estrogen hippocampus

Journal of Steroid Biochemistry & Molecular Biology 131 (2012) 37 51

Contents lists available at SciVerse ScienceDirect

Journal of Steroid Biochemistry and Molecular Biology

jo u r n al hom epage: www.elsev ier .com/ locate / j sbmb

Review

Modulation of synaptic plasticity in the hippocampus by hippocampus-derivedestrogen and androgen

Yuuki OoGen Mura Department ob Core Researchc Bioinformaticd Project of Spe

a r t i c l

Article history:Received 30 June 2011Received in revised form27 September 2011Accepted 12 October 2011

Keywords:EstradiolEstrogenAndrogenNeurosteroidSynaptic plastEstrogen recepHippocampusSpineLTDLTP

The hippocampus synthesizes estrogen and androgen in addition to the circulating sex steroids. Synapticmodulation by hippocampus-derived estrogen or androgen is essential to maintain healthy memoryprocesses. Rapid actions (12 h) of 17-estradiol (17-E2) occur via synapse-localized receptors (ER orER), while slow genomic E2 actions (648 h) occur via classical nuclear receptors (ER or ER). The long-term potentiation (LTP), induced by strong tetanus or theta-burst stimulation, is not further enhanced

Contents

1. Introd2. Modu

2.1. 2.2. 2.3.

3. Synth3.1. 3.2. 3.3.

Abbreviatidehydroepiandehydrogenaspostsynaptic d This article Correspon

Tokyo 153-890E-mail add

0960-0760/$ doi:10.1016/j.icitytor

by E2 perfusion in adult rats. Interestingly, E2 perfusion can rescue corticosterone (stress hormone)-induced suppression of LTP. The long-term depression is modulated rapidly by E2 perfusion. Elevationof the E2 concentration changes rapidly the density and head structure of spines in neurons. ER, butnot ER, drives this enhancement of spinogenesis. Kinase networks are involved downstream of ER.Testosterone (T) or dihydrotestosterone (DHT) also rapidly modulates spinogenesis. Newly developedSpiso-3D mathematical analysis is used to distinguish these complex effects by sex steroids and kinases.

It has been doubted that the level of hippocampus-derived estrogen and androgen may not be highenough to modulate synaptic plasticity. Determination of the accurate concentration of E2, T or DHT in thehippocampus is enabled by mass-spectrometric analysis in combination with new steroid-derivatizationmethods. The E2 level in the hippocampus is approximately 8 nM for the male and 0.52 nM for thefemale, which is much higher than that in circulation. The level of T and DHT is also higher than thatin circulation. Taken together, hippocampus-derived E2, T, and DHT play a major role in modulation ofsynaptic plasticity.

2011 Elsevier Ltd. All rights reserved.

uction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38lation of synaptic plasticity by hippocampal sex steroids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Modulation of long-term potentiation (LTP) and long-term depression (LTD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Spinogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40Synaptic (extranuclear) receptors for estrogen and androgen (Fig. 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

esis of sex steroids in the hippocampus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Pathway of synthesis (Fig. 7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Developmental and age-related change of steroidogenic enzymes or sex-steroid receptors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Concentration of estrogen and androgen (Fig. 9) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

ons: ACSF, articial cerebrospinal uid; AMPA, -amino-3-hydroxy-5-methylisoxazole-4-propionic acid; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; DHEA,drosterone; DPN, diarylpropionitrile; DHT, dihydrotestosterone; E2, 17-estradiol; E1, estrone; GPR30, G protein coupled receptor 30; HSD, hydroxysteroide; LCMS/MS, liquid chromatographytandem mass spectrometry; LTD, long-term depression; LTP, long-term potentiation; NMDA, N-methyl-d-aspartate; PSD,ensity; PREG, pregnenolone; PPT, (propyl-pyrazole-triyl)tris-phenol; RIA, radioimmunoassay; StAR, steroidogenic acute regulatory protein; T, testosterone.

