Inhibition of the enzyme autotaxin reduces cortical excitability and ameliorates the outcome in stroke

Abstract

Stroke penumbra injury caused by excess glutamate is an important factor in determining stroke outcome; however, several therapeutic approaches aiming to rescue the penumbra have failed, likely due to unspecific targeting and persistent excitotoxicity, which continued far beyond the primary stroke event. Synaptic lipid signaling can modulate glutamatergic transmission via presynaptic lysophosphatidic acid (LPA) 2 receptors modulated by the LPA-synthesizing molecule autotaxin (ATX) present in astrocytic perisynaptic processes. Here, we detected long-lasting increases in brain ATX concentrations after experimental stroke. In humans, cerebrospinal fluid ATX concentration was increased up to 14 days after stroke. Using astrocyte-specific deletion and pharmacological inhibition of ATX at different time points after experimental stroke, we showed that inhibition of LPA-related cortical excitability improved stroke outcome. In transgenic mice and in individuals expressing a single-nucleotide polymorphism that increased LPA-related glutamatergic transmission, we found dysregulated synaptic LPA signaling and subsequent negative stroke outcome. Moreover, ATX inhibition in the animal model ameliorated stroke outcome, suggesting that this approach might have translational potential for improving the outcome after stroke.

Full text

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
1 of 12
STROKE
Inhibition of the enzyme autotaxin reduces cortical
excitability and ameliorates the outcome in stroke
Lynn Bitar1†, Timo Uphaus1†, Carine Thalman1†, Muthuraman Muthuraman1, Luzia Gyr2,
Haichao Ji1,3, Micaela Domingues1, Heiko Endle1,3, Sergiu Groppa1, Falk Steffen1, Nabin Koirala1,
Wei Fan4, Laura Ibanez5, Laura Heitsch6, Carlos Cruchaga5, Jin-Moo Lee7, Florian Kloss2,
Stefan Bittner1, Robert Nitsch8‡, Frauke Zipp1*‡, Johannes Vogt1,3*‡
Stroke penumbra injury caused by excess glutamate is an important factor in determining stroke outcome; however,
several therapeutic approaches aiming to rescue the penumbra have failed, likely due to unspecific targeting and
persistent excitotoxicity, which continued far beyond the primary stroke event. Synaptic lipid signaling can
modulate glutamatergic transmission via presynaptic lysophosphatidic acid (LPA) 2 receptors modulated by the
LPA-synthesizing molecule autotaxin (ATX) present in astrocytic perisynaptic processes. Here, we detected
long-lasting increases in brain ATX concentrations after experimental stroke. In humans, cerebrospinal fluid ATX
concentration was increased up to 14 days after stroke. Using astrocyte-specific deletion and pharmacological
inhibition of ATX at different time points after experimental stroke, we showed that inhibition of LPA-related
cortical excitability improved stroke outcome. In transgenic mice and in individuals expressing a single-nucleotide
polymorphism that increased LPA-related glutamatergic transmission, we found dysregulated synaptic LPA signaling
and subsequent negative stroke outcome. Moreover, ATX inhibition in the animal model ameliorated stroke outcome,
suggesting that this approach might have translational potential for improving the outcome after stroke.
INTRODUCTION
Stroke is the leading cause of disability and the second leading cause
of death worldwide (1). Ischemic stroke results from an insufficient
blood supply, leading to disruption of neuronal function and cell
death, which eventually results in persistent loss of brain tissue and
consequent disability (2). Stroke treatment aims at salvaging func-
tionally impaired yet still viable tissue in the ischemic penumbra,
the area surrounding the ischemic core (3–5). In the penumbra,
metabolically compromised neurons are in jeopardy due to second-
ary glutamatergic signaling events (6,7), which lead to excitotoxic
cell death involving glutamatergic receptors and subsequent calcium
overload (8). So far, reducing penumbra loss by direct glutamatergic
inhibition has not been successful; thus, targeting excitotoxicity is
still an urgent need in stroke therapy (9,10). Moreover, recent data
suggest that, whereas basal glutamatergic transmission is important
for neuronal survival, pathophysiological glutamate overload leads
to excitotoxicity (11). Therefore, therapies maintaining physiological,
regenerative glutamatergic transmission but inhibiting pathological
glutamate overload are critically missing.
Under physiological conditions, astrocytes not only maintain
ionic homeostasis but are also involved in the pathology of neuro-
logical diseases including stroke (12–14). We have recently shown
that astrocytes are involved in the regulation of glutamatergic trans-
mission via synaptic expression of the lysophosphatidic acid (LPA)–
synthesizing molecule autotaxin (ATX), whereby ATX expression is,
in turn, triggered by glutamatergic stimulation (15) and is enhanced
upon brain lesion (16). Our own studies have shown that synaptic
bioactive phospholipids like LPA are critical factors in homeostatic
regulation of glutamatergic transmission and cortical network ex-
citability. LPA, which is locally synthetized at the synaptic cleft
of glutamatergic synapses by ATX from lysophosphatidylcholine
(LPC), is a short-lived but potent ligand binding to G protein–
coupled receptors in the nanomolar range (LPA receptors 1 to 6)
(17). Synaptic LPA activates high-affinity LPA2 receptors, which
are located on presynaptic terminals of glutamatergic synapses and
regulate glutamatergic release probabilities (18). On the other hand,
synaptic LPA is regulated by the postsynaptic molecule plasticity-
related gene 1 (PRG-1), which is expressed at the postsynaptic densi-
ty of excitatory synapses on glutamatergic neurons not only in the
cortex (18) but also in the basal ganglia (19). PRG-1 modulates glu-
tamatergic synaptic transmission from the postsynaptic site by reg-
ulating LPA via uptake into intracellular compartments (18,20). In
line with these findings, disruption of PRG-1 function was shown
to result in cortical hyperexcitability and an altered excitation/
inhibition (E/I) balance in cortical networks (18). This synaptic
lipid signaling regulating glutamatergic transmission and leading to
cortical network hyperexcitability was confirmed in different corti-
cal regions (21,22) and was shown to be a critical factor in animal
models for psychiatric disorders (15,23) and in the aging brain (24).
Moreover, we described a PRG-1 single-nucleotide polymorphism
1Department of Neurology, Focus Program Translational Neuroscience (FTN) and
Immunotherapy (FZI), Rhine Main Neuroscience Network (rmn2), University Medical
Center of the Johannes Gutenberg University Mainz, 55131 Mainz, Germany.
2Transfer Group Anti-Infectives, Leibniz Institute for Natural Product Research and
Infection Biology, Hans Knoell Institute, 07745 Jena, Germany. 3Department of
Molecular and Translational Neuroscience, Cologne Excellence Cluster for Stress
Responses in Aging-Associated Diseases (CECAD), Center for Molecular Medicine
Cologne (CMMC), University of Cologne, Faculty of Medicine and University Hospital
Cologne, 50937 Cologne, Germany. 4Focus Program Translational Neuroscience (FTN),
University Medical Center of the Johannes Gutenberg University Mainz, 55131 Mainz,
Germany. 5Department of Psychiatry, Department of Neurology, NeuroGenomics
and Informatics Center, Washington University School of Medicine, St. Louis,
MO 63110, USA. 6Department of Emergency Medicine, Department of Neurology,
Washington University School of Medicine, St. Louis, MO 63110, USA. 7Department
of Neurology, Radiology, and Biomedical Engineering, Washington University School
of Medicine, St. Louis, MO 63110, USA. 8Institute of Translational Neuroscience,
Westfälische Wilhelms-Universität Münster, 48149 Münster, Germany.
*Corresponding author. Email: johannes.vogt@uk-koeln.de (J.V.); zipp@uni-mainz.
de (F.Z.)
†These authors contributed equally to this work as first authors.
‡These authors contributed equally to this work as senior authors.
Copyright © 2022
The Authors, some
rights reserved;
exclusive licensee
American Association
for the Advancement
of Science. No claim
to original U.S.
Government Works
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
2 of 12
(SNP; rs138327459, PRG-1R345T), which disrupts the regulation of
synaptic LPA concentration and affects in a monoallelic version up
to 0.6% of the population, leading to increased cortical network
hyperexcitability (20). Previous data suggest that dysregulated
synaptic lipid signaling may be detrimental under pathological con-
ditions, leading to cortical hyperexcitability and, eventually, epileptic
seizures or to post-epileptic brain damage (18,25). We therefore
hypothesized that synaptic lipid regulation of glutamatergic trans-
mission might play an important role in stroke pathology, where
the combination of glutamatergic overload, cortical hyperexcitability,
and excitotoxic cell death could lead to overactivation of synaptic
lipid signaling and to persistent overexcitation, causing enhance-
ment and spread of tissue injury in stroke. ATX released by astro-
cytes might play a central role in determining lipid signaling and
could represent a potential therapeutic target for reduction of
LPA-related cortical hyperexcitability. Using different genetic ani-
mal models, pharmacological and small-molecule approaches, and
translational data from patients after stroke, we aimed to assess the
role of synaptic lipid-related cortical excitability in preservation of
the penumbra and thus in determining ischemic stroke outcome.
RESULTS
ATX released by astrocytes is associated with infarct volume
and stroke outcome
After ischemic stroke, reactive astrocytes, characterized by increased
glial fibrillary acidic protein (GFAP) expression and coarse mor-
phology, accumulate in the peri-infarct area, subsequently forming
a glial scar (26,27). Consistently, in a model of ischemic stroke
induced by occlusion of the middle cerebral artery (MCAO) in
mice, we detected increased numbers of strongly GFAP-positive
astrocytes in the peri-infarct area and increased GFAP expression
in the cortex of the lesioned hemisphere (Fig.1,AandB). Because
synaptic lipid signaling is able to modulate glutamatergic transmis-
sion relevant for damage spreading in stroke, we aimed at analyzing
the amount of the LPA-synthesizing molecule ATX. We found an
abundant presence of ATX-positive punctae on and in close
proximity to activated astrocytes after MCAO when compared to
naïve astrocytes, which is in line with increased astrocyte-related
ATX expression and released ATX surrounding GFAP-positive
astrocytes (Fig.1C). This expression pattern is in agreement with
previous reports, suggesting that ATX transports LPA, delivering it
close to its effector site where LPA is released and signals via LPA
receptors (28–30). Western blot analysis in the lesioned hemisphere
confirmed a significant increase in ATX production after MCAO
(P<0.05; Fig.1D).
