Electroacupuncture Ameliorates Learning and Memory and Improves Synaptic Plasticity via ...

12 Pages·2016·4.48 MB·English

Hindawi Publishing Corporation
Evidence-Based Complementary and Alternative Medicine
Volume 2016, Article ID 7893710, 11 pages
http://dx.doi.org/10.1155/2016/7893710
Research Article
Electroacupuncture Ameliorates Learning and Memory and
Improves Synaptic Plasticity via Activation of the PKA/CREB
Signaling Pathway in Cerebral Hypoperfusion
Cai-XiaZheng,1MinLu,1Ya-BiGuo,2Feng-XiaZhang,1HuaLiu,3FengGuo,1
Xiao-LinHuang,1andXiao-HuaHan1
1DepartmentofRehabilitationMedicine,TongjiHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,
Wuhan,HubeiProvince,China
2DepartmentofRehabilitationMedicine,XiangyangCentralHospital,HubeiCollegeofArtsandScience,Xiangyang,
HubeiProvince,China
3CollegeofHealthScience,WuhanInstituteofPhysicalEducation,Wuhan,HubeiProvince,China
CorrespondenceshouldbeaddressedtoXiao-LinHuang;[email protected];[email protected]
Received2June2016;Accepted18September2016
AcademicEditor:YiuW.Kwan
Copyright©2016Cai-XiaZhengetal.ThisisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense,
whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
Electroacupuncture (EA) has shown protective effects on cognitive decline. However, the underlying molecular mechanisms
areill-understood.Thepresentstudywasundertakentodeterminewhetherthecognitivefunctionwasamelioratedincerebral
hypoperfusionratsfollowingEAandtoinvestigatetheroleofPKA/CREBpathway.Weusedarat2-vesselocclusion(2VO)model
anddeliveredEAatBaihui(GV20)andDazhui(GV14)acupoints.Morriswatermaze(MWM)task,electrophysiologicalrecording,
Golgisilverstain,Nisslstain,Westernblot,andreal-timePCRwereemployed.EAsignificantly(1)amelioratedthespatiallearning
andmemorydeficits,(2)alleviatedlong-termpotentiation(LTP)impairmentandthereductionofdendriticspinedensity,(3)
suppressedthedeclineofphospho-CREB(pCREB)protein,brain-derivedneurotrophicfactor(BDNF)protein,andmicroRNA132
(miR132), and (4) reduced the increase of p250GAP protein of 2VO rats. These changes were partially blocked by a selective
proteinkinaseA(PKA)inhibitor,N-[2-(p-bromocinnamylamino)ethyl]-5-isoquinoline-sulfonamide(H89),suggestingthatthe
PKA/CREB pathway is potentially involved in the effects of EA. Moreover, any significant damage to the pyramidal cell layer
ofCA1subregionwasabsent.TheseresultsdemonstratedthatEAcouldamelioratelearningandmemorydeficitsandalleviate
hippocampalsynapticplasticityimpairmentofcerebralhypoperfusionrats,potentiallymediatedbyPKA/CREBsignalingpath-
way.
1.Introduction prevent cognitive deficits in rats with cerebral ischemia [8–
11].However,theunderlyingmolecularmechanismsarenot
Duetoanincreaseinagingpopulation,age-relatedcognitive entirelyunderstood.
impairment, in particular vascular cognitive impairment cAMP response element-binding protein (CREB)/CRE-
(VCI),becomesincreasinglychallengingworldwide,without mediated gene transcription and protein synthesis have
effectivemedications[1].Asanovelcombinationalapproach, been speculated to be essential in long-term hippocampal
electroacupuncture (EA) therapy, consisting of traditional synaptic plasticity and memory formation [12–16]. Several
acupunctureandmodernelectrotherapytechnology,wasfre- pieces of evidence have confirmed that protein kinase A
quentlyreportedtoalleviatecognitivedeclineinpatientswith (PKA) is one of the major kinases to activate CREB and
stroke[2],Alzheimerdisease(AD)[3,4],ormildcognitive thePKA/CREBpathwayiscriticalinlearningandmemory
impairment (MCI) [5–7]. In addition, EA was suggested to [17]. Previous studies had revealed that transcription of
2 Evidence-BasedComplementaryandAlternativeMedicine
BDNF gene was CREB-regulated in an activity-dependent GV14
GV20
manner, wherein its expression was involved in neuronal
development, synaptic plasticity, and neuroprotection [13].
Forexample,BDNFwasrelatedtotheselectivevulnerability
of hippocampal neuronal populations during brain injury
[18]. It also played a vital role in long-term potentiation
(LTP),acellularmodelforlearningandmemory[19,20].Of
note,recently,miRNA132(miR132)wasidentifiedasanovel
CREB target gene and was suggested to regulate neuronal
morphogenesisthroughtranslationalinhibitionofaGTPase- Figure1:Ratschematicshowingthelocationoftheacupuncture
activatingprotein,p250GAP[21]. pointsusedinthestudy.GV14represents“Dazhui,”whichislocated
Hitherto, only a few studies about the molecular alter- ontheposteriormidlineandinthedepressionbelowthespinous
ationsofPKA/CREBpathwayduringEAtreatmenthavebeen processoftheseventhcervicalvertebra;GV20represents“Baihui,”
reported. In agreement, acupuncture treatment at Zusanli whichislocatedattherightmidpointoftheparietalbone.
(ST36) was shown to increase cAMP concentration, PKA
activity, pCREB protein, and pERK protein expressions,
accompaniedbyalleviationofmemoryimpairmentincere- group: rats were given H89 at 30min before the operation
bralmulti-infarctionrats[8].Inaddition,oneofourprevious of2VOandthenreceivedEAsimilarlytotheEAgroup;and
studieshadshownthatEAattenuatedneuronalapoptosisand (5)EA+normalsaline(NS)solutiongroup:ratswereinjected
amelioratedlearningandmemoryin2VOrats,viaincreasing withequalvolumeofNSinsteadofH89,at30minbeforethe
theexpressionsofpCREBandBCL-2protein[22]. 2VOoperationandthesubsequentEAtreatment.
