Cellular/Molecular

In Vivo Evidence for Lactate as a Neuronal Energy Source

Matthias T. Wyss,
1,2
Renaud Jolivet,
1
Alfred Buck,
2
Pierre J. Magistretti,
3
and Bruno Weber
1
1
Institute of Pharmacology and Toxicology, University of Zurich, 8091 Zurich, Switzerland,
2
Department of Nuclear Medicine, University Hospital, 8091
Zurich, Switzerland, and
3
Brain Mind Institute, Ecole Polytechnique Fe de rale de Lausanne, 1015 Lausanne, Switzerland
Cerebral energy metabolismis a highly compartmentalized and complex process in which transcellular trafficking of metabolites plays a
pivotal role. Over the past decade, a role for lactate in fueling the energetic requirements of neurons has emerged. Furthermore, a
neuroprotective effect of lactate during hypoglycemia or cerebral ischemia has been reported. The majority of the current evidence
concerning lactate metabolism at the cellular level is based on in vitro data; only a few recent in vivo results have demonstrated that the
brain preferentially utilizes lactate over glucose. Using voltage-sensitive dye (VSD) imaging, beta-probe measurements of radiotracer
kinetics, and brain activation by sensory stimulation in the anesthetized rat, we investigated several aspects of cerebral lactate metabo-
lism. The present study is the first in vivo demonstration of the maintenance of neuronal activity in the presence of lactate as the primary
energysource. The loss of the voltage-sensitive dye signal foundduringsevere insulin-inducedhypoglycemiais completelypreventedbylactate
infusion. Thus, lactate has a direct neuroprotective effect. Furthermore, we demonstrate that the brain readily oxidizes lactate in an activity-
dependent manner. The washout of 1-[
11
C]L-lactate, reflectingcerebral lactate oxidation, was observedtoincrease duringbrainactivationfrom
0.077 0.009 to 0.105 0.007 min
1
. Finally, our data confirm that the brain prefers lactate over glucose as an energy substrate when both
substratesareavailable. Using[
18
F]fluorodeoxyglucose(FDG) tomeasurethelocal cerebral metabolicrateof glucose, wedemonstratedalactate
concentration-dependent reductionof cerebral glucose utilizationduring experimentally increasedplasma lactate levels.
Introduction
For a long time, it was considered that blood-borne glucose is the
sole energy substrate in the adult healthy brain. However, several
studies have reported a different situation in the developing and
diseased brain. For example, in breast-fed newborns and suckling
rats, the utilization of monocarboxylates, such as lactate, and
ketone bodies is increased (Cremer, 1982; Dombrowski et al.,
1989; Nehlig and Pereira de Vasconcelos, 1993). In addition, a
neuroprotective role for exogenous lactate in pathologic situa-
tions, such as prolonged starvation, diabetes, and cerebral isch-
emia, has been documented on several occasions (Gjedde and
Crone, 1975; Schurr et al., 2001; Mason et al., 2006; Berthet et al.,
2009). However, it is still unclear whether the neuroprotective
effect of lactate is indirect [e.g., via increased cerebral blood flow
(CBF) caused by hyperlactemia (Shackford et al., 1994)] or whether
lactate can serve as a direct neuronal energy substrate.
One of the first indications that lactate is not only a product of
the metabolic chain but also a potentially important alternative
neuronal energy substrate was documented more than two de-
cades ago from the demonstration that lactate is able to support
synaptic function in the absence of glucose in rat hippocampal
slices (Schurr et al., 1988). For the living brain, the situationis less
clear. No information about the capability of lactate to sustain
neuronal action in vivo could be provided. Recently, independent
groups have reported lactate utilization by the brain in humans
(Gallagher et al., 2009; van Hall et al., 2009; Boumezbeur et al.,
2010). There is also evidence that lactate oxidation is preferen-
tially located in neurons. Bouzier et al. (2000) used 3-[
13
C]lactate
to investigate cerebral lactate metabolism. From the accumula-
tion of labeled metabolites, they concluded that lactate metabo-
lism is located in a compartment devoid of pyruvate carboxylase,
e.g., neurons.
Furthermore, astrocytic production of lactate, which serves as
an energy pool for neurons, was proposed (Pellerin and Magis-
tretti, 1994), but neuronal lactate use has still to be demonstrated
in situ.
In the present study, we followed a multimodal approach.
First, using optical voltage-sensitive dye (VSD) imaging to di-
rectly monitor the synaptic activity in the somatosensory cortex,
we report for the first time that lactate is capable of sustaining
neuronal activity for a time period of hours in the quasi-absence
of glucose in the living organism. Second, applying radiotracer
kinetic measurements of 1-[
11
C]L-lactate accumulation and clas-
sical cerebral glucose utilizationmeasurements using FDG, we dem-
onstrate that lactate is metabolized by the intact brain in an activity-
dependent manner.
Materials and Methods
Animal preparation
All the animal experiments were approved by the local veterinary author-
ities and were performed by licensed investigators. In total, 64 animals
(male adult Sprague Dawley rats; weighing 220350 g) were included in
the study. Before measurements, the animals were fasted overnight to
Received Jan. 17, 2011; revised March 10, 2011; accepted April 1, 2011.
Author contributions: M.T.W., A.B., and B.W. designed research; M.T.W. performed research; M.T.W. and A.B.
analyzed data; M.T.W., R.J., A.B., P.J.M., and B.W. wrote the paper.
M.T.W. and B.W. are supported by Swiss National Foundation Grants 31003A-124739/1 and PP0033-110751.
R.J. is supported by grants from the Hartmann Muller Foundation, the Olga Mayenfisch Foundation, and the EMDO
Foundation. We thank Florent Haiss for help in setting up the voltage-sensitive dye instrument and Felipe Barros for
valuable discussions.
Correspondence should be addressed to Dr. Bruno Weber, Institute of Pharmacology and Toxicology, University
of Zurich, Ramistrasse 100, 8091 Zurich, Switzerland. E-mail: [email protected].
DOI:10.1523/JNEUROSCI.0415-11.2011
Copyright 2011 the authors 0270-6474/11/317477-09$15.00/0
The Journal of Neuroscience, May 18, 2011 31(20):74777485 7477
standardize plasma levels of glucose. Surgery was performed under 2.5
3.5% isoflurane anesthesia and involved the placement of an arterial and
a venous catheter and tracheotomy for mechanical ventilation. For the
cortical beta-probe acquisitions, the cranial bone above somatosensory
cortices was thinned to translucency. For the voltage-sensitive dye imag-
ing experiments, a craniotomy was performed. Staining with the voltage-
sensitive dye RH1691 (Optical Imaging) was done through the intact
dura mater for 90 min. Leaving the dura intact reduces the movement
artifact during recording and decreases the risk of cortex irritations (Lip-
pert et al., 2007). After staining, the dye was removed by washing with
dye-free Ringers solution. Plasma glucose and lactate levels were
determined periodically (Ektachem DT; Eastman Kodak). The actual
experiments were performed under -chloralose anesthesia (44 mg/kg
bodyweight, s.c.). Several blood gas variables (pH, P
CO
2
, P
O
2
) (AVL,
Compact 3; Roche Diagnostics) and metabolic measures such as plasma
glucose and lactate levels (Ektachem DT; Eastman Kodak), which are
relevant parameters influencing lactate uptake in the brain, were mea-
sured in all animals before the start of measurements. Ventilation was
adjusted to reach physiological blood gas values.