is part of a Special Issue entitled Neurosteroids.ding author at: Department of Biophysics and Life Sciences, Graduate School of Arts and Sciences, The University of Tokyo, 3-8-1 Komaba, Meguro,2, Japan. Tel.: +81 3 5454 6517; fax: +81 3 5454 6517.resses: [email protected], [email protected] (S. Kawato).

see front matter 2011 Elsevier Ltd. All rights reserved.jsbmb.2011.10.004ishia, Suguru Kawatoa,b,c,d,, Yasushi Hojoa,b,c, Yusuke Hatanakaa, Shimpei Higoa,akamia,c, Yoshimasa Komatsuzakia, Mari Ogiue-Ikedaa,d, Tetsuya Kimotoa, Hideo Mukaia,b,c

f Biophysics and Life Sciences, Graduate School of Arts and Sciences, The University of Tokyo, 3-8-1 Komaba, Meguro, Tokyo 153-8902, Japan for Evolutional Science and Technology Project of Japan Science and Technology Agency, The University of Tokyo, Japans Project of Japan Science and Technology Agency, The University of Tokyo, Japancial Coordinate Funds for Promoting Science and Technology, The University of Tokyo, Japan

e i n f o a b s t r a c t

38 Y. Ooishi et al. / Journal of Steroid Biochemistry & Molecular Biology 131 (2012) 37 51

4. Difference between classical slow genomic modulation and rapid synaptic modulation by estrogen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484.1. Slow action via genomic pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484.2. Gene transcription as a downstream event of rapid synaptic action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

5. Model explanation of modulation of synaptic plasticity by sex steroids in relation to synaptocrine and intracrine mechanisms (Fig. 10) . . . . . . 49Refer . . . . . .

1. Introdu

Occurrenadult hippomemory prcadian rhytsynaptic sygen could opotentiatio

Not onlspines havesynaptic pbecause synsynaptic str

For decmones havehormones bloodbrainor electropreplacemensynapses orupon estrogslow modu[611]. Thegen receptThe slow/greplacemen

The acutlating spineslices [6,7,1via synapti[16,18]. Sininto nuclei genomic pr

These accesses, favogonadal hothe brain. Rsteroid prousage withiThis intracrhigh expressupply sterhippocampmay be theand cogniti

2. Modulasteroids

2.1. Moduladepression (

Nanomosynaptic traelectrophys

Interestisynaptic po

. In apprM E2t CA3s, weasal

handpoc6,20,servrague

on tve efold (to thinvesr LToxy-

kinpproductiss is2 ind

acut. Stresupp

E2,hyl-dlyze d supssionect b

MAd NMntag

wereidly ssionlvediol aray, pationssionn of Aicallycopind LTncenem

butt mic

impts thrmedences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ction

ce of local synthesis of estrogen and androgen in thecampus supports estrogen-dependent regulation ofocesses which occur rapidly [15]. Different from cir-hm-dependent synthesis that occurs in ovary or testis,nthesis (transient and rapid) of estrogen and andro-ccur dependent on synaptic events including long-termn (LTP) and long-term depression (LTD) [15].y electrophysiological properties but also dendritic

been studied in relation to memory processes andlasticity which are regulated by neurotransmitters,apse is a site of memory storage and spine is a post-ucture.ades, neuromodulatory actions of gonadal sex hor-

been extensively investigated, because circulating sexcan penetrate into the hippocampus by crossing the

barrier. Genomic slow modulation of synaptogenesishysiological properties is investigated by estrogent for ovariectomized female rats [611]. An increase of

an enhancement of synaptic transmission is observeden replacement of ovariectomized animals. Genomiclation of spines is also observed in slice culturesse slow genomic effects are mediated via nuclear estro-ors ER/ER which initiate transcription processes.enomical modulation of NR2B by 17-estradiol (E2)t enhances LTP in ovariectomized rat [12,13].e/rapid effect of E2 (within 12 h) also occurs by modu-

