nikolakopoulou et al., 2006a

Passive avoidance training decreases synapse density inthe hippocampus of the domestic chick

A. M. Nikolakopoulou,1,2 H. A. Davies1 and M. G. Stewart11The Open University, Biological Sciences, Walton Hall, Milton Keynes MK7 6AA, UK2University of California, Irvine, Neurobiology and Behavior, 2205 McGaugh Hall, Irvine, CA 92697, USA

Keywords: axo-spinous density, dorsal, learning, stress, ventral

Abstract

The bird hippocampus (Hp), although lacking the cellular lamination of the mammalian Hp, possesses comparable roles in spatialorientation and is implicated in passive avoidance learning. As in rodents it can be divided into dorsal and ventral regions based onimmunocytochemical, tracing and electrophysiological studies. To study the effects of passive avoidance learning on synapsemorphometry in the Hp, spine and shaft synapse densities of 1-day-old domestic chicks were determined in dorsal and ventral Hp ofeach hemisphere by electron microscopy, 6 and 24 h following training to avoid pecking at a bead coated with a bitter-tastingsubstance, methyl anthranilate (MeA). The density of asymmetric spine and shaft synapses in MeA-trained birds at 6 h post-trainingwas significantly lower in the dorsal and ventral Hp of the right hemisphere relative to control (untrained) chicks, but by 24 h thisdifference was absent. A hemispheric asymmetry was apparent in the ventral Hp where the water-trained group showed enhancedshaft and spine synapse density in the left hemisphere, whilst in the MeA-trained group only asymmetric shaft synapses follow thesame pattern in relation to the right hemisphere. There were no differences in asymmetric shaft synapses in the dorsal Hp at 6 h post-training, but at 24 h post-training there was a reduction in the density of shaft synapses in the right hemisphere in MeA compared withcontrol birds. These data are discussed in relation to the pruning effects of stress and learning on synapse density in chick Hp.

Introduction

Alterations in connectivity via changes in synaptic efficacy arebelieved to underlie learning and memory storage (Hebb, 1949; Bliss& Collingridge, 1993). Changes in synaptic morphology can be veryrapid, with alterations in shape occurring within seconds (Hering &Sheng, 2001). A number of studies have focused on synaptogenesis inthe chick brain mainly after passive avoidance learning (PAL) wherethe aversive experience is exposure to a bitter-tasting substance,methyl anthranilate (MeA) (Stewart et al., 1987), and in filial (Hornet al., 1985; Horn, 2004) and acoustic imprinting (Thode et al., 2005).Increases in spine and synapse density have been demonstrated in theleft medial striatum (MSt) (Stewart et al., 1987; Lowndes & Stewart,1994) and intermediate medial mesopallium (IMM) (Patel & Stewart,1988) at 24 h after PAL in MeA-trained chicks compared with watercontrol birds.The chick hippocampus (Hp) has been suggested to be homologous

to the mammalian Hp (Kallen, 1962; Erichsen et al., 1991; Atoji et al.,2002). Previous studies have shown that as in mammals the avian Hpplays a key role in spatial memory (Bingman et al., 1990; Regolin &Rose, 1999; Kahn & Bingman, 2004) and demonstrates synapticplasticity with some similarities to mammalian long-term potentiation(LTP) (Margrie et al., 1998). This region has also been implicated inthe passive avoidance paradigm (Sandi et al., 1992). Unal et al. (2002)demonstrated an increase in density of shaft and spine synapses withtime after PAL in the dorsolateral Hp of chicks, with an increase at

both 24 and 48 h after training in the density of shaft synapses inMeA- compared with water-trained birds.In the present study we have examined the effects of PAL on

synaptic morphometry in the ventral and dorsal Hp of the domesticchick, the divisions based on the published findings (Casini et al.,1986; Krebs et al., 1991; Szekely, 1999). Each area was studiedindividually, because previous studies from our group have demon-strated synapse density reduction in the dorsal Hp after ischaemia(Horner et al., 1996), and in rat brain the dorsal and ventral Hp havebeen shown to play different roles in learning tasks (Moser et al.,1993; Hock & Bunsey, 1998; Moser & Moser, 1998). Our studieswere conducted at 6 h and then 24 h after PAL and included twocontrol groups, a completely untrained control group and birds trainedto peck a bead identical to that used with the aversive tasting MeA, butcoated with water. The 6 h time-point was chosen because a proteincascade takes place by this time that enables the structural changesnecessary for short-term memory to be consolidated into long-termmemory (Rose, 1991, 1995a,b), and by 24 h this process is welladvanced (Rose & Stewart, 1999). Each hemisphere was studiedseparately as chicks show evidence of hemispheric asymmetry(Stewart et al., 1987; Sandi et al., 1993; Gagliardo et al., 2001).

Materials and methods

Animals and training

Commercially obtained Ross Chunky eggs (domestic chick Gallusdomesticus) were incubated and hatched in our own brooders until18 ± 6 h old. Chicks were placed in pairs in small aluminium pensilluminated by red bulbs at a temperature of 25–30 �C.The animalswere

Correspondence: Dr A.M. Nikolakopoulou, as above.2

E-mail: [email protected]

Received 5 August 2005, revised 7 December 2005, accepted 10 December 2005

European Journal of Neuroscience, Vol. 23, pp. 1054–1062, 2006 doi:10.1111/j.1460-9568.2006.04619.x

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd

separated into three groups, naı̈ve (undisturbed), water-trained (W) andmethyl anthranilate-trained (MeA). Chicks were first pretrained bybeing presented a small white bead 3 mm in diameter three times withintervals of 5 min between presentations, essentially as describedpreviously (Mileusnic et al., 2000; Dermon et al., 2002). Animals thatsuccessfully pecked the bead were noted as ‘peck’, while those that didnot peckwere noted as ‘no peck’. In order for the chicks to be included inthe study, they should have pecked the white bead three times. Tenminutes after the last presentation of the white bead, a chrome bead4 mm in diameter dipped in either water or MeA was presented to theanimals (the training session). Chicks peck once, and having tastedMeAthey exhibit a disgust response by shaking their heads, emitting stresssounds and beating their beaks on the ground. Chicks were tested 6 h(naı̈ve group n ¼ 7, water group n ¼ 5, MeA group n ¼ 5) or 24 h(naı̈ve group n ¼ 6, water group n ¼ 6, MeA group n ¼ 5) aftertraining by the presentation of a dry chrome bead. Chicks in the watergroup pecked the dry bead whilst those in the MeA group exhibitedrecall of the bitter taste and avoided the chrome bead. Chicks that failedto give the correct response to the task within 20 s, and pecked at the drybead although previously exposed to MeA, were excluded from thestudy. All experimental procedures took place under UKHome Licenceand were also in agreement with the European Communities directive(86 ⁄ 609 ⁄ EEC) for the care and use of laboratory animals.

