Dietary Omega-3 Deficiency from Gestation Increases Spinal Cord Vulnerability to Traumatic Brain Injury-Induced Damage
Gomez-Pinilla F (2012) Dietary Omega-3 Deficiency from Gestation Increases Spinal Cord Vulnerability to
Traumatic Brain Injury-Induced Damage. PLoS ONE 7(12): e52998. doi:10.1371/journal.pone.0052998
Dietary Omega-3 Deficiency from Gestation Increases Spinal Cord Vulnerability to Traumatic Brain Injury- Induced Damage
Zhe Ying 0
Cameron Feng 0
Rahul Agrawal 0
Yumei Zhuang 0
Fernando Gomez-Pinilla 0
Wolf-Hagen Schunck, Max Delbrueck Center for Molecular Medicine, Germany
0 1 Department of Integrative Biology & Physiology, University of California Los Angeles , Los Angeles , California, United States of America, 2 Department of Neurosurgery, UCLA Brain Injury Research Center , Los Angeles, California , United States of America
Although traumatic brain injury (TBI) is often associated with gait deficits, the effects of TBI on spinal cord centers are poorly understood. We seek to determine the influence of TBI on the spinal cord and the potential of dietary omega-3 (n-3) fatty acids to counteract these effects. Male rodents exposed to diets containing adequate or deficient levels of n-3 since gestation received a moderate fluid percussion injury when becoming 14 weeks old. TBI reduced levels of molecular systems important for synaptic plasticity (BDNF, TrkB, and CREB) and plasma membrane homeostasis (4-HNE, iPLA2, syntaxin-3) in the lumbar spinal cord. These effects of TBI were more dramatic in the animals exposed to the n-3 deficient diet. Results emphasize the comprehensive action of TBI across the neuroaxis, and the critical role of dietary n-3 as a means to build resistance against the effects of TBI.
Although deficit in motor function is a common consequence of
traumatic brain injury (TBI), not much is known about the
influence of brain injury on motor centers in the spinal cord. We
are starting to understand that TBI reduces the expression of
molecules important for synaptic plasticity in the spinal cord ,
and are thus arguing for the need of broad therapeutic strategies to
influence the brain and spinal cord. Here we have studied the
capacity of foods to promote increased spinal cord resilience to the
type of diffuse injury caused by brain trauma, in particular, the
essential omega-3 (n-3) poly-unsaturated fatty acid (PUFA)
docosahexaenoic acid (DHA, 22:6n-3), which is gaining
recognition for supporting neuronal function and plasticity. Inadequate
consumption of dietary DHA during CNS development results in
aberrations in neuronal function, and learning ability [2,3], while
dietary DHA supplementation in the adult brain aids recovery
after brain injury [4,5]. In the present study, we seek to determine
whether dietary supplementation of DHA could influence the
capacity of the spinal cord to cope with the effects of injury to the
We used fluid-percussion injury (FPI) as an animal model of
TBI since this injury promotes circuit dysfunction without
extensive neuronal death . Specifically, FPI results in significant
reductions of brain-derived neurotrophic factor (BDNF) and its
downstream effectors. BDNF is important for many aspects of
neuronal function and plasticity, influencing adult neurogenesis,
and providing protection after neuronal injury . As we know
that TBI promotes oxidative damage of the plasma membrane ,
likely influencing the membranes phospholipid composition, such
as DHA, we used the lipid peroxidation marker 4-HNE to assess
the status of the plasma membrane in response to TBI and DHA
interventions . We have also assessed syntaxin-3 based on its
role as a modulator of neuronal membrane expansion, especially
during synaptic growth , and assessed calcium-independent
phospholipase A-2 (iPLA-2) based on its influence on membrane
phospholipid biosynthesis and turnover .
Synaptic Proteins (Fig. 1)
BDNF levels were significantly reduced in the animals fed the
n3 deficient diet (n-3 def/sham, 74%, p,0.01, n = 5) as compared
to the animals fed adequate n-3 (n-3 adq/sham, n = 6) (Fig. 1A).