density or synaptic transmission of the hippocampal417]. Acute modulation of synapses by E2 occursc ER/ER which drives kinases in their downstreamce kinases not only work within synapses but also moveto drive gene transcription, acute E2 effects also driveocesses which may slowly enhance synaptic contacts.ute modulations, relating to memory formation pro-r locally synthesized steroids rather than circulatingrmones which travel a long distance before reachingather than being a limiting factor, a weak activity of sexduction in the hippocampus is sufcient for the localn small volume of neurons (i.e., an intracrine system).ine system contrasts with the endocrine organs in whichsion levels of steroidogenic enzymes are necessary tooids to many other organs via the blood circulation. Forus-derived sex hormones, one of the essential functions

rapid and repetitive modulation of synaptic plasticityve functions, in addition to genomic slow modulation.

tion of synaptic plasticity by hippocampal sex

tion of long-term potentiation (LTP) and long-termLTD)

lar level of E2 exerts an acute inuence (0.51 h) on the

of rats350 g (110 nLTP) atar ratrapid bother the hiprats [1also obold Speffectsnot haweek-down these used fo3-hydrby CaMtion (aLTP-in

StreCan Eduringeffectsrapid of 1 nMN-metto anainducesupprecue effby ErkLTP an(GR) acessesE2 rapsuppreis invoestradpathwphorylsupprerylatiosynaptmicrosweakeCa2+ co

In manism)Mutanshownsuggesries fonsmission of hippocampal slices, as demonstrated byiological investigations.ngly, the effects of E2 on the basal excitatory post-tential (EPSP) or LTP may strongly depend on the age

informationwas rapidlhippocampare perform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

the hippocampus from 4- to 6-week-old or 200 toox. 68 week-old) Sprague-Dawley rats, perfusion of

rapidly increases the basal EPSP (thereby enhancesCA1 synapses [6,7,19]. Concerning 4 week-old Wis-

sometimes (less than 20% probability) observed the EPSP elevation upon E2 perfusion [14,20,21]. On the

110 nM E2 did not affect the basal EPSP and LTP inampus from 12 week-old (3 month-old) adult Wistar22]. No rapid basal EPSP elevation upon E2 perfusion ised in case of 35 month-old as well as 1824 month--Dauley rats [23]. Therefore, E2 may have signicanthe basal EPSP of younger rats (48 weeks old) and mayfect on older adult rats (12 weeks-old or elder). In 34early puberty) rats, E2 even suppresses LTP-inductione similar level to that of 12 week-old rats [9,22]. Intigations, high-frequency tetanic stimulation (1 s) isP-induction, in which the phosphorylation of -amino-5-methylisoxazole-4-propionic acid (AMPA) receptorsase II is a dominant process at high Ca2+ concentra-x. 10 M). E2 may not play an important role on theon processes during tetanic stimulation.

known to suppress the function of the hippocampus.uce neuroprotective effects on synaptic transmissione stress? may be more interesting topic than E2 alonessful level of corticosterone (CORT) at 1 M inducesression of LTP at CA3CA1 synapses. Upon perfusion

LTP-suppression by CORT was abolished (Fig. 1) [24].-aspartate (NMDA) receptor-derived EPSP was usedthe mechanisms of these events. E2 abolished CORT-pression of NMDA-R-EPSP (Fig. 1). This CORT-induced

was abolished by calcineurin inhibitor, and the res-y E2 on the CORT-induced suppression was inhibitedP kinase inhibitor. The CORT-induced suppressions ofDA-R-EPSP were abolished by glucocorticoid receptor

onist, and the restorative effects by E2 on these pro- mimicked by ER and ER agonists. Taken together,rescues LTP and NMDA-R-EPSP from CORT-induceds. Synaptic GR (see Fig. 6E) calcineurin pathway

in these suppressive effects. The rescue effects bye driven via synaptic ER or ER Erk MAP kinaserobably induces activation of NMDA receptor by phos-

of NR2B rescue of Ca2+ inux from CORT-induced during tetanic stimulation restoration of phospho-MPA receptor LTP restoration (Fig. 2) [24]. Note that

localized GR, ER were observed by immunoelectronc analysis (Fig. 6) [16,25]. Under these conditions ofP by CORT, ER may play signicant role due to lowertration (may be 5 M).ory processing, not only LTP (memory forming mech-

also LTD (memory erasing mechanism) are essential.e, which show enhanced LTP and suppressed LTD, haveaired learning of the Morris water maze [26]. Thisat LTD may be required to correct wrong memo-

by some LTP processes, which store not only correct

but also incorrect information. We found that LTDy enhanced by a 110 nM E2 perfusion (for 1 h) inal slices from adult rats (Fig. 3) [16,20]. Recordingsed using custom multielectrodes. LTD can be induced

