kenny JD, Pei J, Guidera JA, Vlasov KY, Lee JT,
Boyden ES, Brown EN, Solt K (2016) Optogenetic activation of dopa-
mine neurons in the ventral tegmental area induces reanimation from
general anesthesia. Proc Natl Acad Sci U S A.
Tononi G, Boly M, Massimini M, Koch C (2016) Integrated information
theory: from consciousness to its physical substrate. Nat Rev Neurosci
17:450– 461. CrossRef Medline
Vijayan S, Ching S, Purdon PL, Brown EN, Kopell NJ (2013) Thalamocor-
tical mechanisms for the anteriorization of alpha rhythms during pro-
pofol-induced unconsciousness. J Neurosci 33:11070–11075. CrossRef
Medline
Voss LJ, Young BJ, Barnards JP, Sleigh J (2005) Differential anaesthetic ef-
fects following microinjection of thiopentone and propofol into the pons
of adult rats: a pilot study. Anaesth Intensive Care 33:373–380. Medline
Watson CJ, Lydic R, Baghdoyan HA (2011) Sleep duration varies as a func-
tion of glutamate and GABA in rat pontine reticular formation. J Neuro-
chem 118:571–580. CrossRef Medline
Weber F, Chung S, Beier KT, Xu M, Luo L, Dan Y (2015) Control of REM
sleep by ventral medulla GABAergic neurons. Nature 526:435–438.
CrossRef Medline
Xi MC, Morales FR, Chase MH (2001) Induction of wakefulness and inhi-
bition of active (REM) sleep by GABAergic processes in the nucleus pontis
oralis. Arch Ital Biol 139:125–145. Medline
Zecharia AY, Yu X, Go¨ tz T, Ye Z, Carr DR, Wulff P, Bettler B, Vyssotski AL,
Brickley SG, Franks NP, Wisden W (2012) GABAergic inhibition of his-
taminergic neurons regulates active waking but not the sleep–wake switch
or propofol-induced loss of consciousness. J Neurosci 32:13062–13075.
CrossRef Medline
Minert et al. •Transient Loss of Consciousness J. Neurosci., September 20, 2017 •37(38):9320 –9331 • 9331
