D-amino acid oxidase controls motoneuron degeneration through D-serine.
ABSTRACT Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder involving an extensive loss of motoneurons. Aberrant excitability of motoneurons has been implicated in the pathogenesis of selective motoneuronal death in ALS. D-serine, an endogenous coagonist of N-methyl-D-aspartate receptors, exacerbates motoneuronal death and is increased both in patients with sporadic/familial ALS and in a G93A-SOD1 mouse model of ALS (mSOD1 mouse). More recently, a unique mutation in the D-amino acid oxidase (DAO) gene, encoding a D-serine degrading enzyme, was reported to be associated with classical familial ALS. However, whether DAO affects the motoneuronal phenotype and D-serine increase in ALS remains uncertain. Here, we show that genetic inactivation of DAO in mice reduces the number and size of lower motoneurons with axonal degeneration, and that suppressed DAO activity in reactive astrocytes in the reticulospinal tract, one of the major inputs to the lower motoneurons, predominantly contributes to the D-serine increase in the mSOD1 mouse. The DAO inactivity resulted from expressional down-regulation, which was reversed by inhibitors of a glutamate receptor and MEK, but not by those of inflammatory stimuli. Our findings provide evidence that DAO has a pivotal role in motoneuron degeneration through D-serine regulation and that inactivity of DAO is a common feature between the mSOD1 ALS mouse model and the mutant DAO-associated familial ALS. The therapeutic benefit of reducing D-serine or controlling DAO activity in ALS should be tested in future studies.
- SourceAvailable from: Ludo Van Den Bosch[show abstract] [hide abstract]
ABSTRACT: Unfortunately and despite all efforts, amyotrophic lateral sclerosis (ALS) remains an incurable neurodegenerative disorder characterized by the progressive and selective death of motor neurons. The cause of this process is mostly unknown, but evidence is available that excitotoxicity plays an important role. In this review, we will give an overview of the arguments in favor of the involvement of excitotoxicity in ALS. The most important one is that the only drug proven to slow the disease process in humans, riluzole, has anti-excitotoxic properties. Moreover, consumption of excitotoxins can give rise to selective motor neuron death, indicating that motor neurons are extremely sensitive to excessive stimulation of glutamate receptors. We will summarize the intrinsic properties of motor neurons that could render these cells particularly sensitive to excitotoxicity. Most of these characteristics relate to the way motor neurons handle Ca(2+), as they combine two exceptional characteristics: a low Ca(2+)-buffering capacity and a high number of Ca(2+)-permeable AMPA receptors. These properties most likely are essential to perform their normal function, but under pathological conditions they could become responsible for the selective death of motor neurons. In order to achieve this worst-case scenario, additional factors/mechanisms could be required. In 1 to 2% of the ALS patients, mutations in the SOD1 gene could shift the balance from normal motor neuron excitation to excitotoxicity by decreasing glutamate uptake in the surrounding astrocytes and/or by interfering with mitochondrial function. We will discuss point by point these different pathogenic mechanisms that could give rise to classical and/or slow excitotoxicity leading to selective motor neuron death.Biochimica et Biophysica Acta 01/2006; 1762(11-12):1068-82. · 4.66 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Although D amino acids are prominent in bacteria, they generally are thought not to occur in mammals. Recently, high levels of D-serine have been found in mammalian brain where it activates glutamate/N-methyl-D-aspartate receptors by interacting with the "glycine site" of the receptor. Because amino acid racemases are thought to be restricted to bacteria and insects, the origin of D-serine in mammals has been puzzling. We now report cloning and expression of serine racemase, an enzyme catalyzing the formation of D-serine from L-serine. Serine racemase is a protein representing an additional family of pyridoxal-5' phosphate-dependent enzymes in eukaryotes. The enzyme is enriched in rat brain where it occurs in glial cells that possess high levels of D-serine in vivo. Occurrence of serine racemase in the brain demonstrates the conservation of D-amino acid metabolism in mammals with implications for the regulation of N-methyl-D-aspartate neurotransmission through glia-neuronal interactions.Proceedings of the National Academy of Sciences 12/1999; 96(23):13409-14. · 9.74 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: D-serine occurs at high levels in the brain, where it is an endogenous coagonist at the "glycine site" of NMDA receptors. However, D-serine action has not been previously compared with that of endogenous glycine, and the relative importance of the two coagonists remains unclear. We now investigated the efficiencies of the two coagonists in mediating NMDA receptor neurotoxicity in organotypic hippocampal slices. Removal of endogenous D-serine from slices was achieved by pretreating the tissue with recombinant D-serine deaminase enzyme. This enzyme is several orders of magnitude more efficient than previous methods to remove D-serine. We report that complete removal of D-serine virtually abolished NMDA-elicited neurotoxicity but did not protect against kainate. Although levels of glycine were 10-fold higher than D-serine, endogenous glycine was ineffective in mediating NMDA receptor neurotoxicity. The effect of endogenous glycine could be observed only after simultaneous removal of endogenous D-serine and blockage of the glycine transporter GlyT1. Our data indicate that D-serine is the dominant coagonist for NMDA receptor-elicited neurotoxicity, mediating all cell death elicited by NMDA in organotypic slices. The results suggest an essential role for this unusual D-amino acid, with implications for the mechanism of neuronal death in the nervous system.Journal of Neuroscience 11/2005; 25(41):9413-7. · 6.91 Impact Factor
D-Amino acid oxidase controls motoneuron
degeneration through D-serine
Jumpei Sasabea,1, Yurika Miyoshib, Masataka Suzukia, Masashi Mitac, Ryuichi Konnod, Masaaki Matsuokae,
Kenji Hamaseb, and Sadakazu Aisoa
aDepartment of Anatomy, Keio University School of Medicine, Tokyo 160-8582, Japan;bGraduate School of Pharmaceutical Sciences, Kyushu University,
Fukuoka 812-8582, Japan;cInnovative Science Research and Development Center, Shiseido Co., Ltd., Yokohama 236-8643, Japan;dCenter for Medical Science,