Basedontheabovefindings,wehypothesizedthattreat- H89(2𝜇g/𝜇L×10𝜇L,Sigma-Aldrich,Shanghai,China)
ment of EA might enhance hippocampal synaptic plasticity oravehicleofsterileNSwasintracerebroventricularly(icv)
and improve cognitive function of cerebral ischemic rats injectedintoratsofEA+H89grouporEA+NSgroup.Under
andthatPKA/CREBsignalingpathwaymightbeapotential intraperitoneal anesthesia (10% chloral hydrate, 350mg/kg,
mechanismunderlyingtheseeffects.Firstly,spatiallearning intraperitoneally(ip)),theheadwasmountedinastereotaxic
andmemory,LTPanddendriticspinedensity,andneuronal instrument (SN-3, Narishige, Japan). Injection coordinates
viabilityaswellasexpressionofpCREB,BDNF,miR132,and were0.8mmposteriorlytothebregma,1.5mmlaterallytothe
p250GAP proteins were examined to test effects of EA in midlineintherighthemisphere,and4.5mmbelowthedural
cerebralhypoperfusion.Furthermore,totestthehypothesis surface.
thatthePKA/CREB pathwayisinvolvedinthosebeneficial Through a ventral midline cervical incision, both com-
effects of EA, rats were pregiven a selective PKA inhibitor, moncarotidarterieswerecarefullyisolatedfromtheirsheaths
H89 (it can strongly inhibit the biological action of PKA), andvagalnervesanddoublyligatedwith4-0silkthread.After
andsimilarlymonitoredfortheabove-mentionedindicators. occlusion,mostanimalsappearednormal,butfewratsthat
Thus, the present study was aimed at determining whether displayedepilepticseizures(2/87)orabsenceofweightgain
andhowthecognitivefunctionandthehippocampalsynaptic (3/87) or intestinal obstruction (2/87) were excluded from
plasticity were altered by EA in cerebral hypoperfusion. subsequentexperiments.
Also, we investigated the probable underlying molecular
machinerycomprisedofthePKA/CREBpathway. 2.3. EA Treatment. EA was commenced one day after the
operationinconsciousrats.Thelocationof“Baihui”acupoint
2.MaterialsandMethods (GV20)and“Dazhui”acupoint(GV14)wasdescribedprevi-
ously[23](Figure1).
2.1. Animals. A total of 87 adult male Sprague-Dawley rats
weighing between 250g and 300g were used (SPF grade, “Acupunctureneedleswereinsertedintothemus-
Experimental Animal Center, Huazhong University of Sci- cle at a depth of 0.5mm of GV20 to serve as a
enceandTechnology,Wuhan,China).Animalswerehoused cathode while another electrode was placed on
inagroupof4-5/cageat24±1∘Cinalight-controlled(12h the GV14 serving as an anode. Then electrical
light/dark cycle) room, providing free access to water and stimulationwasdeliveredusingaG6805-IIelec-
food. All the animal studies were approved by the Review troacupuncturetherapeuticapparatus(Shanghai
Committee for the Care and Use of Laboratory Animals of Medical Electronic Apparatus Co., China), with
TongjiMedicalCollege,HuazhongUniversityofScienceand continuous current at 20Hz for 20min daily.
Technology. Stimulus intensity was based on the visible light
facialmusclestwitching.”
2.2. Establishment of 2VO Model. Animals were randomly
assigned into five groups: (1) control group: rats were anes- 2.4. Morris Water Maze Task. Morris water maze (MWM)
thetized and exposed to surgery but did not deal with task was carried out to test spatial learning and memory.
actual ligation; (2) model group: rats have undergone 2VO Briefly,acircularwatertank(150cmindiameterand50cm
operation;(3)EAgroup:ratsweregivenEAtreatmentafter deep) was filled with 23 ± 2∘C water to a depth of 21cm.
2VO operation, for seven consecutive days; (4) EA+H89 A circular platform of 15cm in diameter and 20cm height
Evidence-BasedComplementaryandAlternativeMedicine 3
wasplacedinthecenterofthetargetquadrant(quadrantI). dendritic branches at 1000x magnification. The density of
Severalvisualcueswerelocatedonthewallofthetestroom. dendritic spines along the length was estimated by spine
Ratsweresubjectedtotwosessionsoffourplacenavigation number/10𝜇m.
trials per day with an interval of at least 4h for three con-
secutivedays(day5–day7,trainingfor5times).Thestarting 2.7. Histology. Under anesthesia, rats were perfused with
pointswere changed forevery trial.On day 8, theplatform 0.9%salinesolutionfollowedby4%PFAin0.1Mphosphate
wasremovedforspatialprobetrialtotestthespatialmemory. buffer, pH 7.4. Brains were then removed and fixed for 6–
∘
The latency to find the submerged platform and the dwell 8h at 4 C. Following paraffin embedding, from each rat
time in quadrant I and the swimming paths were recorded werecollectedthreecoronalslicescontainingthedorsalhip-
automaticallyusingacomputer-basedimageanalyzerMWM pocampus at 5𝜇m thickness. Toluidine blue staining (Nissl
tracking system MT-200 (ChengDu Technology & Market staining) was performed. Representative photomicrographs
Co.,Ltd.,Chengdu,SichuanProvince,China). of pyramidal cell layer of CA1 region were microscopically
captured.Quantitativeanalysisoftheratiosofviableneurons
2.5. Electrophysiological Recording. Rats were anesthetized was processed by Image-Pro Plus (Leica DMLB) software
withurethane(1.5g/kg,ip)andplacedinastereotaxicframe at 400x magnification. An average of three Nissl-stained
(SN-3,Narishige,Japan)forsurgeryandrecording.Abipolar sections was calculated to yield the single parameter per
stimulating electrode was placed into Schaffer collateral rat.Allmicroscopicanalyseswereconductedbyanobserver
pathway of the dorsal hippocampus (4.5mm posteriorly to blindedtothegroups.
bregma, 3.7mm laterally to the midline, and at 2.5mm
depth from the brain surface). A recording electrode was 2.8. Western Blot. Rats in each group were sacrificed by
positionedintheipsilateralstratumradiatumunderneaththe decapitation. Hippocampus tissues were quickly removed
CA1area(3.5mmposteriorlytobregma,2.5mmlaterally,and from the brain and homogenized in ice-cold lysate buffer
at2.0mmdepthfromthebrainsurface).A30minrestperiod containing 50mmol/L Tris-HCl (pH = 8.0), 150mmol/L
wasimplementedaftertheelectrodesinsertion. NaCl, 1% Triton X-100, 100𝜇g/mL PMSF, and phosphatase
Baselineresponseswerecollectedbylow-frequencystim- inhibitors (Sigma, USA). Proteins were fractionated on
ulation(0.05Hz,width0.1ms,and50%ofmaximalresponse) 10%SDS-PAGEgelsandthentransferredtopolyvinylidene
and recorded for 20min. Subsequently, four 100Hz trains difluoride(PVDF)membrane(MerckMillipore,Germany).