FPI reduced levels of BDNF in the n-3 deficient group (n-3 def/
FPI, p,0.01, n = 5). Although FPI reduced BDNF levels in the n-3
adq rats (p,0.02, n = 7), these levels were still higher than the n-3
def group (p,0.05, Fig. 1A). The n-3 deficient diet reduced the
levels of pTrkB/TrkB when compared to the n-3 adq/sham group
(p,0.01, Fig. 1B). Although FPI reduced levels of pTrkB/TrkB in
the group receiving adequate levels of n-3, pTrkB/TrkB levels
were still higher than the n-3 def group (n-3 adq/FPI vs. n-3 def/
FPI, p,0.05, Fig. 1B). Levels of pCREB/CREB were reduced in
the animals exposed to the n-3 deficient diet (n-3 def/sham vs. n-3
adq/sham, p,0.05, Fig. 1C) and FPI had a tendency to reduce
pCREB/CREB levels even further (Fig. 1C). Fig. 1D showed the
representative western bands corresponding to protein markers
and animal groups.
Figure 1. Synaptic plasticity markers BDNF (A), pTrkB (B), and pCREB (C) protein levels were assessed in the lumbar spinal cord of
rats exposed to FPI, using western blot assay. (D). Representative western blot bands from experimental groups. Results were expressed as
mean 6 standard error of the mean (SEM), *P,0.05, **P,0.01.FPI, fluid percussion injury; n-3 def, omega 3 fatty acids deficient; n-3 adq, omega 3
fatty acid adequate. n-3 def/sham: n = 5; n-3 adq/sham: n = 6; n-3 def/FPI: n = 5; n-3 adq/FPI: n = 7.
Membrane Homeostasis (Fig. 2)
We assessed levels of 4-HNE which is a suitable marker of
plasma membrane lipid peroxidation. Results showed that the n-3
def diet increased levels of 4-HNE in the spinal cord as compared
to n-3 adq diet (p,0.01, Fig. 2A, 2B). FPI elevated 4-HNE levels
even further in the n-3 def animals (p,0.01, Fig. 2A, 2B).
Although FPI also elevated levels of 4-HNE in the n-3 adq group,
4-HNE levels were lower than in the n-3 def group (p,0.01,
Fig. 2A, 2B).
We measured iPLA2 levels based on its involvement in the
metabolism of membrane phospholipids  FPI significantly
reduced the levels of iPLA2 in the animals fed n-3 deficient diet
(n3 def/FPI vs. n-3 def/sham, p,0.01). FPI had no effects on levels
of iPLA-2 in the n-3 adq group suggesting a counteractive effect
(Fig. 2C) such that levels of iPLA2 in the n-3 def rats exposed to
FPI rats were significantly lower than their counterpart in the n-3
adq group (p,0.01, Fig. 2C).
Although the exposure to the n-3 deficient diet did not affect
levels of syntaxin-3 in the sham rats relative to the adq group, FPI
strongly reduced syntaxin-3 levels in the n-3 def group (n-3 def/
FPI group as compared to n-3 adq/FPI group (P,0.01) and n-3
def/sham (P,0.05) groups (Fig. 2D).
Fatty Acids in Spinal Cord (Fig. 3)
Levels of docosahexaenoic acid (DHA, 22:6n-3) and
arachidonic acid (AA, 20:4n-6) were measured in the spinal cord region
using gas chromatography. Results showed that the levels of DHA
significantly decreased in animals fed on n-3 deficient diet (n-3
def/sham). FPI did not affect levels of DHA in the n-3 def group
(n-3 def/FPI) or the n-3 adq group (Fig. 3A). In turn, levels of AA
were increased significantly in the sham and FPI groups exposed
to the n-3 deficient diet (p,0.01, Fig. 3B). FPI also increased AA
levels in the n-3 adq rats (P,0.05) (Fig. 3B).
We found that brain concussive injury reduces molecular
substrates of plasticity in the spinal cord, and these effects were
dependent on the availability of DHA in the diet. These results
emphasize the comprehensive action of TBI across the neuroaxis,
and the critical role of diet as a means to build resistance against
the effects of TBI. According to our results, proper exposure to n-3
fatty acids during gestation and throughout maturation of the
CNS is crucial for building neural resilience during adulthood.
The effects of diet and TBI were observed on levels of molecules
Figure 2. Levels of molecules related to plasma membrane homeostasis 4-HNE (A, B), iPLA2 (C), and syntaxin 3 (D) in the lumbar
spinal cord of rats exposed to FPI. Results were expressed as mean 6 standard error of the mean (SEM), *P,0.05, **P,0.01. FPI, fluid percussion
injury; n-3 def, omega 3 fatty acids deficient; n-3 adq, omega 3 fatty acids adequate. n-3 def/sham: n = 5; n-3 adq/sham: n = 6; n-3 def/FPI: n = 5; n-3
adq/FPI: n = 7.
associated with the function of BDNF on synaptic plasticity, and
plasma membrane homeostasis in the spinal cord.