Y. Ooishi et al. / Journal of Steroid Biochemistry & Molecular Biology 131 (2012) 37 51 39

Fig. 1. Estradiol (E2) rapidly rescues LTP from CORT-induced suppression in 12 week-old male hippocampus. (A) One micromolar CORT suppresses LTP and 1 nM E2 rescuedLTP from the CORT-induced suppression at CA3CA1 synapses of acute slices. Curve a, control slices (black closed circle); curve b, 1 M CORT treated slices (open circle);curve c, slices treated by 1 M CORT plus 1 nM E2 (red closed circle). Illustrated data points and error bars represent the mean SEM. Note that co-perfusion of RU486(GR inhibitor) with CORT abolished the CORT-induced suppression of LTP. Vertical axis indicates the maximal EPSP slope. Baseline is set at 100%. Open bar above the graphindicates the period of time during which CORT was administered. Red closed bar above the graph indicates the period of time during which E2 was administered. (B) Onenanomolar estradiol alone does not affect LTP at all at CA3CA1 synapses of acute slices. Curve a, control (black closed circle), curve b, 1 nM E2 (blue closed square). (C)E2 rescued NMDA-receptor-derived EPSP slope from the CORT-induced suppression. Time dependence of NMDA-R-EPSP slope is shown for 20 min. Rapid abolition of theCORT-induced suppression occurs by 1 nM E2 perfusion. The period of CORT perfusion is indicated by the open bar, and that of E2 perfusion by the closed bar. (D) Abolitionof the E2-induced rescue effect on CORT-induced suppression of NMDA-R-EPSP slope following the perfusion of 1 M Ro25-6981, an NR2B inhibitor (perfused as indicatedby the green bar). Statistical signicance (*p < 0.05; **p < 0.01). (For interpretation of the references to color in this gure legend, the reader is referred to the web version ofthe article.)

Modied from [24].

Fig. 2. Hypothetical mechanisms of modulation of LTP by CORT and E2. (A) CORT-induced LTP suppression. An application of 1 M CORT drives the following signaling path-way: CORT GR calcineurin dephosphorylation of NMDA receptor subunit NR2A. Upon tetanic stimulation, AMPA receptor (AMPAR) is not sufciently phosphorylatedat the residue of serine 831, because CORT already suppressed NMDA receptor (NMDAR)-derived Ca2+ current by dephosphorylation of NR2A, resulting in the suppressionof LTP-induction. (B) Rescue by E2 of LTP from CORT-induced suppression. The application of 1 nM E2 drives the following signaling pathway: E2 ER or ER Erk MAPkinase NR2B phosphorylation. E2 nally phosphorylates an NMDA receptor subunit NR2B. Upon tetanic stimulation, estradiol-induced restoration of NMDAR derived Ca2+current by phosphorylation of NR2B causes sufcient phosphorylation of AMPAR at the residue of serine 831, resulting in the restoration of LTP-induction.

Modied from [24].


Fig. 3. Rapid e ximal 0 nM (open cir Here, NMDA stimula at timgraph indicate e (MEcircle) and recaxis represent17-E2 (E2), Penhancement.on electrophyinvolved in thadult hippocamconrmed at 6of the article.)