of high-frequency stimulation (HFS, 80% of maximum Theblotswerethenincubatedwithphospho-CREB(Ser133)
response) were delivered, including 20 pulses and 30s of rabbitanti-ratmonoclonalantibody(mAb)(1:1000,CellSig-
intertrain intervals, lasting for an additional 20min. Fol- nalingTechnology,Danvers,MA,USA),rabbitanti-ratpoly-
lowingHFS,fieldexcitatorypostsynapticpotentials(fEPSPs) clonal antibody for BDNF (1:300, Boster, Wuhan, China),
were recorded, being set at 5Hz at the same stimulus goatanti-ratpolyclonalantibodyforp250GAP(1:200,Santa
intensity as the baseline pulse, over 60min. fEPSP slope Cruz Biotechnology, Santa Cruz, CA, USA), or 𝛽-tubulin
levelwascalculatedbytheratiooftheabsolutefEPSPslope mouseanti-ratmAb(1:1000,AffinityBioscience,USA)anti-
∘
to baseline value. It was defined that LTP was successfully bodies overnight at 4 C. The immunocomplexes were visu-
induced if the fEPSP slope level was ≥120%. Evoked field alizedwithhorseradishperoxidase(HRP)conjugatedmouse
responseswereacquired,amplified,monitored,andanalyzed anti-rabbit,mouseanti-goat,orrabbitanti-mousesecondary
withRM6240BDbiologysignalprocessingsystem(Chengdu antibodies(1:5000,ProteintechGroupInc.,Wuhan,China)
InstrumentFactory,China). by using an enhanced chemiluminescence (ECL) system
(Millipore,Billerica,MA,US).Opticaldensityofthebands
2.6.GolgiSilverStain. Underanesthesia(10%chloralhydrate, was scanned and quantified with NIH ImageJ software and
350mg/kg, ip), rats underwent transcardial perfusion with the results were normalized to the quantity of 𝛽-tubulin in
0.9% saline solution (containing 0.5% sodium nitrite), fol- eachsamplelane.Allassayswererepeatedatleastthreetimes.
lowed by 4% paraformaldehyde (PFA) in 0.1M phosphate
buffer (pH 7.4) for 2h. Then the rats were perfused with a 2.9.QuantitativeReal-TimePCR. Hippocampaltissueswere
mordant dye comprising a mixture of 5% chloral hydrate, processedwithRNAlater(cid:2) Solution(Ambion(cid:2)RNAbyLife
5% potassium dichromate, 4% formaldehyde, and distilled Technologies(cid:3),NewYork,USA)tostabilizeandprotectRNA
water for another 1-2h in dark until toes of the perfused before storage in liquid Nitrogen. Total RNA was extracted
ratsturnedtobeheavytangerine.Brainsweretakenoutand usingTRIzolreagent(Invitrogen,Carlsbad,CA,USA).10ng
coronallysectionedtothreepieces,ofwhichthemiddleones RNA was reverse-transcribed into cDNA with the Toyobo
weresoakedinthemordantagentfor3days.Consequently, First-Strand cDNA Synthesis Kit (USA). U6 RNA served
thespecimenswereimmersedin1.5%silvernitratesolution as an endogenous reference. The reactionswere carried out
for another 3 days. Finally, the brain sections were cut into in triplicate in a StepOne(cid:3) Real-Time PCR System (Life
50𝜇mslicesonavibratome(CampdenInstrument,MA752, Technologies, New York, USA) in a 10𝜇L reaction mixture.
Leicester, UK) and rinsed with 2% potassium dichromate The relative change of miR132 expression was determined
solution,dehydrated,cleared,andcoverslipped. using2−ΔΔCTmethod[24].Theprimersforpre-miR132were
󸀠
Using Olympus BX51 microscope, images of dendrites asfollows:forwardprimer:5-ACCGTGGCTTTCGATTGT-
󸀠 󸀠
of pyramidal neurons in the hippocampal CA1 region were TAC-3 and reverse primer: 5 -TGGTGTCGTGGAGTCG-
traced for at least 50𝜇m from secondary and tertiary 3󸀠, and those for U6 were as follows: forward primer:
4 Evidence-BasedComplementaryandAlternativeMedicine
5󸀠-CCTGCTTCGGCAGCACA-3󸀠 and reverse primer: 5󸀠- HFS: 208.46 ± 5.91% of baseline values). Contrastingly, in
󸀠
AACGCTTCACGAATTTGCGT-3 . themodelgroup,fEPSPslopewassignificantlyreduced(20–
50min after HFS: 144.88 ± 7.46% of baseline values; 50–
2.10.StatisticalAnalysis. Alldatawerepresentedasmean± 80minafterHFS:111.37±4.66%ofbaselinevalues,𝑃<0.01
SEMandstatisticallyanalyzedwithSPSS19.0software(IBM versusthecontrolgroup).EAreversedthe2VO-inducedLTP
Corporation,Somers,NewYork,USA).Theescapelatencies impairment(20–50minafterHFS:204.19±7.58%ofbaseline
inMWMtestweretestedbytwo-wayANOVAwithrepeated values; 50–80min after HFS: 162.45 ± 6.37% of baseline
measures.Theotherdatawereanalyzedbyone-wayANOVA values, 𝑃 < 0.01 versus the model group). The normalized
followed by post hoc test for multiple comparisons among fEPSP slope was significantly decreased in EA+H89 group
the control group, model group, and EA group and were (20–50minafterHFS:144.92±6.50%ofbaselinevalues;50–
evaluated by independent-samples Student’s 𝑡-test between 80minafterHFS:123.52±7.05%ofbaselinevalues;𝑡(60)=
the EA+H89 group and EA+NS group. 𝑃 < 0.05 was −10.534, 𝑃 = 0.000 versus EA+NS group; and 𝑛 = 10 per
consideredasstatisticallysignificant. group,Figures3(c)and3(d))comparedtothatintheEA+NS
group(20–50minafterHFS:207.48±6.42%ofbaselineval-
3.Results ues;50–80minafterHFS:171.56±8.16%ofbaselinevalues).
Theanalysisofdendriticspinedensityshowedasignifi-
3.1. EA Ameliorated Spatial Learning and Memory Deficits cantdifferenceamongthecontrolgroup,modelgroup,and
Induced by 2VO, and the Effect Was Partially Inhibited by EAgroup(𝐹(2,54) = 58.820,𝑃 = 0.000;𝑛 = 6pergroup).