International University of Health and Welfare, Tochigi 324-8501, Japan; andeDepartment of Pharmacology, Tokyo Medical University, Tokyo 160-8402,
Edited by Solomon H. Snyder, The Johns Hopkins University School of Medicine, Baltimore, MD, and approved December 2, 2011 (received for review
September 6, 2011)
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative
disorder involving an extensive loss of motoneurons. Aberrant
excitability of motoneurons has been implicated in the pathogen-
esis of selective motoneuronal death in ALS. D-Serine, an endoge-
nous coagonist of N-methyl-D-aspartate receptors, exacerbates moto-
neuronal death and is increased both in patients with sporadic/
familial ALS and in a G93A-SOD1 mouse model of ALS (mSOD1
mouse). More recently, a unique mutation in the D-amino acid
oxidase (DAO) gene, encoding a D-serine degrading enzyme, was
reported to be associated with classical familial ALS. However,
whether DAO affects the motoneuronal phenotype and D-serine
increase in ALS remains uncertain. Here, we show that genetic
inactivation of DAO in mice reduces the number and size of lower
motoneurons with axonal degeneration, and that suppressed DAO
activity in reactive astrocytes in the reticulospinal tract, one of the
major inputs to the lower motoneurons, predominantly contrib-
utes to the D-serine increase in the mSOD1 mouse. The DAO in-
activity resulted from expressional down-regulation, which was
reversed by inhibitors of a glutamate receptor and MEK, but not
by those of inflammatory stimuli. Our findings provide evidence
that DAO has a pivotal role in motoneuron degeneration through
D-serine regulation and that inactivity of DAO is a common fea-
ture between the mSOD1 ALS mouse model and the mutant
DAO-associated familial ALS. The therapeutic benefit of reducing
D-serine or controlling DAO activity in ALS should be tested in
excitotoxicity|motor neuron disease|neurodegeneration|
motoneurons in the spinal cord and brain leading to fatal pa-
ralysis. Approximately 90% of all cases are sporadic, and the
remaining cases are inherited. Of inherited cases, 20% are as-
sociated with mutations in superoxide dismutase 1 (SOD1), and
10% involves 43-kDa transactivation response DNA-binding
protein (TDP-43) and fused in sarcoma/translocated in lip-
osarcoma (FUS/TLS). Despite extensive studies of previously
identified ALS-causing genes, the mechanism underlying the
selective motoneuronal loss in ALS remains uncertain. Given
that the mechanism is at least, in part, common between sporadic
and familial ALS, identification of the common pathology is
a clue to conquering ALS. Among numerous etiological hy-
potheses, motoneuronal vulnerability to excitotoxicity is one of
the most intensely investigated targets for the treatment of ALS
because it is observed in both sporadic and familial ALS with SOD1
mutations (1, 2). For motoneurons, glutamate is the main excitatory
transmitter, and excessive motoneuron excitability by glutamate
through ionotropic glutamate receptors has been demonstrated.
The N-methyl-D-aspartate (NMDA) receptor (NMDAR) is
a subtype of the ionotropic glutamate receptors and exhibits
relatively higher permeability to the calcium ion (Ca2+) than
myotrophic lateral sclerosis (ALS) is a progressive neuro-
degenerative disorder characterized by selective loss of
non-NMDARs, which links it to a variety of physiological and
pathological processes. Unlike non-NMDARs, glutamate does
not activate the NMDARs unless a coagonist binding site is
occupied. D-Serine, an unusual D-amino acid found in mamma-
lian brain, is a physiological ligand of the coagonist site of the
NMDARs (3, 4); hence, it is pivotal in determining excitability of
glutamatergic neurons. D-Serine is endogenously converted from
L-serine by serine racemase (SRR) (5) and exists at a high level
in the forebrain, where it has a critical role in long-term poten-
tiation (6) and is required for memory formation (6). D-Serine is
also involved in NMDAR-mediated neurotoxicity, a process that
plays a pathophysiological role in stroke and neurodegenerative
diseases (7–9). We previously reported that D-serine is increased
in the spinal cord in both patients with sporadic/familial ALS and
in a G93A-SOD1 mouse model of ALS (mSOD1 mouse) (10).
D-Serine is progressively increased with expressional elevation of
SRR caused by glial activation (10). Intriguingly, a point muta-
tion that diminishes enzyme activity in D-amino acid oxidase
(DAO), a D-serine degrading enzyme, is associated with familial
ALS (11). However, whether DAO is related to D-serine increase
in ALS, or DAO inactivity is relevant to motoneuronal de-
generation in vivo, remains uncertain. In this study, we found
that DAO activity is strikingly suppressed in the reticulospinal
tract of mSOD1 mice, which plays a central role in D-serine in-
crease in mSOD1 mice, and that loss of DAO activity results in
DAO is highly expressed in mammalian CNS, liver, and kidney
and catalyzes the oxidative deamination of D-amino acids. ddY/
DAO−mice, found in outbred ddY mice (12), lack DAO activity
because of a natural point mutation (G181R) (13) and show
abnormal locomotor behavior related to enhanced NMDAR
function (14). Because hypofunction of the NMDAR and genetic
association of DAO have been implicated in schizophrenia, most
studies using ddY/DAO−mice have focused on a therapeutic
approach to schizophrenia through D-serine increase. No phe-
notypic analysis of spinal motoneurons in the mice, however, has
yet been conducted. To study whether inactivation of DAO
affects the motoneuronal phenotype in vivo, ddY/DAO−mice
were backcrossed with C57BL/6J and maintained as homo-
zygotes (B6DAO−/−mice). The
abnormal limb reflex characterized by retraction of hindlimbs
B6DAO−/−mice developed an
Author contributions: J.S. designed research; J.S., Y.M., M.S., and K.H. performed
research; M.M., R.K., and K.H. contributed new reagents/analytic tools; J.S., R.K., M.M.,
K.H., and S.A. analyzed data; and J.S. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| January 10, 2012
| vol. 109
| no. 2
toward the trunk when they were lifted up by their tails, whereas
C57BL/6J mice showed normal extension of the hindlimbs
(Fig. 1A). Motoneurons of the lumbar ventral horn inB6DAO−/−
mice exhibited morphology of degeneration (Fig. 1B), including
the reduction in number by 24% (Fig. 1C and Fig. S1 A–C) and
in size of their somata by 15% on average (Fig. 1D and Fig. S1D)
compared with age-matched C57BL/6J mice. In agedB6DAO−/−
mice, ubiquitin-positive aggregations were observed in the ven-
tral neurons (Fig. 1 E and F), and the ubiquitin-immunoreactive
smear was increased in the immunoblotting of spinal cords (Fig.
