R

CN

CD

h

���

a

ARRAA

KNHSPWR

1

eid(Neows[i

0h

Behavioural Brain Research 235 (2012) 326– 333

Contents lists available at SciVerse ScienceDirect

Behavioural Brain Research

j ourna l ho me pa ge: www.elsev ier .com/ locate /bbr

esearch report

orrelation between hippocampal levels of neural, epithelial and inducibleOS and spatial learning skills in rats

¸ igdem Gökc ek-Sarac , Serdar Karakurt, Orhan Adalı, Ewa Jakubowska-Dogru ∗

epartment of Biological Sciences, Middle East Technical University, Ankara 06800, Turkey

i g h l i g h t s

There are strain-dependent differences in hippocampal levels of NOS isoforms.Learning rate correlated positively with nNOS and negatively with iNOS level.In inbred Wistar rats, learning score correlated negatively with eNOS level.

r t i c l e i n f o

rticle history:eceived 5 April 2012eceived in revised form 27 June 2012ccepted 6 August 2012vailable online 14 August 2012

eywords:OS isoformsippocampuspatial memory

a b s t r a c t

In the present study, to better understand the role of different nitric oxide synthase (NOS) isoformsin hippocampus-dependent forms of learning, we examined the expression of neural, endothelial, andinducible NOS in the hippocampus of young-adult rats classified as “poor” and “good” learners onthe basis of their performance in the partially baited 12-arm radial maze. Taking into considerationstrain-dependent differences in learning skills and NOS expression, experiments were performed on twodifferent lines of laboratory rats: the inbred Wistar (W) and the outcrossed Wistar/Spraque-Dawley (W/S)line. The hippocampal levels of NOS proteins were assessed by Western Blotting. In the present study,genetically more homogenous W rats showed a slower rate of learning compared to the genetically lesshomogenous outcrossed W/S rats. The deficient performance in the W rat group compared to outcrossed

artially baited 12-arm radial mazeestern Blot

ats

W/S rats, and in “poor” learners of both groups compared to “good” learners was due to a higher percent-age of reference memory errors. The overall NOS levels were significantly higher in W group comparedto outcrossed W/S rats. In both rat lines, the rate of learning positively correlated with hippocampallevels of nNOS and negatively correlated with iNOS levels. Hippocampal eNOS levels correlated nega-tively with animals’ performance but only in the W rats. These results suggest that all 3 NOS isoforms are

fferen

implemented but play di

. Introduction

First discovered in 1992, nitric oxide (NO) is a short-lived,ndogenously produced gas that has a signaling function in themmune, cardiovascular, and nervous systems [1–4]. Its productionepends on the expression and activity of enzymes, NO synthasesNOSs) catalyzing the conversion of l-arginine to l-citrulline andO [5]. By modulating NO production in the nervous system byither administration of NO precursors, donors, or scavengers [6],r by pharmacological or genetic NO synthase inactivation [7], itas revealed that this gaseous neurotransmitter plays a role in

ynaptic plasticity and thus in learning and memory formation4]. Many studies demonstrated that in several brain areas includ-ng the cortex, cerebellum, and hippocampus, NO is implemented

∗ Corresponding author. Tel.: +90 312 2105186; fax: +90 312 2107976.E-mail address: bioewa@metu.edu.tr (E. Jakubowska-Dogru).

166-4328/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.bbr.2012.08.005

t roles in neural signaling.© 2012 Elsevier B.V. All rights reserved.

in the induction of long-term potentiation (LTP) and long-termdepression (LTD) which are considered cellular models of learn-ing and memory [8–18]. Parallel to this, in many studies carriedout on different animal species from invertebrates to mammals, afacilitatory effect of NO was observed in several types of behav-ioral learning [19,20]. This included olfactory learning [21–23],cerebellum-dependent motor learning [24], acquisition of activeand passive avoidance [25–28], and hippocampus-dependent spa-tial learning, the latter one accepted as an animal model of humanepisodic memory [29].

The brain levels of NO are determined by the expression and/oractivity of NO synthase which is present in three isoforms: oneinducible (iNOS) and two constitutively expressed: endothelial(eNOS) and neuronal (nNOS). The inducible form of NOS was

found in activated macrophages, astroglia and microglia, the con-stitutively expressed endothelial form was originally found inmicrovessels and glial cells [30–32] while neural form was detectedin different types of neurons of the cerebellum, mesencephalon,

al Brain Research 235 (2012) 326– 333 327

hSNtepifaiatbHfpii(Ncow[oholNDvadacpastiddkts

bha,“pdelW

2

2

oAt

id

C . Gökc ek-Sarac et al. / Behaviour

ypothalamus, striatum, cerebral cortex and hippocampus [33].ome studies, however, have demonstrated location of all threeOS isoforms in brain neurons [34–38] which suggests that all

hree NOS isoforms may be implicated in neural plasticity. How-ver, the role of different isoforms of NO synthase in the learningrocess is still not clear and the data are inconsistent. In rats, an

ncreased expression of nNOS was noted in the dentate gyrus androntal cortex after a water-rewarded spatial alternation task [29]nd increased expression of both, nNOS and eNOS was observedn hippocampus and prefrontal cortex during learning of an oper-nt conditioning task [39]. Interestingly, after place learning inhe water maze, a strong induction of iNOS was also revealedy genome-scale screening at 1, 6, and 24 h after training [40].owever, numerous studies examining the potential role of dif-

erent NO synthases in learning and memory by either genetic orharmacological NOS inactivation during training brought highly

nconsistent results. Systemic or local (intrahippocampal) admin-stration of NOS inhibitors such as l-nitroarginine methyl esterl-NAME), N omega-nitro-l-arginine (l-NA), or 7-nitroindazole (7-I) was reported to impair different forms of learning includingonditioned eyeblink response acquisition [41], the acquisitionf passive avoidance task [42,43] and place learning in eitherater maze [41,44–48], fully or partially baited radial arm maze

45,49,50] and also in a shock-motivated 14 unit T-maze [51]. Somether studies, however, indicated no effects of systemic or subtotalippocampal inhibition of NO synthesis by pharmacological agentsn learning and memory [52–55] and some reported facilitation ofearning process [56]. Few studies that used genetic inactivation ofO synthases (knockout studies) also brought discrepant results.espite the previously reported detrimental effect of eNOS inacti-ation on the induction of LTP in the cortex and hippocampal CA1rea, eNOS knockout in mice was shown not to affect hippocampus-ependent spatial learning in the radial arm maze (RAM) task [57]nd facilitate place learning in the water maze [7]. These discrepan-ies in the results obtained in the studies investigating the effect ofharmacological NOS inactivation on learning and memory mightrise from the fact that applied NOS inhibitors are generally lesselective and their effects depend, among others, on the dose andhe way of administration. On the other hand, a drawback of knock-ng out a gene (unless it is a conditional knockout) is the lack ofetailed studies on how the changes produced by gene knockouturing embryonic development affect the adult phenotype. It isnown that a gene knockout may cause changes different thanhose observed when gene expression is manipulated in the adultubjects.