Tissue fixation

Animals were anaesthetized by intraperitoneal (i.p.) injection with0.2 mL of sodium pentobarbital immediately after testing. They werethen transcardially perfused with heparin in 0.9% saline followed by3.75% acrolein (TAAB, UK) in 2% paraformaldehyde in 0.1 m

phosphate buffer (PB), pH 7.4, followed by 2% paraformaldehyde inPB. Brains were removed from the skull and postfixed at 4 �Covernight in the latter fixative.

Tissue processing

The left hemisphere was marked and the brains were cut at 100 lmthickness with a vibrating microtome (Leica, UK); sections of theantero-posterior level A 8.2 were used for this study (Kuenzel &Masson, 1988) (Fig. 1). The sections were postfixed with 2% osmiumtetroxide in 0.1 m PB for 1 h, washed in 0.1 m PB dehydrated througha graded ethanol series followed by propylene oxide and Epon (AgarScientific, Stansted, UK) before being flat-embedded between twosheets of Aclar (Agar Scientific, Stansted, UK). Sections werepolymerized at 60 �C for 48 h. With the use of a stereoscope to viewthe sections, the Hp was divided into ventral and dorsal parts (Szekely& Krebs, 1996; Szekely, 1999), as indicated in Fig. 1.

Preparation for electron microscopy

The tissue blocks were coded and all subsequent procedures wereperformed blind. Blocks were cut on a UCT ultramicrotome (Leica),and ribbons of sections 70 nm thick were collected on formvar ⁄ car-bon films on copper slot grids. The sections were counterstained inuranyl acetate and Reynolds lead citrate before examination in a JEOL1010 electron microscope. Digital images were collected at amagnification of 12 000 using a Gatan BioScan CCDTV.

Electron microscopy and synapse counting

Synapse identification

Synapses were identified on the basis of a thickening of the pre-and postsynaptic membranes and the presence of vesicles in close

proximity to the presynaptic zone. Synapses with a prominentpostsynaptic density (PSD) are termed asymmetric and mostcommonly occur on dendritic spines, though some may be foundon dendritic shafts (Fig. 2A–D). Synapses where the pre- andpostsynaptic membranes are of equal thickness are termed sym-metric and are typically found on dendritic shafts (Fig. 2B).Symmetric synapses overall comprised less than 2% of the total andwere too infrequent for a statistical analysis and are therefore notincluded in the present analyses. Similarly, perforated synapses,which are those with a split PSD, were not taken into account dueto their rarity.

Synapse counting

The physical dissector was used for estimation of synapse density(Sterio, 1984), where synapses are counted if seen in one section(the nominated) and not in the other (reference section) (Fig. 3).The mathematical formula for calculation of the numerical densityis: Nsynapse ¼ SQ–

syn ⁄ tA, where SQ–syn is the total number of counted

synapses only in the nominated sections, t is the section thickness(distance between the two sections) and A is the area of thecounting frame. Section thickness was determined as described inearlier studies (de Groot & Bierman, 1986; De Groot, 1988).Briefly, the relative electron transmission (PET) was applied todetermine the thickness of a section by using the thicknessestimation diagram that occurs from the application of Small’sminimal fold technique (de Groot & Bierman, 1986; De Groot,1988; Tigges et al., 1996). Synapses are not counted if they touchthe forbidden lines (left side and bottom lines).

Synapse height

Synapse height (Hsyn, which is a measure of the size of the PSD) wasestimated using parameters derived from the dissector method as

Hsyn ¼ ðX

Qsyn=X

Q�synÞ � t

where SQ–syn is the total number of counted synapses only in the

nominated sections and SQsyn is the total number of synapses in boththe nominated and reference fields.

Fig. 1. Coronal sections of the chick brain showing the location of the ventraland dorsal hippocampus (vHp and dHp). AA, anterior arcopallium; AD, dorsalarcopallium; CPi, piriform cortex; GP, globus pallidus; LSt, lateral striatum;M, mesopallium; N, nidopallium; SL, lateral septal nucleus; SM, medial septalnucleus; TnA, nucleus taeniae of amygdala; black dots A14, dopaminergicneurons in the paraventricular nucleus (PVN).

Effects of PAL on the chick hippocampus 6 and 24 h after training 1055

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience, 23, 1054–1062

Statistics

Four-way analysis of variance (anova) for hemisphere (right, left), timeof study (6 h, 24 h), brain area (dorsal, ventral) and training (control,water-trained, MeA-trained) was used to check statistically significantdifferences between the groups tested, and values P < 0.05 were takenas significant. When a P-value was not significant for the main factorsbut only for their interaction, a three-way anovawas conducted for thefactors in the interaction to determine any significant differences. If theP-value was significant, a Fisher least significant difference (LSD)post hoc test was performed in order to find specific differences.

Results

Synapse density alterations in shaft and spine synapses (Fig. 2A–D)were examined in both the ventral and dorsal subdivisions of the chick

Hp 6 and 24 h after PAL. Although the borders of ventral and dorsalHp in chicks have not been clearly defined by previous studies, thedorsal Hp in this study was taken to correspond to that described asarea 3 and 4 by Erichsen et al. (1991) (dorsomedial Hp and part ofdorsolateral as in Szekely & Krebs, 1996). This region has beensuggested to be homologous to the dentate gyrus and hilus,respectively, whereas the ventral Hp in chicks is related to area 2,which shows homology with Ammon’s horn (CA).