1G), but no expressional or localization changes of NMDARs
were detected except for localization shift of NR2A (Fig. S1 E
and F). Furthermore, inactivation of DAO triggers axonal de-
generation with muscle atrophy in aged mice (Fig. 1 H and I and
Fig. S1G). These findings demonstrate that DAO activity influ-
ences the motoneuronal phenotype.
To elucidate the pathophysiological significance of DAO in
ALS, we visualized DAO activity in cryostat sections by modi-
fying an enzyme histochemical (EHC) technique for vibratome
sections (15). EHC is based on oxidation of diaminobenzidine
with peroxidase by using D-proline as a substrate of DAO. In-
tense enzymatic reactivity was observed in the cerebellum, and
relatively weak reactivity was in the pons, medulla oblongata,
and spinal cord, whereas there was no signal at all in the fore-
brain region (Fig. S2 A and B). The specificity of the EHC was
verified by using L-proline as a substrate or tissue sections from
B6DAO−/−mice (Fig. S2C). In the spinal cord, DAO was mainly
distributed in the anterior columns, where motor axons descend
from upper neurons, and in lamina VIII and IX of ventral spinal
gray matter (Fig. 2A). This distribution pattern suggested a cor-
relation between DAO activity and the motor tract. Because
D-serine is essential for the neurotoxicity through NMDARs (7),
DAO has been thought to associate with diseases that involve
NMDAR malfunction. A R199W mutation in DAO found in
patients with familial ALS resulted in dramatic deactivation of
DAO (Fig. S3 A and B) and motoneuronal death in vitro (11). In
mSOD1 mice (16), the most studied ALS animal model, DAO
activity was significantly decreased compared with that in age-
matched wild-type mice (control mice) especially in the ventral
part of lumbar spinal cords (Fig. 2A). Quantification of DAO
activity in the whole tissue of the lumbar spinal cords from
mSOD1,B6DAO+/−, andB6DAO−/−mice showed that the DAO
activity in mSOD1 mice was suppressed to 57.9% of that in
control mice and was almost identical to that inB6DAO+/−mice
[F(3, 8)= 19.42, P = 0.0005] (Fig. 2B). This suppressed activity of
DAO resulted from marked reduction of DAO protein expres-
sion in Western blotting of spinal cords of mSOD1 mice at the
end stage, although significant reduction was not observed at the
onset (Fig. 2C and Fig. S2 D and E).
Using FITC-conjugated tyramide, which reacts sensitively with
peroxidase and enhances the sensitivity of EHC (Fig. S2 A and
B), we performed double staining of DAO activity and glial
fibrillary acidic protein (GFAP; an astrocytic marker), Iba1 (a
microglial marker), or nonphosphorylated neurofilament H
(npNFH, a marker for motoneuron) and found that DAO ac-
tivity was located in a portion of quiescent astrocytes (Fig. 2 D–F
and Fig. S2 F and G), but not in microglia (Fig. 2G) or somata/
axons of motoneurons (Fig. 2 H and I). Pyramidal neurons such
as Purkinje cells in the cerebellum and large neurons in the
brain-stem reticular formation did not exhibit any activity of
DAO (Fig. S2 H and I). Of note is that in mSOD1 mice, DAO
activity was severely suppressed in reactive astrocytes (Fig. 2 D
and E, Right). To further observe the DAO activity in the upper
motor tract in mSOD1 mice, sagittal sections were stained with
EHC enhanced with FITC-tyramide. DAO activity was strikingly
decreased in reactive astrocytes in brainstem reticular formation
of mSOD1 mice compared with control mice (Fig. 2 J and K),
whereas the activity in the cerebellum, pons, and the dorsal part
of the medulla oblongata (vestibular nuclei) remained un-
changed (Fig. 2J). Together with the finding in Fig. 2A, in
mSOD1 mice, DAO suppression was restricted in the retic-
ulospinal tract, which mainly regulates motoneuronal excitability
Number of lumbar MNs
MNs’ major axis (µm)D
Ubiquitin / Nissl / DAPI
Ubiquitin / Nissl / DAPI
Ubiquitin / Nissl / DAPI
(# / 100-µm thick spinal cord)
B6DAO−/−and wild-type control (control) mice (8 mo) were immunolabeled with a choline acetyltransferase (ChAT) antibody. Insets are enlarged ChAT-
positive neurons. (C) Number of motoneurons (MN) in lumbar spinal cord ofB6DAO−/−(n = 4) and control (n = 5), normalized to thickness of coronally sliced
sections. **P = 0.0028 (Student’s t test). (D) The longest diameter of MN’s soma (major axis) was measured (B6DAO−/−, n = 275; control, n = 272). ***P < 0.0001
(Student’s t test). (E and F) Motoneurons in the spinal cords ofB6DAO−/−and control mice (15 mo) were costained with fluorescent analysis of Nissl (red),
ubiquitin (green), and DAPI (blue). Arrows indicate ubiquitin aggregation-positive cells. Squared region was enlarged in F. (G) Western blotting of spinal cord
lysates fromB6DAO−/−and control (15 mo) was performed by using antibodies to ubiquitin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). (H)
Motor fibers in tibial nerve fromB6DAO−/−and control (15 mo) were visualized with immunofluorescent analysis of myelin basic protein (MBP) and npNFH. (I)
Gastrocnemius fromB6DAO−/−and control (15 mo) was stained with H&E. Data are plotted as mean ± SEM (C). (Scale bars: 100 μm.)