To evaluate the role of astrocytic ATX during MCAO, we used
animals with cell type–specific ATX deletion in activated astrocytes
generated by crossing an Atx fl/fl line with a Gfap-Cre+ mouse line
(Atx fl/fl Gfap-Cre+) (15,31). Here, we observed a significant decrease
in stroke infarct volume and an improvement in animal motor
behavior assessed by the modified neurological severity score (mNSS)
72hours after MCAO, suggesting a critical role of ATX expressed
by activated astrocytes in functional stroke outcome (P<0.01;
Fig.1,EtoG). To further assess cellular loss, we quantified activated
caspase-3–induced apoptosis, revealing a significant reduction in
astrocytic ATX-deficient mice, which further supports the impor-
tant role of ATX secretion by activated astrocytes in stroke pathology
(P<0.05; Fig.1,HandI).
ATX and LPA are increased in the CSF of patients after stroke
We assessed the presence of ATX in the proximity of astrocytic
processes and glutamatergic synapses in the human cortex and
measured the concentration of ATX in the cerebrospinal fluid
(CSF) and serum of patients after stroke. Here, we found similar
localization of ATX at glutamatergic synapses (fig. S1A), as previ-
ously described in the mouse brain (15), and significantly elevated
ATX concentrations in the CSF of patients up to 14 days after the
acute stroke event when compared to age- and sex-matched control
donors (P<0.05; Fig.2A). This increase was limited to the central
nervous system, because the concentration of ATX in serum was
rather reduced in patients after stroke when compared to control
subjects (Fig.2B). To further analyze the role of ATX in stroke
outcome, we correlated the amount of ATX with changes in the
National Institutes of Health Stroke Scale (NIHSS) score within the
first 24hours after admission and detected a negative correlation
with high CSF ATX concentrations (Fig.2C). Further analysis of
patients with no improvement in the NIHSS within 24hours after
admission displayed significantly higher CSF ATX concentrations
compared to patients with NIHSS improvement at 24hours after
admission (P< 0.01; Fig.2D). Because NIHSS changes within the
first 24 hours were previously shown to be a good predictor for
long-term stroke outcome (32), our data suggest a deleterious effect
of high amounts of ATX in human stroke pathology. Further-
more, we found a positive correlation between perfusion mismatch
(pointing to stroke penumbra) and ATX concentrations (Fig.2E),
indicating a role of increased ATX in penumbra expansion. ATX is
the main extracellular synthetizing molecule of LPA, a potent but
short-lived bioactive phospholipid shown to substantially increase
presynaptic release probabilities (18). We therefore quantified the
concentration of LPA and its precursor LPC in the CSF by mass
spectrometry and found significantly increased quantities in patients
after stroke when compared to control subjects (P< 0.05 for LPA
and P<0.01 for LPC; Fig.2F). These data suggest that higher ATX
production—previously shown to be stimulated by glutamate (15)
and to result in higher LPA amounts, leading to altered E/I
balance—is involved in synaptic lipid-induced cortical excitability
long after the acute stroke event also in patients.
ATX inhibition rescues penumbra and improves neurological
deficits after MCAO
Our data suggest that stroke-induced up-regulation of ATX in
astrocytes promoted cell death and that ATX is the major source of
LPA. Therefore, we examined whether the observed effect upon
ATX deletion was mediated by reduced LPA production. We evalu-
ated the effects of pharmacological ATX inhibition after MCAO as
a strategy to rescue tissue within the penumbra following stroke. To
this end, we used a potent and specific small-molecule ATX inhibitor
(PF8380), which specifically targets the LPA synthesis function of
ATX (33). We have previously shown that PF8380 accumulates in
the brain in therapeutic concentrations after systemic intraperitoneal
application where it effectively blocked LPA synthesis (15). To
achieve continuous ATX inhibition, mice received an intraperitoneal
injection of PF8380 (30 mg/kg body weight) every 24hours (Fig.3A),
which was shown to effectively reduce ATX activity (34). PF8380
treatment significantly diminished infarct volumes at 72hours after
MCAO and significantly improved mNSS (P<0.05; Fig.3,BtoD).
Mass spectrometry measurements of LPA in CSF revealed a signifi-
cant increase in LPA 24hours after MCAO and a significant decrease
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
3 of 12
in LPA upon ATX inhibition in MCAO animals (P<0.01; Fig.3E).
In line with the described LPA-induced cortical hyperexcitability (18),
ATX inhibition significantly decreased active caspase- 3–associated cell
loss when compared to nontreated animals (P<0.05; Fig.3,FandG).
These data confirm that the LPA-synthesizing enzyme ATX plays
an important LPA-mediated role in the pathology of focal cerebral
ischemia and that systemic targeting of ATX may be a potent strategy
to improve functional outcome after ischemic stroke in mice.
Synaptic LPA-mediated cortical hyperexcitability
determines MCAO outcome
To further characterize the role of LPA signaling in stroke pathology,
we used mice expressing a human SNP in PRG-1 (Prg-1R346T, the
mouse homolog of human PRG-1R345T), the gene encoding for
PRG-1R346T, leading to loss of PRG-1 function. Here, a single–
amino acid mutation impedes PRG-1’s LPA uptake function, lead-
ing to higher synaptic LPA concentrations, increased cortical
excitability, and altered E/I balance in mice and humans (15,20).
Our data show that Prg-1R346T mice displayed higher spontaneous
glutamatergic events, which is the underlying cause of increased
LPA-mediated cortical excitability (Fig.4,AandB). MCAO in
Prg-1R346T mice had increased infarct size and an adverse behavioral
outcome 24hours after stroke (P<0.01; Fig.4,CtoE). In line with
this, assessment of apoptotic cells, as shown by immunostaining against
active caspase-3, revealed a significant increase in caspase- positive
neurons in Prg-1R346T animals, thereby supporting the detrimental
Fig. 1. ATX deletion in astrocytes is associated with an infarct volume reduction and a better stroke outcome. (A) ATX and GFAP expression in the (left) naïve cortex
and (right) peri-infarct area of the cortex 72 hours after MCAO. Scale bar, 5 m. (B) Western blot assessment of GFAP expression 72 hours after stroke (n = 5 naïve animals
and n = 5 MCAO, two-tailed t test). Tubulin was used as a loading control. (C) Higher magnification of GFAP-positive astrocytes and ATX expression in the (left) naïve
cortex and (right) lesioned cortex 72 hours after MCAO. Note the numerous ATX punctae in the lesioned cortex. Scale bar, 1 m. (D) ATX expression in naïve animals and
in the infarcted hemisphere 72 hours after MCAO assessed by Western blot (n = 5 naïve and n = 6 MCAO, one-tailed t test). -Actin was used as a loading control. (E) Representative
images of brain sections from WT and Atxfl/fl Gfap-Cre+ animals 72 hours after MCAO. Infarct area is outlined by dashed line. Scale bar, 500 m. (F) Infarct volume 72 hours
after MCAO in Atxfl/fl Gfap-Cre+ animals (n = 9) compared to WT littermates (n = 12, two-tailed t test). (G) mNSS in Atxfl/fl Gfap-Cre+ animals after MCAO compared to WT
littermates (n = 12 per group, Bayesian analysis). (H) Representative activated caspase-3 staining in WT animals and Atxfl/fl Gfap-Cre+ animals 72 hours after MCAO. Red
arrows indicate activated caspase-3–positive cells. Scale bars, 50 and 5 m (inset). (I) Quantification of activated caspase-3–positive cells per mm2 in the penumbra of
Atxfl/fl Gfap-Cre+ animals (n = 8) compared to WT littermates (n = 6, two-tailed t test). Data presented as means ± SEM (*P < 0.05, **P < 0.01, and ***P < 0.001 or accuracy
of **>90% for Bayesian analysis).
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
4 of 12
role of cortical excitability in stroke (P < 0.05; Fig. 4, F and G).
Moreover, neurofilament light (NfL) chain, a neuronal scaffolding
protein released into the extracellular space upon neuroaxonal
damage (35,36) and shown to be correlated with stroke mortality
(37), was significantly increased in Prg-1R346T mice when compared
to Prg-1 wild-type (Prg-1WT) littermates, indicating a more profound
neuronal damage in these mice after MCAO (P<0.05; Fig.4H).
Together, our data suggest that a mild preexisting LPA- related cor-
tical hyperexcitability, as present in Prg-1R346T–expressing mice and
human PRG-1R345T mutation carriers (20), has detrimental effects
in stroke pathology.
Last, to determine whether synaptic LPA signaling led to cortical
hyperexcitability and thereby to worse stroke outcome, we analyzed
Prg-1−/−/Lpa2−/− animals. Here, deletion of the presynaptic LPA2
receptor was shown to completely rescue the LPA-induced cortical
hyperexcitability caused by PRG-1 disruption (18). Whereas
Prg-1R346T animals showed increased cortical excitability and in-
creased stroke volumes, Prg-1−/−/Lpa2−/− animals, lacking the
presynaptic LPA2 receptor and thereby displaying disrupted synap-
tic LPA signaling, showed normal cortical excitability with no
difference compared to Prg-1WT/Lpa2WT conditions. In Prg-1−/−/
Lpa2−/− animals, we did not find differences in stroke volumes or in
neurological scores when compared to Prg-1WT/Lpa2WT littermates
(Fig.4,I to K). These data strongly argue for the involvement of
synaptic LPA signaling acting via regulation of cortical network
excitability in stroke pathology.