Synapse ultrastructure

As described above, two different types of synapses were examined;asymmetric dendritic shaft and asymmetric spine (also termedasymmetric axo-dendritic and axo-spinous, respectively) (Fig. 2Aand D).

Fig. 2. Photomicrographs of synaptic contacts in the chick Hp. (A–D). (A) An asymmetric shaft synapse (on dendrites) in dorsal 1-day-old chick Hp indicated with ablack asterisk. The presynaptic part can be clearly distinguished by the presence of vesicles (ves). (B) Representative symmetric shaft synapse marked with anarrowhead from the dorsal Hp of the MeA-trained group. Asterisk indicates spine synapses. (C) Representation of asymmetric spine synapses marked with blackasterisks from the right hemisphere of the ventral Hp of the water-trained group. (D) Two dendrites in the ventral Hp of a control bird 24 h post-training receiving twoasymmetric synapses from a presynaptic axon terminal (At) resulting in axo-dendritic synapses. At, axon terminal; Den, dendrite; mit, mitochondrion; sp, spine. Scalebars, 200 nm.

1056 A.M. Nikolakopoulou et al.


Effects of PAL 6 h post-training

Two sets of data are presented in Fig. 4, for synapse density 6 h afterPAL in ventral and dorsal Hp from each hemisphere.

Asymmetric shaft density

Data for asymmetric shaft synapses in the ventral and dorsal Hp 6 hpost-training for control, water and MeA birds are presented inFig. 4A and B. Four-way anova showed statistical differences forbrain regions examined (ventral and dorsal Hp) (F1,109 ¼ 4.648,P ¼ 0.033), and the interaction between hemisphere and area of study(F1,109 ¼ 5.546, P ¼ 0.02). There is a 48% difference in theasymmetric axo-dendritic (asym shaft) synapses in the left hemisphereof the ventral Hp of the water-trained group compared with the righthemisphere. LSD post hoc analysis demonstrated that 6 h post-training the ventral Hp of the right hemisphere shows significantlylower synapse density (25%) in relation to the left hemisphere of thesame area (P ¼ 0.045) in the water-trained group.

Asymmetric spine density

Mean data for asymmetric spine synapse density in left and right dorsaland ventral Hp of the three bird groups are shown in Fig. 4A andB. Four-way anova revealed statistically significant differences for the inter-action of time after training and training group (F2,109 ¼ 3.377,P ¼ 0.038), but not for single factors. Therefore, a three-way anova

only for the 6 h time course (training group, hemisphere, area) wasperformed, which revealed a statistically significant difference betweenthe training groups (F2,52 ¼ 3.856, P ¼ 0.027) and hemisphere.Post hoc tests showed that 6 h post-training the dorsal Hp of the righthemisphere of the MeA-trained group has significantly fewer asym-metric spine synapses (35% less) in comparison to control birds(P ¼ 0.00084). It also has fewer of this type of synapse than the ventralpart of the right hemisphere of controls (P ¼ 0.04). Furthermore, the

water- and MeA-trained groups exhibit significant hemispheric differ-ences 6 h after training, with the right ventral Hp having fewerasymmetric spine synapses in comparison to the left hemisphere (44%for water, P ¼ 0.017; 33% for MeA-trained birds, P ¼ 0.04).

Effects of PAL 24 h post-training


Data for asymmetric axo-dendritic synapses in control, water- andMeA-trained birds 24 h post-training are shown in Fig. 5A and B.A four-way anova shows statistically significant differences for brainregions examined (ventral vs. dorsal Hp) (F1,109 ¼ 4.648, P ¼ 0.033),and the interaction between hemisphere and region examined(F1,109 ¼ 5.546, P ¼ 0.02). Fisher LSD post hoc tests showed thatin the dorsal Hp of the right hemisphere there is a 33% decrease insynapse density in the MeA-trained group in comparison with controlbirds (P ¼ 0.038; Fig. 5A), whilst the density of axo-dendriticsynapses (asym shaft) in the right hemisphere of the dorsal Hp ofcontrol birds is 29% greater than in the left hemisphere (P ¼ 0.031).

Differences in synaptic density in Hp between 6 h and 24 hpost-training groups


After a Fisher post hoc test, a 48% increase was confirmed inasymmetric shaft synapses of water-trained chicks in the ventral Hp ofthe right hemisphere (P ¼ 0.012) at 24 h in comparison to 6 h post-training.

Asymmetric spine density

Post hoc tests showed a contrasting pattern between the control andthe water-trained groups in the dorsal and ventral Hp. At 6 h therewere significantly more asymmetric spine synapses (1.61 vs. 1.31 per

Fig. 3. Example of two images used for synapse density estimation with the dissector method. The image on the right is the ‘nominated or look-up’ image, whilstthe left is the ‘reference’ image. Only synapses that are located within the borders of the lines are counted. The dashed lines are the forbidden lines; a synapsetouching the dashed lines is not counted. An asymmetric spine on a dendrite is marked with an asterisk in both images and it is not counted. The black arrow in thelook-up section indicates an asymmetric synapse onto a dendrite, which is counted as it does not appear in the reference section. The black arrow indicates asymmetric synapse onto a spine (look-up section). The black arrowheads show an asymmetric synapse onto a spine in both images and therefore are not counted. Inthe reference image the star indicates a symmetric axo-dendritic synapse. In this case three synapses would have been counted, one symmetric axo-dendritic and oneasymmetric axo-dendritic synapse in the reference section, and one symmetric axo-spinous in the look-up section.



lm3, a 23% increase) in the dorsal Hp of the right hemisphere of thecontrol group (P ¼ 0.033) than in the right ventral Hp; however, thisdifference disappeared by 24 h. In contrast, in the right ventral Hp ofthe water-trained group 6 h after training there were fewer asymmetricspine synapses (35% less) in comparison to the 24 h water-trainedgroup (P ¼ 0.0024). No other changes were observed.