H89. The potential protective effects of EA against spatial Thedendriticspinedensitywasmarkedlydecreasedinmodel
learning and memory deficits of 2VO rats were assessed grouprats(controlrats:11.29±0.52,modelrats:6.21±0.27,
usingtheMWMtest.Theanalysisofescapelatenciesshowed 𝑃<0.01versusthecontrolgroup,Figure3(f)),anditwassig-
significant differences between the control group, model nificantlyimprovedinEAgrouprats(EArats:9.21±0.22,𝑃<
group, and EA group (group effects: 𝐹(2,65) = 19.489, 𝑃 = 0.01versusthemodelgroup).Inaddition,theEA+H89group
0.000; training day effects: 𝐹(4,260) = 13.239, 𝑃 = 0.000; ratshadsignificantlyfewerdendriticspinesthanthoseofthe
and𝑛=10pergroup).Comparedwiththecontrolgroup,the EA+NSgrouprats(5.20±0.44and7.98±0.67,resp.;𝑡(19)=
modelgroupratsexhibitedsignificantlylongerescapelaten- −3.821,𝑃<0.01versustheEA+NSgroup,Figure3(f)).
ciesfromthe2ndtothe5thtimeduringthetrainingperiod
(𝑃 < 0.05, resp., Figure 2(b)). Compared with the model 3.3. Significant Signs of Neuronal Loss Were Absent in Hip-
group,theEAgroupratsshortenedtheescapelatenciesfrom pocampus in Experimental Rats from the Five Groups. Nissl
the3rdtothe5thtimeduringthetrainingperiod(𝑃 < 0.05, stainwasusedtodistinguishviableneuronsfromtheapop-
resp., Figure 2(b)). The analysis of escape latencies showed totic or the neurotic ones. The former exhibited abundant
significant differences between the EA+H89 group and cytoplasmandNisslsubstance(stainedasdarkblue),obvious
EA+NSgroup(groupeffects:𝐹(1,38) = 18.913,𝑃 = 0.000; oval nuclei (stained as light blue), and prominent nucleoli
trainingdayeffects:𝐹(4,152) = 12.944,𝑃 = 0.000;and𝑛 = whiletheapoptoticornecroticcellsexhibitedpyknoticmor-
10pergroup).TheEA+H89groupratsdemonstratedsigni- phologywithamorphousorfragmentednuclei.
ficantly prolonged escape latencies than the EA+NS group Significantsignsofneuronallosswereabsentaccordingto
ratsfromthe2ndtothe5thtimeduringthetrainingperiod theoverallobservationinthehippocampusofexperimental
(𝑃 < 0.05, resp., Figure 2(c)). Typical swimming paths rats from the five groups (Figures 4(a)–4(e)). Considering
in the spatial probe trial of each group were displayed in previous reports that CA1 region is more vulnerable to
Figure2(d).Analyseswithone-wayANOVArevealedasigni- ischemic damage [25], we focused on the neuronaldamage
ficantdifferencebetweenthecontrolgroup,modelgroup,and inCA1pyramidalcelllayer.AsshowninFigures4(a1)–4(e1),
EAgroup(𝐹(2,27) = 5.470,𝑃 = 0.010;𝑛 = 10pergroup). several damaged neurons (black arrows) in CA1 pyramidal
The time spent in the target quadrant was significantly cell layer existed in the sections from the five groups. In
decreased in the model group (𝑃 < 0.01 versus control Figure4(f),quantitativeandstatisticalanalysisofpyramidal
group,Figure2(e)),whileitwassignificantlyincreasedinthe cell viability in CA1 pyramidal cell layer did not reach
EA group (𝑃 < 0.01 versus model group, Figure 2(e)). The significanceneitheramongcontrolrats,modelrats,andEA
EA+H89groupratsexhibitedsignificantlyshortertimethan rats (control: 65.19 ± 1.25, model: 59.70 ± 3.20, and EA:
that of the EA+NS group (𝑡(9) = −2.902, 𝑃 < 0.01 versus 61.34 ± 1.15; 𝐹(2,109) = 3.075, 𝑃 > 0.05; and 𝑛 = 5 per
EA+NSgroup;𝑛=10pergroup,Figure2(f)). group)norbetweenEA+H89ratsandEA+NSrats(EA+H89:
56.07±1.77,EA+NS:56.69±0.93,𝑡(26)=−0.336,𝑃 > 0.05;
3.2. EA Alleviated LTP Impairment and the Reduction of and𝑛=5pergroup).
Dendritic Spine Density of 2VO Rats, and the Effect Was
Partially Inhibited by H89. The analysis of fEPSP slopes 3.4.EAIncreasedtheExpressionofpCREBProtein,BDNFPro-
showed a significant difference among the control group, tein, and miR132 and Decreased the Expression of p250GAP
modelgroup,andEAgroup(𝐹(2,90) = 112.541,𝑃 = 0.000; Protein in the Hippocampus of 2VO Rats and These Effects
𝑛=10pergroup).AsshowninFigures3(a)and3(b),HFSof Were Partially Inhibited by H89. Analyses with one-way
theSchaffercollateralinputstoCA1pyramidalcellsinduced ANOVA revealed a significant main effect of EA (pCREB:
astableLTPintheslopeoffEPSPincontrolrats(20–50min 𝐹(2,17) = 11.766, 𝑃 < 0.01; BDNF: 𝐹(2,19) = 13.820,
afterHFS:252.22±7.98%ofbaselinevalues;50–80minafter 𝑃=0.000;miR132:𝐹(2,33)=9.106,𝑃<0.01;andp250GAP:
Evidence-BasedComplementaryandAlternativeMedicine 5
2VO EA
Electrophysiological recording
Time (days) Morphological analysis
0 1d 2d 3d 4d 5d 6d 7d 8d Biochemical analysis
MWM test
(a)
60 60
∗
∗ &
&
&
Escape latency (s) 2400 # ∗# ∗# Escape latency (s) 2400 &
0 0
1 2 3 4 5 1 2 3 4 5
(Times) (Times)
Control EA EA+H89
Model EA+NS
(b) (c)
II I II I II I II I II I
III IV III IV III IV III IV III IV
Control Model EA EA+H89 EA+NS
(d)
30 30
me spent in target quadrant (s) 112250505 ∗∗ ## me spent in target quadrant (s) 112250505 &&
Ti Ti
0 0
Control Model EA EA+H89 EA+NS
Control
Model EA+H89
EA EA+NS
(e) (f)
Figure2:EApreventedspatiallearningandmemorydeficitsinducedby2VO,andtheeffectwaspartiallyinhibitedbyH89.(a)Timeline
oftheexperimentstoshowtheEAtherapyanddetection.(bandc)Escapelatencytofindthehiddenplatformfromthe1sttothe5thtime.