ATX inhibition rescues outcome of amplified stroke upon
altered E/I balance
To investigate the translational potential of interrupting this ATX-
induced excitotoxic cycle after stroke, we analyzed whether inhibi-
tion of the LPA-synthesizing enzyme ATX is able to rescue the
notably higher MCAO sequelae observed in Prg-1R346T animals.
After MCAO, Prg-1R346T animals and Prg-1WT littermates were
treated for three consecutive days with the ATX inhibitor PF8380
(Fig.5A) and underwent T2*-weighted magnetic resonance imaging
(MRI) for penumbra and infarct core assessment at 24 and 72hours
after MCAO (Fig.5B). Prg-1WT animals displayed a reduction in
penumbra and infarct core volume after systemic ATX inhibition at
24 and at 72hours after MCAO (Fig.5,CandD). Whereas untreated
Prg-1R346T animals displayed an increase in penumbra and infarct
volume at 24and 72hours after MCAO when compared to Prg-1WT
littermates, PF8380-treated Prg-1R346T animals improved after MCAO,
reaching values that were not different from treated Prg-1WT litter-
mates (Fig.5, Cand D). Reduced stroke size and reduced cortical
excitability also translated into improved functional outcomes
(assessed by mNSS) and improved motor function (assessed by
rotarod) after daily PF8380 treatment up to 72hours after MCAO
(Fig.5,EandF). On a cellular level, PF8380-treated Prg-1R346T mice
after MCAO displayed a reduced number of activated caspase-3–
positive apoptotic cells (fig. S1, B and C). These data strongly
suggest that ATX inhibition reduces excess LPA and normalizes
preexisting cortical excitability in the brain [previously demonstrated
to be present in Prg-1−/− animals (15,20)] even in the presence of a
detrimental pathology like ischemic stroke. To further corroborate
the role of synaptic lipid signaling in glutamatergic transmission in
stroke, we performed electrophysiological assessment at the single-
neuron level and invivo at the cortical network level. We found
significantly enhanced spontaneous excitatory postsynaptic currents
Fig. 2. High ATX concentrations are associated with worse stroke outcome in
patients after stroke. (A) ATX concentrations in the CSF at different time points
after stroke [n = 6, within 24 hours; n = 12, >24 hours to 7 days; n = 3, >7 to 14 days
after stroke; and n = 20 controls (C); Bayesian analysis]. (B) Serum ATX amounts after
stroke (n = 47, within 24 hours; n = 32, >2 to 7 days; n = 5, >7 to 14 days after stroke;
and n = 30 controls, Bayesian analysis). (C) Linear regression analysis correlating CSF
ATX concentrations with NIHSS improvement over the first 24 hours (NIHSS6h–24h)
after admission [correlation coefficient (r) = −0.619; P = 0.001]. Confidence intervals
(95%) are shown by dashed lines, and regression line is shown in bold in (C) and (E)
(n = 19 stroke patients). (D) Comparison of ATX CSF amounts in patients after stroke
with no improvement in NIHSS within 24 hours after admission (NIHSS6h–24h ↓, n = 8)
and in patients with NIHSS improvement during the first 24 hours (NIHSS6h–24h ↑,
n = 11, Bayesian analysis). (E) Linear regression analysis correlating (r = 0.710; P = 0.0172)
higher perfusion mismatch volumes and increasing ATX concentrations in the CSF
(n = 7 patients with stroke). (F) LPA and its precursor LPC in the CSF of patients after
stroke (n = 21 LPC and 20 LPA) compared to control individuals (n = 21 LPC and
18 LPA, Bayesian analysis). If not stated otherwise, data are presented as scatter
dot plot with all data points; horizontal lines (black or gray for better visibility)
represent medians (*P < 0.05 and **P < 0.01 or accuracy of *>85% or **>90% for
Bayesian analysis).
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
5 of 12
(sEPSCs) 24hours after MCAO in cortical neurons in the stroke
penumbra (marked by an asterisk in Fig.5B) compared to control
conditions and a decrease in sEPSC in these neurons when ATX
was inhibited by PF8380 (P<0.05; Fig.5,GtoJ). In line with these
results, invivo electrophysiology revealed a significantly enhanced
number of active neurons and an increased cortical neuron firing
rate after MCAO in the cortical layers 4 and 5 (P<0.01; Fig.5K), a
penumbra region, as shown by MRI analyses. After ATX inhibition
in animals that underwent MCAO, we observed a decreased firing
rate in Prg-1WT and Prg-1R346T mice (P<0.01; Fig.5L). The number
of actively firing neurons decreased after ATX inhibition in both
Prg-1WT and Prg-1R346T animals (P<0.01; Fig.5M). Moreover, we
have analyzed the effect of ATX inhibition by PF8380 at the cortical
network level. PF8380 decreased the power and the coherence of
neuronal network oscillations in the theta and in the gamma range,
which were described to reflect cortical network excitability (38)
and were shown to be mediated by synaptic LPA signaling (fig. S1,
D and E) (15). Together, these data support the importance of
ATX-produced LPA to continuous neu-
ronal loss in the penumbra driven by
LPA-induced cortical hyperexcitability.
Synaptic lipid signaling and stroke
outcome in mouse and human
To evaluate the putative therapeutic
potential of ATX inhibition in stroke
outcome, which necessarily has to be
effective after the initial stroke event,
we assessed the effect of PF8380 appli-
cation at different time points (30, 60,
90, and 180 min) after MCAO (Fig.6A).
Single application of PF8380 after
MCAO followed by subsequent daily
PF8380 injection resulted in significant
reduction of stroke volumes 72 hours
after MCAO for all analyzed time points
(P< 0.05 for 30 and 180 min, P<0.01
for 90 min, and P<0.001 for application
60min after MCAO; Fig.6,BandC). In
addition to a reduction in stroke volume,
stroke outcome in all treated groups
was markedly improved over the full
72-hour analysis period. Compared to
control animals, PF8380-treated animals
(treatment started at 30, 60, or 90min
after MCAO) revealed a better neuro-
logical status already 3hours after MCAO,
which continuously improved over
72 hours (Fig. 6D). In line with these
data, animals receiving PF8380 180min
after MCAO—a large interval compared
to all other treatments for stroke so
far—did not directly show differences in
their neurological status at 3hours after
MCAO (at the time point of PF8380
application) but significantly improved
at 72hours (P<0.01; Fig.6D, right) in
comparison to untreated animals.
To assess the role of synaptic lipid
signaling in patients with stroke, we analyzed those patients carry-
ing the above-described loss-of-function SNP in PRG-1 in a
monoallelic version (PRG-1R345T), which abolishes PRG-1’s ability
to take up LPA from the synaptic cleft, leading to increased gluta-
matergic transmission in the cortex (20). Here, we assessed stroke
impairment using the NIHSS at baseline [measured within 6hours
of stroke onset (Fig.6E); see also Supplementary Materials and
Methods] and at 24hours afterward as described in (32). These first
24hours after stroke seem to be of pathological importance and an
independent predictor for 90-day stroke outcomes (32). Patients
carrying PRG-1R345T had a significantly worse clinical stroke severity
(higher NIHSS24h) 24hours after admission (P<0.05; Fig.6F) and
a significantly smaller improvement in clinical stroke severity within
the first 24hours after admission than age- and sex-matched
patients (P <0.05; Fig.6G). These data suggest that synaptic lipid
signaling is an important pathological factor affecting long-term
stroke outcome in patients, thus highlighting ATX inhibition as a
potential therapeutic option for treating stroke in the clinic.
Fig. 3. ATX inhibition reduces infarct volume and neurological deficits after MCAO. (A) Experimental paradigm:
PF8380 treatment (or vehicle) started just before MCAO and was continued by daily intraperitoneal administration
up to 72 hours. (B) Exemplary cresyl violet–stained brain slices at 72 hours after MCAO from control and PF8380-treated
animals. Infarct area is outlined by dashed line. Scale bar, 500 m. (C) Quantification of infarct volume 72 hours after
MCAO (n = 11 vehicle and n = 12 PF8380, two-tailed t test). (D) mNSS after MCAO (n = 11 vehicle and n = 12 PF8380,
Bayesian analysis). (E) LPA concentrations in mice CSF 24 hours after stroke with and without ATX inhibition com-
pared to control (n = 12 controls; and for MCAO, n = 10 vehicle treated and 5 PF8380 treated, Bayesian analysis).
Scatter dot plot presents all data points, and line shows median. a.u., arbitrary units. (F) Activated caspase-3–positive
cells (red arrows) in vehicle- and PF8380-treated animals. Scale bars, 50 and 5 m (inset). (G) Quantitative analysis of
activated caspase-3–positive cells after ATX inhibition by PF8380 (n = 10 vehicle and n = 11 PF8380, one-tailed t test).
If not otherwise stated, data are presented as means ± SEM (*P < 0.05 or accuracy of **>90% for Bayesian analysis).
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
6 of 12
DISCUSSION
After the primary stroke event, the penumbra—a potentially viable
but hypoperfused brain area that may make up as much as half of
the total lesion volume (39)—is at risk due to secondary glutamatergic
signaling, which is still active long after the primary stroke event
(7). One reason for the lack of success regarding antiglutamatergic
therapeutic strategies may be that secondary effects of excessive
glutamate are difficult to target without dampening necessary basal
glutamate activity. Blocking glutamatergic signaling did inhibit not
only glutamate-related hyperexcitability but also baseline glutama-
tergic transmission known to have neuroprotective effects (8).