DAO inactivation triggers motoneuron degeneration. (A) Abnormal limb reflex inB6DAO−/−mice (8 mo). (B) Motoneurons in spinal cords (L2) of
| www.pnas.org/cgi/doi/10.1073/pnas.1114639109Sasabe et al.
in rodents (17). In contrast to the brainstem and spinal cord,
DAO activities in the tissues rich in DAO, the cerebellum and
kidney, were not altered in mSOD1 mice (Fig. S2J), suggesting
that the DAO inactivation in the reticulospinal tract was not
caused simply by expression of G93A-SOD1, but was affected by
Does the suppression of DAO activity affect D-serine levels?
The net amount of D-serine is controlled in balance with syn-
thesis and degradation. To evaluate each contribution, we
measured the amount of D-/L-serine with a 2D-HPLC system,
a highly selective and sensitive method (Fig. S4 and Materials and
Methods) (18, 19). The D-serine level in the spinal cord is neg-
atively correlated with the activity of DAO (Fig. 3A), suggesting
that DAO crucially determines the D-serine level in this region.
In the spinal cords of mSOD1 mice, D-serine was progressively
increased (Fig. 3B) (10), whereas no significant alteration of
D-serine was detected in those of control mice (Fig. 3B) or in the
cortices between mSOD1 and control mice (Fig. S4 F–H).
Although the L-serine level was also increased (Fig. 3C) consis-
tent with a previous report (20), the D-/L-serine ratio was still
higher in mSOD1 mice than in control mice (Fig. 3D). The
elevated D-/L-ratio implies three possibilities: D-serine synthesis
was increased, D-serine degradation was decreased, or both.
To evaluate the contribution of synthesis and degradation to the
D-serine increase, we generated mSOD1 mice lacking DAO ac-
tivity (B6DAO−/−/mSOD1 mice). D-Serine content in the lumbar
spinal cords ofB6DAO−/−/mSOD1 mice showed only a 12.9%
increase compared with that in
D-serine synthesis through L-serine increase and up-regulation
of SRR (Fig. S5) (10) exceeded its degradation by DAO
inmSOD1 mice, the level
mSOD1 mice should have been distinctly higher than in
B6DAO−/−mice. Therefore, this result demonstrates that DAO
inactivation is a dominant contributing factor for D-serine in-
crease in mSOD1 mice.
B6DAO−/−mice (Fig. 3E). If
DAO catalyzes oxidative deamination of neutral and basic
D-amino acids. Among D-amino acids, free D-serine and D-alanine
are good intrinsic substrates of DAO in mammalian tissues. To
study whether the inactivation of DAO affects D-amino acids
other than D-serine, we measured D-serine and D-alanine as well
as D-aspartate (18, 19, 21), as a control D-amino acid that is not
metabolized by DAO. Genetic inactivation of DAO markedly
increased the D-alanine level, whereas D-aspartate was not af-
fected at all (Fig. S6A). In contrast to D-serine, the levels of
D-alanine and D-aspartate in the spinal cord of mSOD1 mice did
not differ significantly from control (Fig. S6 A–C). We speculate
that the D-alanine level was not increased in mSOD1 mice be-
cause it is a better substrate of DAO than D-serine (22): D-serine
was not fully degraded in heterozygotes (B6DAO+/−) with half
the DAO activity of control mice (B6DAO+/+) (Fig. 2B and
Fig. S6A), whereas D-alanine inB6DAO+/−mice was kept as low
as that in
activation increased solely D-serine because DAO activity was
half, but not fully, inactivated in mSOD1 mice (Fig. 2B).
Is D-serine also increased in other neurodegenerative dis-
eases? We examined mouse models for ALS/frontotemporal
lobar degeneration (FTLD), sporadic Parkinson’s disease, and
familial Alzheimer’s disease: A315T-TDP-43 transgenic mice
(mTDP-43 mice), 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
(MPTP)-treated mice, and Tg2576 mice, respectively. D-Serine
was specifically increased in mSOD1 mice, but not in other
(Fig. S7A). Although the mTDP-43 mice exhibit a motoneuronal
phenotype as do the mSOD1 mice (23), DAO activity in mTDP-
43 mice was almost identical to that in control mice (Fig. S7 B
and C). Reactive astrocytes, also observed in mTDP-43 mice
(23), did not show DAO activity, but a portion of quiescent
astrocytes were supposed to retain the normal level of DAO
activity in spinal cords of mTDP-43 mice (Fig. S7D). These
B6DAO+/+mice (Fig. S6A). Therefore, DAO in-
DAO / GFAP / DAPI
DAO activity (nmol/min mg protein)
DAO / GFAP / DAPI
DAO / GFAP / DAPI
DAO / npNFH / DAPI
DAO / Iba1 / DAPI
DAO / npNFH / DAPI
DAO / GFAP / DAPIDAO / GFAP / DAPI
4 months5 months
DAO / GFAP / DAPI
DAO / GFAP / DAPI
(5 mo). The dashed line indicates a border between the gray and white matter. DAO activity was suppressed in ventral gray and white matter (arrows). (B)
DAO activity in spinal cords of control,B6DAO+/−,B6DAO−/−, and mSOD1 mice (5 mo) was assayed (n = 3 each). P = 0.0005 (one-way ANOVA), *P < 0.05, **P <
0.01, N.S., not significant (followed by Tukey’s multiple comparison test). Data are plotted as mean ± SEM. (C) Immunoblot of spinal cords of mSOD1 and
control mice (n = 3 each) at 4 mo (onset) and 5 mo (end stage) was performed with antibodies to DAO, GAPDH, and human SOD1. Arrowheads are DAO with
a size of 38 kDa. (D–K) Coronal sections of lumbar spinal cord (D, F, G, and I, ventral white matter; E and H, ventral gray matter, lamina IX) or sagittal ones of
hindbrain (J) and brainstem reticular formation (K) of control and mSOD1 mice (5 mo) were costained with DAO enzyme histochemistry (green), immuno-
fluorescent analysis of GFAP/Iba1/npNFH (red), and DAPI (blue). Squared areas in (E and K) are high magnification images. (J) Shown are joint images.