Voltage-sensitive dye imaging
The VSD was excited with 630 nm light from an LED (Aculed; PerkinEl-
mer Life and Analytical Sciences). A custom-built tandem lens system
consisting of two 50 mmobjectives and dichroic and emission filters was
used. Data were collected with a high-speed CMOS-based camera (Mi-
cam Ultima; Scimedia). Images were collected with 1 ms temporal reso-
lution. Images were analyzed using custom-written Matlab routines and
the software package PMOD (Mikolajczyk et al., 1998). Bleaching of
fluorescence was corrected by subtraction of a best-fit double exponen-
tial. Time courses of fluorescence changes were quantified as F/F
0
from
circular regions of interest of constant diameter manually placed over the
activated area. To compare VSD signal from different animals, region of
interests (ROIs) were centeredonthe locationof the earliest responses. In
addition, to estimate the spatial extent of the neuronal activation at the
time point of the highest activation(peak amplitude), the area containing
voxels 70% of the maximum within the experiment was determined.
The amplitudes of sensory-evoked responses were calculated as the
change in the VSD signal (F/F
0
) over a fixed time interval for each
experiment. Finally, the time until peak response was determined as the
time spanbetweenstimulationonset andpeak amplitude withinthe ROI.
Baseline was the 10 ms period immediately before the stimulus.
In VSD experiments, five experimental groups were examined: (1)
effect of hyperlactemia (n 4) on neuronal integrity, (24) evolution of
the VSD signal after induction of severe hypoglycemia (insulin at 20
IU/kg, i.p.) and subsequent continuous infusion of saline (n 5), glu-
cose (n 5), or lactate (n 5), respectively, and (5) control experiments
with saline infusion only (n 3).
Radiotracer experiments
Radiotracer synthesis: production of 1-[
11
C]L-lactate. Racemic 1-[
11
C]lactic acid
was made by [
11
C]cyanohydrine synthesis starting from [
11
C]HCN
(trapped as [
11
C]KCN) and acetaldehyde bisulfite adduct. The quantita-
tively formed 1-[
11
C]DL-lactonitrile was hydrolyzed by reflux in concen-
trated HCl, and the reaction mixture was introduced directly into a
semipreparative polymeric HPLCcolumn (250 10 mm, 0.03%H
3
PO
4
as the mobile phase; Polymerx 10; Phenomenex) in which the product
was isolated from the aggressive reaction matrix. Then the collected ra-
cemic 1-[
11
C]lactic acid was automatically introduced into a preparative
chiral HPLC column coated with a penicillamine-derived chiral selector
and eluted with 1 mM CuSO
4
solution as the mobile phase in which both
enantiomers were separated by ligand exchange chromatography. The
fractions corresponding to [
11
C]D-lactic acid or [
11
C]L-lactic acid were
passed through anion exchanger Sep-Pak cartridges (Accell Plus QMA,
in CO
3
2
form) in which the Cu
2
ions (as insoluble carbonate) and the
[
11
C]lactate are retained. The latter was selectively washed out from the
cartridges with 2 mmol/L sodium phosphate buffer, pH 7.4, to obtain
injectable solutions of pure Na-[
11
C]D-lactate and L-lactate enantiomers.
The quality control of the final product was done by chiral ligand ex-
change HPLC. The procedure led to a product of 99% chemical and
enantiomeric purity. The specific activity at the end of synthesis was
400 GBq/mol. FDGwas obtained fromthe daily in-house production
by the Radiochemistry Department of the University Hospital Zurich.
Beta-probe experiments using 1-[
11
C]L-lactate and FDG. A recently de-
veloped two-channel system with surface probes was used for the mea-
surement of the radioactivity concentration in brain cortices (Wyss et al.,
2009). The used devices consist of scintillator crystals with a thickness of
0.2 mm and a diameter of 3 mm. The beta scintillators were made light-
tight by applying a uniform coating of silver particles. The scintillations
were measured using a photomultiplier tube and counting electronics
(PerkinElmer Life and Analytical Sciences). The limited range of beta
particles within biological tissues leads to a limited detection volume
centered around the scintillating tip of the probe (Wyss et al., 2009). The
sensitivities of the used scintillators were 0.310.60 cps/kBq/cc.
For beta-probe measurements, the arterial and venous lines were con-
nected to form an arterio-venous (av) shunt. The av shunt was used for
continuous monitoring of arterial blood pressure, injection of the radio-
tracers, continuous measurement of the total arterial blood radioactivity,
and the collection of blood samples. For the recording of the total blood
radioactivity, the shunt ran through a coincidence counter (GE Medical
Systems) (for details, see Weber et al., 2002). Before the actual experi-
mental measurements for 1-[
11
C]L-lactate, the kinetic properties and the
buildup of metabolites of the novel radiotracer were characterized in 10
animals. In addition, in four of these animals, the first pass extraction
fraction was determined with additional measurements of CBF (see be-
low). All subsequent 1-[
11
C]L-lactate experiments consisted of two ac-
quisitions. For each acquisition, 100150 MBq of radiotracer was
intravenously injected. First, the change of 1-[
11
C]L-lactate kinetics from
baseline conditions to increased neuronal activity [electrical infraorbital
nerve stimulation (ION): 2 mA, 2 Hz, 1 ms pulses applied during the
whole acquisition period; n 5] and the effect of monocarboxylic acid
transporter blockade using -cyano-4-hydroxy-cinnamate (CIN) (n
5) was investigated. CIN was injected intraperitoneally (90 mg/kg) 30
min before tracer injection.
FDG was used to determine the local cerebral metabolic rate of glucose
(LCMR
glu
). For eachFDGacquisition, 5080MBqof tracer was injected. All
FDG experiments consisted of two measurements. Accumulation of FDG
was measured at baseline (SAL) and at artificially increased plasma lactate
levels (LAC) (100 mol kg
1
min
1
during the first 20 min; thereafter,
50 mol kg
1
min
1
, i.v.) during rest conditions (n 4 for SAL; n 13
for LAC) and during activation (n 2 for SAL; n 5 for LAC).
Characterization of 1-[
11
C]lactate. The total radioactivity in arterial
blood was continuously recorded over 40 min using a coincidence coun-
ter (GE Medical Systems). Whole blood activity was then corrected for
(1) a different tracer concentration in whole blood and plasma and (2)
the buildup of labeled metabolites. The ratio
11
C
plasma
/
11
C
whole blood
was
determined in all animals at different time points (three to four blood
samples per animal). The data of all animals were then pooled, and the
time course of the ratio was approximated by fitting a quadratic polyno-
mial to the data. This function was then used to convert counts in whole
blood to counts in plasma.
Samples (about 400 l) were collected at different time points after
tracer injection, with a maximum of four blood samples per animal, to
determine the time course of the ratio of the
11
C activity in plasma to
whole blood and for analysis of authentic tracer and metabolites. These
samples were first centrifuged for 3 min at 2000 rpm. Proteins were then
precipitated with 75 l of acetonitrile in 50 l of plasma. After centrifu-
gation for 3 min at 2000 rpm, the composition of the
11
C-derived radio-
activity in the supernatant (80 l) was analyzed by HPLCon a polymeric
column (5 m, 250 4.1 mm inner diameter; PRP-1; Hamilton) with 3
mmol/L phosphoric acid in water, pH 2.67, as the mobile phase (1 ml/
min). The retention times of [
11
C]HCO
3

(3.3 min) and lactic acid (5.1

min) were determined by using aqueous solution of NaHCO
3
and DL-
lactic acid as reference compounds, detected by UV absorption at 220
nm. The amount of authentic tracer was expressed as a fraction of total
plasma
11
C counts.