Accordingly, graded intervention into glutamatergic transmission—
inhibiting prolonged post-stroke glutamatergic hyperexcitability and
subsequent cell death but maintaining physiological glutamatergic
transmission and thereby survival programs—may be an effective
paradigm in stroke therapy (11). We have previously shown (i) that
glutamate stimulation acting via ionotropic receptors induces
up-regulation of the LPA-synthesizing molecule ATX, which is
Fig. 4. Loss-of-function SNP in the Prg-1 gene (Prg-1R346T) increased infarct size after MCAO. (A) Original traces of sEPSCs from Prg-1R346T neurons and Prg-1WT (WT)
neurons. (B) Comparison of sEPSC frequencies and amplitudes of Prg-1R346T neurons compared to neurons from WT littermates (n = 10 WT neurons from five mice and
n = 10 Prg-1R346T neurons from five mice, two-tailed t test). (C) Images of corresponding brain slices of WT and Prg-1R346T animals showing infarct size 24 hours after
MCAO. Scale bar, 500 m. (D) Quantification of infarct volume 24 hours after MCAO in Prg-1R346T mice compared to Prg-1WT littermates (n = 12 WT and n = 11 Prg-1R346T,
two-tailed t test). (E) Behavioral outcome in Prg-1R346T animals as measured by mNSS compared to WT littermates (n = 12 WT and n = 12 Prg-1R346T, Bayesian analysis).
(F) Activated caspase-3 immunostaining in the penumbra in Prg-1R346T animals and in Prg-1WT littermates. Scale bars, 50 and 5 m (inset). (G) Quantification of activated
caspase-3–positive cells in Prg-1R346T animals compared to Prg-1WT littermates (n = 9 WT and n = 9 Prg-1R346T, two-tailed Mann-Whitney test). (H) Analysis of NfL amounts
in the CSF of Prg-1R346T animals compared to WT littermates (n = 6 WT and n = 8 Prg-1R346T, two-tailed t test). (I) Images of corresponding brain slices from Prg-1−/−/Lpa2−/−
and Prg-1WT/Lpa2WT littermates showing infarct size 72 hours after MCAO. Scale bar, 500 m. (J) Quantification of infarct sizes in Prg-1−/−/Lpa2−/− mice and Prg-1WT/Lpa2WT
littermates (n = 11 WT and n = 11 Prg-1−/−/Lpa2−/−, two-tailed Mann-Whitney test). (K) Behavioral outcome in Prg-1−/−/Lpa2−/− animals as measured by mNSS compared to
Prg-1WT/Lpa2WT littermates (n = 12 WT and n = 11 Prg-1−/−/Lpa2−/−, Bayesian analysis). Data are presented as means ± SEM (*P < 0.05 and **P < 0.01 or accuracy of **>90%
for Bayesian analysis).
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
7 of 12
Fig. 5. ATX inhibition de-
creases cortical hyperex-
cit ability after MCAO and is
associated with improved
stroke outcome. (A) Protocol
for ATX inhibition by PF8380,
for MRI assessment, and for
electrophysiological assessmen t.
Arrows indicate time points for
MRI imaging, electrophysiological
assessment (Ephys), or PF8380
application. (B) T2*-weighted
MRI images at 24 and 72 hours
after MCAO show penumbra
(yellow) and infarct (gray) zones
in Prg-1WT and Prg-1R346T ani-
mals with and without ATX
inhibition by PF8380. Penumbra
and infarct core were assessed
as described in (48) (see also
Supplementary Materials and
M ethods). Asterisk marks penum-
bra in the lower cortical layers
of the somatosensory cortex.
(C and D) Penumbra and inf arct
zones at (C) 24 and (D) 72 hou rs
after MCAO in Prg-1R346T ani-
mals and Prg-1WT littermates
with and without ATX inhibi-
tion (n = 6 WT, n = 5 WT PF8380,
n = 5 Prg-1R346T, and n = 5
Prg-1R346T PF8380, Bayesian
analysis). (E) mNSS in PF8380-
treated WT animals and in
Prg-1R346T mice compared to
vehicle-treated littermates (n = 13
WT, n = 18 WT PF8380, n = 16
Prg-1R346T, and n = 15 Prg-1R346T
PF8380, Bayesian analysis).
(F) Rotarod performance in
PF8380-treated Prg-1WT and
Prg-1R346T mice compared to
vehicle-treated littermates (n = 5
WT, n = 5 WT PF8380, n = 4
Prg-1R346T, and n = 5 Prg-1R346T
PF8380, Bayesian analysis).
(G) Left image shows a patched
layer 4 (L4) medium spiny
neuron (MSN) in the putative
penumbra in the somatosensor y
cortex [marked by an asterisk
in the MRI image in (B)] 24 hours
after MCAO. Scale bar, 25 m.
Examples of sEPSCs of MSNs under control conditions 24 hours after MCAO and 24 hours after MCAO with PF8380 treatment are shown on the right. (H) sEPSC frequencies
in L4 MSNs after MCAO (n = 8 controls and 5 MCAO from at least three animals per condition, Bayesian analysis). (I) sEPSCs in the same L4 MSNs 24 hours after MCAO
before and 10 min after PF8380 application. sEPSCs were normalized for better comparison to the sEPSC frequencies before application (n = 5 MCAO before and after
PF8380 application, Bayesian one sample t test). (J) Amplitudes of sEPSCs in L4 MSN from (C) and MCAO before and after PF8380 (n = 8 controls and 5 MCAO, Bayesian
analysis). (K) In vivo measurements of firing rates 24 hours after MCAO in L4 and L5 MSNs of the somatosensory cortex in the putative penumbra compared to control
littermates (n = 7 controls, n = 9 MCAO for 0 to 5 Hz, and n = 10 for 5 to 10 Hz, Bayesian analysis). (L and M) Analysis of (L) firing rates in L4 and L5 MSNs and (M) the number
of firing neurons in L4 MSNs firing in specific frequencies ranges in PF8380-treated Prg-1R346T mice and Prg-1WT littermates compared to vehicle-treated animals (n = 4
WT + vehicle, n = 5 WT + PF8380, n = 5 Prg-1R346T + vehicle, n = 5 Prg-1R346T + PF8380, Bayesian analysis). Data are presented as means ± SEM or as box with the median and
whiskers including all data points, which are depicted by dots (*P < 0.05 and **P < 0.01 or accuracy of *>85% or **>90% for Bayesian analysis).
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
8 of 12
released by cortical astrocytes in close proximity to excitatory
synapses and (ii) that ATX-dependent synaptic lipid signaling is in-
volved in pathological glutamatergic hyperexcitability but does not
affect basal glutamatergic transmission (15). Our data suggest that
up-regulation of ATX, which is secreted upon glutamatergic stimu-
lation, may be triggered by stroke-induced increased glutamatergic
transmission and produced by activated, GFAP-expressing astro-
cytes in the lesioned cortex; is associated with secondary neuronal
(excitotoxic) injury; and may play a crucial role in ischemic stroke
long-term outcome. In carriers of an SNP mutation that affects
synaptic LPA signaling (PRG-1R345T in humans and Prg-1R346T
homolog in mice), increased cortical excitability leads to worse
Fig. 6. Translational potential of synaptic lipid signaling in stroke
outcome in mice and in humans. (A) ATX inhibition protocol for differ-
ent PF8380 treatment strategies after MCAO. (B) Representative images of
brain sections from WT animals 72 hours after MCAO with or without
PF8380 treatment started at different time points after MCAO. Infarct area
is outlined by dashed line. Scale bar, 500 m. (C) Quantification of brain
infarct volumes at 72 hours after MCAO after ATX inhibition started at 30,
60, 90, or 180 min after MCAO (two-tailed t test). (D) Neurologic score
assessed during the first 72 hours after MCAO after PF8380 treatment
started at 30, 60, 90, or 180 min after MCAO (Bayesian analysis). (C and D)
Treatment 30 min after MCAO: n = 9 WT + vehicle and n = 10 WT + PF8380;
60 min after: n = 12 WT + vehicle and n = 11 WT + PF8380; 90 min after:
n = 12 WT + vehicle and n = 14 WT + PF8380; and 180 min after: n = 11 WT + vehicle and n = 11 WT + PF8380. (E and F) NIHSS assessed (E) at baseline (within 6 hours after
stroke, NIHSS6h) and (F) 24 hours after baseline (NIHSS24h) in patients expressing the PRG-1R345T human mutation and in age- and sex-matched control (Cstroke) patients
with stroke (n = 27 Cstroke and 14 PRG-1R345T, Bayesian analysis). n.s., not significant. (G) Change in NIHSS between baseline and 24 hours later (NIHSS6h–24h), a predictive
factor for the 90-day stroke outcome (32), in PRG-1R345T mutation carriers and in age- and sex-matched control patients after stroke (n = 25 control and 12 PRG-1R345T
patients, Bayesian analysis). Data are presented as means ± SEM (*P < 0.05, **P < 0.01, and ***P < 0.001 or accuracy of *>85% for Bayesian analysis).
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
9 of 12
stroke outcome, thus demonstrating that synaptic lipid signaling is
important in stroke pathology in both mice and humans. Moreover,
we found markedly increased ATX concentrations in the CSF of
patients during the first 2 weeks after stroke. Here, high CSF ATX
was negatively correlated with NIHSS improvement during the first
24hours after stroke (a recently reported predictor for 90-day stroke
outcome) (32) and positively correlated with higher penumbra
volume. Increased ATX expression was in line with higher LPA
production in the CSF of patients after stroke and with increased
concentrations of the LPA precursor LPC, suggesting a continuous
fueling of the dysregulated synaptic lipid signaling by high ATX
production after stroke. On the other hand, pharmacological ATX
inhibition after experimental stroke and genetically modified mouse
models designed to reduce LPA-related cortical excitability were
effective in ameliorating morphological and functional stroke out-
come. The importance of synaptic lipid-regulated cortical activity was
proven by cell type–specific deletion of ATX in reactive GFAP-
positive astrocytes in the lesioned cortex, which improved stroke
outcome.