In light of all these contradictory results, to contribute to aetter understanding of the role of different NOS isoforms inippocampus-dependent forms of learning, we applied a differentpproach to these studies and examined the expression of n-, e-

and iNOS in the hippocampus of young-adult rats classified asgood” and “poor” learners on the basis of their performance in theartially baited 12-arm radial maze. Taking into consideration theifferences between rodent strains in learning skills [57] and NOSxpression [58,59], we carried out our experiments on two differentines of laboratory rats, the inbred Wistar (W) and the outcrossed

istar/Spraque-Dawley (W/S) line.

. Materials and methods

.1. Subjects

Experiments were conducted on 27 young male Wistar (W) and 33 young maleutcrossed Wistar/Spraque-Dawley rats (W/S) obtained from GATA Animal Facility,

nkara, Turkey. Rats were kept in home cages in groups of three under a constant

emperature (21 ◦C) and a 12/12 h light/dark cycle.Rats were food restricted one week before training and throughout the exper-

ment to maintain 85% of their ad libitum body weight. Body weight was recordedaily. Subjects were taken into experiments once a day, in the same order and at

Fig. 1. The scheme of 12-arm radial maze with baited arms marked by black dots.

about the same time. The animal care procedures and all experimental manipu-lations in the proposed study were pursued in accordance with the METU EthicCommittee Regulations (Protocol No: 2009/01).

2.2. Apparatus

The apparatus was a twelve arm radial maze made of plywood, painted flatgrey and elevated 80 cm above the floor of a room containing a variety of distinctdistal cues. The maze consisted of a central platform measuring 40 cm in diameterand twelve 60 cm long and 9 cm wide arms. Each arm was framed by 15 cm highsidewalls made of clear plexiglas, preventing rats from directly crossing from onearm to the other, but allowing them to see the cues in the room. A guillotine doorwas placed at the entrance to each arm. The doors could be moved separately one byone, or in concert. At the end of each arm, there was a food well (2 cm wide and 2 cmdeep). To unify the food odor traces throughout the maze, in all food wells, 1 cmfrom the bottom, a perforated partition was inserted beneath which food pellets(two chocolate flavored rice puffs) were placed. The same pellets were used as abait in the course of the training. The maze was illuminated with a dim homogenouslight.

2.3. Procedure

The behavioral procedure was adopted from Jakubowska-Dogru et al. [60]. Priorto the experiments, for six consecutive days, all rats were daily handled, for 5 mineach. At the beginning of the experiments, rats were given 5 days of habituation andshaping trials, in which they were allowed to explore the maze for 10 min each dayand eat all the food pellets scattered throughout the maze. On each successive day,the number of pellets was reduced and they were placed closer to the ends of thearms. At the end of the shaping training, food (2 g chocolate flavored rice puffs) wasplaced only in the food wells of six semi-randomly selected arms (Fig. 1). A partiallybaited RAM such as used in the present study allowed simultaneous estimation ofreference and working memory errors. In the daily trials, rats were placed on thecentral platform facing different directions. The guillotine doors were raised and ratswere allowed to make their first choice by entering one of the arms. Each time, therat returned to the central platform after making a choice, the guillotine doors wereshut for 5 s and only then the animal was allowed to make the next choice. Rats werepermitted to choose among the arms until they completed the trial by either eatingall the pellets or making 12 choices, or 10 min had passed, whichever came first. Theentry to the arm was counted only when the rat crossed the midpoint of the alleywith its two forepaws. All rats were trained to the criterion of three consecutive dailytrials with maximum 3 incorrect arm entries out of the first 18 choices (6 per trial).In the course of the experiments the following measures were recorded: (a) numberof choices to the acquisition criterion; (b) total number of entries to unbaited arms

(reference memory errors, RMEs); (c) total number of re-entries to either baited orunbaited arms (working memory errors, WMEs).

On the basis of their performance in the radial maze, rats were classifiedas “good” (number of days to the acquisition criterion ≤ group mean − 3 SEM),

3 al Brain Research 235 (2012) 326– 333

“r

rpa

2

cioE(wowrt(

2

eaapoewmwemiotaCfaiatgIu

2

WWpowwehhs

3

3

coa(etil

Fig. 2. Mean (±SEM) number of choices to the arbitrary acquisition criterion (A) andmean (±SEM) numbers of working and reference memory errors in inbred W andoutcrossed W/S rat groups (B). Error bars denote ± SEM. Asterisks denote the level

28 C . Gökc ek-Sarac et al. / Behaviour

poor” (number of days to the acquisition criterion ≥ group mean + 3 SEM), and theemaining “intermediate” learners.

After all rats had reached the task acquisition criterion, they were given oneetraining trial and decapitated 15 min later. The brains were rapidly removed andlaced on dry ice, hippocampi dissected and immediately frozen in liquid nitrogennd then stored at −80 ◦C until the biochemical assays were performed.

.4. Tissue sample preparation

Left and right hippocampi of “good” and “poor” learners from both W and out-rossed W/S rats were homogenized by ultrasonication in the presence of proteasenhibitors. Tissue samples were transferred to a centrifuge tube containing 800 �lf Triton lysis buffer (1% Triton X-100, 20 mM Tris–HCL pH 7.4, 150 mM NaCl, 5 mMGTA, 10 mM EDTA, 10% glycerol, 20 mM HEPES and protease inhibitor cocktailRoche)). Samples were then subjected to sonication. Sonication was performedith a sonic dismembrator in an ice-bath with a microtip probe set to a power

utput of 40 W for three cycles of 5 s each (15 s interval between sonications). After-ards they were centrifuged at 13,500 rpm for 30 min at 4 ◦C, then supernatant was

emoved and stored at −80 ◦C until use. Protein concentrations were determined inhe supernatant by Lowry et al. [61] method using crystalline bovine serum albuminBSA) as a standard.