The fraction data of all animals was then pooled, and the time course
of the fraction was approximated by a decaying biexponential function.
This function was subsequently used to convert the total plasma activity
7478 J. Neurosci., May 18, 2011 31(20):74777485 Wyss et al. Lactate as a Neuronal Energy Substrate
to the time course of authentic [
11
C]lactate (input curve). At the end of
these experiments, the rats were perfused with PBS, and brains were
prepared for measurements with the HPLC system. Each brain was first
homogenized before adding acetonitrile (150%). The subsequent proce-
dure was the same as with the blood samples described above, except that
the amount of supernatant injected into the HPLC system was 200 l.
CBF measurements (used for estimating the first-pass extraction fraction
of lactate). The basis of the CBF calculation was the one-tissue compart-
ment model, including a partition coefficient for H
2
15
O. The change of
the radioactivity concentration in tissue C
tiss
is then defined by the fol-
lowing differential equation:
dC
tiss
/dt CBF(C
a
(t) C
tiss
(t)/p), (1)
where C
a
is the arterial tracer concentration, and p is the tissue partition
coefficient, e.g., the fraction of tissue that is H
2
15
O permeable. In this
configuration, C
tiss
is the concentration of H
2
15
O in 1 ml of brain, and it
is assumed that H
2
15
O immediately reaches a homogeneous concentra-
tion in permeable space and no division into a vascular and a tissue
compartment is necessary. The analytical solu-
tion of Equation 1 is as follows:
C
tiss
CBF * e
(CBF/pt )
VC
a
, (2)
where Vrepresents mathematical convolution.
Equation 2 was fitted to the data using least-
squares fitting (Marquardt algorithm) imple-
mentedby the software PMOD(Mikolajczyk et
al., 1998). Before data analysis, tissue time ac-
tivity curves and arterial input curves were cor-
rected for the physical decay of
15
O.
Analysis of the radiotracer data
The calculations and parameter fitting were
performed using the software PMOD(Mikola-
jczyk et al., 1998). Data were corrected for
physical decay and multiplied by a calibration
factor taking into account differences in sensi-
tivities of the probe systemand the coincidence
counter. Before analysis, data were down-
sampled to obtain a 10 or 30 s resolution for
11
Cand
18
F data, respectively. The investigated
methods consisted of standard compartmental
modeling using an arterial input function.
1-[
11
C]L-Lactate. Tracer kinetic modeling
was performed using the one-tissue compart-
ment model. The parameters are as follows: K
1
describes the uptake of tracer across the blood
brainbarrier and is related to CBF and the first-
pass extraction fraction EF (K
1
CBF * EF);
and k
2
represents the back-diffusion of label
from the tissue to the vascular compartment.
Label exchange between the compartments is
described by the following differential equation:
dC
tiss
/dt K
1
C
plasma
(t) k
2
C
tiss
(t).
(3)
Because the total activity measured in a region
is composed of counts from tissue and blood,
all models contained a parameter () correct-
ing for blood activity:
C
VOI
(1)C
tiss
C
blood
, (4)
where C
VOI
is
11
C activity measured by the tip
of the beta scintillator, is the percentage of
intravascular space in the tissue, C
tiss
is activity
in the extravascular compartment, C
blood
is to-
tal blood activity, and C
tiss
was calculated by
numerical integration of the differential equa-
tions. The vascular fraction was included as a fit parameter that im-
proved the least-squares fit.
The kinetic model was adjusted in two ways for the analysis of dual-
injection data arising from use of the surface probe. (1) Two sets of rate
constants (K
1
, k
2
) were used in the calculation of the operational equa-
tion: the first set for the time until the secondinjection, andthe secondset
thereafter. The least-squares fit procedure resulted in estimates for all
four rate constants (K
1
1, k
2
1 and K
1
2, k
2
2) and the vascular
fraction. (2) The correction function to derive plasma activity from
whole blood activity was adjusted to distinguish between the contribu-
tions fromthe two injections. To this end, the bloodactivity fromthe first
injection was extrapolated using an exponential function, which was
fitted to the blood activity 10 min before the second injection.
The interpretation of the 1-[
11
C]L-lactate experiments is based on the
following assumptions, which are discussed in detail below: (1) parame-
ter k
2
is reflecting the activity of the enzyme lactate dehydrogenase
(LDH), and (2) the total amount of lactate that is oxidized is the product
of k
2
and intracellular lactate concentration.
Figure 1. Lactate is able to sustain neuronal activity in the absence of glucose. A, A single electrical pulse applied to the rats
hindpaw evoked a transient increase in VSD fluorescence in the primary somatosensory cortex that was abolished after only 150
min in animals receiving saline after the insulin injection (top row) but remained stable in animals infused with glucose (middle
row) or lactate (bottomrow) after 240 min. BD, The time courses of the VSD signal displayed a similar result in animals supplied
with glucose (C; gluc) or lactate (D; lac) with sustained amplitude 240 min after insulin injection. In contrast, removal of the signal
was observed in animals receiving saline infusion only (B; sal). In the group supplied with lactate, the signal onset was delayed
(example shown in D). The vertical black line represents the time point of hindpawstimulation. E, Comparison of amplitudes in all
three groups injected with insulin at 0 min (mean SE; n 5). FH, Mean plasma levels of glucose and lactate plotted for saline
(F ), glucose (G), and lactate (H) animals.
Wyss et al. Lactate as a Neuronal Energy Substrate J. Neurosci., May 18, 2011 31(20):74777485 7479
FDG. FDGdata were used for calculating LCMR
glu
. The procedure for
the quantification of the probe studies followed the [
14
C]deoxyglucose
method described by Sokoloff et al. (1977). Acompartmental model with
two tissue compartments and four kinetic rate constants (K
1
to k
4
) was
used. The operational equation, which was fitted to the probe time activ-
ity curve, additionally included a vascular fraction as a fit parameter.
This improved the least-squares fit compared with a fixed value of
5%. LCMR
glu
values were calculated from the fitted rate constants:
LCMR
glu
K
1
* k
3
/(k
2
k
3
)C
plasma
/LC, where C
plasma
represents the
plasma glucose concentration, and LC represents the lumped constant
(LC of 0.437).
The kinetic model was adjusted in two ways for the dual-injection data
analysis. (1) Two sets of four rate constants (K
1
to k
4
) were used in the
operational equationcalculation: the first set for the time until the second
injection, and the second set thereafter. The least-squares fit procedure
resulted in estimates for all eight rate constants and the vascular fraction.
(2) The correction function to derive plasma activity from whole blood
activity was adjusted to distinguish between the contributions from the
two injections. To this end, the blood activity fromthe first injection was
extrapolated using an exponential function that was fitted to the blood
activity 10 min before the second injection.
Statistics
Results are presented as mean SD if not otherwise stated. To test
differences between different interventions for inter-individual compar-
isons, the nonparametric Wilcoxons signed rank test was applied. A p
value of 0.05 was taken as the significance limit.