Modied from

pharmacolois induced binux throuinduced byadult hippocannot indution can indA 30 min peing in a decapplicationInvestigatiocontributioER agonisER agoniscontributioLTD. Taken the moderaNMDA applreceptors re

On the modulation512 M unphorylationprocess at t

E2-induexistent systrogen modulation of LTD (adult male rat). [Upper left] Time dependence of macle), 1 nM (yellow closed triangle) and 10 nM (red closed diamond), respectively. tion, irrespective of the test condition. LTD was induced by 30 M NMDA perfusions the E2 administration period. [Upper right] Custom-made 64 multielectrode prob

ording (blue circle) electrodes are indicated. [Lower left] Comparison of modulatory effecs relative EPSP amplitude at t = 60 min, where EPSP amplitude at t = 60 min of the controPT (ER agonist) and DPN (ER agonist) at indicated concentrations. Co-perfusion of 10 n

Co-perfusion of 1 M ICI with 10 nM 17-E2 did not suppress the enhancing effect of LTsiological signals, such as LTP, LTD or kainate current presented in previous reports [16,2e E2-modulation of synaptic transmission. An NMDA-induced chemical LTD is used for adpal slices. In contrast, an electrical LTD is inducible for younger hippocampus which is

0 min via statistical analysis (*p < 0.05; **p < 0.01). (For interpretation of the references to

Mukai et al. [16].

gically by the transient application of NMDA. This LTDy the activation of phosphatase due to a moderate Ca2+

gh NMDA receptors [27]. Note that LTD is effectively the transient application of NMDA (chemical LTD) forcampus, whereas low frequency electrical stimulationce LTD in adult slices. Low frequency electrical stimula-uce LTD in slices from rats younger than 4 weeks of age.rfusion of 10 nM E2 signicantly enhances LTD result-rease in plateau EPSP amplitude (at 60 min after NMDA), for example, 80% 60% (CA3CA1 synapses) [16,20].ns using specic estrogen agonists indicated that then of ER but not ER is essential for these E2 effects.t exhibits a signicant LTD enhancement in CA1, whilet induces a suppression of LTD in CA1, implying that then of ER is opposite to that of ER in the E2 effect oncollectively, E2-bound ER may activate phosphatase atte Ca2+ concentration of 0.71 M induced upon 30 Mication [28], and facilitate dephosphorylation of AMPAsulting in the enhancement of LTD.other hand, E2-bound ER is not functional in LTP

at the transiently high Ca2+ concentration of approx.der tetanic stimulation [20,2830], because the phos-

of AMPA receptors by CaM kinase II is the dominanthe high Ca2+ concentration.ced rapid modulation of LTD or LTP occurs only in pre-napses, because newly generated spines induced by

E2-treatmeno increaseperfusion [1enhances Lgen supplemelectrophysmission are

2.2. Spinog

Dendritilated by E2which form(spines wittion of spincreate sitesulated upoanalysis of from adult mment with total densitto 1.31 spinphenol (PPspine denspropionitriBlocking EREPSP amplitude in CA1 of hippocampal slices. E2 concentration was100% EPSP amplitude refers to the EPSP value at t = 40 min prior toe t = 03 min (closed red bar above the graph). Hatched bar above theD64, Panasonic, Japan) with the hippocampal slice. Stimulation (red

t on LTD by E2 and agonists in the CA1 of hippocampal slices. Verticall slice without drug application is taken as 100%. From left to right,M 17-E2 with 1 nM 17-E2 ( + ) blocks the 17-E2 effect of LTDD by estradiol. Importantly, ICI does not inhibit any type of E2 effects0,60]. The ICI results suggest that dimer formation of ER/ER is notult hippocampal slices, because an electrical LTD is not inducible foryounger than 3 week-old. The signicance of the estradiol effect was

color in this gure legend, the reader is referred to the web version

nts do not form new synapses within 2 h, as judged from in the baseline magnitude of EPSP signal during 2 h of E26]. On the other hand, the slow effect of E2 (24 days)TP via formation of new synaptic contacts for estro-ented ovariectomized 8-week female rats [12,31]. The

iological properties of NMDA receptor-mediated trans- altered by estrogens [32].