.5. Western Blot protein analysis

nNOS, eNOS and iNOS protein levels in the hippocampus were assessed by West-rn Blot assays. The hippocampal samples of “good” and “poor” learners from Wnd outcrossed W/S rats were processed on the same blot. To run the Western Blotssay, 20 �l (2 �g/�l protein) samples of hippocampal protein homogenates werelaced on 7.5% polyacrylamide gels and the separated proteins were transferrednto nitrocellulose paper by means of electroblotting with Mini Protean Tetra Cellquipment (Bio-Rad Laboratories, Richmond, CA, USA). Afterwards the membranesere incubated at room temperature in blocking solution made of 5% dried non-fatilk in TBS-T buffer (20 mM Tris–HCl, 0.5 M NaCl and 0.05% Tween 20, pH 7.6). Blotsere then incubated for 2 h at room temperature with primary antibodies against

ach NOS isoforms dissolved in 5% dried non-fat milk in TBS-T buffer. Monoclonalouse anti nNOS (Invitrogen #37-2800), anti eNOS (SantaCruz #136977), and anti

NOS (SantaCruz #7271) were used. Primary antibodies were applied at a dilutionf 1:1000 (nNOS and iNOS) and 1:250 (eNOS). After being rinsed with TBST buffer,he membranes were incubated for 1 h at room temperature with the secondarylkaline phosphatase conjugated antibody (goat anti-rat IgG-ALP conjugate, Santaruz #sc-3824) used at a dilution of 1:2000 for both nNOS and iNOS and of 1:1000

or eNOS in 5% dried non-fat milk in TBS-T buffer. Blots were rinsed with TBST buffernd incubated with substrate solution as described by Ey and Ashman [62] to visual-ze the specifically bound antibodies. Blots were then incubated with a monoclonalntibody directed against �-actin, which was used as a control to normalize NOS pro-ein expression levels. The final images were photographed using a computer-basedel imaging instrument (VilberLourmat) with Infinity-Capt version 12.9 software.mmunoreactive protein bands were quantified by densitometric scanning methodsing an Image J software package.

.6. Statistical analysis

To evaluate the behavioral data, one-way ANOVA with either strain (W versus/S) or learning skills (“poor” versus “good”) as independent variable was applied.ilcoxon test for paired comparisons was used to assess the potential differences in

rotein levels between left and right hippocampi. Two-way ANOVA was conductedn the NOS data to assess the main effects of strain and animals’ learning skills asell as a potential strain x learning skills interaction. Additionally, one-way ANOVAith a learning group as independent factor was performed for each rat strain and

ach NOS isoform separately. To correlate the individual learning performance withippocampal levels of three different NOS isoforms, the Pearson’s correlation testas been applied. A p value less than or equal to 0.05 was considered as statisticallyignificant. The statistical package SPSS v.18 was used for statistical analyses.

. Results

.1. Radial-arm maze task

The number of choices to the arbitrary acquisition criterion of 3onsecutive daily trials with maximum 3 incorrect arm entries outf the first 18 choices (6 per trial) and the total numbers of WMEsnd RMEs were compared between W (n = 27) and outcrossed W/Sn = 33) rats (Fig. 2). One-way ANOVA revealed a significant differ-

nce between W and outcrossed W/S rats in the number of choiceso criterion (F(1;59) = 6.919, p ≤ 0.01) with W rats executing signif-cantly more reference (F(1;59) = 14.59, p ≤ 0.001) and significantlyess working memory errors (F(1;59) = 4.41, p ≤ 0.04) as compared

of significance: *p < 0.05, ***p < 0.001.

to outcrossed W/S rats. In both rat groups, the mean total numberof RMEs (49.48 ± 3.53 and 36.5 ± 1.02, respectively) was signifi-cantly (p ≤ 0.001) higher than the mean total number of WMEs(25.92 ± 1.40 and 29.72 ± 1.16, respectively).

According to their performance in the RAM task, 9 W and 8 out-crossed W/S rats were classified as “good” learners (total numberof choices to the acquisition criterion ≤ group mean − 3 SEM) while8 Wistar and 5 outcrossed W/S rats were classified as “poor” learn-ers (total number of choices to the acquisition criterion ≥ groupmean + 3 SEM) with the remaining rats classified as “intermedi-ate” learners. Fig. 3 presents the mean numbers of total choices tocriterion (A) and mean numbers of working and reference mem-ory errors (B) estimated for “good” and “poor” learners from Wand outcrossed W/S rat groups. One-way ANOVA revealed a signif-icant difference between “poor” learners from W and outcrossedW/S rat groups in the mean number of total choices to criterion(F(1;12) = 9.225, p ≤ 0.011). No such difference was found betweenW and outcrossed W/S “good” learners.

In all animal groups, RMEs significantly outnumbered WMEs.Within both W and outcrossed W/S rat groups, “poor” learnerscompared to “good” learners, showed significantly higher numberof both WMEs (F(1;16) = 11.684, p ≤ 0.004; F(1;12) = 14.461, p ≤ 0.003,respectively) and RMEs (F(1;16) = 65.435, p ≤ 0.001; F(1;12) = 43.700,p ≤ 0.001, respectively). Highly significant difference was observedbetween W and outcrossed W/S “poor” learners in the number ofRMEs (F(1;12) = 14.58, p ≤ 0.003), with “poor” learners from the out-crossed W/S rat group making much less RMEs than the “poor”learners from the W rat group. Conversely, there was no significant

difference between W and outcrossed W/S “poor” learners in thenumber of WMEs. There was also no significant difference in thecount of either RMEs or WMEs between “good” learners from Wand outcrossed W/S groups.