Arrowheads indicate diminished DAO activity in brainstem reticular formation. (Scale bars: D, E, and K, 100 μm; F–I, 25 μm.)
DAO activity is suppressed in the reticulospinal tract. (A) DAO enzyme histochemistry of lumbar spinal cord sections of mSOD1 and control mice
Sasabe et al. PNAS
| January 10, 2012
| vol. 109
| no. 2
results suggested that the pathological significance of DAO was
different between models of ALS and ALS/FTLD.
What decreased DAO protein level in mSOD1 mice? The
mRNA level of DAO was suppressed even at the preonset stage
of mSOD1 mice and continued to decrease over the course of
the disease (Fig. 4A). Because DAO activity was suppressed in
reactive astrocytes (Fig. 2 D, E, and K), we speculated that some
extracellular soluble factor, such as proinflammatory cytokines
or glutamate (24), influenced DAO expression in mSOD1 mice.
Incubation of primary cultured glia with spinal cord lysates of the
end-staged mSOD1 mouse reduced mRNA expression of DAO
in glia by 34% compared with that of the control mouse
(Fig. 4B). Among specific inhibitors for principal intracellular
signaling pathways [PD98059: for MAPK/extracellular signal-
regulated kinase (MEK) 1, SP600125: for c-Jun N-terminal ki-
nase, SB203580: for p38, wortmannin: for phosphatidylinositol-3
kinase, AG490: for Janus kinase 2, caffeic acid phenethyl ester
(CAPE): for NF-κB; ref. 25], only PD98059 recovered the down-
regulation of DAO caused by the mSOD1 lysate (Fig. 4B). In-
deed, the mSOD1-lysate treatment activated extracellular signal-
regulated kinase (ERK) 1/2 and cAMP responsive element
binding protein (CREB) for a longer duration than did the
control lysate treatment (Fig. 4C). In contrast, CAPE rather
decreased DAO expression (Fig. 4B), and proinflammatory
stimuli such as TNF-α or lipopolysaccharide elevated it without
affecting ERK activation (Fig. S8 A and B), suggesting proin-
flammatory factors potentially counteract repression of DAO
caused by ERK activation. However, ERK-mediated down-reg-
ulation of DAO caused by mSOD1 lysate was far more potent
than TNF-α–induced up-regulation (Fig. S8C). In support of
these findings, astrocytes with phosphorylated ERK1/2 were in-
creased and did not show DAO activity in the spinal ventral horn
of the mSOD1 mouse (Fig. 4 D and E). Furthermore, incubation
of primary cultured glia with PD98059 lowered the D-/L-serine
ratio in the cultured medium (Fig. 4F), suggesting that the MEK/
ERK signal is pivotal in DAO down-regulation.
In agreement with these results and the fact that stimulation
of NMDAR activates the MAPK pathway, glutamate reduced
DAO expression in primary cultured glia (Fig. S8D). MK-801,
a selective noncompetitive NMDAR antagonist, suppressed the
decrease of DAO expression in primary cultured glia treated
with the mSOD1 lysate, whereas 6,7-dinitroquinoxaline-2,3-
dione (DNQX) (a non-NMDAR antagonist) or a peptide an-
tagonist (WP9QY) for TNF receptor did not affect it (Fig.
4G). The glial expressions of NMDARs were observed mainly
in fibrous astrocytes in ventral white matter of control mice,
whereas NMDARs were expressed also in reactive astrocytes in
mSOD1 mice (Fig. S8 E–H), supporting our view that gluta-
matergic action mediated by the NMDAR-MEK/ERK pathway
seemed to be responsible for astrocytic DAO down-regulation in
Our data highlight the pathologic relevance of D-serine increase
derived from reduced DAO activity in ALS. Using inbred mice
lacking DAO activity, we provide evidence that loss of DAO
activity triggers motoneuron degeneration. Moreover, using
biochemical assays and sensitive quantitative methods to detect
DAO activity and D-amino acids in the tissues, we have shown
that (i) DAO activity is drastically suppressed in reactive astro-
cytes of the reticulospinal tract of mSOD1 mice, (ii) decreased
degradation due to DAO inactivation contributes dominantly to
the increase of D-serine, and (iii) the reduction of DAO activity
is caused by progressive repression of DAO gene expression that
mediates the NMDAR/ERK pathway.
In this study, we show that D-serine homeostasis is disrupted in
the reticulospinal tract in mSOD1 mice (Fig. S9 A and B). The
reticulospinal tract is a major descending motor pathway in
mammals and is assumed to be responsible for coordinated gross
movements primarily of proximal muscles, whereas the cortico-
spinal tract mediates fine movements, particularly of the hand
(26). The corticospinal tract is especially well developed in pri-
mates (26, 27), and its sclerosis observed in the lateral columns
of the spinal cord is a major characteristic of ALS. In nonprimate
mammals including rodents, however, there are no direct cor-
tico-motoneuronal connections. The reticulospinal tract, located
in the anterior columns, primarily relays cortical input to spinal
motoneurons (17, 28, 29). Therefore, D-serine degradation by
DAO in the reticulospinal tract is speculated to have a physio-
logical significance in controlling motoneuronal excitability in
rodents. Although, in humans, it is unknown whether DAO also
exists in the reticulospinal tract or the other descending motor
pathways, considering the R199W mutation in DAO is associ-
ated with familial ALS, D-serine homeostasis by DAO might also
be physiologically important in the excitability of motoneurons
Because of the hindbrain-shifted distribution of DAO activity,
physiological D-serine level in the spinal cord is strictly main-
tained at ≈1/50th of that in the forebrain (30). In the mSOD1
mice, reduced DAO activity in the reactive astrocytes caused by
repressions of its mRNA and protein expression contributes
dominantly to the progressive D-serine increase and raises D-
serine level by nearly three times at the end stage. Previously, we
showed that proinflammatory factors also contribute to D-serine
increase through SRR in activated microglia (10). Although such
D-Ser content (nmol/g)
D-Ser content (nmol/g)
D-Ser content (nmol/g)
L-Ser content (nmol/g)
(A and E) D-Serine levels in lumbar spinal cords of control (5 mo; A),B6DAO+/−
(5 mo; A),B6DAO−/−(5 mo; A and E), mSOD1 (end stage; E), andB6DAO−/−/
mSOD1 mice (end stage; E) (n = 3 each) were analyzed with 2D-HPLC. A
typical chromatogram of D-serine (arrows) in each group is shown on the
right. P < 0.0001 (one-way ANOVA). *P < 0.05, ***P < 0.001 (followed by
Tukey’s multiple comparison test). (B–D) Amounts of D-/L-serine (B and C)
and their ratio (D) in lumbar spinal cords of control and mSOD1 mice were
quantified with 2D-HPLC at 3 mo (preonset stage around day 80, mSOD1, n =
6; control, n = 5), 4 mo (onset stage around day 120, mSOD1, n = 7; control,
n = 7) and 5 mo (end stage around day 150, mSOD1, n = 7; control, n = 5) of
age. P < 0.0001 (one-way ANOVA). Data are plotted as the mean ± SEM.