Results
Lactate can sustain neuronal integrity as an alternative
energy substrate
Severe glucose deprivation was induced by insulin injection (20
IU/kg, injection at time 0 min) while monitoring neural activity
with VSD imaging. In three experimental groups, infusion of
saline, glucose, or lactate was immediately initiated after the in-
sulin injection. The total recording time for VSD imaging is lim-
ited because of dye washout and excitation light-related
amplitude reductions of the signal (Lippert et al., 2007). There-
fore, imaging commenced with delay, 100 min after the insulin
injection. In animals infused with saline only, blood glucose lev-
els reached 1.5 mmol/L already after 60 min. Plasma lactate levels
varied on average between 1.0 and 1.5 mmol/L throughout the
whole experimental period (Fig. 1F). Severe hypoglycemia re-
duced the amplitude of the VSD signal by 50% from the base-
line after only 140 min after the application of insulin (Fig.
1A, B,E). Thereafter, the signal was further decreased as a result of
severe glucose deprivation. In animals supplied continuously
with glucose, plasma levels of glucose were not reduced by insulin
action and remained at normoglycemic values (Fig. 1G). Not
surprisingly, the neuronal signal did not show a significant de-
cline throughout the period of data acquisition [270 min after the
insulin injection; n 5; not significant (n.s.) compared with
baseline acquisition at 120 min (Fig. 1A, C,E)]. In the group in
which sodium lactate was infused over the whole time period
(plasma lactate levels between 5.5 and 9.7 mmol/L), plasma levels
of glucose also exhibited a rapid decline to values 1 mmol/L
within 60 min (Fig. 1H). Nevertheless, the amplitude of the VSD
signal reflecting neuronal activity remained high (n 5; n.s.
compared with the 120 min time point) (Fig. 1A, D,E). The
spread of activation remained constant in the glucose- and
lactate-treated animals within the 240 min after the insulin injec-
tion, whereas the extent of activation in the animals receiving
saline infusion only decreased considerably (data not shown). In
the latter group, four of five animals showed no pixels surpassing
70% of the maximal amplitude after 150 min. In the fifth, the
number decreased by 50%. The time course of the VSD signal
revealed an accentuated delay of the signal in animals receiving
lactate infusion compared with glucose-infused rats (Fig. 1C,D).
In the animals receiving glucose infusion after insulin injection
(n 5), the times-to-peak were 21.8 1.6, 21.6 1.9, and
20.7 1.5 ms after 120, 180, and 240 min, respectively. In the
lactate group (n 5), delays at the corresponding time points
were 21.2 0.9, 23.5 4.4, and 36.2 10.7 ms. The difference
after 240 min was significant between the two groups ( p 0.05).
Lactate is oxidized by the brain in an
activity-dependent manner
To directly assess lactate use in the adult rodent brain in vivo, we
synthesized radiolabeled 1-[
11
C]L-lactate. Characterization of
1-[
11
C]L-lactate revealed suitable properties for use in brain
studies, allowing cerebral tracer kinetic measurements with high
temporal resolution. An extraction fraction (EF K
1
/CBF) of
30 10% (n 4) was calculated, which is sufficient to study
physiological processes in the brain and is in good agreement
with previous studies reporting the uptake of L-lactate to be 25
50% of that of glucose (Knudsen et al., 1991; Hassel and Brthe,
2000). Kinetic modeling showed that a one-tissue compartment
model was sufficient for the analysis of 1-[
11
C]L-lactate data,
yielding K
1
and k
2
values measuring the uptake and the release of
radiolabel, respectively. The time courses of the fraction of au-
thentic tracer and metabolites inplasma are showninFigure 2. As
in brain, the main metabolite detected in blood was [
11
C]CO
2
.
After 20 min, the fractionof true tracer fell to 40%. Inthe brain,
Figure 2. 1-[
11
C]L-lactate is a suitable tracer to study cerebral lactate oxidation. A, Mea-
sured radioactivity concentration in the brain (open circles), model fit (black line), and arterial
input curve (gray line). The inset displays the residuals of the fitting to a one-tissue compart-
ment model. The absence of any distribution bias supports the adequacy of the applied one-
tissue compartment model. B, Fraction of native radiolabeled lactate over 40 min after
intravenous injection of 1-[
11
C]L-lactate in blood. The filled circles represent data points from
baseline experiments (n7) performedtocharacterize the radiotracer, andthe solidline is the
corresponding fit of a biexponential curve.
7480 J. Neurosci., May 18, 2011 31(20):74777485 Wyss et al. Lactate as a Neuronal Energy Substrate
the analysis revealed 58 8% lactate and 42 8% CO
2
after 40
min (n 7). Using for the first time
11
C-radiolabeled lactate to
measure cerebral lactate oxidation, we observed a noticeable in-
crease of both rate constants K
1
and k
2
during increased neuronal
activity (Fig. 3B, Table 1). Delivery of the tracer increased during
electrostimulation of the infraorbital nerve as reflected by an in-
crease of the CBF-dependent K
1
from 0.09 0.01 to 0.12 0.03
ml min
1
ml
1
tissue (n 5; change n.s.).
In parallel to the elevated delivery, radioactivity washout (k
2
)
increased from0.077 0.009 to 0.105 0.007 min
1
(n 5; p
0.05). In Discussion, we will argue that this is most likely related
to increased LDH activity during stimulation.
There is strong evidence in our and other studies that the label
leaves the tissue as CO
2
, which is cleaved off during the conver-
sion of pyruvate to acetyl-CoA by pyruvate dehydrogenase
(PDH) (Fig. 3A). In the present study, the metabolite analysis
revealed only lactate and CO
2
as labeled metabolites. This is in
line with a study by van Hall et al. (2009). Using 1-[
13
C]lactate,
they reported that 86 15%of the label was released as CO
2
. The
situation is completely different, if lactate is labeled at position 2
or 3. In that case, most of the label ends up in metabolites of the
TCA cycle (van Hall, 1999; Bouzier et al., 2000; van Hall et al.,
2009).
Increased uptake of blood-borne lactate and efficiency of
lactate transport during activation
It is possible to calculate cerebral uptake of blood-borne lactate as
the product of K
1
and blood lactate concentration (in this group
of experiments, plasma lactate levels ranged from 1.7 to 2.8
mmol/L) (Table 2). Based on these values, baseline mean cerebral
lactate uptake was estimated to be 0.17 mol g
1
min
1
and
increased to 0.22 mol g
1
min
1
during stimulation.
A detailed analysis of K
1
revealed furthermore that the
stimulation-induced increase cannot be completely explained by
an increase in CBF only but must at least be partly attributable to
an increased transport rate for lactate. We estimated this increase
to be in the order of 30% using the following relationships:
K
1
CBF * EF, (5)
where EF can be expressed according to Renkin (1959) and Crone
(1963), and
EF 1 e
PS/CBF
. (6)
Combining Equations 5 and 6 yields the following:
PS CBF * ln(1 K
1
/CBF), (7)
where PS is the permeabilitysurface area product. We per-
formed a simulation assuming a constant PS from baseline to
ION stimulation. In previous experiments, the applied electro-
stimulationof the IONled to a meanincrease of CBF from0.35 to
0.50 ml min
1
ml
1
tissue (Weber et al., 2003). In the present
series of 1-[
11
C]L-lactate experiments, K
1
increased from 0.09
0.01 to 0.12 0.03 ml min
1
ml
1
tissue during ION. Insert-
ing these numbers into Equation 7 demonstrates a 32% increase
of PS, from 0.104 to 0.137 ml min
1
ml
1
. This suggests that
monocarboxylate transporter (MCT) efficiency increases during
increased demand.