C . Gökc ek-Sarac et al. / Behavioural Brain Research 235 (2012) 326– 333 329

Fig. 3. Mean number (±SEM) of total choices to criterion (A) and mean number(±SEM) of working and reference memory errors (B) emitted by “good” and “poor”learners from inbred W and outcrossed W/S rat groups, respectively. Error barsd*

3

pTnn

Fig. 5. Western Blot analysis of the protein levels of nNOS, eNOS, and iNOS in thehippocampus of inbred W and outcrossed W/S rats. Band quantification is expressed

p ≤ 0.001 and F(1;59) = 45.179, p ≤ 0.001, respectively) and a signifi-

Fre

enote ± SEM. Asterisks denote the level of significance: *p < 0.05, **p < 0.01, and**p < 0.001.

.2. Western Blot protein expression assays of NOS isoforms

Western Blot analysis for NOS expression was carried out forroteins extracted from the left and right hippocampi separately.

herefore, for each group, the number of samples was twice theumber of subjects: W “good” n = 18; W “poor” n = 16; W/S “good”

= 16; W/S “poor” n = 10. The differences in band quantification

ig. 4. Immunoreactive protein bands representing nNOS (the upper lane, 170 kDa), eNOight (R) hippocampi of “good” (G) and “poor” (P) learners in inbred W (A) and outcrossedach gel.

as the mean ± SEM of the relative intensity with respect to that of �-Actin used as theinternal control. Error bars denote ± SEM. Asterisks denote the level of significance:*p < 0.05 and ***p < 0.001.

between left and right hippocampi were statistically insignificant(Wilcoxon test for paired comparisons, p ≥ 0.1). Therefore, the leftand the right hippocampal Western Blot data were pooled for thefurther statistical analyzes. In the Western Blot analysis, nNOS,eNOS, and iNOS immunoreactive proteins were detected as bandsof 170 kDa, 133 kDa, and 130 kDa, respectively (Fig. 4).

Expression levels of NO synthases were determined bycomparing band intensities with that of �-Actin, which wasdetected at the position corresponding to a molecular weightof 42 kDa and was used as the internal control (Figs. 5 and 6).Two-way ANOVA applied to nNOS data revealed highly signif-icant main effect of both, the strain and the learning skills(F(1;59) = 388.117, p ≤ 0.001 and F(1;59) = 13.947, p ≤ 0.001, respec-tively), however, the strain × learning skills interaction was yieldedinsignificant. The same analysis applied to eNOS data revealedsignificant main effects of the strain and the learning skills(F(1;59) = 11.386, p ≤ 0.001 and F(1;59) = 31.886, p ≤ 0.001, respec-tively), but this time, strain × learning skills interaction wasalso significant (F(1;59) = 17.014, p ≤ 0.001). Similarly, two-wayANOVA conducted on iNOS data confirmed a highly significantmain effect of the strain and the learning skills (F(1;59) = 68.522,

cant strain × learning skills interaction (F(1;59) = 20.517, p ≤ 0.001).The comparison of simple effects was done by one-way ANOVAwith either strain or group (“good” versus “poor” learners) as

S (the middle lane, 133 kDa), and iNOS (the bottom lane, 130 kDa) from left (L) and W/S (B) rat groups. The �-Actin band (42 kDa) was used as an internal control for

330 C . Gökc ek-Sarac et al. / Behavioural Brain Research 235 (2012) 326– 333

F OS, eNo meanc and

iaorinF

NoApflacee(

rtptosdtale

ac(cwWpnt

4

dW

ig. 6. Comparison of the results of Western Blot analysis of the protein levels of nNutcrossed W/S rat groups, independently. Band quantification is expressed as the

ontrol. Error bars denote ± SEM. Asterisks denote the level of significance: *p < 0.05

ndependent factor. As shown in Fig. 5, quantitative immunoblotnalysis of nNOS, eNOS, and iNOS levels in the hippocampusf inbred W and outcrossed W/S rats done by one-way ANOVAevealed significantly lower expression of all three NOS isoformsn W rats as compared to W/S rats with the greatest difference inNOS level (F(1;59) = 346.852, p ≤ 0.001; F(1;59) = 5.610, p ≤ 0.021 and(1;59) = 26.204, p ≤ 0.001, for nNOS, eNOS and iNOS, respectively).

The Fig. 6 illustrates the differences in the levels of threeOS isoforms between “good” and “poor” learners in W andutcrossed W/S group, independently. In both groups, one-wayNOVA yielded a significantly higher expression of hippocam-al nNOS (F(1;33) = 7.129, p ≤ 0.012 and F(1;25) = 6.487, p ≤ 0.018or W and outcrossed W/S groups, respectively) and significantlyower expression of hippocampal iNOS (F(1;33) = 6.896, p ≤ 0.013nd F(1;25) = 30.671, p ≤ 0.001, respectively) in “good” learners asompared to “poor” learners. “Good” learners of both groups, alsoxhibited lower hippocampal levels of eNOS, however, this differ-nce was yielded statistically significant only in W group of ratsF(1;33) = 56.094, p ≤ 0.001).

Pearson’s correlation analysis performed on the data from Wat group confirmed a significant positive correlation between theotal number of choices to criterion and both eNOS (r = 0.853**,

≤ 0.001) and iNOS expression (r = 0.487**, p ≤ 0.003). In W group,he negative correlation between the nNOS level and the numberf choices to reach the performance criterion was only marginallyignificant (r = −0.274, p = 0.116). Similar analysis applied to theata from outcrossed W/S rat group revealed a significant nega-ive correlation between the total number of choices to criterionnd nNOS expression (r = −0.530**, p ≤ 0.005) and a positive corre-ation between the total number of choices to criterion and iNOSxpression (r = 0.609**, p ≤ 0.001).

Comparison of NOS isoforms’ levels in the hippocampi of “good”nd “poor” learners between the two strains confirmed signifi-antly higher level of nNOS (F(1;33) = 404.295, p ≤ 0.001) and iNOSF(1;33) = 6.865, p ≤ 0.013) in outcrossed W/S “good” learners asompared to W “good” learners. The levels of nNOS and iNOSere also higher in outcrossed W/S “poor” learners as compared to

“poor” learners (F(1;25) = 102.761, p ≤ 0.001 and F(1;25) = 99.430, ≤ 0.001, respectively), however, W/S “poor” learners showed sig-ificantly lower level of eNOS compared to “poor” learners fromhe W rat strain (F(1;25) = 14.254, p ≤ 0.001).