Ratsinthemodelgrouphadlongerescapelatenciesthanthecontrolgroup(b).EAcouldreversethespatiallearningimpairmentinduced
by2VO.RatsintheEA+H89grouphadlongerescapelatencythantheEA+NSgrouprats(c).H89couldpartlyreversethebeneficialeffect
ofEA.(d)Thetypicalswimmingpathsofeachgroupinthespatialprobetrial.(eandf)Timespentinthetargetquadrantintheprobetest.
Theamountoftimeexpendedinthetargetquadrantwassignificantlyshorterinthemodelgroupratsthaninthecontrolgrouprats,and
ratstreatedwithEAexhibitedsignificantlymoretimeinthetargetquadrant(e).RatsintheEA+H89groupspentashortertimeinthetarget
quadrantcomparedtoEA+NSgrouprats(f).Eachvaluerepresentsmean±SEM,𝑛 = 10foreachgroup.∗∗𝑃 < 0.01,comparedwiththe
controlgroup;∗𝑃 < 0.05,comparedwiththecontrolgroup;##𝑃 < 0.01,comparedwiththemodelgroup;#𝑃 < 0.05,comparedwiththe
modelgroup;&&𝑃<0.01,comparedwiththeEA+NSgroup;and&𝑃<0.05,comparedwiththeEA+NSgroup.
6 Evidence-BasedComplementaryandAlternativeMedicine
400 300
)ne )ne
% of baseli 230000 % of baseli 200 ∗∗ ## ##
P slope ( 100 HFS P slope ( 100 ∗∗
PS PS
E E
f 0 f 0
−20 0 20 40 60 80 20–50 50–80
(min) (min)
Control Control
Model Model
EA EA
(a) (b)
300 300
)ne )ne
baseli 200 baseli 200
ope (% of 100 HFS ope (% of 100 && &&
P sl P sl
PS PS
E E
f 0 f 0
−20 0 20 40 60 80 20–50 50–80
(min) (min)
EA+H89 EA+H89
EA+NS EA+NS
(c) (d)
20
m
𝜇
0
1
/ 15
y
nsit
e
d
es 10 ##
n
pi
d s ∗∗ &
e
aliz 5
m
or
N
m 0
𝜇
0
1
Control EA+H89
Model EA+NS
Control Model EA EA+H89 EA+NS EA
(e) (f)
Figure3:EAalleviatedLTPimpairmentandthereductionofdendriticspinedensityof2VOrats,andtheeffectswerepartiallyinhibitedby
H89.(aandc)ThelineargraphofthenormalizedfEPSPslope.ThedownwardfilledarrowindicatesHFS.(bandd)Thehistogramofaverage
fEPSPslopeat20–50minand50–80minafterHFS.ThenormalizedfEPSPslopewassignificantlydecreasedinthemodelgroup.EAreversed
the2VO-inducedLTPimpairment,andthiseffectwaspartlyinhibitedbyH89.Eachvaluerepresentsmean±SEM,𝑛=10foreachgroup.(e)
RepresentativedendriticsegmentsofCA1pyramidalneuronsinthehippocampus(1000x,scalebar=10𝜇m).(f)Normalizedmeandendrite
spinedensitycounts.EAattenuates2VO-induceddendriticspinelossofCA1regioninthehippocampus,andthiseffectwaspartlyinhibited
byH89.Eachvaluerepresentsmean±SEM,𝑛=6foreachgroup.∗∗𝑃<0.01,comparedwiththecontrolgroup;##𝑃<0.01,comparedwith
themodelgroup;&&𝑃<0.01,comparedwiththeEA+NSgroup;and&𝑃<0.05,comparedwiththeEA+NSgroup.
Evidence-BasedComplementaryandAlternativeMedicine 7
CA1
CA1
500𝜇m Control 50𝜇m Control EA+H89 EA+H89
(a) (a1) (d) (d1)
CA1
CA1
Model Model EA+NS EA+NS
(b) (b1) (e) (e1)
100
ns
CA1 uro%) 80
nen (
of viable A1 regio 60
atios in C 40
R
EA EA 20
(c) (c1)
Control EA+H89
Model EA+NS
EA
(f)
Figure4:Significantsignsofneuronallosswereabsentinhippocampusinexperimentalratsfromthefivegroups.Photomicrographsof
toluidineblue-stainedsectionsfromcontrolgroup(aanda1),modelgroup(bandb1),EAgroup(candc1),EA+H89group(dandd1),and
EA+NSgroup(eande1).(a–e)40x,scalebar=500𝜇m.(a1–e1)400x,scalebar=50𝜇m.Significantsignsofneuronallosswereabsentbasedon
theoverallobservationofhippocampusinexperimentalratsfromthefivegroups.Severaldamagedneurons(blackarrows)inCA1pyramidal
celllayerwereobserved.(f)QuantitativeanalysisofpyramidalcellviabilityinCA1pyramidalcelllayerwasnotsignificantlydifferentneither
amongthecontrolrats,themodelrats,andtheEArats(control:65.19±1.25,model:59.70±3.20,andEA:61.34±1.15;𝐹(2,109)=3.075,
𝑃 > 0.05;and𝑛 = 5pergroup)norbetweentheEA+H89ratsandtheEA+NSrats(EA+H89:56.07±1.77,EA+NS:56.69±0.93;𝑡(26)=
−0.336,𝑃>0.05;and𝑛=5pergroup).Eachvaluerepresentsmean±SEM.