enesis

c spines (presynaptic structures) can be rapidly modu- [33]. Spines consist of not only spine-synapses (spines

synapses, approx. half of spines) but also free spineshout forming synapses, approx. half of spines). Modula-ogenesis involves the production of new spines that

for new neuronal contacts. Spines are rapidly mod-n E2 application, which is observed by single spineLucifer-Yellow injected neurons in hippocampal slicesale rats (3 months) [17,30,34,35]. Following a 2 h treat-

1 nM E2 in the stratum radiatum of the CA1 region, they of spines signicantly increases from 0.85 spines/mes/m (Fig. 4) [16]. ER agonist propyl-pyrazole-trinyl-T) [36] also induces a signicant enhancement of theity to 1.20 spines/m. However, ER agonist, diaryl-le (DPN) [36], increases the spine density only slightly.s by ICI 182,780 completely suppresses the enhancing


Fig. 4. Flow of spine head diameter determination by Spiso-3D. We use the spine brightness function IA(x1) which is obtained by subtraction of traced dendrite from thetotal brightness function. The spine head region points are extracted as points where both 1 and 2 yield negative values (1D and 2A), since the spine head is an isolatedclosed volume with a closed surface. (1) Locating spine center at an optical slice with z = zj . (1A) A schematic illustration of the pixel P(x,y) that is a true spine center. Fournearest neighbors in X or Y direction with all A are negative. (1B) P(x,y) that is not a true spine center with one positive A. (1C) A schematic example of gradient vectorimage. Painted tiles indicate pixels of spine center candidate, i.e. ||grad A|| > 0, white tiles are pixels with grad A = 0. (1D) Calculation of the inner product of gradient vectors at pixels where 1 and 2 are both negative. (1E) Spine center detection image created by selecting pixels with negative A. (1F) Digitized spine center detection image.In spine center candidates, minimum number of pixels to reach the center pixel from perimeter (edge) pixel is counted and assigned on each pixel. A perimeter pixel has avalue of 1. The pixel having maximum number rc = 2 is of the spine center C. (2) Spine diameter determination. Determination of spine head diameter is performed using thedistance image that is the digitized radius detection image. (2A) Spine center detection image. Eigenvalues of Hessian matrix are calculated at each pixel. indicatesa pixel where (1) eigenvalues 1 and 2 are both negative and (2) 12 > S (S: sensitivity set in the Spiso program by user). k indicates pixel with 12 S. (2B) Image ofinformation of spines. pixels are marked for spine center candidates, while k pixels are omitted. (2C) Gradient vector image in (1C). (2D) By superimposing (2B) and(2C), the connected area in gradient vector image with pixel inside is dened as spine. The spine pixels are indicated as orange pixels. (2E) The radius detection image.To create the digitized distance image, the minimum number of pixels to reach the center pixel from the perimeter pixel is assigned on each pixel. (2F) Spine diameteris determined by combining the spine center detection image (1F) with (2E), by superimposing both center C. The combined area is again digitized to create the distanceimage. The maximum distance number R (assigned for the center C) is adopted as a spine radius (a spine diameter Dj = 2R). Finally, 3-dimensional integration of spines isperformed, resulting in determination of maximum Dj as the spine diameter D. (3) Results of calculated dendritic spines by Spiso-3D. (3A) Original image of dendrite. (3B)Traced dendrite (connected series of red circles) and spines (yellow circles) superimposed on the image. (3C) Calculated diameters of spines are superimposed on the spineimages. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of the article.)

Modied from Mukai et al. [38].


effect of E2 on the spine density. Interestingly, blocking ERKMAP kinase by PD98059 or U0126 completely prevents the E2-induced spinogenesis [35]. Taken together, the rapid E2-inducedenhancement of the CA1 spine density probably occurs by theactivation othat when the basal lenhancing E2-bound Emately 0.1

Compariful, howevefor discrimieffects of Eguishable. Idistributionand E2 treaessary.