Synaptic height

Asymmetric shaft synapse height

Mean data for asymmetric shaft synaptic height (Hsyn) in the threechick groups at 6 and 24 h post-training are shown in Table 1.Four-way anova showed significant differences for the interactionof time after training and the Hp area examined (F1,108 ¼ 5.88,P ¼ 0.017).Post hoc tests at 6 h post-training show that in the ventral Hp of the

left hemisphere, Hsyn in control birds is greater than in water-trainedchicks (P ¼ 0.0096), but is reduced relative to MeA-trained birds(P ¼ 0.036). Additionally, the MeA-trained group demonstratesincreased Hsyn relative to water-trained chicks (P > 0.0001). In thedorsal part of the right hemisphere Hsyn is higher in the MeA-trainedgroup than in water (P ¼ 0.0001) or control birds (P ¼ 0.005).

At 24 h post-training there are no differences in Hsyn between thegroups in either hemisphere of ventral or dorsal Hp. However,comparison between values at 6 and 24 h shows that Hsyn in the MeA-trained group is greater at 6 than at 24 h in the left ventral Hp(P ¼ 0.0059), whilst Hsyn in the right dorsal Hp of the water-trainedgroup is greater than at 6 h (P ¼ 0.036).

Asymmetric spine synaptic height

The data for the height (Hsyn) of the major class of synapses examined,asymmetric spine synapses, are presented in Table 2. Four-way anovashowed no differences in spine synaptic height for any of the factorsexamined.

Discussion

Our data reveal marked alterations in the density of asymmetricsynapses (both on spines and dendritic shafts) in the chick Hp afterpassive avoidance training. In general, training results in a decreaserather than an increase in synapse density. In the ventral Hp of theright hemisphere of the water-trained group at 6 h post-training,asymmetric shaft synaptic density was reduced in comparison to the

Fig. 4. Histograms showing asymmetric synapsedensities (Nv ⁄ lm3) in the dorsal (A) and ventral(B) hippocampus (Hp) of the right and left hemi-sphere of chicks 6 h after avoidance training, andin water-trained and naı̈ve controls (control n ¼ 6,water n ¼ 5, MeA n ¼ 5). Vertical bars on thehistogram blocks represent means ± SEM.(A) The asterisks indicate significant reductions inasymmetric spine synapse density in the dorsal Hpof the right hemisphere of MeA-trained chicks(P ¼ 0.0008). (B) Significant differences are in-dicated between the right and left hemisphere inthe water-trained group (n ¼ 5) (P ¼ 0.045) forasymmetric shaft synapses (�), whilst � and +show differences for asymmetric spine synapsedensities between right and left hemisphere ofwater and MeA-trained groups (P ¼ 0.017 andP ¼ 0.04, respectively).



left hemisphere, and asymmetric axo-spinous synapse density was alsoreduced in the right hemisphere of both the water- and MeA-trainedgroups. Similarly, in the dorsal Hp of the right hemisphere, there was adecline in synaptic density of asymmetric axo-spinous synapses 6 hafter training in the MeA-trained group in comparison to the controlgroup. Our data differ from that of Unal et al. (2002), where anincrease in the density of shaft synapses was found in Hp at both 24

and 48 h post-training, with small decreases in spine synapse densityat 24 h post-MeA training. A reason for the differences may be thedissimilarity in the Hp regions examined (Unal et al., 2002 studied thedorsolateral Hp only), and also the control group in the Unal et al.study did not include naı̈ve (completely untrained) animals. Further-more, unlike the findings in the study by Unal et al. (2002), our dataindicate that synaptic changes occur without alteration in PSD size

Fig. 5. Histograms showing asymmetric synapsedensities (Nv ⁄ lm3) in the dorsal (A) and ventral(B) hippocampus (Hp) in the right and left hemi-sphere of chicks 24 h after avoidance training, andin water-trained and naı̈ve controls (control n ¼ 6,water n ¼ 6, MeA n ¼ 6). Vertical bars on thehistogram blocks represent means ± SEM.(A) There are significantly fewer (�) asymmetricshaft synapses in the dorsal Hp of the left hemis-phere of control animals, compared with the righthemisphere (P ¼ 0.031). Also the MeA-trainedgroup has significantly fewer asymmetric shaftsynapses (*) in comparison to the control group(P ¼ 0.038). (B) None of the differences in theventral Hp between the groups or hemispheres at24 h for either asymmetric shaft or spine synapsesare significant.

Table 1. Asymmetric shaft synapse height

Asymmetric shaft synapse height 6 h after training (in lm) Asymmetric shaft synapse height 24 h after training (in lm)

Ventral hippocampus Dorsal hippocampus Ventral hippocampus Dorsal hippocampus

R L R L R L R L

Control 0.12 ± 0.006 0.12 ± 0.008 0.11 ± 0.1 0.12 ± 0.009 0.12 ± 0.008 0.12 ± 0.006 0.12 ± 0.01 0.13 ± 0.01Water 0.13 ± 0.01 0.10 ± 0.008** 0.095 ± 0.01 0.11 ± 0.007 0.11 ± 0.006 0.11 ± 0.006 0.13 ± 0.009� 0.12 ± 0.006MeA 0.12 ± 0.01 0.14 ± 0.016*,� 0.13 ± 0.01* 0.11 ± 0.01 0.12 ± 0.01 0.10 ± 0.006 0.12 ± 0.01 0.13 ± 0.02

Data are presented as means (± SEM) showing asymmetric shaft synaptic height (Hsyn) in right (R) and left hemisphere (L) of the three chick groups 6 and 24 h afterpassive avoidance training. *P < 0.05 and **P < 0.01, comparing control and trained birds at 6 h after training, but there were no significant differences in Hsynbetween the groups in either hemisphere of ventral or dorsal hippocampus. �MeA-trained group 6 h Ventral (R) > MeA-trained group 24 h Ventral (R)(P ¼ 0.0059). �Water-trained group 6h Dorsal (R) < water-trained group 24 h Dorsal (R) (P ¼ 0.036).