According to our results, FPI and the diet deficient in DHA
reduced protein levels of BDNF and its receptor TrkB in the SC,
as well as elements related to the action of BDNF on synaptic
plasticity such as syntaxin 3 and CREB, which have recognized
roles in synaptic plasticity and learning and memory . These
results suggest that FPI reduces the capacity of the SC for
plasticity. The action of the BDNF system seems crucial for
mediating the action of DHA in the brain as a diet deficient in
DHA has been shown to reduce activation of the BDNF TrkB
receptors , and the capacity of the SC for learning a
motorsensory task . Therefore, the reduction of BDNF
because of the DHA def or TBI in our study may have negative
implications for the potential of the SC to functionally recover
after brain or SC injury. On the other hand, the fact that DHA
supplementation is related to higher levels of BDNF argues in
favor of a therapeutic potential of DHA. Indeed, DHA has shown
protective capacity when provided after hemisection or
compression spinal cord injury by increasing the survival of neurons and
improving locomotor performance .
DHA is a structural component of plasma membrane, and
membrane bound DHA supports membrane fluidity , which is
instrumental for neuronal signaling. The high contents of DHA
and other phospholipids in the plasma membranes make the
membrane a vulnerable target to lipid peroxidation. Lipid
peroxidation has been linked to a disruption in membrane
homeostasis and impairment of synaptic plasticity. Here, we
found that FPI increased lipid peroxidation in the SC as evidenced
by increased levels of 4-HNE. The phospholipase A2 (PLA2) family
is involved in the metabolism of membrane phospholipids ,
and the calcium-independent PLA2 (iPLA2) plays an important
Figure 3. Gas chromatography was used to assess levels of DHA (A) and AA (B) in the cervical spinal cord of FPI rats. An n-3 def diet
significantly decreased DHA and increased AA levels. FPI increased AA levels of n-3 adq group (p,0.05) but had no effects in n-3 def group. Data are
shown as ratio of fatty acid(mg)/tissue(g). *P,0.05, **P,0.01. DHA, docosahexaenoic acid; AA, arachidonic acid. n-3 def/sham: n = 5; n-3 adq/sham:
n = 6; n-3 def/FPI: n = 5; n-3 adq/FPI: n = 7.
role in synaptic plasticity [16,17]. Therefore, our results showing
significant changes in iPLA2 levels in the n-3 def animals
undergoing FPI provide an indication for the compromise of
membrane homeostasis. In turn, STX-3 is a membrane-bound
synaptic protein which function is influenced by DHA . The
fact that the diet deficient in DHA increased lipid peroxidation
and decreased syntaxin 3, suggests how a lack of membrane DHA
promotes membrane instability . Syntaxin 3 is positioned in
the presynaptic plasma membrane to detect local changes in
PUFA  and plays a crucial role in the docking and fusion of
vesicles during synaptic transmission . Therefore, our results
showing that FPI and dietary n-3 affect levels of 4-HNE, iPLA2,
and STX-3, suggest a potential mechanism by which TBI and diet
can influence membrane homeostasis required for functional
recovery after spinal cord injury.
It is important to consider that although DHA (1.2%) is the
most abundant omega-3 fatty acid in our diet, there are other less
abundant fatty acids as well such as eicosapentaenoic acid (EPA).
EPA has also been reported to support neural repair events such as
reducing axonal injury after spinal cord compression; however, its
action appears less effective than DHA . Given the large
difference in contents of DHA (1.2%) Vs. EPA (0.24%), it is likely
that the main effects of the diet are related to DHA.
The effects of the n-3 feeding can also be perceived at the levels
of AA and DHA in the spinal cord, as evidenced by results
showing that n-3 deficiency increased AA levels but reduced DHA
levels. These results are consistent with the possibility that AA
could replace DHA in the membrane. A reduction in membrane
fluidity can affect transmembrane receptors such as TrkB, and this
may explain why the n-3 def diet reduced TrkB activity in our
study. DHA modifies the characteristics of lipid rafts by
incorporating into raft domains of the membrane and influencing
signaling across embedded receptors  such as TrkB receptors
The current results emphasize the pervasive effects of brain
trauma impacting CNS regions, which are distant from the lesion.