. Discussion

In the present study, a significant difference in hippocampus-ependent spatial learning was shown between an inbred line ofistar (W) rats and the outcrossed Wistar/Sprague Dawley (W/S)

OS, and iNOS in the hippocampus of “good” and “poor” learners from inbred W and ± SEM of the relative intensity with respect to that of �-Actin, used as the internal***p < 0.001.

rats. W rats, classified as a genetically more homogenous line oflaboratory rats, had a higher percent of “poor” learners and man-ifested a slower rate of acquisition of a spatial memory task dueto the significantly higher number of reference memory errorscompared to the genetically less homogenous outcrossed W/S rats.Between-group comparison of learning scores done for “good”and “poor” learners independently, revealed a lack of a significantdifference in the rate of learning between W and outcrossed W/S“good” learners. However, the number of RMEs was significantlyhigher and thus, the rate of reaching performance criterion signif-icantly slower in the inbred line of W “poor” learners compared tooutcrossed W/S “poor” learners.

Interestingly, the levels of all three NOS isoforms were sig-nificantly higher in the hippocampus of outcrossed W/S ratscompared to inbred W rats with the most prominent differencein the constitutive expression of nNOS. These results demonstratethat learning skills and the expression of NOS isoforms varymarkedly in different strains. Strain-dependent variation in NOSexpression was also previously reported by other authors [57,58].Assuming that depending on its concentration, NO may act as aexcitatory or inhibitory modulator of learning-related neuroplas-ticity [27,56], strain-dependent variation in NOS expression maypartially account for the inconsistent results regarding the role ofdifferent NOS isoforms in learning and memory formation obtainedin the experiments wherein NO synthesis was either inhibitedor elevated by pharmacological agents. In our experiments too,strain-dependent differences in NOS expression make the inter-pretation of the obtained results more difficult. Certainly, the loweroverall levels of all three NOS isoforms in W rat group cannot beheld responsible for the deficient learning in these rats comparedto the outcrossed W/S rats. Our attempt to correlate the expres-sion of NOS isoforms with individual learning skills of animalsrepresenting random rat populations based on detailed analysis ofbehavioral and Western Blot data in the subclasses of “good” and“poor” learners provided an evidence for the existence of a directrelationship between nNOS level and the rate of place learning inthe RAM task. This notion is supported by significantly higher nNOSlevels in “good” learners” as compared to “poor” learners observedin both W and outcrossed W/S rat groups and a positive correlationbetween nNOS expression and individual animal performanceassessed for each group independently by Pearson’s correlationtest. These results are consistent with the previous findingsdemonstrating deficient learning in rats trained in spatial memory

tasks, a water maze and an 8-arm radial maze, after administrationof a selective inhibitor of nNOS, 7-nitro indazole (7-NI), which wasshown to reduce neuronal NO synthase by 85% without affectingblood pressure [45]. Administration of 7-NI was also reported

al Brai

tmtCows[piCpmbsfAaSiaotsac

oe“lalclcWwttcotwg

misppHtd

snofbnsioef

C . Gökc ek-Sarac et al. / Behaviour

o impair memory retention as assessed by the probe trial of aorris water maze (MWM) task in mice [43]. Also, bilateral injec-

ions of l-NAME, a preferential nNOS inhibitor, to the hippocampalA1 region were reported to result in dose-dependent impairmentf place learning in the water maze in Wistar rats. This impairmentas reversed by co-administration of l-arginine, the NO precursor,

uggesting involvement of nNOS in the water maze place learning46]. MWM is a typical reference memory task and deficienterformance in this learning task observed after NOS inactivation

ndicates the role of NO in long-term reference memory formation.onversely, a partially baited 12-arm RAM task applied in theresent study includes components of both reference and workingemory. The data collected in this study show that although in

oth W and outcrossed W/S rat groups “poor” learners emittedignificantly more working and reference memory errors, therequency of RMEs was significantly higher than that of WMEs.lso when comparing W and outcrossed W/S laboratory rat lines,

highly significant difference was noted only in RME counts.imilar results were reported by Zou et al. [50] who observedmpaired performance in partially baited 8-arm RAM in ratsdministered NOS inhibitors l-NAME and 7-IN with greater impactn reference than on working memory errors. Taken together,hese observations confirm that local production of NO duringynaptic activation, in addition to short-lasting functional changest synapses, also contributes to long-term remodeling of neuralonnections responsible for establishing long-term memories [63].

Despite the higher overall levels of all three NOS isoforms inutcrossed W/S rats compared to W rat line, in W “poor” learners,NOS level was significantly higher than that in outcrossed W/Spoor” learners. eNOS was also significantly higher in W “poor”earners compared with W “good” learners. Pearson’s correlationnalysis confirmed an inverse relationship between hippocampalevel of eNOS and the learning rate in inbred W rats but not in out-rossed W/S rats. The negative correlation between hippocampalevel of the constitutively expressed eNOS and the spatial learningapability observed in the present study in the laboratory strain of

rats is consistent with some previous reports by other authorsho demonstrated accelerated place learning, better retention of

he acquired place preference, and improved reversal learning inhe water maze in the eNOS-/-mice [7]. However, the latter resultould also be due to the impact of eNOS knockout during embry-nic development on some other aspects of behavior relevant tohe learning process such as activity or anxiety. NOS-like activityas found in regions involved in activity and anxiety, such as basal

anglia and amygdala [35,64,65].In contrast to these results, the in vitro experiments on LTP in

ice hippocampus demonstrated the most profound attenuationn CA1 LTP in animals with double nNOS-/eNOS- mutation whichuggested that both NOS isoforms play a role in long-term synapticotentiation (a cellular model of memory formation) and may com-ensate for each others’ effects in mice with single mutation [66].owever, a direct relationship between LTP and memory was ques-

ioned by some authors on the basis of anatomical and functionalata [67–69].