(Hsyn), at least in asymmetric axo-spinous synapses, which are themajor class of synapses in chick Hp (three times more numerous thanasymmetric shaft). There are alterations in the size of asymmetric shaftsynapses at 6 h post-training, with Hsyn larger in the left hemisphere ofMeA birds than in water or control birds, but these differencesdisappear by 24 h post-training.The synapse reduction in both the dorsal and ventral Hp would

appear to suggest that the training process per se may havecontributed to a large extent to the decrease in synapse density inthe chick Hp, as it occurs both in the water- and MeA-trained groups.Conversely, in the rat, the passive avoidance task (O’Malley et al.,1998) or water maze training (Moser et al., 1994) results in anincrease in dendritic spine density in Hp dentate gyrus and CA1,respectively, whilst the enriched environment has similar effects inCA3 (Altschuler, 1979). One may argue that in chicks passiveavoidance training is not a spatial task and therefore the Hp is affecteddifferently from mammals, resulting in reduced synaptic connectivityin the MeA-trained group. However, recent studies in rat Hp have alsoidentified synapse density alterations after non-spatial tasks such asolfactory learning (Knafo et al., 2004). They emphasize that anincrease in axo-spinous synapse density in CA1 apical dendritesoccurs only if odour exposure is accompanied by olfactory learning.Although chicks associate the intense smelling MeA with learning(Marples & Roper, 1997; Richard & Davies, 2000; Dermon et al.,2002), no synapse density enhancement has been monitored in ourstudies. On the other hand, studies in the IMM 1 h after PAL haveshown increases in spine density in the right hemisphere in relation tothe left in the MeA-trained group as well as in comparison tountrained animals (Doubell & Stewart, 1993). It is important to notethat at 1 h post-training the biochemical cascade for memoryformation is at an early stage, and is different than at 6 h (Rose,1995a; Rose & Stewart, 1999), as cell adhesion molecules, which areessential for memory formation, are activated only 5–8 h post-training(Scholey et al., 1993, 1995). The c-fos and c-jun proteins show a peakin expression 2 h after imprinting (McCabe & Horn, 1994; Amba-lavanar et al., 1999; Suge & McCabe, 2004), and 1–2 h after passiveavoidance training (Freeman & Rose, 1995) in IMM. Consequently,6 h may be the time-point when the procedures for long-term memoryformation start to take place, which may be reflected by increasedsynaptogenesis. In contrast to these studies, auditory filial imprintingin the chick has been detected to cause spine density reduction in thedorsocaudal nidopallium (Bock & Braun, 1999) and mediorostralnidopallium ⁄mesopallium (Bock & Braun, 1998), implying thatlearning is not always accompanied by synapse density increases,which is consistent with our results from the trained groups reportedhere. One possible explanation for these findings is that the synapsereduction and the likely synaptic remodelling may relate to the

process of selective stabilization of synapses, meaning that thetraining process prunes any overexpression. The functional conse-quences are likely to be pathways that are more specified, but wecannot make further assumptions about this on the basis of ourinformation alone. However, synapse elimination or pruning has beensuggested to be a natural procedure occurring during neuronalactivation and synaptic remodelling (Goda & Davis, 2003).Twenty-four hours post-training, a reduction in asymmetric shaft

synapses in the MeA-trained group was demonstrated in the dorsalpart of the right Hp. One hypothesis may be that shaft synapsesdevelop into spine synapses as there is no difference at 24 h in spinesynaptic density between the untrained (control) and MeA-trainedbirds. However, the overall synaptic density is not notably elevated inthe MeA-trained group as this significant reduction of shaft synapsesis not compensated by substantial increases in spine synapse density.The decline 6 h after training in axo-spinous synapse density in

MeA chicks in the dorsal hemisphere of the right hemisphere could beexplained firstly either by late spine formation, or secondly by branchor synaptic elimination. In the first case, it is known from mammalianstudies that axo-dendritic synapses appear first and give rise todendritic spines (Mates & Lund, 1983; Fiala et al., 1998). The presentdata, however, have not shown any differences in the number of shaftsynapses 6 h post-training in the MeA-trained group in relation tocontrols. The second hypothesis could be that branch eliminationresults in decreased dendritic spine formation. Several explanationscould be given for this phenomenon; apoptosis may occur to eliminatethe dendritic tree together with synaptic connections. Thus, apoptosismay modulate synaptic remodelling by inducing cell death of old ornewly formed neurons after training so that new contacts take place totransform short- to long-term memory, keeping the synaptic balance inthe chick brain.Another more plausible explanation based on prior data could be

that the passive avoidance training is a stressful experience. Sandi &Rose (1997) have demonstrated that plasma corticosterone levelsincrease 5 min after MeA tasting, but return to basal levels by 15 min,whilst recent studies from our lab (Nikolakopoulou, 2005) have shownthat the levels of cortisol are higher in the chick Hp of the MeA-trained group in relation to controls 20 min after passive avoidancetraining. Although in the Sandi & Rose (1997) study it was shown thatthe levels of corticosterone return to normal, the elevated corticoster-one levels may affect synaptic plasticity by acting on brain-derivedneurotrophic factor (BDNF), which has reduced expression after stress(Smith et al., 1995; Ueyama et al., 1997), thus influencing synapticdismantling (Hu et al., 2005). BDNF has also been shown to mediatesynaptic plasticity, as levels are increased after LTP (Castren et al.,1993) and regulate axonal remodelling and branching (Inoue & Sanes,1997; Lom & Cohen-Cory, 1999; McAllister et al., 1999), synapse

Table 2. Asymmetric spine synapse height

Asymmetric spine synapse height 6 h after training (in lm) Asymmetric spine synapse height 24 h after training (in lm)