These results have important implications for the design of
potential treatments directed to counteract the effects of TBI.
Based on results showing the comprehensive effects of brain injury
in the brain and spinal cord, it appears that interventions that have
the capacity to influence the entire neuroaxis can be particularly
effective. As discussed above, the broad spectrum of action of the
omega-3 fatty acid DHA positively influencing the brain and
spinal cord appears particularly suitable for this purpose. It is
critical to complement our molecular data with behavioral studies.
It is known that the type of TBI used in the current study promotes
deficits in cognition and gait  and that post-injury treatment
with DHA counteracts some these deficits . A period as short as
12 days of DHA following FPI has been shown to be sufficient to
counteract deficits in hippocampal-dependent learning . The
unique aspect of our results is the demonstration that dietary DHA
during CNS maturation confers resilience to neurological damage
in adult life. These results have important implications to appraise
the role of diet as a vulnerability factor for the outcome of TBI. It
is a common observation that healing after brain or spinal cord
injury is not often predictable based on the extent of the
neurological damage. This implies that vulnerability factors
associated with the environment and genetics have great potential
to determine the outcome of CNS injured patients.
Our results show that exposure to n-3 fatty acids during
gestation and throughout maturation of the CNS is important for
building resilience to neurological damage incurred later on in life.
Further studies are required to define whether shorter dietary
DHA exposure can confer CNS protection. In conclusion, these
results are important to define the broad and positive action of n-3
diet on counteracting the effects of concussive brain injury on the
Materials and Methods
All experimental procedures were performed in accordance
with the United States National Institutes of Health Guide for the
Care and Use of Laboratory Animals and were approved by the
University of California at Los Angeles (UCLA) Chancellors
Animal Research Committee (ARC).
Two-day pregnant female Sprague-Dawley rats weighing
between 280 and 300 g were obtained from Charles River
Laboratories, Inc. (Portage, MI, USA). The animals were housed
with 12-h light/dark cycles and a maintained temperature of 22
24uC. Animals were randomly divided into two dietary groups:
omega 3 fatty acids adequate (n-3 adq) diet vs omega 3 fatty acids
deficient (n-3 def) diet. The dietary treatment started with the
mothers and the male offspring were weaned to the same diet as
their dam. After 14 weeks, the animals were subjected to moderate
FPI, and continued their diet until sacrificed one week later.
The two custom diets (n-3 def, n-3 adq) were prepared
commercially (Dyets, Bethlehem, PA, USA) and contained the
same basal macronutrients, vitamins, minerals, and basal fats
(hydrogenated coconut and safflower oils). Vitamin-free casein
Alacid 710 (NZMP North America Inc., CA, USA) was included.
The n-3 adq diet contained an extra 0.5% flaxseed oil (linoleic
acid), 1.2% DHA and 0.24% EPA (Nordic Naturals, Inc.
Watsonville, CA, USA), relative to the n-3 def diet. (Table 1).
The individuals daily intake of DHA was about 480 mg per
kilogram of animal weight.