𝐹(2,17) = 10.539,𝑃 < 0.01;and𝑛 = 5pergroup).Asillus- 4.Discussion
trated in Figure 5, the expressions of the pCREB protein,
BDNF protein, and miR132 were decreased 7 days after the ThepresentstudyhasrevealedthatEAatGV20andGV14for
2VOoperation,accompaniedbyanincreasedexpressionof 7 days could significantly ameliorate learning and memory
p250GAP protein (pCREB, BDNF, miR132, and p250GAP: deficitsandLTPimpairmentintheSchaffercollateralpath-
𝑃 < 0.01,resp.,versusthecontrolgroup).TreatmentofEA way.Inaddition,itcouldalsorestorethedendriticspineloss
impededthereductionofpCREBprotein,BDNFprotein,and in the hippocampal CA1 region in cerebral hypoperfusion
miR132 expression in 2VO rats, as well as suppressing the rats.PKA/CREBsignalingpathwaywaspotentiallyinvolved
increase of p250GAP protein (pCREB, BDNF, miR132, and in the neuroprotective effects, including the regulation of
p250GAP:𝑃 < 0.05,resp.,versusthemodelgroup).More- proteinssuchaspCREB,BDNF,p250GAP,andmiR132.
over, this protective effect of EA was partially reversed in The 2VO rat model is widely used to study the chronic
EA+H89grouprats,whencomparedwiththeratsinEA+NS cerebral hypoperfusion- (CCH-) related cognitive impair-
group(pCREB:𝑡(4)=−3.290,𝑃<0.05;BDNF:𝑡(7)=−2.711, ment. The subchronic phase of this model was defined as 3
𝑃 < 0.05;miR132:𝑡(17)=−2.109,𝑃 < 0.05;andp250GAP: days–8weeksaftertheocclusion,causingadramaticreduc-
𝑡(7)=2.722,𝑃 < 0.05,versustheEA+NSgroup;and𝑛 = 5 tionincerebralbloodflow(CBF)andhypoxic-ischemiccon-
pergroup,Figure5). ditions[26].EveniftheCBFgraduallyreturnstothebaseline
8 Evidence-BasedComplementaryandAlternativeMedicine
pCREB BDNF
𝛽-Tubulin 𝛽-Tubulin
Control Model EA 89EA+H EA+NS Control Model EA 89EA+H EA+NS
EB 0.6 NF 0.6 2.0
Relative band intensity of pCR 000...024 ∗∗ # & Relative band intensity of BD 000...024 ∗∗ # & Fold change of miR132 0011....0505 ∗∗ ## &
Control EA+H89 Control EA+H89 Control EA+H89
Model EA+NS Model EA+NS Model EA+NS
EA EA EA
(a) (b) (c)
p250GAP
𝛽-Tubulin
Control Model EA 89A+H EA+NS
E
P
A 0.6
G
0
5
2
y of p 0.4 ∗∗ &
nsit
e
nt
d i 0.2 #
n
a
b
e
v
ati 0.0
el
R
Control EA+H89
Model EA+NS
EA
(d)
Figure5:EAincreasedtheexpressionofthepCREBprotein,BDNFprotein,andmiR132anddecreasedtheexpressionofp250GAPproteinin
thehippocampusof2VOratsandtheseeffectswerepartiallyinhibitedbyH89.(a)ImmunoblotanalysisforpCREBproteininhippocampus
andhistogramofitsrelativebandintensity.(b)ImmunoblotanalysisofBDNFproteininhippocampusandhistogramofitsrelativeband
intensity. (c) Histogram of fold change of miR132 in the hippocampus. (d) Immunoblot analysis for p250GAP protein in hippocampus
andhistogramofitsrelativebandintensity.QuantitativeanalysisdemonstratedthathippocampalpCREB,BDNF,andmiR132levelswere
decreasedby2VOandincreasedbyEAtreatment.Concurrently,EAreducedtheelevationofp250GAPlevelinthehippocampusof2VO
rats.TheseprotectiveeffectswereatleastpartiallyinhibitedbyH89administration.Eachvaluerepresentsmean±SEM,𝑛=5foreachgroup.
∗∗𝑃<0.01,comparedwiththecontrolgroup;##𝑃<0.01,comparedwiththemodelgroup;#𝑃<0.05,comparedwiththemodelgroup;and
&𝑃<0.05,comparedwiththeEA+NSgroup.
withthecompensatorymechanismofvascularplasticity,such to the dendritic spines loss in the hippocampus, which
cerebral hypoperfusion may be a risk factor for subsequent contributed to the learning and memory dysfunctions in
neuronal degeneration and cognitive impairment [27]. In 2VOrats[29].Therefore,interventionsthatreverse,atleast
Schaffer-CA1synapses,theLTPwassignificantlyinhibitedat partly,earlyneuropathologicaloutcomesofCCHmightbean
both1dayand4daysafterclampingofthebilateralcommon effectivewaytopreventtheprogressivedeclineofcognitive
carotidarteries;thatis,aso-calleddelayeddysfunctionmight functions. In this study, EA could significantly reverse the
existinthehippocampalneurons[28].Consecutively,itwas LTPimpairmentanddendriticspineloss,indicatingaben-
demonstratedthatthesynaptictransmissionreductionofthe eficialroleinimprovingtheneuronalsynaptictransmission
hippocampus and prefrontal cortex was partially attributed andfunctionalreconstructionincerebralhypoperfusionrats.
Evidence-BasedComplementaryandAlternativeMedicine 9
Significant pathological changes in the brain may be rats. These results revealed the important role of miR132 in
absent at early stages of bilateral common carotid ligation. theimprovementoflearningandmemorybyEA.
A previous study had shown that the pyramidal cell loss in AgroupofratswerepregivenH89[45]throughicvinjec-
theunilateralhippocampalCA1regionwasonlyobservedin tionandthenreceivedoperationof2VOandEAtreatments.
a few 2VO rats (1/6) on day 7 after the surgical procedure InterestingresultswerefoundthatH89administrationcould
[30]. Similarly, it was reported that 2VO did not influence reverse the beneficial effects of EA on spatial learning and
the hippocampal CA1 pyramidal cell number or density of memory, including LTP facilitation and restoration of den-
glial fibrillary acidic protein (GFAP) 14 days after surgery, driticspines.Simultaneously,theeffectsofEAonthemolec-
though a late-emerging CA1 cell loss was detected at 190 ular level of pCREB protein, BDNF protein, miR132, and
days after the operation [31]. Together, these findings have p250GAP protein were also blocked by H89 in EA-treated
prompted a consensus that initial 2VO does not cause 2VOrats.Alltheseresultsimpliedapotentialinvolvementof
significantneuronalloss.Consistentwiththesereports,our PKA/CREBsignalpathwayintheeffectsofEAoncognitive
experimentaloutcomesalsodetectedminimaldifferencesin function and hippocampal synaptic plasticity in cerebral
viablehippocampalneuronsamongtheexperimentalrats. hypoperfusion.