Ventral hippocampus Dorsal hippocampus Ventral hippocampus Dorsal hippocampus

R L R L R L R L

Control 0.11 ± 0.005 0.13 ± 0.007 0.1 ± 0.009 0.12 ± 0.007 0.12 ± 0.007 0.12 ± 0.004 0.12 ± 0.006 0.13 ± 0.008Water 0.13 ± 0.01 0.1 ± 0.008 0.1 ± 0.02 0.11 ± 0.004 0.12 ± 0.004 0.12 ± 0.009 0.12 ± 0.005 0.13 ± 0.009MeA 0.11 ± 0.004 0.13 ± 0.006 0.13 ± 0.01 0.11 ± 0.005 0.11 ± 0.01 0.11 ± 0.004 0.11 ± 0.01 0.12 ± 0.007

Data are presented as means (± SEM) of asymmetric spine Hsyn in the three chick groups 6 and 24 h after passive avoidance training. There were no significantdifferences in Hsyn at either 6 or 24 h after training in either left (L) or right (R) hemispheres of any of the three chick groups.



formation and stability (Poo, 2001) and synaptic transmission(Boulanger & Poo, 1999). Stress reduces neurogenesis (Gould &Tanapat, 1999), causes axon degeneration by Ca2+ excitotoxicity(Choi, 1995; Rothstein et al., 1996) and synapse reduction in CA3(Sandi et al., 2003; Stewart et al., 2005). Thus, stress-inducedapoptosis may cause dendritic atrophy, resulting in reduced spinedensity.

However, the fact that other chick brain areas, notably the IMM andstriatal regions (MSt), are affected positively in terms of increasedsynapse formation after PAL (Rose et al., 1980; Rose & Csillag, 1985;Stewart et al., 1987; Stewart & Rusakov, 1995) may indicate thatalthough Hp participates in PAL, its role is in the early stages ofmemory acquisition (Sandi et al., 1992) but not the longer termmemory storage stages, which in PAL are centred in the mesopalliumand medial striatum (Rose & Stewart, 1999).

Acknowledgements

The authors would like to thank Mrs Frances Colyer for technical support,Dr Jose Julio Rodriguez for helpful comments, Dr Mark Gardener for his helpwith the statistical analyses, and the personnel of the animal unit for animalwelfare.

Abbreviations

BDNF, brain-derived neurotrophic factor; CA, Ammon’s horn; Hp, hippocam-pus; IMM, intermediate medial mesopallium; LSD, least significant difference;LTP, long-term potentiation; MeA, methyl anthranilate; MSt, medial striatum;PAL, passive avoidance learning; PB, phosphate buffer; PSD, postsynapticdensity.

References

Altschuler, R.A. (1979) Morphometry of the effect of increased experience andtraining on synaptic density in area CA3 of the rat hippocampus.J. Histochem. Cytochem., 27, 1548–1550.

Ambalavanar, R., McCabe, B.J., Potter, K.N. & Horn, G. (1999) Learning-related fos-like immunoreactivity in the chick brain: time-course andco-localization with GABA and parvalbumin. Neuroscience, 93, 1515–1524.

Atoji, Y., Wild, J.M., Yamamoto, Y. & Suzuki, Y. (2002) Intratelencephalicconnections of the hippocampus in pigeons (Columba livia). J. Comp.Neurol., 447, 177–199.

Bingman, V.P., Ioale, P., Casini, G. & Bagnoli, P. (1990) The avianhippocampus: evidence for a role in the development of the homing pigeonnavigational map. Behav. Neurosci., 104, 906–911.

Bliss, T.V. & Collingridge, G.L. (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature, 361, 31–39.

Bock, J.&Braun, K. (1998)Differential emotional experience leads to pruning ofdendritic spines in the forebrain of domestic chicks. Neural Plast., 6, 17–27.

Bock, J. & Braun, K. (1999) Filial imprinting in domestic chicks is associatedwith spine pruning in the associative area, dorsocaudal neostriatum. Eur. J.Neurosci., 11, 2566–2570.

Boulanger, L. & Poo, M.M. (1999) Presynaptic depolarization facilitatesneurotrophin-induced synaptic potentiation. Nat. Neurosci., 2, 346–351.

Casini, G., Bingman, V.P. & Bagnoli, P. (1986) Connections of the pigeondorsomedial forebrain studied with WGA-HRP and 3H-proline. J. Comp.Neurol., 245, 454–470.

Castren, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lindholm, D., Thoenen,H. & Riekkinen, P.J. (1993) The induction of LTP increases BDNF and NGFmRNA but decreases NT-3 mRNA in the dentate gyrus. Neuroreport, 4,895–898.

Choi, D.W. (1995) Calcium: still center-stage in hypoxic-ischemic neuronaldeath. Trends Neurosci., 18, 58–60.

De Groot, D.M. (1988) Comparison of methods for the estimation of thethickness of ultrathin tissue sections. J. Microsc., 151, 23–42.

Dermon, C.R., Zikopoulos, B., Panagis, L., Harrison, E., Lancashire, C.L.,Mileusnic, R. & Stewart, M.G. (2002) Passive avoidance training enhancescell proliferation in 1-day-old chicks. Eur. J. Neurosci., 16, 1267–1274.

Doubell, T.P. & Stewart, M.G. (1993) Short-term changes in the numericaldensity of synapses in the intermediate and medial hyperstriatum ventrale

following one-trial passive avoidance training in the chick. J. Neurosci., 13,2230–2236.

Erichsen, J.T., Bingman, V.P. & Krebs, J.R. (1991) The distribution ofneuropeptides in the dorsomedial telencephalon of the pigeon (Columbalivia): a basis for regional subdivisions. J. Comp. Neurol., 314, 478–492.

Fiala, J.C., Feinberg, M., Popov, V. & Harris, K.M. (1998) Synaptogenesis viadendritic filopodia in developing hippocampal area CA1. J. Neurosci., 18,8900–8911.

Freeman, F.M. & Rose, S.P. (1995) MK-801 blockade of Fos and Junexpression following passive avoidance training in the chick. Eur. J.Neurosci., 7, 563–569.

Gagliardo, A., Ioale, P., Odetti, F., Bingman, V.P., Siegel, J.J. & Vallortigara, G.(2001) Hippocampus and homing in pigeons: left and right hemisphericdifferences in navigational map learning. Eur. J. Neurosci., 13, 1617–1624.