DAO dominantly contributes to D-serine increase in the spinal cord.
| www.pnas.org/cgi/doi/10.1073/pnas.1114639109Sasabe et al.
proinflammatory stimuli induce compensatory DAO expression
in astrocytes (Fig. S8A), the potent activation of ERK1/2
diminishes the action and rather results in reduction of DAO
expression (Fig. S8C). In ALS, excessive glutamate remains in
the synaptic cleft due to loss of a glutamate transporter
(EAAT2) (31–33) and is thought to activate ERK1/2 in reactive
astrocytes through NMDARs but not AMPARs (Fig. 4G and
Fig. S9C). The notion that activity of NMDARs affects DAO
expression is supported by an in vivo study that systemic ad-
ministration of MK-801 significantly up-regulates mRNA ex-
pression of DAO in the rat hindbrain (34). Although research
into astrocytic NMDARs is still controversial, some findings
support the involvement of astrocytic NMDARs in rodent and
human astrocytic cell signaling (35, 36). Whether the NMDARs
observed in reactive astrocytes in mSOD1 mice (Fig. S8G) are
functional and have pathological significance awaits future
studies, but our findings shed light on the aberrant DAO regu-
lation in mSOD1 mice.
DAO inactivation in mice results in pathological reflex
(Fig. 1A), spinal motoneuron degeneration (Fig. 1 B–D), and
abnormal locomotor activity (14). Although more detailed
studies are required for full characterization of the motor
phenotype in theB6DAO−/−mice, theB6DAO−/−mice may bear
a pathophysiological resemblance to the classical adult onset
familial ALS associated with R199W-DAO, in which DAO ac-
tivity is at trace level. Given that reduced DAO activity in the
motoneuron input is crucial for motoneuron degeneration,
undegraded D-amino acids might reasonably be associated with
the mechanism of the degeneration. Moderate loss of DAO
activity increases exclusively D-serine in the spinal cord, whereas
its complete loss concomitantly affects D-alanine (Fig. S6A).
Both D-serine and D-alanine, albeit a lower affinity ligand than
D-serine, bind to a coagonist site in NR1 subunit of NMDARs
(37); therefore, the increase of both D-amino acids in the mo-
toneuron input leads to elevating the occupancy of the
coagonist site. Because coagonist binding is not only essential
for NMDAR activity, but also increases the receptor’s affinity
for glutamate (38) and decreases its desensitization (39), the
high occupancy of the site is assumed to elevate motoneuron
excitability. Thus, D-serine homeostasis in the spinal cord is
considered to be physiologically important in motoneuronal
excitability, and our findings give rise to the view that inactivity
of DAO is pathologically relevant to the vulnerability of
motoneurons to excitotoxicity in ALS.
The reduced DAO activity observed commonly in mSOD1
mice and familial ALS with R199W-DAO may not represent the
whole ALS pathology because DAO activity is not suppressed in
mTDP-43 mice. However, together with the findings in our
earlier study that D-serine is increased in both sporadic and fa-
milial ALS with A4T-SOD1 (10), it is likely that a group of
sporadic ALS and familial ALS with DAO and SOD1 mutations
share the pathology of increased D-serine.
In conclusion, this study provides a unique understanding of
the role of DAO and D-serine in motoneuron physiology as well
as in ALS pathophysiology as a putative enhancer of motoneu-
ron excitability. Our data also stress the potential use of regu-
lators of DAO activity or D-serine antagonists as a therapeutic
strategy in ALS.
Materials and Methods
Materials. For information about the materials used in this study, please refer
to SI Materials and Methods.
Animals. All experiments on animals were carried out in accordance with
institutional guidelines. The study protocol was approved by the Animal
Experiment Committee of KEIO University. Animals used in this study are
detailed in SI Materials and Methods.