CREB is best known for its roles in learning and
memory. Over the past few decades, accumulated evidence 5.Conclusions
established that pCREB-mediated gene transcription and
protein synthesis contributed to the development of long- Takentogether,ourdataindicatedthatEAcouldameliorate
termmemory[32].Thiswasbroughtaboutbyparticipatingin learning and memory deficits and alleviate hippocampal
processessuchaslong-termpotentiationofsynapticstrength synapticplasticityimpairmentofcerebralhypoperfusionrats,
[15, 32] and structural synaptic changes. In a gerbil global withaprobableregulationofPKA/CREBsignalingpathway,
ischemiamodel,preconditioningischemiainducedischemic including the expression of pCREB protein, BDNF protein,
tolerance by a transient increase of pCREB and subsequent miR132,andp250GAPprotein.
upregulation of BCL-2 expression [33]. However, ongoing
CompetingInterests
chronicischemiareducedtheexpressionofpCREB,resulting
intheselectivevulnerabilityofCA1pyramidalcellsat48and
The authors declare that there is no conflict of interests
72hfollowingmildhypoxic-ischemic(HI)injuryinrats[25].
regardingthepublicationofthispaper.
In the present study, we found that EA could increase the
expressionofthepCREBproteininthehippocampusof2VO
Acknowledgments
rats,suggestingthatEAmightimprovelearningandmemory
through activation of CREB mediating gene transcription
ThisworkwassupportedbytheresearchgrantfromNational
andproteinsynthesis.
NaturalScienceFoundationofChina(NSFC,no.81171858).
SufficientevidenceshavesuggestedthatCREBisamajor
TheauthorsaregratefultoPeng-ChengHuangandPanLuo
regulator of BDNF-induced neuronal responses. On one
(Department of Pharmacology, School of Basic Medicine,
hand, the transcription and translation of BDNF gene are
TongjiMedicalCollege,HuazhongUniversityofScienceand
CREB-dependent [34, 35]; on the other hand, BDNF could
Technology) for their expert technical assistance in elec-
stimulateSer133phosphorylationandactivateCREBthrough
trophysiologicalrecordings.
CaMKIV [36] and RSKs [37]. Thus, a positive feedback
loop might have connected BDNF and CREB. Meanwhile, References
BDNF was demonstrated to play a crucial role in synapto-
genesis [38] andLTP[39, 40]. Moreover,itwas reportedto [1] The Lancet Neurology, “G8 dementia summit: a chance for
enhance high-frequency transmission [41] and mediate the unitedaction,”TheLancetNeurology,vol.13,no.1,p.1,2014.
redistributionofthesynapticproteinswithinthepresynaptic [2] S.Zhang,B.Wu,M.Liuetal.,“Acupunctureefficacyonischemic
terminals.Similarfindingswereobservedinthisstudy:that stroke recovery: multicenter randomized controlled trial in
hippocampal BDNF protein expression was decreased after China,”Stroke,vol.46,no.5,pp.1301–1306,2015.
the 2VO procedure and increased after the EA treatment, [3] J.Zhou,W.Peng,M.Xu,W.Li,andZ.Liu,“Theeffectiveness
accompanied by alleviation of LTP impairment and partial andsafetyofacupunctureforpatientswithalzheimerdisease:a
systematicreviewandmeta-analysisofrandomizedcontrolled
restorationofdendriticspinedensity.
trials,”Medicine,vol.94,no.22,p.e933,2015.
We also observed another newly recognized CREB-
[4] Z. Wang, P. Liang, Z. Zhao et al., “Acupuncture modulates
driventarget,miR132,whichwasindicatedinmediatingden-
restingstatehippocampalfunctionalconnectivityinAlzheimer
driticplasticitythroughsuppressingtranslationofp250GAP,
disease,”PLoSONE,vol.9,no.3,ArticleIDe91160,2014.
a member of the Rac/Rho family of GAPs [21, 42]. As
[5] H.Zhang,L.Zhao,S.Yangetal.,“Clinicalobservationoneffect
reported, the transcription of miR132 was induced in
of scalp electroacupuncture for mild cognitive impairment,”
response to neurotrophins [43] or synaptic activity [44].
JournalofTraditionalChineseMedicine,vol.33,no.1,pp.46–
BDNF could regulate miR132 transcription via the ERK1/2
50,2013.
signaling, together with kinase MSK1 and the phosphory-
[6] Y. Feng, L. Bai, Y. Ren et al., “FMRI connectivity analysis of
lation of CREB [43]. Herein, we found that EA altered the acupunctureeffectsonthewholebrainnetworkinmildcogni-
morphologyofdendriticspines,accompaniedbyanincrease tiveimpairmentpatients,”MagneticResonanceImaging,vol.30,
ofmiRNA-132andareductionofp250GAPproteinin2VO no.5,pp.672–682,2012.
10 Evidence-BasedComplementaryandAlternativeMedicine
[7] J. Yu, X. Zhang, C. Liu, Y. Meng, and J. Han, “Effect of [23] X.Han,X.Huang,Y.Wang,andH.Chen,“Astudyofastrocyte
acupuncture treatment on vascular dementia,” Neurological activationintheperiinfarctregionaftercerebralischemiawith
Research,vol.28,no.1,pp.97–103,2006. electroacupuncture,” Brain Injury,vol.24, no.5,pp.773–779,
[8] Q.-Q. Li, G.-X. Shi, J.-W. Yang et al., “Hippocampal cAMP/ 2010.
PKA/CREBisrequiredforneuroprotectiveeffectofacupunc- [24] K. J. Livak and T. D. Schmittgen, “Analysis of relative gene
ture,”PhysiologyandBehavior,vol.139,pp.482–490,2015. expression data using real-time quantitative PCR and the
2(−DeltaDeltaC(T))Method,”Methods,vol.25,no.4,pp.402–
[9] Y.ZhuandY.Zeng,“Electroacupunctureprotectedpyramidal
408,2001.
cellsinhippocampalCA1regionofvasculardementiaratsby
inhibitingtheexpressionofP53andNoxa,”CNSNeuroscience [25] M. Walton, E. Sirimanne, C. Williams, P. Gluckman, and M.
andTherapeutics,vol.17,no.6,pp.599–604,2011. Dragunow, “The role of the cyclic AMP-responsive element
bindingprotein(CREB)inhypoxic-ischemicbraindamageand
[10] T.Wang,C.-Z.Liu,J.-C.Yu,W.Jiang,andJ.-X.Han,“Acupunc-
repair,”MolecularBrainResearch,vol.43,no.1-2,pp.21–29,1996.
ture protected cerebral multi-infarction rats from memory
[26] E. Farkas, P. G. M. Luiten, and F. Bari, “Permanent, bilateral
impairment by regulating the expression of apoptosis related
commoncarotidarteryocclusionintherat:amodelforchronic
genesBcl-2andBaxinhippocampus,”PhysiologyandBehavior,
cerebral hypoperfusion-related neurodegenerative diseases,”
vol.96,no.1,pp.155–161,2009.