Effect of MCT blockade on lactate uptake
Injection of CIN 30 min before [
11
C]lactate tracer kinetic mea-
surements reduced K
1
from 0.13 0.01 to 0.09 0.01
Figure 3. The brain oxidizes lactate in an activity-dependent manner. A, Schematic of the
biochemical pathways involved in the degradation of 1-[
11
C]L-lactate (
11
C-Lac) and the pro-
posedinterpretationof therateconstants K
1
andk
2
describingthekinetics of theradiotracer (for
details, see Results). Briefly, K
1
represents delivery of 1-[
11
C]L-lactate (orange arrow), and k
2
reflects kinetically the loss of radiolabel after conversion of lactate to pyruvate and of pyruvate
to acetyl-coenzyme A (purple arrows). During activation (B; S; n 5), delivery (K
1
) and wash-
out (k
2
) increased, whereas duringMCT blockade(C; CIN; n5), thetransfer of 1-[
11
C]L-lactate
slowed down in both directions (Pyr, pyruvate; Ac CoA, acetyl coenzyme A; TCAcycle, tricarbox-
ylic acid cycle). *p 0.05, n.s., not significant.
Table 1. 1-
1 1
CL-lactate kinetics
Animal number
K
1
(ml min
1
ml
1
tissue) k
2
(min
1
)
Baseline Stimulation Baseline Stimulation
1 0.081 0.091 0.073 0.082
2 0.085 0.111 0.074 0.105
3 0.097 0.101 0.09 0.107
4 0.086 0.12 0.069 0.112
5 0.111 0.128 0.076 0.095
Mean 0.092 0.110 0.076 0.100
SD 0.01 0.01 0.01 0.01
Baseline CIN Baseline 4-CIN
6 0.138 0.076 0.116 0.052
7 0.139 0.083 0.115 0.071
8 0.132 0.107 0.102 0.082
9 0.128 0.091 0.109 0.076
10 0.115 0.08 0.098 0.081
Mean 0.130 0.087 0.108 0.072
SD 0.01 0.01 0.01 0.01
Table 2. Plasma lactate levels and calculated cerebral lactate uptake
Animal number
Plasmalactatelevel (mmol/L)
Calculatedcerebral lactateuptake
(K
1
*lac; mol g
1
min
1
Baseline Stimulation Baseline Stimulation
1 2 2.2 0.162 0.200
2 2 2.9 0.170 0.322
3 2 2.1 0.194 0.212
4 1.7 1.6 0.146 0.192
5 1.5 1.5 0.167 0.192
Mean 1.84 2.06 0.17 0.22
SD 0.2 0.6 0.02 0.06
lac, Plasma lactate level.
Wyss et al. Lactate as a Neuronal Energy Substrate J. Neurosci., May 18, 2011 31(20):74777485 7481
ml min
1
ml
1
tissue (n 5; p 0.05)
and k
2
from 0.11 0.01 to 0.07 0.01
min
1
(n 5; p 0.05) (Fig. 3C, Table 1).
CIN has been shown to permeate the
bloodbrain barrier and to be effective 30
min after intraperitoneal injection (Schurr
et al., 2001). The reduction in K
1
after CIN
administration is most likely attributable to
the reduction of lactate transport by a par-
tial blockade of MCTs. The concomitant
reduction in k
2
cannot be reliably inter-
preted without additional biochemical
measurements. The fact that the relative
decrease of k
2
is of the same magnitude as
the decrease of K
1
(33%) points to the
possibility that k
2
under CIN blockade
represents back-diffusion of lactate into
vascular space.
Increased blood lactate reduces the
local metabolic rate of glucose
To address the question of how the brain
processes blood-borne lactate compared
with glucose, we artificially increased the
supply of lactate to the brain by continu-
ous intravenous infusion of sodium lac-
tate (Fig. 4A). This led to a significant
increase of plasma lactate levels (from
1.3 0.6 to5.8 1.6 mmol/L; n10; p
0.05). Plasma glucose levels were not af-
fected by the continuous lactate infusions.
In all examined animals, a clear reduction
of LCMR
glu
was observed at resting con-
ditions (decreases ranging from 25 to
49%; n 13; p 0.05) (Fig. 4B). The
reduction of the cerebral glucose utiliza-
tion was attributable to both a decreased
delivery of glucose into the tissue (re-
flected by a change of K
1
from 0.16 0.05 to 0.12 0.03
ml min
1
ml
1
tissue; n 13; p 0.05) and a reduced phos-
phorylation by hexokinase (k
3
decreased from 0.04 0.02 to
0.03 0.01 min
1
; n13; p 0.05). The replacement of glucose
by lactate as an energy substrate occurred in a dose-dependent
manner (Fig. 4C).
Reduction in glucose utilization is balanced by the excess of
lactate consumption
The observedmeanreductioninglucose utilizationwas 21 mol/
100 g/min (Fig. 4B). Because glucose has a six-carbon skeleton,
this corresponds to 42 mol/100 g/min three-carbon substrates
such as pyruvate. The following estimation demonstrates that the
decrease in glucose utilization is quantitatively matched by an
increase in lactate use. For this purpose, we made the following
assumptions: (1) lactate transport from blood into the brain
equals K
1
* plasma lactate concentration ([Lac]), (2) K
1
(0.09
min
1
at baseline) does not markedly change within the range of
lactate concentrations found in our experiments, and (3) all the
lactate, which is taken up, is ultimately oxidized by the brain.
Thus, lactate uptake during baseline conditions in the FDG
experiments was 12 mol/100 g/min ( 0.09 min
1
* 1.3
mmol/L; mean K
1
determined in baseline 1-[
11
C]lactate experi-
ments (Fig. 3B) and mean [Lac] from FDG experiments during
saline infusion) and 52 mol/100 g/min during hyperlactemia
( 0.09 min
1
* 5.8 mmol/L; mean [Lac] during lactate infu-
sion). The difference of lactate uptake between baseline and hy-
perlactemia conditions is therefore 40 mol/100 g/min (52
mol/100 g/min 12 mol/100 g/min). This matches the above
mentioned reduction of three-carbon constituents derived from
glucose during hyperlactemia.
Reduction of glucose consumption during hyperlactemia is
more pronounced during activation
Our results further demonstrate that the reduction of glucose
utilization is not only dependent on blood lactate levels but also
on the level of activation. As shown in Figure 4, D and E, the
decrease of glucose utilization was more pronounced on the ac-
tivated side, as reflected by the decrease of the ratio contralateral/
ipsilateral in the primary somatosensory cortex (i.e., S1
contra
vs
S1
ipsi
). During normolactemia, this ratio was 1.65 0.47, but
during hyperlactemia, it decreased to 1.26 0.37 (n 5; p
0.05) (Fig. 4E). In two control animals, in which both acqui-
sitions were performed under saline infusion, the ratio of LC-
MR
glu
between S1
contra
and S1
ipsi
remained the same during
the first and the second acquisition (Fig. 4E). Plasma levels of
glucose and lactate behaved in a similar manner as in the
baseline group.