The positive correlation found in the present study between thepatial learning capacity and the hippocampal nNOS levels and theegative correlation between the learning scores and the levelsf hippocampal eNOS in W rats may be related to apparently dif-erent neurophysiological roles of these two NOS isoforms. It haseen demonstrated that genetic inactivation of eNOS but not ofNOS led to a strong reduction in NMDA-induced GABA release ineveral brain regions, including the hippocampus, whereas NMDA-

nduced glutamate release was reduced only by the inactivationf the neuronal isoform [70]. According to these results, elevatedNOS could facilitate the inhibitory while elevated nNOS couldacilitate excitatory neural actions. However, it should be noted

n Research 235 (2012) 326– 333 331

that NO is at the same time a potent vasodilator and eNOS locatedin brain microvessels and astrocytes is implicated in the regula-tion of cerebral circulation and thus supply of O2 and glucose tothe active brain regions [71]. Impaired eNOS-dependent regulationof cerebral blood flow is considered one of the cardiovascular riskfactors contributing to stroke and vascular cognitive impairmentduring aging. Our results suggest that in some rat strains includingW rats, the role of eNOS in neural signaling may be greater than itsrole in the regulation of cerebral blood flow.

In both W and outcrossed W/S rat lines the overall level of hip-pocampal eNOS was significantly higher than that of n and iNOS.Interestingly, in both rat groups the level of iNOS expression wasrelatively high (higher than or equal to that of nNOS, Fig. 5). Thisresult is consistent with the strong induction of hippocampal iNOSin rats trained in place learning task previously reported by otherauthors [40]. In addition, in this study, a significant negative corre-lation was found between rate of learning and hippocampal iNOSlevels in both rat lines. Similarly, in aged Long Evans rats trained inthe MWM, decreased nNOS and increased iNOS levels were foundin the hippocampus and the cortex of cognitively impaired animals[72]. All these suggest that iNOS, in addition to its well-known rolein brain defense mechanisms and its deleterious function in neuro-logical disorders, may also play some role in learning and memoryformation.

To our knowledge, this is the first study investigating thecorrelation between hippocampal levels of 3 NOS isoforms andindividual learning skills in random populations of laboratory ratlines. Taken together, our results showed that there are strain-dependent differences in the overall levels of neuronal, endothelialand inducible NOS which may partially explain the discrepan-cies in the literature regarding the role of different NOS isoformsin NO-dependent neuroplasticity. In both rat lines, rate of learn-ing positively correlated with hippocampal levels of nNOS andnegatively correlated with iNOS levels. Hippocampal eNOS levelscorrelated negatively with animals’ performance but only in the Wrats and not in the outcrossed W/S rats. These results suggest thatnot only nNOS and eNOS but also iNOS may impact adult neuralplasticity laying the basis of learning process and memory forma-tion in young subjects although mechanistic explanation of iNOSrole requires further elucidation. The opposite effects of elevatednNOS and eNOS expression on learning observed in W rat line is inaccordance with different localization of these two enzymes in thehippocampal neurons and indicates the different roles of these twoNOS isoforms in neuro-signaling.

Acknowledgement

This work has been supported by the grant from the Turk-ish Scientific and Technical Council (TÜBITAK) to EJD; Project No:109S133.

References

[1] Bredt DS, Snyder SH. Nitric oxide, a novel neuronal messenger. Neuron1992;8:3.

[2] Esplugues JV. NO as a signalling molecule in the nervous system. British Journalof Pharmacology 2002;135:1079–95.

[3] Holscher C. Nitric oxide, the enigmatic neuronal messenger: its role in synapticplasticity. Trends in Neurosciences 1997;20:298–303.

[4] Prast H, Philippu A. Nitric oxide as modulator of neuronal function. Progress inNeurobiology 2001;64:51–68.

[5] Moncada S, Higgs A. The l-arginine-nitric oxide pathway. The New EnglandJournal of Medicine 1993;329:2002.

[6] Calabrese V, Mancuso C, Calvani M, Rizzarelli E, Butterfield DA, Stella AMG.

Nitric oxide in the central nervous system: neuroprotection versus neurotox-icity. Nature Reviews Neuroscience 2007;8:766–75.

[7] Frisch C, Dere E, Silva MADS, Gödecke A, Schrader J, Huston JP. Supe-rior water maze performance and increase in fear-related behavior in theendothelial nitric oxide synthase-deficient mouse together with monoamine

3 al Brai

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

32 C . Gökc ek-Sarac et al. / Behaviour

changes in cerebellum and ventral striatum. The Journal of Neuroscience2000;20:6694–700.

[8] Bon CLM, Garthwaite J. On the role of nitric oxide in hippocampal long-termpotentiation. The Journal of Neuroscience 2003;23:1941–8.

[9] Calabresi P, Gubellini P, Centonze D, Sancesario G, Morello M, Giorgi M, et al.A critical role of the nitric oxide/cGMP pathway in corticostriatal long-termdepression. The Journal of Neuroscience 1999;19:2489–99.

10] Haley JE, Schaible E, Pavlidis P, Murdock A, Madison DV. Basal and apicalsynapses of CA1 pyramidal cells employ different LTP induction mechanisms.Learning and Memory 1996;3:289–95.

11] Haul S, Gödecke A, Schrader J, Haas HL, Luhmann HJ. Impairment of neocorticallong-term potentiation in mice deficient of endothelial nitric oxide synthase.Journal of Neurophysiology 1999;81:494–7.

12] Hawkins RD, Son H, Arancio O. Nitric oxide as a retrograde messengerduring long-term potentiation in hippocampus. Progress in Brain Research1998;118:155–72.

13] Lev-Ram V, Nebyelul Z, Ellisman MH, Huang PL, Tsien RY. Absence of cerebellarlong-term depression in mice lacking neuronal nitric oxide synthase. Learningand Memory 1997;4:169–77.

14] Lu YF, Kandel ER, Hawkins RD. Nitric oxide signaling contributes to late-phaseLTP and CREB phosphorylation in the hippocampus. The Journal of Neuro-science 1999;19:10250–61.

15] O’dell TJ, Huang PL, Dawson TM, Dinerman JL, Snyder SH, Kandel ER, et al.Endothelial NOS and the blockade of LTP by NOS inhibitors in mice lackingneuronal NOS. Science 1994;265:542.

16] Wang HG, Lu FM, Jin I, Udo H, Kandel ER, de Vente J, et al. Presynaptic andpostsynaptic roles of NO, cGK, and RhoA in long-lasting potentiation and aggre-gation of synaptic proteins. Neuron 2005;45:389–403.