Histological Analysis. Histological analysis is detailed in SI Materials
control lysatemSOD1 lysate
p-ERK / GFAP / DAPI
p-ERK / GFAP / DAPI
p-ERK / DAO / DAPI
1.0 10 10 0.9 6.9 8.4 0.8 1.8 4.1 0.7 1.0 0.8
1.0 10 9.2 0.6 2.3 1.8 0.7 1.0 2.0 0.4 0.7 0.3
spinal cords of control and mSOD1 mice at 3 (preonset), 4 (onset), and 5 (end stage) mo of age were analyzed in real-time PCR. P < 0.0001 (one-way ANOVA),
*P < 0.05, **P < 0.01, ***P < 0.001 (followed by Tukey’s multiple comparison test). (B and G) Primary cultured glia were treated with spinal cord lysate of
control or mSOD1 mice or solvent with/without various inhibitors or dimethyl sulfoxide (DMSO). The mRNA expression of DAO standardized with GAPDH in
the cells was analyzed in real-time PCR (n = 3). P < 0.0001 (one-way ANOVA), *P < 0.05, ***P < 0.001, N.S., not significant (followed by Tukey’s multiple
comparison test). (C) Primary cultured glia were stimulated with spinal cord lysate of control (Co) or mSOD1 mice (mS) or solvent (So) for the indicated time,
and Western blotting was performed with phospho (p)-ERK, total (t)-ERK, p-CREB, and t-CREB antibodies. Values indicate relative amounts of p-ERK1/2 or
p-CREB standardized with t-ERK1/2 or t-CREB, respectively. (D and E) Lumbar spinal cords of control and mSOD1 mice (5 mo) were costained with immu-
nofluorescent analysis of phospho-ERK1/2 (green) and GFAP (red, D) or DAO enzyme histochemistry (red, E). Nuclei were stained with DAPI (blue). Shown
regions are lamina IX. Squared areas in D are high magnification images. (F) Primary cultured glia were incubated with D-serine in the presence of PD98059
for 5 d (n = 6). Levels of D-/L-serine in the cultured media were determined by using 2D-HPLC. Data are plotted as the mean ± SEM.
DAO expression is down-regulated through the NMDAR/ERK pathway in mSOD1 mice. (A) mRNA expressions of DAO standardized with GAPDH in
Sasabe et al.PNAS
| January 10, 2012
| vol. 109
| no. 2
Motoneuron Counts and Diameter Measurement. Please refer to SI Materials
and Methods for details regarding motoneuron counts and diameter
Enzyme Assay of DAO. DAO activity was determined as reported by Watanabe
et al. (40) with some modifications. Please see SI Materials and Methods for
2D-HPLC. Amino acids in tissues were derivatized with 4-fluoro-7-nitro-2,1,3-
benzoxadiazole (NBD-F) (Tokyo Kasei), subjected to HPLC (NANOSPACE SI-2
series; Shiseido), separated into each amino acid by a reversed-phase column,
and further separated into enantiomers by an enantioselective column. The
fluorescence intensity was detected at 530 nm with excitation at 470 nm.
Please see SI Materials and Methods for additional details.
RNA Isolation and Real-Time Quantitative PCR. RNA isolation and real-time
quantitative PCR are detailed in SI Materials and Methods.
Cloning, Mutagenesis, and Transfection. See SI Materials and Methods for
Primary Culture of Glia. Primary cultured glia were prepared from cerebellum
of E16 mouse embryos and treated with various inhibitors or spinal cord
lysate or both. Please refer to SI Materials and Methods for additional details.
Western Blot Analysis. Western blot analysis is detailed in SI Materials
Statistical Analysis. All values in the text and figures of this study indicate
means ± SEM. Statistical analyses for the experiments were performed with
two-tailed Student’s t test or one-way ANOVA followed by Tukey’s multiple
comparison test, in which P < 0.05 was assessed as significant. All analyses
were performed by using Prism 5 (GraphPad Software).
ACKNOWLEDGMENTS. We thank K. Yamashita, N. Suzuki, A. Gotoh, and
S. Hayashi for invaluable assistance with animal work; D. Wylie for expert
opinion on the manuscript; and T. Yoshida-Nishimoto and M. Yamamoto for
indispensable support. This work was supported by Grant-in-Aid for
Scientific Research (A) (to S.A.), Grant-in-Aid for Young Scientists (B) (to
J.S.), the Nakabayashi Trust for ALS Research (to J.S.), and Takeda Science
Foundation (to J.S.). We appreciate Shiseido Co. Ltd. (Tokyo, Japan) for
1. Bruijn LI, Miller TM, Cleveland DW (2004) Unraveling the mechanisms involved in
motor neuron degeneration in ALS. Annu Rev Neurosci 27:723–749.
2. Van Den Bosch L, Van Damme P, Bogaert E, Robberecht W (2006) The role of ex-
citotoxicity in the pathogenesis of amyotrophic lateral sclerosis. Biochim Biophys Acta
3. Mothet JP, et al. (2000) D-serine is an endogenous ligand for the glycine site of the
N-methyl-D-aspartate receptor. Proc Natl Acad Sci USA 97:4926–4931.
4. Panatier A, et al. (2006) Glia-derived D-serine controls NMDA receptor activity and
synaptic memory. Cell 125:775–784.
5. Wolosker H, Blackshaw S, Snyder SH (1999) Serine racemase: A glial enzyme synthe-
sizing D-serine to regulate glutamate-N-methyl-D-aspartate neurotransmission. Proc
Natl Acad Sci USA 96:13409–13414.
6. Basu AC, et al. (2009) Targeted disruption of serine racemase affects glutamatergic
neurotransmission and behavior. Mol Psychiatry 14:719–727.
7. Shleper M, Kartvelishvily E, Wolosker H (2005) D-serine is the dominant endogenous
coagonist for NMDA receptor neurotoxicity in organotypic hippocampal slices.
J Neurosci 25:9413–9417.
8. Mustafa AK, et al. (2010) Serine racemase deletion protects against cerebral ischemia
and excitotoxicity. J Neurosci 30:1413–1416.
9. Inoue R, Hashimoto K, Harai T, Mori H (2008) NMDA- and beta-amyloid1-42-induced
neurotoxicity is attenuated in serine racemase knock-out mice. J Neurosci 28:
10. Sasabe J, et al. (2007) D-serine is a key determinant of glutamate toxicity in amyo-
trophic lateral sclerosis. EMBO J 26:4149–4159.
11. Mitchell J, et al. (2010) Familial amyotrophic lateral sclerosis is associated with a mu-
tation in D-amino acid oxidase. Proc Natl Acad Sci USA 107:7556–7561.
12. Konno R, Yasumura Y (1983) Mouse mutant deficient in D-amino acid oxidase activity.
13. Sasaki M, et al. (1992) A single-base-pair substitution abolishes D-amino-acid oxidase
activity in the mouse. Biochim Biophys Acta 1139:315–318.