Figure 4. Lactate is preferred over glucose by the brain. A and D detail the protocol during baseline (A) and stimulation (D)
conditions. Overall, hyperlactemia reduced LCMR
glu
by on average 38% in a dose-dependent manner (B, C) during baseline
conditions. In activated cortex (S1
contra
), cerebral glucose utilization was further decreased. The effectiveness of stimulation was
controlled autoradiographically (D). The ratio S1
contra
/S1
ipsi
decreased from 1.65 0.47 to 1.26 0.37 (E). 1, Acquisition 1; 2,
acquisition 2; LAC, hyperlactemia; SAL, saline. *p 0.05, n.s., not significant. n 13 for LAC group, n 4 for SAL group at
baseline conditions, n 5 for LAC group, and n 2 for SAL at stimulation conditions.
7482 J. Neurosci., May 18, 2011 31(20):74777485 Wyss et al. Lactate as a Neuronal Energy Substrate
Neuronal activity is not reduced by lactate
To verify that the observed reduction in glucose utilization does
not originate from a reduced requirement for energy of the brain
attributable to impaired neural activity, we performed an addi-
tional series of VSD experiments during hyperlactemia. Under
these conditions, increased lactate levels did not affect neuronal
excitability (Fig. 5).
Discussion
For a long time, no functional role has been assigned to cerebral
lactate. Contrarily, the presence of lactate in the brain has been
interpreted as a sign of cerebral harm and of hypoxia (Siesjo,
1981). Over the past few decades, this perspective has changed
and it has been postulated that lactate may also possess physio-
logical functions relevant to the CNS. Important roles have been
attributed to lactate such as the modulation of CBF (Gordon et
al., 2008), as a buffer to sustain neuronal energy supply (Kasis-
chke et al., 2004) and most prominently as a pivotal element in
the neuronglia metabolic cooperation to regulate energy supply
in an activity-dependent manner (for review, see Pellerin and
Magistretti, 2003; Magistretti, 2006). Furthermore, it has been
shown recently that lactate levels are sensed by a specific type of
neuron (orexin neurons in the lateral hypothalamus) to coordi-
nate the balancing of brain activity and energy supply (Lamet al.,
2005; Shimizu et al., 2007; Erlichman et al., 2008; Parsons and
Hirasawa, 2010). Nevertheless, the details of brain energy metab-
olism in general and of lactate as a cerebral energy substrate in
particular are still highly debated.
Lactate serves as a metabolic source
for neurons
It has been shown in vitro that lactate ox-
idation is able to sustain excitatory synap-
tic activity (Schurr et al., 1988; Rouach et
al., 2008), and there have been reports
about lactate protecting cerebral function
during hypoglycemia (Maran et al., 1994;
King et al., 1998). However, the latter in
vivo studies did not use direct neuronal
readouts to assess synaptic activity, and
glucose levels were above 2.5 mmol/L. The
present study is the first in vivo demon-
stration of the maintenance of neuronal
excitability in the presence of lactate as the
primary energy source. The amplitude of
the stimulation-induced increase in neuro-
nal activitywas maintainedwhenlactatewas
supplied during severe hypoglycemia (Fig.
1). However, the increaseddelay of the VSD
signal may indicate that lactate alone may
not quite suffice to completely sustain neu-
ronal functioning. Indeed, it has been
shown that glucose is used by neurons to
maintain their antioxidant status via the
pentose phosphate pathway (PPP), which
cannot be fueled by lactate (Magistretti,
2008; Herrero-Mendez et al., 2009). Be-
cause of the low plasma glucose levels and
subsequent small glucose concentration
gradient, the transport into neurons is in-
adequate (Barros and Deitmer, 2010) and
presumably not enough substrate is avail-
able for the neurons to efficiently stimulate
their antioxidant PPP. Neurons, which
are thought to be especially vulnerable against reactive oxygen
and nitrogen species, might not be capable of completely avoid-
ing oxidative damage, whichmay leadto the observeddelay of the
electrical signal in the absence of glucose. Furthermore, this find-
ing supports the important role in functional neuroenergetics
dedicated to glucose because glucose is required by the astrocyte
to pump glutamate (Magistretti et al., 1999). Second, lactate is
transported by MCTs in a cotransport with protons (Hertz and
Dienel, 2005). Lactate elevations as performed in our study may
lead to changes in lactate influx that result in alterations of intra-
cellular and extracellular pHin nerve tissue. These changes of the
proton gradient could disturb the conductance of action poten-
tials along nerve axons and lead to a delayed response.
Physiological meaning of k
2
Using 1-[
11
C]L-lactate, we demonstrated that the brain readily
oxidizes lactate at normal plasma lactate levels inaccordance with
recent work using 1- and 3-[
13
C]L-lactate (Bouzier et al., 2000;
van Hall et al., 2009; Boumezbeur et al., 2010). Our radiotracer
approach enables the measurement of cerebral lactate oxidation
without altering blood and tissue concentrations of lactate. In
this respect, it is important to consider the physiological meaning
of k
2
. By definition, it represents the loss of label from tissue (Fig.
3A). The model depicted in Figure 3Ademonstrates that the label
can be lost by back-diffusion of labeled lactate into vascular space
and as labeled CO
2
. The latter pathway is more interesting be-
cause it is related to the oxidation of lactate, unless the diffusion
of CO
2
across the bloodbrain barrier is rate limiting. This is
Figure5. Cerebral integrity is not disruptedduringhyperlactemia. A, B, Asingle example demonstratingthe spread(A) andthe
amplitude over time (B) during saline (SAL) and lactate (LAC) infusion is shown. CE, Mean results of changes in amplitude (C),
spread (D), and time-to-peak (E) (n 4; n.s., not significant compared with baseline).
Wyss et al. Lactate as a Neuronal Energy Substrate J. Neurosci., May 18, 2011 31(20):74777485 7483
unlikely because CO
2
back-diffusion does not seem restricted
(Paulson, 2002). Radiolabeled lactate is processed by LDH and
PDHbefore the label is lost as CO
2
. Parameter k
2
can therefore be
related to either of themor to both, depending on whether one of
these steps is rate limiting. A flux estimation suggests that the
processing by LDH and not PDH is the rate-limiting step.
If k
2
represented the conversion of pyruvate to acetyl-CoA, it
would closely be related to the flux of pyruvate (Flux
PDH
), which
in this model would be expressed as follows: Flux
PDH
k
2
* C
pyr
,
or C
pyr
Flux
PDH
/k
2
, where C
pyr
is the intracellular pyruvate
concentration. Inserting a measured Flux
PDH
value of 1.13
mol g
1
min
1
(Oz et al., 2004) and the mean k
2
from our
[
11
C]L-lactate experiments (0.077 min
1
) yields anunphysiolog-
ical pyruvate concentration of 15 mmol/L. Measured pyruvate
concentrations are in the range of 0.10.2 mmol/L (McIlwain,
1985).
In analogy, the conversion of lactate to pyruvate by LDH yields
Flux
LDH
k
2
* C
lac
, where C
lac
is the intracellular lactate con-
centration. Using the numbers measuredinhumans by magnetic
resonance spectroscopy [Flux
LDH
0.06 mol g
1
min
1
)
(Boumezbeur et al., 2010)], the above equation yields an intra-
cellular lactate concentration of 0.06/0.077 0.78 mmol/L. This
is in the same range as the reported number [1 mmol/L at plasma
lactate levels of 1.5 mmol/L (Boumezbeur et al., 2010)]. These
estimations suggest that k
2
is more related to the activity of LDH
than PDH.