In adult distinct hea(approx. 5%very small.distributionplex morphclassicatio

To do a oped Spiso-identies thfrom their images of cis performesurface. Forobtained asbrightness optical slice

I(x + u) =

=

=

with

grad I =

where 1 aextracts spicompletelyray-burstinness to deare almost USA, manuaparison, wespines (0.2head spine

Using Spof testosterE2 (1 nM) oacute hipp(within 2 h)to 1.28 (T)effects of T

were indistinguishable, closer examination spine head diameterrevealed marked differences in the distribution of spine head diam-eter between T, DHT and E2 treatments. DHT treatment was foundto increase large- and middle-head spines, whereas T increased

and ed o

of thimplMPAe de

ductinsity

AMPsed dpotenoderon ththe iuishen.entioitro ed inor 1amation men

[6,7,presy in

rats g/kn cu

otheHT,

rapirkerses t

in ovraph

(14gatey. Foemal

CA1leveenesen [4ly theronspoca

of NMt, as en-inor anitro es fpal sutiond cnoudecr

markd sli

exprer, wf Erk MAP kinase via ER [37]. It should be notedthe Ca2+ concentration in spines is decreased fromevel by blocking NMDA receptors with MK-801, theeffect by E2 is completely suppressed. The function ofR probably requires the basal Ca2+ level of approxi-0.2 nM.son with the total density of spines is facile and use-r in many cases the total spine density is not sensitivenation of different steroid treatments. For example the2, T and DHT on the total spine density are indistin-n such a case, closer examination of spine head diameter

is necessary in order to differentiate between T, DHTtments. A rigorous analysis of spine morphology is nec-

hippocampal slices, the majority of spines (>95%) haveds and necks, while the populations of stubby spines, no neck) and lopodium (approx. 1%, no head) are

Therefore, the determination of spine head diameter is a very powerful method in order to analyze the com-ological changes in spines, instead of the conventionaln of mushroom/thin/stubby/lopodium [30].rigorous spine head diameter analysis, we have devel-3D software which mathematically and automaticallye spine head and determines the diameter of the spinegeometrical features (see Fig. 4-1, -2, -3, ow chartalculation) [38]. The identication of the spine headd by extraction of spine head points within a closed

mathematical analysis, we use Hessian tensor that is 2nd derivatives from Taylor expansion of the spinefunction I(x) in each optical slice, in 2030 z-seriess obtained by confocal images:

I(x) + I(u) + 12

2I(u) +

I(x) + grad I u + 12

2utHu+ = I(x + u)

I(x) + (g1u1 + g2u2) +12

2(1u21 + 2u22)+

I

xI

y

, H =

2I

x22I

xy

2I

yx

2I

y2

, diag H =

(1 0

0 2

)

nd 2 are the eigenvalues of Hessian tensor. Spiso-3Dnes based on geometrical features of spines, therefore

different approach from other methods including theg method [39,40] that exploits information of bright-ne boundaries of spines. Results obtained by Spiso-3Didentical to those by Neurolucida (MicroBrightFeild,l-based analysis software) [38]. For quantitative com-

classify spines into three subclasses, i.e., small-head0.4 m), middle-head spines (0.40.5 m) and large-(0.51.0 m).iso-3D, we clearly distinguished the different effectsone (T, 10 nM), dihydrotestosterone (DHT, 10 nM) andn hippocampal spines in CA1 pyramidal neurons inocampal slices (Fig. 4). These sex-hormones rapidly

increased the total spine density from 0.97 spines/m, 1.32 (DHT) and 1.34 (E2), respectively. While the, DHT and E2 treatment on the total spine density

large- increaseffectstional more Awith ththe inthe desity ofIncreacould only meffect strate distingandrog

Attfor in vdepletACSF fwere dcentrasuppleeffectsmay retherapfemale1050effect iOn theT and D

Theco-woincreasmentsmicrogeffectsinvestitherapadult ftum ofto the spinogestrognot ongic neuthe hipmeansagonisEstrogrecept