Postlesional-reactive astrogliosis, a typical pathological feature
following brain lesions, is characterized by strong GFAP expression
and a coarse morphology of astrocytes. Whereas, in healthy condi-
tions, cortical astrocytes do not express GFAP, activated cortical
astrocytes strongly up-regulate GFAP, thereby delineating lesioned
brain areas (12). In addition to the described role of astrocytes, here,
we were able to show that reactive astrocytes play an active and yet
unknown role in stroke pathology. Our morphological findings of
increased ATX expression in the ischemic brain are in line with the
described ATX properties. ATX is shed from its expression site and
functions as a courier for the synthetized LPA, allowing for extra-
cellular transport (29). Last, ATX binds close to the LPA effector
site via its integrin-binding somatomedin B–like domain (28), acti-
vating specific LPA receptors. This supports the idea of local ATX
production and LPA synthesis after glutamatergic stimulation,
which enhances neuronal excitability in the threatened penumbra
and dynamically drives neuronal cell death.
ATX inhibition effectively reduces cortical hyperexcitability by
reducing synaptic LPA concentrations while not altering basal
glutamatergic activity as shown previously (15). Synaptic LPA
produced by ATX acts as an enhancer of glutamatergic transmission
in the nervous system via increased presynaptic glutamate release
probabilities (18,21). Our intervention inhibiting the accumulation
of LPA in a transient MCAO model proved efficient in rescuing the
jeopardized penumbra and in promoting functional recovery after
stroke. LPA amounts in the CSF were reduced by inhibition of the
LPA-synthesizing enzyme ATX in Prg-1WT mice after MCAO, re-
sulting in an improved overall outcome including penumbra rescue
and improvement of motor and sensory deficits.
To understand the role of LPA-related cortical excitability on
stroke outcome on a molecular level, we further analyzed a mouse
line (Prg-1R346T) expressing a recently reported human SNP in the
PRG-1 gene that affects cortical activity in ~5 million European and
American citizens. This single–amino acid mutation (arginine [R]
346 to threonine [T]) disrupts the regulatory function of PRG-1 on
synaptic LPA by preventing the transport of LPA from the synaptic
cleft into intracellular compartments, thus resulting in increased syn-
aptic LPA concentrations and subsequent cortical hyperexcitability
via increased presynaptic glutamate release (20). Prg-1R346T mice
exhibited larger infarct volumes than control littermates and
presented an overall unfavorable behavioral outcome after MCAO.
To further prove the role of synaptic LPA signaling and cortical
excitability in stroke outcome, we analyzed animals in which synap-
tic LPA signaling and PRG-1 function were both disrupted. These
mice (Prg-1−/−/Lpa2−/− mice) display normal cortical excitability
even in the presence of higher synaptic LPA concentrations due to
the deleted presynaptic LPA2 receptors (18). Here, we found that,
after MCAO, Prg-1−/−/Lpa2−/− mice did not display notable differ-
ences regarding their outcome when compared to their WT litter-
mates. These data strongly argue for a critical role of synaptic
LPA-induced cortical excitability mediated by the presynaptic
LPA2 receptor in stroke.
To further corroborate the ability of ATX inhibition to influence
stroke outcome, we compared Prg-1WT animals and Prg-1R346T mice
treated with the ATX inhibitor PF8380 using MRI 24 and 72hours
after MCAO. Although untreated Prg-1R346T mice displayed higher
penumbra and stroke volumes than their WT littermates, no change
was found when PF8380-treated Prg-1R346T mice were compared to
their similar PF8380-treated Prg-1WT littermates. This suggests that
ATX inhibition is able to completely reverse a preexisting increased
cortical excitability, leading to penumbra loss and excitotoxic cell
death. Together, our data clearly support the hypothesis of a detri-
mental role of phospholipid signaling and the protective role of
ATX inhibition in stroke outcome. Strong up-regulation of ATX in
the CSF of patients within the first 24hours after stroke and continuous
elevation of ATX in the CSF up to 14 days after stroke emphasizes
the concept of glutamate-induced ATX up-regulation and subse-
quent phospholipid-induced cortical hyperexcitability and excito-
toxicity in the lesioned brain. This is in line with previous data
showing critical involvement of ATX-LPA signaling in traumatic
brain injury (16,40,41).
Although our morphological data in the human cortex suggest a
similar expression of ATX at glutamatergic synapses as described in
mice (15), a slightly different neuron/astrocyte ratio in humans may
result in altered effect sizes of an ATX inhibitor when applied
in patients after stroke. Our analysis of patients carrying the
PRG-1R345T SNP, which specifically affects LPA accumulation at the
synaptic cleft, suggests an aggravation of the disease in these patients
and a deterioration of long-term stroke outcome, corroborating the role
of synaptic lipid signaling in stroke pathology in humans. The trans-
lational relevance is further supported by the clear correlation be-
tween high ATX concentrations in the CSF of patients after stroke
and reduced NIHSS improvement during the first 24hours after
admission, a recently described long-term outcome predictor (32).
These findings should be confirmed in future studies with larger
patient cohorts to assess the effect size of altered synaptic lipid
signaling in stroke. Although the translational potential of our
study is clear, our findings are limited by the number of CSF
samples available from patients with stroke and the absence of
longitudinal patient CSF measurements, which would, however,
be difficult to achieve due to ethical considerations. Furthermore,
whereas ATX inhibition was a successful intervention in our animal
stroke model, metabolic stability of an ATX inhibitor and its pene-
tration through the blood-brain barrier may present further chal-
lenges on the way to the clinic (42).
Patients receive treatments at delayed time intervals after stroke
and systemic thrombolysis is applied only in the first hours after the
infarct. With this in mind, we tested this approach to inhibit ATX and
thereby synaptic lipid-induced vicious penumbra hyperexcitability
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
10 of 12
in a clinically relevant manner. Here, we found that single PF8380
application even up to 3hours after the experimental stroke proce-
dure followed by subsequent daily treatment markedly reduced
stroke volume and improved stroke outcome in mice.
Together, our data suggest that glutamate-triggered astrocytic
ATX production leading to a dysregulated synaptic bioactive lipid
signaling, which lasts long after the initial stroke event, is an
important pathological factor in stroke outcome, thus presenting a
potential therapeutic option for stroke therapy. Moreover, in line
with experimental paradigms that aim to reduce excessive gluta-
mate signaling after stroke but maintain basal glutamatergic trans-
mission (11), which has beneficial effects, ATX inhibition is capable of
reducing pathological hyperexcitability while preserving basal neuro-
protective glutamatergic transmission even when applied clearly after
the initial stroke event. Although several experimental therapies for
stroke still need to be validated in humans, recent progress in bringing
ATX inhibitors to human use (43) may constitute an innovative op-
tion to salvage the critically jeopardized ischemic penumbra to pre-
vent an irrevocable widening of the ischemic brain damage.
MATERIALS AND METHODS
Study design
The objective of this study was to assess the role of synaptic lipid
signaling in the regulation of glutamatergic overexcitation during
stroke and to explore the possibility of disrupting the vicious circle
of overexcitation that leads to brain demise, thereby improving
stroke outcome by inhibiting autotaxin, a synaptic lipid-synthetizing
molecule. For specific analysis of the underlying molecular mecha-
nisms, cell type–specific ATX deletion (Atxfl/fl Gfap-Cre+) and phar-
macological ATX inhibition, applied at different time points during
and after MCAO, were used. In addition to histological (infarct
volume) and MRI assessment of penumbra and infarct core after
MCAO, ATX expression in the murine brain (Western blot) and
LPA concentrations in the CSF (mass spectrometry) after stroke
and after ATX inhibition were measured as indicated. Using a
mouse line expressing a human mutation (Prg-1R346T) leading to
higher cortical excitability, we could confirm (including via single-
cell and invivo electrophysiology) the general hypothesis of the
detrimental role of synaptic lipid-induced cortical hyperexcitability
in stroke outcome and corroborate ATX inhibition as a critical
target in improving stroke outcome. To support the translational
relevance of our results, we assessed ATX concentrations in the CSF
and in the blood plasma of patients with stroke and correlated them
with NIHSS development and MRI data. NIHSS improvement
during the first 24hours, a predictor for 90-day stroke outcome, con-
firmed adverse stroke outcome in patients carrying the PRG-1R345T
mutation. Further details of the study are provided in the corre-
sponding sections of the Supplementary Materials. The detailed
number of biological replicates is reported in each figure legend.
Animal experiments were planned using G*Power software (Heinrich-
Heine-Universität Düsseldorf, version 3.1.9.6) to determine sample
size (setting  at 0.05 and power at 0.8) and were approved by the
local ethical committee (Landesuntersuchungsamt Rheinland-Pfalz
G14-1-038 and G19-1-067). They were conducted in accordance
with the national laws for the use of animals in research and followed
the ARRIVE (Animal Research: Reporting of in Vivo Experiments)
guidelines. Adult (8- to 13-week-old) male mice were used for all
MCAO experiments; animals were randomly distributed into two groups
receiving either vehicle or the ATX inhibitor PF8380 (vehicle diluted).
Infarct size assessment was performed by a blinded investigator. For
detailed information, see Supplementary Materials and Methods. End
points for the animal experiments were determined in accordance
with institutionally approved criteria.
The collection of human CSF and blood plasma samples was
approved by the local Ethics Committee of the Medical Faculty of
the Johannes Gutenberg University Mainz (approval 2020-15522)
and was performed following a written informed consent according
to the Declaration of Helsinki. Data from patients expressing the
PRG-1 SNP in a monoallelic variant (GC variant rs138327459;
PRG-1R345T) were collected in the Genetics of Early Neurological
InStability after Ischemic Stroke study as described in detail in (32).
Patients expressing this SNP were age- and sex-matched, when
possible, to two control patients with stroke (GG).