14. Almond SL, et al. (2006) Behavioral and biochemical characterization of a mutant
mouse strain lacking D-amino acid oxidase activity and its implications for schizo-
phrenia. Mol Cell Neurosci 32:324–334.
15. Horiike K, Tojo H, Arai R, Nozaki M, Maeda T (1994) D-amino-acid oxidase is confined
to the lower brain stem and cerebellum in rat brain: Regional differentiation of as-
trocytes. Brain Res 652:297–303.
16. Gurney ME, et al. (1994) Motor neuron degeneration in mice that express a human
Cu,Zn superoxide dismutase mutation. Science 264:1772–1775.
17. Alstermark B, Ogawa J (2004) In vivo recordings of bulbospinal excitation in adult
mouse forelimb motoneurons. J Neurophysiol 92:1958–1962.
18. Morikawa A, Hamase K, Zaitsu K (2003) Determination of D-alanine in the rat central
nervous system and periphery using column-switching high-performance liquid
chromatography. Anal Biochem 312:66–72.
19. Miyoshi Y, et al. (2009) Determination of D-serine and D-alanine in the tissues and
physiological fluids of mice with various D-amino-acid oxidase activities using two-
dimensional high-performance liquid chromatography with fluorescence detection.
J Chromatogr B Analyt Technol Biomed Life Sci 877:2506–2512.
20. Lobsiger CS, Boillée S, Cleveland DW (2007) Toxicity from different SOD1 mutants
dysregulates the complement system and the neuronal regenerative response in ALS
motor neurons. Proc Natl Acad Sci USA 104:7319–7326.
21. Hamase K, et al. (2010) Simultaneous determination of hydrophilic amino acid
enantiomers in mammalian tissues and physiological fluids applying a fully
automated micro-two-dimensional high-performance liquid chromatographic con-
cept. J Chromatogr A 1217:1056–1062.
22. Molla G, et al. (2006) Characterization of human D-amino acid oxidase. FEBS Lett 580:
23. Wegorzewska I, Baloh RH (2011) TDP-43-based animal models of neurodegeneration:
New insights into ALS pathology and pathophysiology. Neurodegener Dis 8:262–274.
24. Philips T, Robberecht W (2011) Neuroinflammation in amyotrophic lateral sclerosis:
Role of glial activation in motor neuron disease. Lancet Neurol 10:253–263.
25. Natarajan K, Singh S, Burke TR, Jr., Grunberger D, Aggarwal BB (1996) Caffeic acid
phenethyl ester is a potent and specific inhibitor of activation of nuclear transcription
factor NF-kappa B. Proc Natl Acad Sci USA 93:9090–9095.
26. Lemon RN (2008) Descending pathways in motor control. Annu Rev Neurosci 31:
27. Lawrence DG, Kuypers HG (1968) The functional organization of the motor system in
the monkey. I. The effects of bilateral pyramidal lesions. Brain 91:1–14.
28. Alstermark B, Ogawa J, Isa T (2004) Lack of monosynaptic corticomotoneuronal EPSPs
in rats: Disynaptic EPSPs mediated via reticulospinal neurons and polysynaptic EPSPs
via segmental interneurons. J Neurophysiol 91:1832–1839.
29. Yang HW, Lemon RN (2003) An electron microscopic examination of the corticospinal
projection to the cervical spinal cord in the rat: Lack of evidence for cortico-moto-
neuronal synapses. Exp Brain Res 149:458–469.
30. Morikawa A, et al. (2001) Determination of free D-aspartic acid, D-serine and D-al-
anine in the brain of mutant mice lacking D-amino acid oxidase activity. J Chromatogr
B Biomed Sci Appl 757:119–125.
31. Rothstein JD, Van Kammen M, Levey AI, Martin LJ, Kuncl RW (1995) Selective loss of
glial glutamate transporter GLT-1 in amyotrophic lateral sclerosis. Ann Neurol 38:
32. Trotti D, Rolfs A, Danbolt NC, Brown RH, Jr., Hediger MA (1999) SOD1 mutants linked
to amyotrophic lateral sclerosis selectively inactivate a glial glutamate transporter.
Nat Neurosci 2:848.
33. Howland DS, et al. (2002) Focal loss of the glutamate transporter EAAT2 in a trans-
genic rat model of SOD1 mutant-mediated amyotrophic lateral sclerosis (ALS). Proc
Natl Acad Sci USA 99:1604–1609.
34. Yoshikawa M, Oka T, Kawaguchi M, Hashimoto A (2004) MK-801 upregulates the
expression of d-amino acid oxidase mRNA in rat brain. Brain Res Mol Brain Res 131:
35. Schipke CG, et al. (2001) Astrocytes of the mouse neocortex express functional
N-methyl-D-aspartate receptors. FASEB J 15:1270–1272.
36. Lee MC, et al. (2010) Characterisation of the expression of NMDA receptors in human
astrocytes. PLoS ONE 5:e14123.
37. Kleckner NW, Dingledine R (1988) Requirement for glycine in activation of NMDA-
receptors expressed in Xenopus oocytes. Science 241:835–837.
38. Fadda E, Danysz W, Wroblewski JT, Costa E (1988) Glycine and D-serine increase the
affinity of N-methyl-D-aspartate sensitive glutamate binding sites in rat brain syn-
aptic membranes. Neuropharmacology 27:1183–1185.
39. Lerma J, Zukin RS, Bennett MV (1990) Glycine decreases desensitization of N-methyl-
D-aspartate (NMDA) receptors expressed in Xenopus oocytes and is required for
NMDA responses. Proc Natl Acad Sci USA 87:2354–2358.
40. Watanabe T, Motomura Y, Suga T (1978) A new colorimetric determination of
D-amino acid oxidase and urate oxidase activity. Anal Biochem 86:310–315.
| www.pnas.org/cgi/doi/10.1073/pnas.1114639109 Sasabe et al.