Goda, Y. & Davis, G.W. (2003) Mechanisms of synapse assembly anddisassembly. Neuron, 40, 243–264.

Gould, E. & Tanapat, P. (1999) Stress and hippocampal neurogenesis. Biol.Psychiatry, 46, 1472–1479.

de Groot, D.M. & Bierman, E.P. (1986) A critical evaluation of methods forestimating the numerical density of synapses. J. Neurosci. Meth., 18, 79–101.

Hebb, D.O. (1949) The Organization of Behavior. A NeuropsychologicalTheory. John Wiley, New York.

Hering, H. & Sheng, M. (2001) Dendritic spines: structure, dynamics andregulation. Nat. Rev. Neurosci., 2, 880–888.

Hock, B.J. Jr & Bunsey, M.D. (1998) Differential effects of dorsal and ventralhippocampal lesions. J. Neurosci., 18, 7027–7032.

Horn, G. (2004) Pathways of the past: the imprint of memory. Nat. Rev.Neurosci., 5, 108–120.

Horn, G., Bradley, P. & McCabe, B.J. (1985) Changes in the structure ofsynapses associated with learning. J. Neurosci., 5, 3161–3168.

Horner, C.H., Davies, H.A., Brown, J. & Stewart, M.G. (1996) Reduction innumerical synapse density in chick (Gallus domesticus) dorsal hippocampusfollowing transient cerebral ischaemia. Brain Res., 735, 354–359.

Hu, B., Nikolakopoulou, A.M. & Cohen-Cory, S. (2005) BDNF stabilizessynapses and maintains the structural complexity of optic axons in vivo.Development, 132, 4285–4298.

Inoue, A. & Sanes, J.R. (1997) Lamina-specific connectivity in the brain:regulation by N-cadherin, neurotrophins, and glycoconjugates. Science, 276,1428–1431.

Kahn, M.C. & Bingman, V.P. (2004) Lateralization of spatial learning in theavian hippocampal formation. Behav. Neurosci., 118, 333–344.

Kallen, B. (1962) II. Embryogenesis of brain nuclei in the chick telencephalon.Ergeb Anat Entwicklungsgesch, 36, 62–82.

Knafo, S., Ariav, G., Barkai, E. & Libersat, F. (2004) Olfactory learning-induced increase in spine density along the apical dendrites of CA1hippocampal neurons. Hippocampus, 14, 819–825.

Krebs, J.R., Erichsen, J.T. & Bingman, V.P. (1991) The distribution ofneurotransmitters and neurotransmitter-related enzymes in the dorsomedialtelencephalon of the pigeon (Columba livia). J. Comp. Neurol., 314, 467–477.

Kuenzel, W. & Masson, M. (1988) A Stereotaxic Atlas of the Brain of the Chick(Gallus Domesticus). The Johns Hopkins University Press, Maryland.

Lom, B. & Cohen-Cory, S. (1999) Brain-derived neurotrophic factordifferentially regulates retinal ganglion cell dendritic and axonal arborizationin vivo. J. Neurosci., 19, 9928–9938.

Lowndes, M. & Stewart, M.G. (1994) Dendritic spine density in the lobusparolfactorius of the domestic chick is increased 24 h after one-trial passiveavoidance training. Brain Res., 654, 129–136.

Margrie, T.W., Rostas, J.A. & Sah, P. (1998) Long-term potentiation of synaptictransmission in the avian hippocampus. J. Neurosci., 18, 1207–1216.

Marples, N.M. & Roper, T.J. (1997) Response of domestic chicks to methylanthranilate odour. Anim Behav., 53, 1263–1270.

Mates, S.L. & Lund, J.S. (1983) Spine formation and maturation of type 1synapses on spiny stellate neurons in primate visual cortex. J. Comp.Neurol., 221, 91–97.

McAllister, A.K., Katz, L.C. & Lo, D.C. (1999) Neurotrophins and synapticplasticity. Annu. Rev. Neurosci., 22, 295–318.

McCabe, B.J. & Horn, G. (1994) Learning-related changes in Fos-likeimmunoreactivity in the chick forebrain after imprinting. Proc. Natl Acad.Sci. USA, 91, 11417–11421.

Mileusnic, R., Lancashire, C.L., Johnston, A.N. & Rose, S.P. (2000) APP isrequired during an early phase of memory formation. Eur. J. Neurosci., 12,4487–4495.

Moser, M.B. & Moser, E.I. (1998) Functional differentiation in thehippocampus. Hippocampus, 8, 608–619.



Moser, E., Moser, M.B. & Andersen, P. (1993) Spatial learning impairmentparallels the magnitude of dorsal hippocampal lesions, but is hardly presentfollowing ventral lesions. J. Neurosci., 13, 3916–3925.

Moser, M.B., Trommald, M. & Andersen, P. (1994) An increase in dendriticspine density on hippocampal CA1 pyramidal cells following spatial learningin adult rats suggests the formation of new synapses. Proc. Natl Acad. Sci.USA, 91, 12673–12675.

Nikolakopoulou, A.M. (2005) Neural and synaptic plasticity in the chick brainafter passive avoidance training. PhD Thesis, The Open University,Biological Sciences, Milton Keynes.

O’Malley, A., O’Connell, C. & Regan, C.M. (1998) Ultrastructural analysisreveals avoidance conditioning to induce a transient increase in hippocampaldentate spine density in the 6 hour post-training period of consolidation.Neuroscience, 87, 607–613.

Patel, S.N. & Stewart, M.G. (1988) Changes in the number and structure ofdendritic spines 25 hours after passive avoidance training in the domesticchick, Gallus domesticus. Brain Res., 449, 34–46.

Poo, M.M. (2001) Neurotrophins as synaptic modulators. Nat. Rev. Neurosci.,2, 24–32.

Regolin, L. & Rose, S.P. (1999) Long-term memory for a spatial task in youngchicks. Anim. Behav., 57, 1185–1191.