Statistical analysis
Experimental data were analyzed using GraphPad Prism 6 and are
presented as means± SEM, unless stated otherwise. All groups of
data were first assessed for Gaussian distribution (using the
Kolmogorov-Smirnov test) and for outliers; comparison between
two groups was performed using a two-tailed unpaired t test for nor-
mally distributed data or a Mann-Whitney U test for nonparametric
data. Asterisks indicate significance with *P<0.05, **P<0.01,
and ***P<0.001 for all datasets.
For behavioral parameters, electrophysiology, LPA measure-
ments, and human patient data, we did an outlier analysis based on
the probability distributions of the measured parameters for each
group, which were plotted as a histogram and interpolated with a
kernel (window) fitting the distribution. The number of bins to
assess the distribution was selected on the basis of the number of
data points in the analyses for them to be comparable over the
different group comparisons. We set a 95% threshold to determine
the outlier in the analyses. For data with an unequal distribution
and/or unequal sample sizes, we used Bayesian posterior distribu-
tion analyses to identify the difference between the two groups. The
Bayesian posterior distribution was performed in R version 3.6.2
(www.R-project.org) with the open-source toolbox BEST (www.
indiana.edu/~kruschke/BEST). This analysis provides complete
distributions of credible values for group means and their differences
(44). This type of analysis was performed on the basis of published
data for classification and effect size estimation (45). To evaluate the
incremental value of the longitudinal measurement with several
time points, we used a two-step procedure. First, we performed a
cluster analysis by grouping all possible combinations of time points
to identify pairs with an area under the curve of>0.5. Cumulative
sums were estimated among all time points by normalizing each
parameter to the mean value before summation. Cumulative sums
were built for the combination of three or four available time points
for each behavioral parameter. As a second step, we built the
composite score by estimating the error for those combinations
that survived the first step and by assigning the weights based on the
least error. This type of analysis was performed on the basis of
previous work by Engel et al. (46). Last, the composite score for
each analyzed behavioral parameter was then tested for group
differences using the above-mentioned Bayesian posterior distribu-
tion analyses. Bayesian posterior analysis was performed using the
Markov chain Monte Carlo (MCMC) approach to compute the
Bayes factor for the choice of priors and the default MCMC sample
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
11 of 12
size of 100,000. Accuracy ranges for the differences of the means
≥85% and an effect size of ≥85% were considered significant and
were labeled with “*” (comparable to P < 0.05), and differences of
the means ≥90% and an effect size of ≥90% were considered highly
significant and were labeled with “**” (comparable to P < 0.01) (47).
SUPPLEMENTARY MATERIALS
www.science.org/doi/10.1126/scitranslmed.abk0135
Materials and Methods
Fig. S1
Table S1
Data file S1
MDAR Reproducibility Checklist
References (48–65)
View/request a protocol for this paper from Bio-protocol.
REFERENCES AND NOTES
1. P. B. Gorelick, The global burden of stroke: Persistent and disabling. Lancet Neurol. 18,
417–418 (2019).
2. E. H. Lo, T. Dalkara, M. A. Moskowitz, Mechanisms, challenges and opportunities in stroke.
Nat. Rev. Neurosci. 4, 399–414 (2003).
3. E. H. Lo, A new penumbra: Transitioning from injury into repair after stroke. Nat. Med. 14,
497–500 (2008).
4. J. C. Baron, Protecting the ischaemic penumbra as an adjunct to thrombectomy for acute
stroke. Nat. Rev. Neurol. 14, 325–337 (2018).
5. L. Catanese, J. Tarsia, M. Fisher, Acute ischemic stroke therapy overview. Circ. Res. 120,
541–558 (2017).
6. P. Lipton, Ischemic cell death in brain neurons. Physiol. Rev. 79, 1431–1568 (1999).
7. M. Fricker, A. M. Tolkovsky, V. Borutaite, M. Coleman, G. C. Brown, Neuronal cell death.
Physiol. Rev. 98, 813–880 (2018).
8. M. A. Moskowitz, E. H. Lo, C. Iadecola, The science of stroke: Mechanisms in search
of treatments. Neuron 67, 181–198 (2010).
9. D. W. Choi, Excitotoxicity: Still hammering the ischemic brain in 2020. Front. Neurosci. 14,
579953 (2020).
10. Y. Ge, W. Chen, P. Axerio-Cilies, Y. T. Wang, NMDARs in cell survival and death:
Implications in stroke pathogenesis and treatment. Trends Mol. Med. 26, 533–551 (2020).
11. J. Yan, C. P. Bengtson, B. Buchthal, A. M. Hagenston, H. Bading, Coupling of NMDA
receptors and TRPM4 guides discovery of unconventional neuroprotectants. Science 370,
eaay3302 (2020).
12. M. Pekny, M. Pekna, A. Messing, C. Steinhauser, J. M. Lee, V. Parpura, E. M. Hol,
M. V. Sofroniew, A. Verkhratsky, Astrocytes: A central element in neurological diseases.
Acta Neuropathol. 131, 323–345 (2016).
13. M. Ito, K. Komai, S. Mise-Omata, M. Iizuka-Koga, Y. Noguchi, T. Kondo, R. Sakai, K. Matsuo,
T. Nakayama, O. Yoshie, H. Nakatsukasa, S. Chikuma, T. Shichita, A. Yoshimura, Brain
regulatory T cells suppress astrogliosis and potentiate neurological recovery. Nature 565,
246–250 (2019).
14. J. Yang, M. D. C. Vitery, J. Chen, J. Osei-Owusu, J. Chu, Z. Qiu, Glutamate-releasing SWELL1
channel in astrocytes modulates synaptic transmission and promotes brain damage
in stroke. Neuron 102, 813–827.e6 (2019).
15. C. Thalman, G. Horta, L. Qiao, H. Endle, I. Tegeder, H. Cheng, G. Laube, T. Sigrudsson,
M. J. Hauser, S. Tenzer, U. Distler, J. Aoki, A. J. Morris, G. Geisslinger, J. Roper, S. Kirischuk,
H. J. Luhmann, K. Radyushkin, R. Nitsch, J. Vogt, Synaptic phospholipids as a new target
for cortical hyperexcitability and E/I balance in psychiatric disorders. Mol. Psychiatry 23,
1699–1710 (2018).
16. N. E. Savaskan, L. Rocha, M. R. Kotter, A. Baer, G. Lubec, L. A. van Meeteren, Y. Kishi, J. Aoki,
W. H. Moolenaar, R. Nitsch, A. U. Bräuer, Autotaxin (NPP-2) in the brain: Cell type-specific
expression and regulation during development and after neurotrauma. Cell. Mol. Life Sci.
64, 230–243 (2007).
17. Y. C. Yung, N. C. Stoddard, H. Mirendil, J. Chun, Lysophosphatidic acid signaling in the
nervous system. Neuron 85, 669–682 (2015).
18. T. Trimbuch, P. Beed, J. Vogt, S. Schuchmann, N. Maier, M. Kintscher, J. Breustedt,
M. Schuelke, N. Streu, O. Kieselmann, I. Brunk, G. Laube, U. Strauss, A. Battefeld, H. Wende,
C. Birchmeier, S. Wiese, M. Sendtner, H. Kawabe, M. Kishimoto-Suga, N. Brose,
J. Baumgart, B. Geist, J. Aoki, N. E. Savaskan, A. U. Bräuer, J. Chun, O. Ninnemann,
D. Schmitz, R. Nitsch, Synaptic PRG-1 modulates excitatory transmission via lipid
phosphate-mediated signaling. Cell 138, 1222–1235 (2009).
19. H. Tokumitsu, N. Hatano, M. Tsuchiya, S. Yurimoto, T. Fujimoto, N. Ohara, R. Kobayashi,
H. Sakagami, Identification and characterization of PRG-1 as a neuronal calmodulin-
binding protein. Biochem. J. 431, 81–91 (2010).
20. J. Vogt, J. W. Yang, A. Mobascher, J. Cheng, Y. Li, X. Liu, J. Baumgart, C. Thalman,
S. Kirischuk, P. Unichenko, G. Horta, K. Radyushkin, A. Stroh, S. Richers, N. Sahragard,
U. Distler, S. Tenzer, L. Qiao, K. Lieb, O. Tüscher, H. Binder, N. Ferreiros, I. Tegeder,
A. J. Morris, S. Gropa, P. Nürnberg, M. R. Toliat, G. Winterer, H. J. Luhmann, J. Huai,
R. Nitsch, Molecular cause and functional impact of altered synaptic lipid signaling
due to a prg-1 gene SNP. EMBO Mol. Med. 8, 25–38 (2016).
21. P. Unichenko, S. Kirischuk, J. W. Yang, J. Baumgart, T. Roskoden, P. Schneider, A. Sommer,
G. Horta, K. Radyushkin, R. Nitsch, J. Vogt, H. J. Luhmann, Plasticity-related gene 1 affects
mouse barrel cortex function via strengthening of glutamatergic thalamocortical
transmission. Cereb. Cortex 26, 3260–3272 (2016).
22. J. Vogt, S. Kirischuk, P. Unichenko, L. Schluter, A. Pelosi, H. Endle, J. W. Yang,
N. Schmarowski, J. Cheng, C. Thalman, U. Strauss, A. Prokudin, B. S. Bharati, J. Aoki,
J. Chun, B. Lutz, H. J. Luhmann, R. Nitsch, Synaptic phospholipid signaling modulates
axon outgrowth via glutamate-dependent Ca2+-mediated molecular pathways. Cereb.
Cortex 27, 131–145 (2017).
23. P. Schneider, S. Petzold, A. Sommer, R. Nitsch, H. Schwegler, J. Vogt, T. Roskoden, Altered
synaptic phospholipid signaling in PRG-1 deficient mice induces exploratory behavior
and motor hyperactivity resembling psychiatric disorders. Behav. Brain Res. 336, 1–7
(2018).