17] Zhuo M, Kandel E, Hawkins R. Nitric oxide and cGMP can produce eithersynaptic depression or potentiation depending on the frequency of presynapticstimulation in the hippocampus. Neuroreport: An International Journal for theRapid Communication of Research in Neuroscience 1994;5:1033–6.

18] Zhuo M, Laitinen JT, Li XC, Hawkins RD. On the respective roles of nitric oxideand carbon monoxide in long-term potentiation in the hippocampus. Learningand Memory 1998;5:467–80.

19] Paul V, Ekambaram P. Involvement of nitric oxide in learning and memoryprocesses. The Indian Journal of Medical Research 2011;133:471.

20] Susswein AJ, Katzoff A, Miller N, Hurwitz I. Nitric oxide and memory. The Neu-roscientist 2004;10:153–62.

21] Kendrick K, Guevara-Guzman R, Zorrilla J, Hinton M, Broad K, MimmackM, et al. Formation of olfactory memories mediated by nitric oxide. Nature1997;388:670–3.

22] Okere CO, Kaba H. Increased expression of neuronal nitric oxide synthase mRNAin the accessory olfactory bulb during the formation of olfactory recognitionmemory in mice. European Journal of Neuroscience 2000;12:4552–6.

23] Samama B, Boehm N. Inhibition of nitric oxide synthase impairs early olfactoryassociative learning in newborn rats. Neurobiology of Learning and Memory1999;71:219–31.

24] Yanagihara D, Kondo I. Nitric oxide plays a key role in adaptive control oflocomotion in cat. Proceedings of the National Academy of Sciences 1996;93:13292.

25] Antonov I, Ha T, Antonova I, Moroz LL, Hawkins RD. Role of nitric oxide in clas-sical conditioning of siphon withdrawal in Aplysia. The Journal of Neuroscience2007;27:10993.

26] Bernabeu R, de Stein ML, Izquierdo I, Medina JH. Role of hippocampal NO inthe acquisition and consolidation of inhibitory avoidance learning. Neurore-port: An International Journal for the Rapid Communication of Research inNeuroscience 1995;6:1498–500.

27] Huang A, Lee EHY. Role of hippocampal nitric oxide in memory retention inrats. Pharmacology Biochemistry and Behavior 1995;50:327–32.

28] Pitsikas N, Rigamonti AE, Cella SG, Sakellaridis N, Muller EE. The nitric oxidedonor molsidomine antagonizes age-related memory deficits in the rat. Neu-robiology of Aging 2005;26:259–64.

29] Zhang S, Chen J, Wang S. Spatial learning and memory induce up-regulationof nitric oxide-producing neurons in rat brain. Brain Research 1998;801:101–6.

30] Barna M, Komatsu T, Reiss CS. Activation of type III nitric oxide synthasein astrocytes following a neurotropic viral infection. Virology 1996;223:331–43.

31] Gabbott P, Bacon S. Localisation of NADPH diaphorase activity and NOSimmunoreactivity in astroglia in normal adult rat brain. Brain Research1996;714:135–44.

32] Iwase K, Miyanaka K, Shimizu A, Nagasaki A, Gotoh T, Mori M, et al.Induction of endothelial nitric-oxide synthase in rat brain astrocytes bysystemic lipopolysaccharide treatment. Journal of Biological Chemistry2000;275:11929.

33] Ihara H, Kuwamura M, Atsuta M, Nihonmatsu I, Okada T, Mukamoto M, et al.Expression of neuronal nitric oxide synthase variant, nNOS-[mu], in rat brain.Nitric Oxide 2006;15:13–9.

34] Chiang LW, Schweizer FE, Tsien RW, Schulman H. Nitric oxide synthaseexpression in single hippocampal neurons. Molecular Brain Research 1994;27:

183–8.

35] Dinerman JL, Dawson TM, Schell MJ, Snowman A, Snyder SH. Endothelialnitric oxide synthase localized to hippocampal pyramidal cells: implicationsfor synaptic plasticity. Proceedings of the National Academy of Sciences1994;91:4214.

[

[

n Research 235 (2012) 326– 333

36] Doyle C, Slater P. Localization of neuronal and endothelial nitric oxide synthaseisoforms in human hippocampus. Neuroscience 1997;76:387–95.

37] Huang Z, Huang PL, Panahian N, Dalkara T, Fishman MC, Moskowitz MA. Effectsof cerebral ischemia in mice deficient in neuronal nitric oxide synthase. Science1994;265:1883.

38] Wood CE, Chen GF, Keller-Wood M. Expression of nitric oxide synthaseisoforms is reduced in late-gestation ovine fetal brainstem. AmericanJournal of Physiology–Regulatory, Integrative and Comparative Physiology2005;289:R613–9.

39] Rapanelli M, Frick LR, Zanutto BS. Modulation of endothelial and neuronal nitricoxide synthases during learning of an operant conditioning task. Journal ofNeurochemistry 2010;113:725–34.

40] Cavallaro S, D’Agata V, Manickam P, Dufour F, Alkon DL. Memory-specifictemporal profiles of gene expression in the hippocampus. Proceedings of theNational Academy of Sciences 2002;99:16279.

41] Chapman PF, Atkins CM, Allen MT, Haley JE. Inhibition of nitric oxide syn-thesis impairs two different forms of learning. Neuroreport: An InternationalJournal for the Rapid Communication of Research in Neuroscience 1992;3:567–70.

42] Harooni HE, Naghdi N, Sepehri H, Rohani AH. The role of hippocampal nitricoxide (NO) on learning and immediate, short-and long-term memory retrievalin inhibitory avoidance task in male adult rats. Behavioural Brain Research2009;201:166–72.

43] Mutlu O, Ulak G, Belzung C. Effects of nitric oxide synthase inhibitors 1-(2-trifluoromethylphenyl)–imidazole (TRIM) and 7-nitroindazole (7-NI) onlearning and memory in mice. Fundamental and Clinical Pharmacology2011;25:368–77.

44] Estall LB, Grant SJ, Cicala GA. Inhibition of nitric oxide (NO) production selec-tively impairs learning and memory in the rat. Pharmacology Biochemistry andBehavior 1993;46:959–62.