Richard, S. & Davies, D.C. (2000) Comparison of methyl anthranilate anddenatonium benzoate as aversants for learning in chicks. Physiol. Behav., 70,521–525.

Rose, S.P. (1991) How chicks make memories: the cellular cascade from c-fosto dendritic remodelling. Trends Neurosci., 14, 390–397.

Rose, S.P. (1995a) Cell-adhesion molecules, glucocorticoids and long-term-memory formation. Trends Neurosci., 18, 502–506.

Rose, S.P. (1995b) Glycoproteins and memory formation. Behav. Brain Res.,66, 73–78.

Rose, S.P. & Csillag, A. (1985) Passive avoidance training results in lastingchanges in deoxyglucose metabolism in left hemisphere regions of chickbrain. Behav. Neural Biol., 44, 315–324.

Rose, S.P., Gibbs, M.E. & Hambley, J. (1980) Transient increase in forebrainmuscarinic cholinergic receptor binding following passive avoidancelearning in the young chick. Neuroscience, 5, 169–178.

Rose, S.P. & Stewart, M.G. (1999) Cellular correlates of stages of memoryformation in the chick following passive avoidance training. Behav. BrainRes., 98, 237–243.

Rothstein, J.D., Dykes-Hoberg, M., Pardo, C.A., Bristol, L.A., Jin, L., Kuncl,R.W., Kanai, Y., Hediger, M.A., Wang, Y., Schielke, J.P. & Welty, D.F.(1996) Knockout of glutamate transporters reveals a major role for astroglialtransport in excitotoxicity and clearance of glutamate. Neuron, 16, 675–686.

Sandi, C., Davies, H.A., Cordero, M.I., Rodriguez, J.J., Popov, V.I. & Stewart,M.G. (2003) Rapid reversal of stress induced loss of synapses in CA3 of rathippocampus following water maze training. Eur. J. Neurosci., 17, 2447–2456.

Sandi, C., Patterson, T.A. & Rose, S.P. (1993) Visual input and lateralization ofbrain function in learning in the chick. Neuroscience, 52, 393–401.

Sandi, C. & Rose, S.P. (1997) Training-dependent biphasic effects ofcorticosterone in memory formation for a passive avoidance task in chicks.Psychopharmacology (Berl.), 133, 152–160.

Sandi, C., Rose, S.P. & Patterson, T.A. (1992) Unilateral hippocampal lesionsprevent recall of a passive avoidance task in day-old chicks. Neurosci. Lett.,141, 255–258.

Scholey, A.B., Mileusnic, R., Schachner, M. & Rose, S.P. (1995) A role fora chicken homolog of the neural cell adhesion molecule L1 in consolida-tion of memory for a passive avoidance task in the chick. Learn. Mem., 2,17–25.

Scholey, A.B., Rose, S.P., Zamani, M.R., Bock, E. & Schachner, M. (1993) Arole for the neural cell adhesion molecule in a late, consolidating phase ofglycoprotein synthesis six hours following passive avoidance training of theyoung chick. Neuroscience, 55, 499–509.

Smith, M.A., Makino, S., Kvetnansky, R. & Post, R.M. (1995) Stress andglucocorticoids affect the expression of brain-derived neurotrophic factorand neurotrophin-3 mRNAs in the hippocampus. J. Neurosci., 15, 1768–1777.

Sterio, D.C. (1984) The unbiased estimation of number and sizes of arbitraryparticles using the disector. J. Microsc., 134, 127–136.

Stewart, M.G., Csillag, A. & Rose, S.P. (1987) Alterations in synaptic structurein the paleostriatal complex of the domestic chick, Gallus domesticus,following passive avoidance training. Brain Res., 426, 69–81.

Stewart, M.G., Davies, H.A., Sandi, C., Kraev, I.V., Rogachevsky, V.V., Peddie,C.J., Rodriguez, J.J., Cordero, M.I., Donohue, H.S., Gabbott, P.L. & Popov,V.I. (2005) Stress suppresses and learning induces plasticity in CA3 of rathippocampus: a three-dimensional ultrastructural study of thorny excres-cences and their postsynaptic densities. Neuroscience, 131, 43–54.

Stewart, M.G. & Rusakov, D.A. (1995) Morphological changes associated withstages of memory formation in the chick following passive avoidancetraining. Behav. Brain Res., 66, 21–28.

Suge, R. & McCabe, B.J. (2004) Early stages of memory formation in filialimprinting: Fos-like immunoreactivity and behavior in the domestic chick.Neuroscience, 123, 847–856.

Szekely, A.D. (1999) The avian hippocampal formation: subdivisions andconnectivity. Behav. Brain Res., 98, 219–225.

Szekely, A.D. & Krebs, J.R. (1996) Efferent connectivity of the hippocampalformation of the zebra finch (Taenopygia guttata): an anterograde pathwaytracing study using Phaseolus vulgaris leucoagglutinin. J. Comp. Neurol.,368, 198–214.

Thode, C., Bock, J., Braun, K. & Darlison, M.G. (2005) The chickenimmediate-early gene ZENK is expressed in the medio-rostralneostriatum ⁄ hyperstriatum ventrale, a brain region involved in acousticimprinting, and is up-regulated after exposure to an auditory stimulus.Neuroscience, 130, 611–617.

Tigges, J., Herndon, J.G. & Rosene, D.L. (1996) Preservation into old age ofsynaptic number and size in the supragranular layer of the dentate gyrus inrhesus monkeys. Acta Anat. (Basel), 157, 63–72.

Ueyama, T., Kawai, Y., Nemoto, K., Sekimoto, M., Tone, S. & Senba, E.(1997) Immobilization stress reduced the expression of neurotrophins andtheir receptors in the rat brain. Neurosci. Res., 28, 103–110.

Unal, B., Bradley, P.M., Sahin, B., Canan, S., Aslan, H. & Kaplan, S. (2002)Estimation of numerical density and mean synaptic height in chickhippocampus 24 and 48 hours after passive avoidance training. Brain Res.Dev. Brain Res., 136, 135–144.



Documents

nikolakopoulou et al., 2006a