24. C. Fischer, H. Endle, L. Schumann, A. Wilken-Schmitz, J. Kaiser, S. Gerber, C. F. Vogelaar,
M. H. H. Schmidt, R. Nitsch, I. Snodgrass, D. Thomas, J. Vogt, I. Tegeder, Prevention
of age-associated neuronal hyperexcitability with improved learning and attention upon
knockout or antagonism of LPAR2. Cell. Mol. Life Sci. 78, 1029–1050 (2021).
25. P. E. Schauwecker, Microarray-assisted fine mapping of quantitative trait loci on
chromosome 15 for susceptibility to seizure-induced cell death in mice. Eur. J. Neurosci.
38, 3679–3690 (2013).
26. K. L. Adams, V. Gallo, The diversity and disparity of the glial scar. Nat. Neurosci. 21, 9–15
(2018).
27. Z. Liu, M. Chopp, Astrocytes, therapeutic targets for neuroprotection and
neurorestoration in ischemic stroke. Prog. Neurobiol. 144, 103–120 (2016).
28. J. Hausmann, S. Kamtekar, E. Christodoulou, J. E. Day, T. Wu, Z. Fulkerson, H. M. Albers,
L. A. van Meeteren, A. J. Houben, L. van Zeijl, S. Jansen, M. Andries, T. Hall, L. E. Pegg,
T. E. Benson, M. Kasiem, K. Harlos, C. W. Kooi, S. S. Smyth, H. Ovaa, M. Bollen, A. J. Morris,
W. H. Moolenaar, A. Perrakis, Structural basis of substrate discrimination and integrin
binding by autotaxin. Nat. Struct. Mol. Biol. 18, 198–204 (2011).
29. S. A. Jethwa, E. J. Leah, Q. Zhang, N. A. Bright, D. Oxley, M. D. Bootman, S. A. Rudge,
M. J. Wakelam, Exosomes bind to autotaxin and act as a physiological delivery
mechanism to stimulate LPA receptor signalling in cells. J. Cell Sci. 129, 3948–3957 (2016).
30. F. Salgado-Polo, A. Fish, M. T. Matsoukas, T. Heidebrecht, W. J. Keune, A. Perrakis,
Lysophosphatidic acid produced by autotaxin acts as an allosteric modulator of its
catalytic efficiency. J. Biol. Chem. 293, 14312–14327 (2018).
31. L. A. van Meeteren, P. Ruurs, C. Stortelers, P. Bouwman, M. A. van Rooijen, J. P. Pradère,
T. R. Pettit, M. J. Wakelam, J. S. Saulnier-Blache, C. L. Mummery, W. H. Moolenaar,
J. Jonkers, Autotaxin, a secreted lysophospholipase D, is essential for blood vessel
formation during development. Mol. Cell. Biol. 26, 5015–5022 (2006).
32. L. Heitsch, L. Ibanez, C. Carrera, M. M. Binkley, D. Strbian, T. Tatlisumak, A. Bustamante,
M. Ribo, C. Molina, A. Davalos, E. Lopez-Cancio, L. Munoz-Narbona, C. Soriano-Tarraga,
E. Giralt-Steinhauer, V. Obach, A. Slowik, J. Pera, K. Lapicka-Bodzioch, J. Derbisz,
T. Sobrino, J. Castillo, F. Campos, E. Rodriguez-Castro, S. Arias-Rivas, T. Segura,
G. Serrano-Heras, C. Vives-Bauza, R. Diaz-Navarro, S. Tur, C. Jimenez, J. Marti-Fabregas,
R. Delgado-Mederos, J. Arenillas, J. Krupinski, N. Cullell, N. P. Torres-Aguila, E. Muino,
J. Carcel-Marquez, F. Moniche, J. A. Cabezas, A. L. Ford, R. Dhar, J. Roquer, P. Khatri,
J. Jimenez-Conde, I. Fernandez-Cadenas, J. Montaner, J. Rosand, C. Cruchaga, J. M. Lee;
International Stroke Genetics Consortium, Early neurological change after ischemic
stroke is associated with 90-day outcome. Stroke 52, 132–141 (2021).
33. D. Castagna, D. C. Budd, S. J. Macdonald, C. Jamieson, A. J. Watson, Development
of autotaxin inhibitors: An overview of the patent and primary literature. J. Med. Chem.
59, 5604–5621 (2016).
34. J. Gierse, A. Thorarensen, K. Beltey, E. Bradshaw-Pierce, L. Cortes-Burgos, T. Hall,
A. Johnston, M. Murphy, O. Nemirovskiy, S. Ogawa, L. Pegg, M. Pelc, M. Prinsen,
M. Schnute, J. Wendling, S. Wene, R. Weinberg, A. Wittwer, B. Zweifel, J. Masferrer, A
novel autotaxin inhibitor reduces lysophosphatidic acid levels in plasma and the site
of inflammation. J. Pharmacol. Exp. Ther. 334, 310–317 (2010).
35. S. Tiedt, M. Duering, C. Barro, A. G. Kaya, J. Boeck, F. J. Bode, M. Klein, F. Dorn, B. Gesierich,
L. Kellert, B. Ertl-Wagner, M. W. Goertler, G. C. Petzold, J. Kuhle, F. A. Wollenweber,
N. Peters, M. Dichgans, Serum neurofilament light. Neurology 91, e1338–e1347 (2018).
36. T. Uphaus, S. Bittner, S. Gröschel, F. Steffen, M. Muthuraman, K. Wasser, M. Weber-Krüger,
F. Zipp, R. Wachter, K. Gröschel, NfL (neurofilament light chain) levels as a predictive
marker for long-term outcome after ischemic stroke. Stroke 50, 3077–3084 (2019).
37. T. F. Gendron, M. K. Badi, M. G. Heckman, K. R. Jansen-West, G. K. Vilanilam, P. W. Johnson,
A. R. Burch, R. L. Walton, O. A. Ross, T. G. Brott, T. M. Miller, J. D. Berry, K. A. Nicholson,
Downloaded from https://www.science.org at Yale University on April 24, 2022

Bitar et al., Sci. Transl. Med. 14, eabk0135 (2022) 20 April 2022
SCIENCE TRANSLATIONAL MEDICINE | RESEARCH ARTICLE
12 of 12
Z. K. Wszolek, B. E. Oskarsson, K. N. Sheth, L. H. Sansing, G. J. Falcone, B. L. Cucchiara,
J. F. Meschia, L. Petrucelli, Plasma neurofilament light predicts mortality in patients
with stroke. Sci. Transl. Med. 12, eaay1913 (2020).
38. G. Buzsaki, X. J. Wang, Mechanisms of gamma oscillations. Annu. Rev. Neurosci. 35,
203–225 (2012).
39. B. R. Broughton, D. C. Reutens, C. G. Sobey, Apoptotic mechanisms after cerebral
ischemia. Stroke 40, e331–e339 (2009).
40. P. J. Crack, M. Zhang, M. C. Morganti-Kossmann, A. J. Morris, J. M. Wojciak, J. K. Fleming,
I. Karve, D. Wright, M. Sashindranath, Y. Goldshmit, A. Conquest, M. Daglas,
L. A. Johnston, R. L. Medcalf, R. A. Sabbadini, A. Pebay, Anti-lysophosphatidic acid
antibodies improve traumatic brain injury outcomes. J. Neuroinflammation 11, 37 (2014).
41. T. Frugier, D. Crombie, A. Conquest, F. Tjhong, C. Taylor, T. Kulkarni, C. McLean, A. Pebay,
Modulation of LPA receptor expression in the human brain following neurotrauma.
Cell. Mol. Neurobiol. 31, 569–577 (2011).
42. Y. Jia, Y. Li, X.-D. Xu, Y. Tian, H. Shang, Design and development of autotaxin inhibitors.
Pharmaceuticals 14, 1203 (2021).
43. Z. Tan, H. Lei, M. Guo, Y. Chen, X. Zhai, An updated patent review of autotaxin inhibitors
(2017-present). Expert Opin. Ther. Pat. 31, 421–434 (2021).
44. J. K. Kruschke, Bayesian estimation supersedes the t test. J. Exp. Psychol. Gen. 142,
573–603 (2013).
45. L. Michels, M. Muthuraman, A. R. Anwar, S. Kollias, S. E. Leh, F. Riese, P. G. Unschuld,
M. Siniatchkin, A. F. Gietl, C. Hock, Changes of functional and directed resting-state
connectivity are associated with neuronal oscillations, ApoE genotype and amyloid
deposition in mild cognitive impairment. Front. Aging Neurosci. 9, 304 (2017).
46. S. Engel, M. Friedrich, M. Muthuraman, F. Steffen, A. Poplawski, S. Groppa, S. Bittner,
F. Zipp, F. Luessi, Intrathecal B-cell accumulation and axonal damage distinguish
MRI-based benign from aggressive onset in MS. Neurology 6, e595 (2019).
47. V. E. Johnson, Revised standards for statistical evidence. Proc. Natl. Acad. Sci. U.S.A. 110,
19313–19317 (2013).
48. R. Bhattacharjee, R. K. Gupta, B. Das, V. K. Dixit, P. Gupta, A. Singh, Penumbra
quantification from MR SWI-DWI mismatch and its comparison with MR ASL PWI-DWI
mismatch in patients with acute ischemic stroke. NMR Biomed. 34, e4526 (2021).
49. T. Chiang, R. O. Messing, W. H. Chou, Mouse model of middle cerebral artery occlusion.
J. Vis. Exp. 2011, 2761 (2011).
50. K. L. Schaar, M. M. Brenneman, S. I. Savitz, Functional assessments in the rodent stroke
model. Exp. Transl. Stroke Med. 2, 13 (2010).
51. S.-Y. Park, S. Marasini, G.-H. Kim, T. Ku, C. Choi, M.-Y. Park, E.-H. Kim, Y.-D. Lee,
H. Suh-Kim, S.-S. Kim, A method for generating a mouse model of stroke: Evaluation
of parameters for blood flow, behavior, and survival [corrected]. Exp. Neurobiol. 23,
104–114 (2014).