Fluid Percussion Injury (FPI)
FPI was performed as previously described . Under deep
anesthesia, A 3.0 mm diameter craniotomy was made 3.0 mm
posterior to the bregma and 6.0 mm lateral (left) to the midline,
and a cap was glued over the craniotomy. The cap was filled with
0.9% saline solution and a mild fluid percussion pulse (1.5 atm)
was administered. The severity of injury was confirmed based on
the unconscious time (less than 120 seconds) of the animal
following injury before the first response to a paw inch. Sham
animals underwent an identical preparation with the exception of
One week following FPI, rats were sacrificed by decapitation,
and the cervical enlargement (C3C6) and lumbar enlargement
(L2L6) regions were collected and stored in 270uC for lipids and
BDNF, phosphorylated tyrosine kinase (pTrkB), phosphorylated
cAMP response element-binding (pCREB), syntaxin-3,
calciumindependent phospholipase A2 (iPLA-2), and 4-Hydroxynonenal
(4-HNE) proteins of the right lumbar SC were analyzed using
western blots as previously described . Tissue was first
homogenized in a lysis buffer and the total protein was separated
by electrophoresis on a polyacrylamide gel and electrotransferred
onto a PVDF (nitrocellulose for BDNF) membrane (Millipore,
Bedford,MA). After blocking, the membranes were rinsed with
TBS-T and incubated with the primary antibody for actin (Santa
Cruz Biotechnology, Santa Cruz, CA, USA), BDNF (1:300, Santa
Cruz Biotechonology, Santa Cruz, CA, USA), pTrkB (1:200, BD
Biosciences, Sparks, MD, USA), TrkB (1:500, Santa Cruz
Biotechnology, Santa Cruz, CA, USA), pCREB (1:1,000,
Millipore, Bedford, MA, USA), CREB (1:200, Millipore, Bedford, MA,
USA), syntaxin-3 (1:300, Santa Cruz Biotechnology, Santa Cruz,
CA, USA), iPLA-2 (1:200, Santa Cruz Biotechnology, Santa Cruz,
CA, USA), or 4-HNE (1:500, Santa Cruz Biotechnology, Santa
Cruz, CA, USA ). Immunocomplexes were visualized by
chemiluminescence using the commercial kit ECL plus
(Amersham Pharmacia Biotech Inc., Piscataway, NJ, USA). Respective
protein sizes were compared to the Benchmark pre-stained protein
ladder (Invitrogen Technology, Carlsbad, CA, USA). Protein
bands were digitally scanned and quantified using the ImageJ
software. Actin was used as an internal control. The
phosphorylated proteins were normalized to their respective
Fatty Acids Analysis by Gas Chromatography
The lipids content of the cervical SC were extracted and
analyzed by gas chromatography. Lipids were first extracted by
homogenizing tissue on ice in a 2:1 (vol:vol) chloroform:methanol
solution with 50 ug/mL of butylated hydroxytoluene added to
prevent lipid oxidation. Tricosanoic acid methylester (C23:0) was
added to each sample to function as an internal standard. After
extraction, lipids were methylated by heating at 90uC for 1 hour
in14% (w/v) boron trifluoride-methanol reagent. The lipid
contents were analyzed using the Clarus 500 gas chromatograph
with a built-in Autosampler (PerkinElmer), and the total runtime
for each sample was 34 min. 1uL FA methyl esters (FAME) was
injected in split injection mode with a 100:1 split ratio, and the
resultant peaks were identified and quantified by comparison with
the standard Supelco 37-component FAME Mix (Sigma-Aldrich,
Protein data were expressed as a mean percentage of the control
(sham/n-3 adq) group DHA and AA levels were expressed as a
ratio of fatty acid (mg)/tissue (g). Significance among groups was
determined using SPSS software. One-way analysis of variance
(ANOVA) was performed followed by Fishers Least Significance
Difference test. Results were expressed as mean 6 standard error
of the mean (SEM), n = 57/group. Significant difference was
considered at P,0.05.
Conceived and designed the experiments: ZY RA FGP. Performed the
experiments: CF RA YZ. Analyzed the data: CF YZ. Wrote the paper: ZY
1. Wu A , Ying Z , Schubert D , Gomez-Pinilla F ( 2011 ) Brain and spinal cord interaction: a dietary curcumin derivative counteracts locomotor and cognitive deficits after brain trauma . Neurorehabil Neural Repair 25 : 332 - 342 .
2. Fedorova I , Hussein N , Baumann MH , Di Martino C , Salem N ( 2009 ) An n-3 fatty acid deficiency impairs rat spatial learning in the Barnes maze . Behav Neurosci 123 : 196 - 205 .
3. Bhatia HS , Agrawal R , Sharma S , Huo YX , Ying Z , et al. ( 2011 ) Omega-3 fatty acid deficiency during brain maturation reduces neuronal and behavioral plasticity in adulthood . PLoS One 6 : e28451 .
4. Wu A , Ying Z , Gomez-Pinilla F ( 2004 ) Dietary Omega-3 fatty acids normalize BDNF levels, reduce oxidative damage, and counteract learning disability after traumatic brain injury in rats . J Neurotrauma 21 : 1457 - 1467 .
5. Mills JD , Hadley K , Bailes JE ( 2011 ) Dietary supplementation with the omega-3 fatty acid docosahexaenoic acid in traumatic brain injury . Neurosurgery 68 : 474 - 481 ; discussion 481.