Evidence for increased lactate oxidation during stimulation
The increase of k
2
can be seen as a facilitation of the flux of
substrates through the oxidative chain. However, an effective in-
crease of the lactate flux furthermore requires that there is no
relevant stimulation-induced drop in another factor driving lac-
tate oxidation, e.g., the intracellular lactate concentration. How-
ever, such a drop seems unlikely for two reasons. First, the
extracellular lactate concentration has been shown to increase
during stimulation (Hu and Wilson, 1997; Caesar et al., 2008).
Second, there is evidence for a translocation of MCT2 to the
membrane surface during stimulation (Pierre et al., 2009), which
would augment lactate transport and lead to higher intracellular
lactate concentration. Such a translocation could also explain the
increased permeability surface product during stimulation ob-
served in this study.
There is another line of evidence supporting an elevated neu-
ronal lactate oxidation during increased neuronal activity. An
increased conversion of lactate to pyruvate is facilitated by an
increase of the lactate/pyruvate ratio. Besides the increase in the
lactate concentration (see above), a decrease of the pyruvate con-
centration or both causes a drop in this ratio. Areduced pyruvate
concentration is achieved by a reduced glycolytic flux. Indeed,
there is evidence fromexperiments in cell cultures that glutamate
inhibits neuronal glucose uptake (Porras et al., 2004; Castro et al.,
2009) and that upregulation of glycolysis under stress conditions
is prevented (Herrero-Mendez et al., 2009). Recently, it has been
found in vivo that increased neuronal firing mediates inhibition
of glucose transport in neurons while stimulating astrocytic glu-
cose uptake (Chuquet et al., 2010).
Concluding remarks
The results presented demonstrate in vivo that lactate is metabo-
lized by neurons. Indeed, VSD imaging enabled us to follow a
neuronal readout, which is supported by the supplied lactate dur-
ing the experimental period. Considering the facts that intrace-
rebral glycogen stores are relatively limited and consumed in the
absence of exogenous glucose within a few minutes (Brown and
Ransom, 2007) and that gluconeogenic activity in the brain is
negligible (Nelsonet al., 1985), it is evident that the neurons must
rely on lactate as an energy substrate under hypoglycemic condi-
tions. In addition, hepatic gluconeogenesis is suppressed by the
insulin action. Thus, it can be suggested that the neuroprotective
effect of lactate observed in previous studies (Gjedde and Crone,
1975; Schurr et al., 2001; Mason et al., 2006; Berthet et al., 2009)
is attributable to direct lactate oxidationinneurons. However, we
cannot exclude astrocytic lactate uptake and oxidation.
It is important to note that the data reported here using radiola-
beled lactate provide evidence for the ability of neurons to increase
lactate use during stimulation. During increased activation, we ob-
servedanincreasedturnover of 1-[
11
C]lactate alsounder normogly-
cemia. Considering the fact that all the experiments were performed
under anesthesia, thus reducingthe overall metabolic rate by a factor
of twofoldtothreefold(Hyder et al., 2006), it is reasonable toassume
that an even larger contribution of lactate to brain metabolismmay
occur in the awake animal.
In summary, our multimodal experiments demonstrate that
(1) lactate is capable of maintaining neuronal integrity to a large
degree in the absence of glucose, (2) lactate is preferred over
glucose if both substrates are available, (3) lactate is readily me-
tabolized by the intact adult rodent brain, and (4) its metabolism
is activity dependent.
References
Barros LF, Deitmer JW (2010) Glucose and lactate supply to the synapse.
Brain Res Rev 63:149159.
Berthet C, Lei H, Thevenet J, Gruetter R, Magistretti PJ, Hirt L (2009) Neu-
roprotective role of lactate after cerebral ischemia. J Cereb Blood Flow
Metab 29:17801789.
Boumezbeur F, Petersen KF, Cline GW, Mason GF, Behar KL, Shulman GI,
Rothman DL (2010) The contribution of blood lactate to brain energy
metabolism in humans measured by dynamic 13C nuclear magnetic res-
onance spectroscopy. J Neurosci 30:1398313991.
Bouzier AK, Thiaudiere E, Biran M, Rouland R, Canioni P, Merle M (2000)
The metabolism of [3-(13)C]lactate in the rat brain is specific of a pyru-
vate carboxylase-deprived compartment. J Neurochem 75:480486.
BrownAM, RansomBR (2007) Astrocyte glycogenand brainenergy metab-
olism. Glia 55:12631271.
Caesar K, Hashemi P, Douhou A, Bonvento G, Boutelle MG, Walls AB, Lau-
ritzen M (2008) Glutamate receptor-dependent increments in lactate,
glucose and oxygen metabolism evoked in rat cerebellum in vivo.
J Physiol 586:13371349.
Castro MA, Beltran FA, Brauchi S, Concha II (2009) A metabolic switch in
brain: glucose and lactate metabolism modulation by ascorbic acid.
J Neurochem 110:423440.
Chuquet J, Quilichini P, Nimchinsky EA, Buzsaki G (2010) Predominant
enhancement of glucose uptake in astrocytes versus neurons during acti-
vation of the somatosensory cortex. J Neurosci 30:1529815303.
Cremer JE (1982) Substrate utilization and brain development. J Cereb
Blood Flow Metab 2:394407.
Crone C (1963) The permeability of capillaries in various organs as deter-
mined by use of the indicator diffusion method. Acta Physiol Scand
58:292305.
Dombrowski GJ Jr, Swiatek KR, Chao KL (1989) Lactate, 3-hydroxybutyrate, and
glucose as substrates for the early postnatal rat brain. Neurochem Res
14:667675.
Erlichman JS, Hewitt A, Damon TL, Hart M, Kurascz J, Li A, Leiter JC (2008)
Inhibition of monocarboxylate transporter 2 in the retrotrapezoid nu-
cleus in rats: a test of the astrocyte-neuron lactate-shuttle hypothesis.
J Neurosci 28:48884896.
Gallagher CN, Carpenter KL, Grice P, Howe DJ, MasonA, Timofeev I, Menon
DK, Kirkpatrick PJ, Pickard JD, Sutherland GR, Hutchinson PJ (2009)
The human brain utilizes lactate via the tricarboxylic acid cycle: a 13C-
labelled microdialysis and high-resolution nuclear magnetic resonance
study. Brain 132:28392849.
7484 J. Neurosci., May 18, 2011 31(20):74777485 Wyss et al. Lactate as a Neuronal Energy Substrate
Gjedde A, Crone C (1975) Induction processes in blood-brain transfer of
ketone bodies during starvation. Am J Physiol 229:11651169.
Gordon GR, Choi HB, Rungta RL, Ellis-Davies GC, MacVicar BA (2008)
Brain metabolismdictates the polarity of astrocyte control over arterioles.
Nature 456:745749.
Hassel B, Brthe A (2000) Cerebral metabolism of lactate in vivo: evidence
for neuronal pyruvate carboxylation. J Cereb Blood Flow Metab 20:
327336.
Herrero-Mendez A, Almeida A, Fernandez E, Maestre C, Moncada S, Bolanos
JP (2009) The bioenergetic and antioxidant status of neurons is con-
trolled by continuous degradation of a key glycolytic enzyme by APC/C-
Cdh1. Nat Cell Biol 11:747752.
Hertz L, Dienel GA (2005) Lactate transport and transporters: general prin-
ciples and functional roles in brain cells. J Neurosci Res 79:1118.