52. S. Zausinger, E. Hungerhuber, A. Baethmann, H. Reulen, R. Schmid-Elsaesser, Neurological
impairment in rats after transient middle cerebral artery occlusion: A comparative study
under various treatment paradigms. Brain Res. 863, 94–105 (2000).
53. M. Tanaka, Y. Kishi, Y. Takanezawa, Y. Kakehi, J. Aoki, H. Arai, Prostatic acid phosphatase
degrades lysophosphatidic acid in seminal plasma. FEBS Lett. 571, 197–204 (2004).
54. X. Liu, J. Huai, H. Endle, L. Schlüter, W. Fan, Y. Li, S. Richers, H. Yurugi, K. Rajalingam, H. Ji,
H. Cheng, B. Rister, G. Horta, J. Baumgart, H. Berger, G. Laube, U. Schmitt, M. J. Schmeisser,
T. M. Boeckers, S. Tenzer, A. Vlachos, T. Deller, R. Nitsch, J. Vogt, PRG-1 regulates synaptic
plasticity via intracellular PP2A/1-integrin signaling. Dev. Cell 38, 275–290 (2016).
55. A. Brureau, V. Blanchard-Bregeon, C. Pech, S. Hamon, P. Chaillou, J.-C. Guillemot,
P. Barneoud, P. Bertrand, L. Pradier, T. Rooney, N. Schussler, NF-L in cerebrospinal fluid
and serum is a biomarker of neuronal damage in an inducible mouse model
of neurodegeneration. Neurobiol. Dis. 104, 73–84 (2017).
56. J. D. Rohrer, I. O. C. Woollacott, K. M. Dick, E. Brotherhood, E. Gordon, A. Fellows,
J. Toombs, R. Druyeh, M. J. Cardoso, S. Ourselin, J. M. Nicholas, N. Norgren, S. Mead,
U. Andreasson, K. Blennow, J. M. Schott, N. C. Fox, J. D. Warren, H. Zetterberg, Serum
neurofilament light chain protein is a measure of disease intensity in frontotemporal
dementia. Neurology 87, 1329–1336 (2016).
57. I. A. Mulder, A. Khmelinskii, O. Dzyubachyk, S. de Jong, M. J. H. Wermer, M. Hoehn,
B. P. F. Lelieveldt, A. M. J. M. van den Maagdenberg, MRI mouse brain data of ischemic
lesion after transient middle cerebral artery occlusion. Front. Neuroinform. 11, 51 (2017).
58. T. F. Chan, L. A. Vese, Active contours without edges. IEEE Trans. Image Process. 10,
266–277 (2001).
59. A. F. Frangi, W. J. Niessen, K. L. Vincken, M. A. Viergever, Multiscale vessel enhancement
filtering. Lect. Notes Comput. Sci 1496, 130–137 (1998).
60. F. J. Chaure, H. G. Rey, R. Quian Quiroga, A novel and fully automatic spike-sorting
implementation with variable number of features. J. Neurophysiol. 120, 1859–1871
(2018).
61. K. Dolan, H. C. Martens, P. R. Schuurman, L. J. Bour, Automatic noise-level detection
for extra-cellular micro-electrode recordings. Med. Biol. Eng. Comput. 47, 791–800 (2009).
62. N. Koirala, L. Serrano, S. Paschen, D. Falk, A. R. Anwar, P. Kuravi, G. Deuschl, S. Groppa,
M. Muthuraman, Mapping of subthalamic nucleus using microelectrode recordings
during deep brain stimulation. Sci. Rep. 10, 19241 (2020).
63. Z. Zhao, Y. Xu, An extremely simple method for extraction of lysophospholipids
and phospholipids from blood samples. J. Lipid Res. 51, 652–659 (2010).
64. M. Okudaira, A. Inoue, A. Shuto, K. Nakanaga, K. Kano, K. Makide, D. Saigusa, Y. Tomioka,
J. Aoki, Separation and quantification of 2-acyl-1-lysophospholipids and 1-acyl-2-
lysophospholipids in biological samples by LC-MS/MS. J. Lipid Res. 55, 2178–2192 (2014).
65. K. Kano, H. Matsumoto, A. Inoue, H. Yukiura, M. Kanai, J. Chun, S. Ishii, T. Shimizu, J. Aoki,
Molecular mechanism of lysophosphatidic acid-induced hypertensive response. Sci. Rep.
9, 2662 (2019).
Acknowledgments: We thank C. Ernest for proofreading and editing the manuscript,
C. Liefländer for technical support, D. Cleppien for MRI imaging, and M. Dichgans for advice on
stroke genetics. Funding: This work was supported by the German Research Foundation
(DFG) grant SFB 1080 (to R.N., F.Z., and J.V.), grant SFB 1451 (to R.N. and J.V.), grant SFB 1292
(to F.Z.), and grant SFB-TR-128 (to F.Z., S.B., S.G., and T.U.); DFG Excellence Strategy–EXC
2030–390661388 (to J.V.); ERC PoC “PsychAID” (to R.N.); ERC AdV “LiPsyD” (to R.N.); and Else
Kröner-Fresenius-Stiftung (EKFS) grant 2018_EKMS.21 (to T.U.). The funders of the study had
no role in the collection, analysis, or interpretation of data; in the writing of the report; or in
the decision to submit the paper for publication. Author contributions: R.N., F.Z., and J.V.
designed the study. L.B. and T.U. conducted experimental stroke and behavioral studies. C.T.,
M.D., H.E., and W.F. did the electrophysiological recordings. L.B., C.T., H.J., and F.S. performed
molecular biology experiments. M.M., N.K., and S.G. performed penumbra and infarct analyses
in MCAO animals. L.G. and F.K. performed analyses of LPA and LPC. L.I., L.H., C.C., and J.-M.L.
analyzed and provided stroke patient genetic and clinical data from the GENISIS study. M.M.,
N.K., L.B., and J.V. performed statistical analysis. All authors were involved in data
interpretation. J.V., L.B., C.T., S.B., R.N., and F.Z. wrote the paper. Competing interests: T.U. has
received honoraria from Merck Serono and advanceCOR and personal fees from Pfizer. S.G. has
received lecture fees from Abbott, AbbVie, Bial, Ipsen, and UCB and research funding from
BMBF, DFG, Abbott, Boston Scientific, Boehringer Foundation, MagVenture, Medtronic,
Precisis, and Innovationsfonds GBA. C.C. receives research support from Biogen, Eisai, Alector,
and Parabon NanoLabs and is also a member of the advisory board of Vivid Genetics, Halia
Therapeutics, and ADx Healthcare. J.-M.L. received research support from Biogen and personal
fees from Regenera. S.B. has received honoraria from Biogen Idec, Bristol Myers Squibb, Merck
Healthcare, Novartis, Roche, Sanofi Genzyme, and Teva and research funding from DFG, Hertie
Foundation, and the Hermann and Lilly Schilling Foundation. F.Z. has recently received
research grants and/or consultation funds from Biogen, Federal Ministry of Education and
Research (BMBF), Bristol Myers Squibb, Celgene, German Research Foundation (DFG), Janssen
Pharmaceuticals, Max Planck Society (MPG), Merck Serono, Novartis, Progressive MS Alliance
(PMSA), Roche, Sanofi Genzyme, and Sandoz. J.V. has received research support from the DFG
and Boehringer Foundation. All author disclosures are outside of the submitted work. J.V. and
R.N. hold a European patent for “LPA level reduction for treating central nervous system
disorders” (EP 3368035A1), which is not related to the present work. All other authors declare
that they have no competing interests. Data and materials availability: All data associated
with this study are present in the paper or the Supplementary Materials. Materials used in this
study will be available for the scientific community by contacting the corresponding authors
and, when appropriate, completion of a material transfer agreement. Analysis scripts are
available via public GitHub repositories (https://github.com/MuthuramanMuthuraman/
Strokeproject and https://github.com/HeEndle/MiniMonster).
Submitted 16 June 2021
Resubmitted 13 January 2022
Accepted 8 March 2022
Published 20 April 2022
10.1126/scitranslmed.abk0135
Downloaded from https://www.science.org at Yale University on April 24, 2022

Use of this article is subject to the Terms of service
Science Translational Medicine (ISSN ) is published by the American Association for the Advancement of Science. 1200 New York Avenue
NW, Washington, DC 20005. The title Science Translational Medicine is a registered trademark of AAAS.
Copyright © 2022 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works
Inhibition of the enzyme autotaxin reduces cortical excitability and ameliorates the
outcome in stroke
Lynn BitarTimo UphausCarine ThalmanMuthuraman MuthuramanLuzia GyrHaichao JiMicaela DominguesHeiko
EndleSergiu GroppaFalk SteffenNabin KoiralaWei FanLaura IbanezLaura HeitschCarlos CruchagaJin-Moo LeeFlorian
KlossStefan BittnerRobert NitschFrauke ZippJohannes Vogt
Sci. Transl. Med., 14 (641), eabk0135. • DOI: 10.1126/scitranslmed.abk0135
Reducing excitation surrounding stroke
After stroke, excessive glutamate release from damaged cells contributes to secondary injury in the stroke penumbra.
Preserving the penumbra is critical for improving stroke outcome; however, current pharmacological approaches
have not been successful. Here, Bitar et al. studied the mechanisms mediating excitotoxicity in the penumbra and
showed that autotaxin (ATX) was increased in astrocytes after stroke in mice, and ATX increase was also found
in the cerebrospinal fluid of patients after stroke. ATX increase mediated lysophosphatidic acid (LPA)–induced
hyperexcitability in rodents, and its inhibition ameliorated stroke outcome, suggesting that the ATX-LPA signaling might
be a potential target for treating stroke.
View the article online
https://www.science.org/doi/10.1126/scitranslmed.abk0135
Permissions
https://www.science.org/help/reprints-and-permissions
Downloaded from https://www.science.org at Yale University on April 24, 2022