45] Hölscher C, McGlinchey L, Anwyl R, Rowan MJ. 7-Nitro indazole, a selectiveneuronal nitric oxide synthase inhibitor in vivo, impairs spatial learning in therat. Learning and Memory 1996;2:267–78.

46] Majlessi N, Choopani S, Bozorgmehr T, Azizi Z. Involvement of hippocampalnitric oxide in spatial learning in the rat. Neurobiology of Learning and Memory2008;90:413–9.

47] Mogensen J, Wortwein G, Hasman A, Nielsen P, Wang Q. Functional and neuro-chemical profile of place learning after l-nitro-arginine in the rat. Neurobiologyof Learning and Memory 1995;63:54–65.

48] Prendergast MA, Buccafusco JJ, Terry AV. Nitric oxide synthase inhibitionimpairs spatial navigation learning and induces conditioned taste aversion.Pharmacology Biochemistry and Behavior 1997;57:347–52.

49] Yamada K, Noda Y, Nakayama S, Komori Y, Sugihara H, Hasegawa T, et al. Roleof nitric oxide in learning and memory and in monoamine metabolism in therat brain. British Journal of Pharmacology 1995;115:852.

50] Zou LB, Yamada K, Tanaka T, Kameyama T, Nabeshima T. Nitric oxide synthaseinhibitors impair reference memory formation in a radial arm maze task in rats.Neuropharmacology 1998;37:323–30.

51] Ingram DK, Spangler EL, Meyer RC, London ED. Learning in a 14 unit T-maze isimpaired in rats following systemic treatment with Nw-nitro–arginine. Euro-pean Journal of Pharmacology 1998;341:1–9.

52] Bannerman D, Chapman P, Kelly P, Butcher S, Morris R. Inhibition of nitricoxide synthase does not impair spatial learning. The Journal of Neuroscience1994;14:7404–14.

53] Blokland A, De Vente J, Prickaerts J, Honig W, Van Ittersum MM, SteinbuschH. Local inhibition of hippocampal nitric oxide synthase does not impair placelearning in the Morris water escape task in rats. European Journal of Neuro-science 1999;11:223–32.

54] Markvartová V, Vozeh F. The effect of nitric oxide synthase inhibitors nitro-l-arginine and 7-nitroindazole on spatial learning and motor functionsin lurcher mutant and wild type mice. Prague Medical Report 2008;109:40–9.

55] Tobin JR, Gorman LK, Baxter MG, Traystman RJ. Nitric oxide synthase inhibi-tion does not impair visual or spatial discrimination learning. Brain Research1995;694:177–82.

56] Du W, Harvey JA. The nitric oxide synthesis inhibitor l-NAME facilitates associa-tive learning. Progress in Neuro-Psychopharmacology and Biological Psychiatry1996;20:1183–95.

57] Dere E, Frisch C, De Souza Silva M, Gödecke A, Schrader J, Huston J. Unalteredradial maze performance and brain acetylcholine of the endothelial nitric oxidesynthase knockout mouse. Neuroscience 2001;107:561–70.

58] Blackshaw S, Eliasson M, Sawa A, Watkins C, Krug D, Gupta A, et al. Species,strain and developmental variations in hippocampal neuronal and endothelialnitric oxide synthase clarify discrepancies in nitric oxide-dependent synapticplasticity. Neuroscience 2003;119:979–90.

59] Hölscher C. Different strains of rats show different sensitivity to block of long-term potentiation by nitric oxide synthase inhibitors. European Journal ofPharmacology 2002;457:99–106.

60] Jakubowska-Dogru E, Umut Gumusbas U, Kara F. Individual variation in thespatial reference and working memory assessed under allothetic and idio-thetic orientation cues in rat. Acta Neurobiologiae Experimentalis 2002;63:

17–24.

61] Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with theFolin phenol reagent. Journal of Biological Chemistry 1951;193:265–75.

62] Ey PL, Ashman LK. The use of alkaline phosphatase-conjugated anti-immunoglobulin with immunoblots for determining the specificity of

al Brai

[

[

[

[

[

[

[

[

[

C . Gökc ek-Sarac et al. / Behaviour

monoclonal antibodies to protein mixtures. Methods in Enzymology1986;121:497–509.

63] Inglis FM, Furia F, Zuckerman KE, Strittmatter SM, Kalb RG. The role of nitricoxide and NMDA receptors in the development of motor neuron dendrites. TheJournal of Neuroscience 1998;18:10493–501.

64] Dun N, Dun S, Förstermann U. Nitric oxide synthase immunoreactivity in ratpontine medullary neurons. Neuroscience 1994;59:429–45.

65] West AR, Grace AA. Striatal nitric oxide signaling regulates the neuronalactivity of midbrain dopamine neurons in vivo. Journal of Neurophysiology2000;83:1796.

66] Son H, Hawkins RD, Martin K, Kiebler M, Huang PL, Fishman MC, et al. Long-term potentiation is reduced in mice that are doubly mutant in endothelial andneuronal nitric oxide synthase. Cell 1996;87:1015–23.

67] Amaral D, Witter M. The three-dimensional organization of the hippocampalformation: a review of anatomical data. Neuroscience 1989;31:571–91.

[

n Research 235 (2012) 326– 333 333

68] McEachern JC, Shaw CA. An alternative to the LTP orthodoxy: aplasticity-pathology continuum model. Brain Research Reviews 1996;22:51–92.

69] Moser E, Moser MB, Andersen P. Spatial learning impairment parallels the mag-nitude of dorsal hippocampal lesions, but is hardly present following ventrallesions. The Journal of Neuroscience 1993;13:3916–25.

70] Kano T, Shimizu-Sasamata M, Huang P, Moskowitz M, Lo E. Effects of nitric oxidesynthase gene knockout on neurotransmitter release in vivo. Neuroscience1998;86:695–9.

71] Faraci FM. Protecting the brain with eNOS. Circulation Research 2006;99:

1029–30.

72] Law A, O’donnell J, Gauthier S, Quirion R. Neuronal and inducible nitric oxidesynthase expressions and activities in the hippocampi and cortices of youngadult, aged cognitively unimpaired, and impaired Long-Evans rats. Neuro-science 2002;112:267–75.