BrainResearchReviews,vol.54,no.1,pp.162–180,2007.
[11] J. Yu, C. Liu, X. Zhang, and J. Han, “Acupuncture improved
[27] Z.Jing,C.Shi,L.Zhuetal.,“Chroniccerebralhypoperfusion
cognitiveimpairmentcausedbymulti-infarctdementiainrats,”
induces vascular plasticity and hemodynamics but also neu-
PhysiologyandBehavior,vol.86,no.4,pp.434–441,2005.
ronaldegenerationandcognitiveimpairment,”JournalofCere-
[12] E.R.Kandel,“Themolecularbiologyofmemory:cAMP,PKA, bralBloodFlowandMetabolism,vol.35,no.8,pp.1249–1259,
CRE, CREB-1, CREB-2, and CPEB,” Molecular Brain, vol. 5, 2015.
article14,2012.
[28] K.Mori,M.Yoshioka,N.Sudaetal.,“Anincompletecerebral
[13] K. Sakamoto, K. Karelina, and K. Obrietan, “CREB: a multi- ischemiaproducedadelayeddysfunctionintherathippocam-
facetedregulatorofneuronalplasticityandprotection,”Journal palsystem,”BrainResearch,vol.795,no.1-2,pp.221–226,1998.
ofNeurochemistry,vol.116,no.1,pp.1–9,2011. [29] H.Jia,X.M.Zhang,B.A.Zhang,Y.Liu,andJ.M.Li,“Den-
[14] J. Viosca, M. Lopez de Armentia, D. Jancic, and A. Barco, driticmorphologyofneuronsinmedialprefrontalcortexand
“Enhanced CREB-dependent gene expression increases the hippocampusin2VOrats,”NeurologicalSciences,vol.33,no.5,
excitability of neurons in the basal amygdala and primes the pp.1063–1070,2012.
consolidation ofcontextual andcued fear memory,”Learning [30] H. Ohta, H. Nishikawa, H. Kimura, H. Anayama, and M.
andMemory,vol.16,no.3,pp.193–197,2009. Miyamoto, “Chronic cerebral hypoperfusion by permanent
[15] A.Barco,J.M.Alarcon,andE.R.Kandel,“Expressionofcon- internalcarotidligationproduceslearningimpairmentwithout
stitutivelyactiveCREBproteinfacilitatesthelatephaseoflong- braindamageinrats,”Neuroscience,vol.79,no.4,pp.1039–1050,
termpotentiationbyenhancingsynapticcapture,”Cell,vol.108, 1997.
no.5,pp.689–703,2002. [31] B.A.Pappas,J.C.DeLaTorre,C.M.Davidson,M.T.Keyes,
[16] M.Mizuno,K.Yamada,N.Maekawa,K.Saito,M.Seishima,and andT.Fortin,“Chronicreductionofcerebralbloodflowinthe
T.Nabeshima,“CREBphosphorylationasamolecularmarker adultrat:late-emergingCA1celllossandmemorydysfunction,”
ofmemoryprocessinginthehippocampusforspatiallearning,” BrainResearch,vol.708,no.1-2,pp.50–58,1996.
BehaviouralBrainResearch,vol.133,no.2,pp.135–141,2002. [32] R.Bourtchuladze,B.Frenguelli,J.Blendy,D.Cioffi,G.Schutz,
andA.J.Silva,“Deficientlong-termmemoryinmicewithatar-
[17] T.Abel,P.V.Nguyen,M.Barad,T.A.S.Deuel,E.R.Kandel,and
getedmutationofthecAMP-responsiveelement-bindingpro-
R.Bourtchouladze,“GeneticdemonstrationofaroleforPKA
tein,”Cell,vol.79,no.1,pp.59–68,1994.
inthelatephaseofLTPandinhippocampus-basedlong-term
memory,”Cell,vol.88,no.5,pp.615–626,1997. [33] T.Mabuchi,K.Kitagawa,K.Kuwabaraetal.,“Phosphorylation
of cAMP response element-binding protein in hippocampal
[18] M.R.WaltonandM.Dragunow,“IsCREBakeytoneuronal
neuronsasaprotectiveresponseafterexposuretoglutamatein
survival?” Trends in Neurosciences, vol. 23, no. 2, pp. 48–53,
vitroandischemiainvivo,”TheJournalofNeuroscience,vol.21,
2000.
no.23,pp.9204–9213,2001.
[19] B.LuandW.Gottschalk,“Modulationofhippocampalsynaptic
[34] X. Tao, S. Finkbeiner, D. B. Arnold, A. J. Shaywitz, and M.
transmissionandplasticitybyneurotrophins,”ProgressinBrain 2+
E.Greenberg,“Ca influxregulatesBDNFtranscriptionbya
Research,vol.128,pp.231–241,2000.
CREBfamilytranscriptionfactor-dependentmechanism,”Neu-
[20] B.LuandA.Chow,“Neurotrophinsandhippocampalsynaptic ron,vol.20,no.4,pp.709–726,1998.
transmissionandplasticity,”JournalofNeuroscienceResearch, [35] P. B. Shieh, S.-C. Hu, K. Bobb, T. Timmusk, and A. Ghosh,
vol.58,no.1,pp.76–87,1999. “Identificationofasignalingpathwayinvolvedincalciumregu-
[21] N.Vo,M.E.Klein,O.Varlamovaetal.,“AcAMP-responseele- lationofBDNFexpression,”Neuron,vol.20,no.4,pp.727–740,
ment binding protein-induced microRNA regulates neuronal 1998.
morphogenesis,” Proceedings of the National Academy of Sci- [36] S. Finkbeiner, S. F. Tavazoie, A. Maloratsky, K. M. Jacobs, K.
encesoftheUnitedStatesofAmerica,vol.102,no.45,pp.16426– M.Harris,andM.E.Greenberg,“CREB:amajormediatorof
16431,2005. neuronal neurotrophin responses,” Neuron, vol. 19, no. 5, pp.
[22] X.Han,X.Zhao,M.Luetal.,“Electroacupunctureameliorates 1031–1047,1997.
learningandmemoryviaactivationoftheCREBsignalingpath- [37] J. Xing, D. D. Ginty, and M. E. Greenberg, “Coupling of the
way in the hippocampus to attenuate apoptosis after cerebral RAS-MAPK pathway to gene activation by RSK2, a growth
hypoperfusion,” Evidence-Based Complementary and Alterna- factor-regulatedCREBkinase,”Science,vol.273,no.5277,pp.
tiveMedicine,vol.2013,ArticleID156489,8pages,2013. 959–963,1996.