Hu Y, Wilson GS (1997) Atemporary local energy pool coupled to neuronal
activity: fluctuations of extracellular lactate levels in rat brain monitored
with rapid-response enzyme-based sensor. J Neurochem 69:14841490.
Hyder F, Patel AB, Gjedde A, Rothman DL, Behar KL, Shulman RG (2006)
Neuronal-glial glucose oxidation and glutamatergic-GABAergic func-
tion. J Cereb Blood Flow Metab 26:865877.
Kasischke KA, Vishwasrao HD, Fisher PJ, Zipfel WR, Webb WW (2004)
Neural activity triggers neuronal oxidative metabolismfollowed by astro-
cytic glycolysis. Science 305:99103.
King P, Kong MF, Parkin H, MacDonald IA, Barber C, Tattersall RB (1998)
Intravenous lactate prevents cerebral dysfunction during hypoglycaemia
in insulin-dependent diabetes mellitus. Clin Sci (Lond) 94:157163.
Knudsen GM, Paulson OB, Hertz MM (1991) Kinetic analysis of the human
blood-brain barrier transport of lactate and its influence by hypercapnia.
J Cereb Blood Flow Metab 11:581586.
LamTK, Gutierrez-Juarez R, Pocai A, Rossetti L (2005) Regulation of blood
glucose by hypothalamic pyruvate metabolism. Science 309:943947.
Lippert MT, Takagaki K, Xu W, Huang X, Wu JY (2007) Methods for
voltage-sensitive dye imaging of rat cortical activity with high signal-to-
noise ratio. J Neurophysiol 98:502512.
Magistretti PJ (2006) Neuron-glia metabolic coupling and plasticity. J Exp
Biol 209:23042311.
Magistretti PJ (2008) Brain energy metabolism. In: Fundamental neurosci-
ence (Squire L, Berg D, Bloom FE, du Lac S, Ghosh A, Spitzer N, eds), pp
271293. San Diego: Academic.
Magistretti PJ, Pellerin L, Rothman DL, Shulman RG (1999) Energy on de-
mand. Science 283:496497.
Maran A, Cranston I, Lomas J, Macdonald I, Amiel SA (1994) Protection by
lactate of cerebral function during hypoglycaemia. Lancet 343:1620.
Mason GF, Petersen KF, Lebon V, Rothman DL, Shulman GI (2006) In-
creased brain monocarboxylic acid transport and utilization in type 1
diabetes. Diabetes 55:929934.
McIlwain H, Bachelard HS (1985) Biochemistry and the central nervous
system, Ed 5. New York: Churchill Livingstone.
Mikolajczyk K, Szabatin M, Rudnicki P, Grodzki M, Burger C (1998) A
JAVAenvironment for medical image data analysis: initial application for
brain PET quantitation. Med Inform (Lond) 23:207214.
Nehlig A, Pereira de Vasconcelos A (1993) Glucose and ketone body utili-
zation by the brain of neonatal rats. Prog Neurobiol 40:163221.
Nelson T, Lucignani G, Atlas S, Crane AM, Dienel GA, Sokoloff L (1985)
Reexamination of glucose-6-phosphatase activity in the brain in vivo: no
evidence for a futile cycle. Science 229:6062.
Oz G, Berkich DA, Henry PG, Xu Y, LaNoue K, Hutson SM, Gruetter R
(2004) Neuroglial metabolism in the awake rat brain: CO
2
fixation in-
creases with brain activity. J Neurosci 24:1127311279.
Parsons MP, Hirasawa M (2010) ATP-sensitive potassium channel-
mediated lactate effect on orexin neurons: implications for brain energet-
ics during arousal. J Neurosci 30:80618070.
Paulson OB (2002) Blood-brain barrier, brain metabolism and cerebral
blood flow. Eur Neuropsychopharmacol 12:495501.
Pellerin L, Magistretti PJ (1994) Glutamate uptake into astrocytes stimu-
lates aerobic glycolysis: a mechanism coupling neuronal activity to glu-
cose utilization. Proc Natl Acad Sci U S A 91:1062510629.
Pellerin L, Magistretti PJ (2003) Food for thought: challenging the dogmas.
J Cereb Blood Flow Metab 23:12821286.
Pierre K, Chatton JY, Parent A, Repond C, Gardoni F, Di Luca M, Pellerin L
(2009) Linking supply to demand: the neuronal monocarboxylate trans-
porter MCT2 and the alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-
propionic acid receptor GluR2/3 subunit are associated in a common
trafficking process. Eur J Neurosci 29:19511963.
Porras OH, Loaiza A, Barros LF (2004) Glutamate mediates acute glucose
transport inhibition in hippocampal neurons. J Neurosci 24:96699673.
Renkin EM (1959) Transport of potassium-42 from blood to tissue in iso-
lated mammalian skeletal muscles. Am J Physiol 197:12051210.
Rouach N, Koulakoff A, Abudara V, Willecke K, Giaume C (2008) Astro-
glial metabolic networks sustain hippocampal synaptic transmission. Sci-
ence 322:15511555.
Schurr A, West CA, Rigor BM (1988) Lactate-supported synaptic function
in the rat hippocampal slice preparation. Science 240:13261328.
Schurr A, Payne RS, Miller JJ, Tseng MT, Rigor BM (2001) Blockade of
lactate transport exacerbates delayed neuronal damage in a rat model of
cerebral ischemia. Brain Res 895:268272.
Shackford SR, Schmoker JD, Zhuang J (1994) The effect of hypertonic re-
suscitation on pial arteriolar tone after brain injury and shock. J Trauma
37:899908.
Shimizu H, Watanabe E, Hiyama TY, Nagakura A, Fujikawa A, Okado H,
Yanagawa Y, Obata K, Noda M (2007) Glial Nax channels control lac-
tate signaling to neurons for brain [Na] sensing. Neuron 54:5972.
Siesjo BK (1981) Cell damage in the brain: a speculative synthesis. J Cereb
Blood Flow Metab 1:155185.
Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, PettigrewKD,
Sakurada O, Shinohara M (1977) The [14C]deoxyglucose method for
the measurement of local cerebral glucose utilization: theory, procedure,
and normal values in the conscious and anesthetized albino rat. J Neuro-
chem 28:897916.
van Hall G (1999) Correction factors for 13C-labelled substrate oxidation at
whole-body and muscle level. Proc Nutr Soc 58:979986.
van Hall G, Strmstad M, Rasmussen P, Jans O, Zaar M, Gam C, Quistorff B,
Secher NH, Nielsen HB (2009) Blood lactate is an important energy
source for the human brain. J Cereb Blood Flow Metab 29:11211129.
Weber B, Burger C, Biro P, Buck A (2002) A femoral arteriovenous shunt
facilitates arterial whole blood sampling in animals. Eur J Nucl Med Mol
Imaging 29:319323.
Weber B, Spath N, Wyss M, Wild D, Burger C, Stanley R, Buck A (2003)
Quantitative cerebral blood flow measurements in the rat using a beta-
probe and H2 15O. J Cereb Blood Flow Metab 23:14551460.
Wyss MT, Obrist NM, Haiss F, Eckert R, Stanley R, Burger C, Buck A, Weber
B (2009) A beta-scintillator for surface measurements of radiotracer ki-
netics in the intact rodent cortex. Neuroimage 48:339347.
Wyss et al. Lactate as a Neuronal Energy Substrate J. Neurosci., May 18, 2011 31(20):74777485 7485