In vincreaspocamcontribRune aendogecantly (spineculturephysinHowevsmall-head spines (Fig. 5). In contrast, E2 treatmentnly small-head spines. The observed differences in thee hormones on spine subpopulations may have func-ications, for example, large-head spines may contain

receptors, since spine-head size positively correlatesnsity of AMPA-type glutamate receptors [41,42]. Sinceon of long-term potentiation (LTP) is dependent on

of AMPA receptors in spines [41,42], increased den-A receptors in large-head spines could facilitate LTP.ensity of large-head spines following DHT treatmentstially facilitate LTP induction, in contrast to T, whichately increased large-head spines, or E2 which had noe density of large-head spines. These ndings demon-mportance of the consideration of spine diameter to

different types of neurotrophic effects of estrogen and

n must be paid to acute hippocampal slices often usedinvestigations of spinogenesis or LTP/LTD. Steroids are

acute slices during recovery incubation in steroid-free2 h, in order to remove damaged surface cells whichged by vibratome slicing (Fig. 9-2). Because the con-of E2, T and DHT is less than 0.5 nM in acute slices,tation of E2, T or DHT (110 nM) induces signicant9,16,22,43]. In this sense, these acute slice experimentsent an in vitro model of in vivo estrogen replacementwhich circulating estrogen-depleted ovariectomized

(hippocampal E2 of 0.2 nM) are injected with E2 atg rat weight. Probably E2 (110 nM) cannot induce anyltured slices which keep endogenous 510 nM E2 level.r hand, the hippocampus in vivo contains 510 nM E2,

preventing action of such a nanomolar E2, T and DHT.d effect of estrogen is also observed in vivo. Leranth and

demonstrate that the E2 (60 g/kg) rapidly (30 min)he spine-synapse density due to synaptic rearrange-ariectomized adult rats after E2 injection, using electronic analysis [44]. On the other hand, the slow genomic

days) of E2 on spine plasticity have been extensivelyd in vivo from the view point of estrogen replacementr example, a supplement of estrogen to ovariectomizede rats increase the density of spines in the stratum radia-

pyramidal neurons, resulting in a recovery of spinesl of intact rat [4447]. These enhancement effects onis are also observed as rapidly as 4 h after s.c. injection of4]. Results from in vivo investigations of rats may reect

direct but also the indirect effects of E2 on glutamater- via cholinergic or serotonergic neurons, projecting tompus [44,48]. Estrogen mediates slow spine changes byDA receptors. Estradiol increases the binding of NMDA

well as the NR1 subunit levels in CA1 neurons [49,50].duced increases in spine density are blocked by NMDAtagonists [11,51].investigations also show that the CA1 spine densityollowing several days treatment of cultured hip-lices with 0.5 M (=0.2 g/mL) exogenous E2 [8]. Then of hippocampus-derived E2 has been examined byo-workers. They demonstrate that the suppression ofs E2 synthesis by letrozole treatments for 4 days signi-eases the density of spines, spine-synapses, spinophiliner) and synaptophysin (presynaptic marker) in CA1 ofces [5]. Application of 100 nM E2 rescues the synapto-ession that was once decreased by letrozole treatments.hen slice cultures are not treated with letrozole,


Fig. 5. Rapid m wereradiatum of C t horm1.34 spines/m roids(C) Density of ds (op1 nM E2 (dotte 0.01 vlegend, the rea

Modied from

applicationsity, spine-scan be expE2 in slices[52].

Compareon the formhas been poworkers repdensity in mof T propiomidal neurreduced wiof T on dendated by loca(within 2 h[55].

2.3. Synapt(Fig. 6)

What is(12 h) on stic membraattempts hators. At thsynaptic (e[18]. Althoumay work kinases, me[16,18,37,5odulation of spines by estrogen and androgen (adult male rat, 12 weeks). SpinesA1 pyramidal neurons. (A) Total spine density. A 2-h treatment in ACSF withou. (B) Histogram of spine head diameters. After a 2-h treatment in ACSF without ste

three subtypes of spines with DHT, T and E2. From left to right, ACSF without steroid column). Vertical axis is the average number of spines per 1 m. *p < 0.05, **p

estrogen hippocampus

Documents