6. Wu A , Molteni R , Ying Z , F G-P ( 2003 ) A saturated-fat diet aggravates the outcome of traumatic brain injury on hippocampal plasticity and cognitive function by reducing brain-derived neurotrophic factor . Neuroscience 119 ( 2 ): 365 - 375 .
7. Griesbach GS , Hovda DA , Gomez-Pinilla F ( 2009 ) Exercise-induced improvement in cognitive performance after traumatic brain injury in rats is dependent on BDNF activation . Brain Res 1288 : 105 - 115 .
8. Wu A , Ying Z , Gomez-Pinilla F ( 2011 ) The salutary effects of DHA dietary supplementation on cognition, neuroplasticity, and membrane homeostasis after brain trauma . J Neurotrauma 28 : 2113 - 2122 .
9. Chytrova G , Ying Z , Gomez-Pinilla F ( 2010 ) Exercise contributes to the effects of DHA dietary supplementation by acting on membrane-related synaptic systems . Brain Res 1341 : 32 - 40 .
10. Sharma S , Ying Z , Gomez-Pinilla F ( 2010 ) A pyrazole curcumin derivative restores membrane homeostasis disrupted after brain trauma . Exp Neurol 226 : 191 - 199 .
11. Farooqui AA , Horrocks LA ( 2006 ) Phospholipase A2-generated lipid mediators in the brain: the good, the bad, and the ugly . Neuroscientist 12 : 245 - 260 .
12. Benito E , Barco A ( 2010 ) CREB's control of intrinsic and synaptic plasticity: implications for CREB-dependent memory models . Trends Neurosci 33 : 230 - 240 .
13. Joseph MS , Ying Z , Zhuang Y , Zhong H , Wu A , et al. ( 2012 ) Effects of Diet and/or Exercise in Enhancing Spinal Cord Sensorimotor Learning . PLoSOne 7: e41288 .
14. King VR , Huang WL , Dyall SC , Curran OE , Priestley JV , et al. ( 2006 ) Omega3 fatty acids improve recovery, whereas omega-6 fatty acids worsen outcome, after spinal cord injury in the adult rat . J Neurosci 26 : 4672 - 4680 .
15. Suzuki H , Park SJ , Tamura M , Ando S ( 1998 ) Effect of the long-term feeding of dietary lipids on the learning ability, fatty acid composition of brain stem phospholipids and synaptic membrane fluidity in adult mice: a comparison of sardine oil diet with palm oil diet . Mech Ageing Dev 101 : 119 - 128 .
16. Fitzpatrick JS , Baudry M ( 1994 ) Blockade of long-term depression in neonatal hippocampal slices by a phospholipase A2 inhibitor . Brain Res Dev Brain Res 78 : 81 - 86 .
17. Wolf MJ , Izumi Y , Zorumski CF , Gross RW ( 1995 ) Long-term potentiation requires activation of calcium-independent phospholipase A2 . FEBS Lett 377 : 358 - 362 .
18. Darios F , Davletov B ( 2006 ) Omega-3 and omega-6 fatty acids stimulate cell membrane expansion by acting on syntaxin 3 . Nature 440 : 813 - 817 .
19. Ansari MA , Roberts KN , Scheff SW ( 2008 ) A time course of contusion-induced oxidative stress and synaptic proteins in cortex in a rat model of TBI . J Neurotrauma 25 : 513 - 526 .
20. McMahon HT , Sudhof TC ( 1995 ) Synaptic core complex of synaptobrevin, syntaxin, and SNAP25 forms high affinity alpha-SNAP binding site . J Biol Chem 270 : 2213 - 2217 .
21. Hall JC , Priestley JV , Perry VH , Michael-Titus AT ( 2012 ) Docosahexaenoic acid, but not eicosapentaenoic acid, reduces the early inflammatory response following compression spinal cord injury in the rat . J Neurochem 121 : 738 - 750 .
22. Shaikh SR , Rockett BD , Salameh M , Carraway K ( 2009 ) Docosahexaenoic acid modifies the clustering and size of lipid rafts and the lateral organization and surface expression of MHC class I of EL4 cells . J Nutr 139 : 1632 - 1639 .
23. Nagappan G , Lu B ( 2005 ) Activity-dependent modulation of the BDNF receptor TrkB: mechanisms and implications . Trends Neurosci 28 : 464 - 471 .