B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
ava i l ab l e a t www.sc i enced i rec t . com
www.e l sev i e r. com/ loca te /b ra in res
Research Report
Idazoxan increases perforant path-evoked EPSP slope pairedpulse inhibition and reduces perforant path-evokedpopulation spike paired pulse facilitation in rat dentate gyrus
John Knighta, Carolyn W. Harleya,b,⁎aDivision of Basic Medical Sciences, Faculty of Medicine, Memorial University of Newfoundland, St. John's, Newfoundland, Canada A1B 3X9bDepartment of Psychology, Memorial University of Newfoundland, St. John's, Newfoundland, Canada A1B 3X9
A R T I C L E I N F O
⁎ Corresponding author. Department of PsychFax: +1 709 737 4000.
Article history:Accepted 5 December 2005Available online 19 January 2006
Norepinephrine, acting via β-adrenoceptors, enhances the perforant path-evoked potential indentate gyrus. Using systemic idazoxan to increase norepinephrine, and paired perforant pathpulses to probe early inhibition, previous investigators reported that idazoxan increased initialspike amplitude and increased somatic feedback inhibition. Here, feedback inhibition was re-examined in idazoxan-treated (5 mg/kg) rats under urethane anesthesia. To control for initialincreased spike amplitude after idazoxan, evoked potentials were matched, pre- and post-idazoxan, on initial population spike. Input–output current profiles were also compared pre-and post-idazoxan. Saline- and timolol-filled micropipettes permitted evaluation of acontribution of local β-adrenoceptors. As previously observed, initial spike amplitude waspotentiated by idazoxan. Comparable spike potentiation was not seen on the timololmicropipette. Paired pulse inhibition of spike amplitude apparently increased, but input–outputcurve comparisons revealed a loss of feedback facilitation rather than an increase in feedbackinhibition. Initial EPSP slopes were depressed after idazoxan in input–output curve data. EPSPslope feedback ratioswere significantly reduced following idazoxan.These data suggest idazoxanhas multiple effects on perforant path input to the dentate gyrus. Spike potentiation followingidazoxan has previously been shown to depend on intact norepinephrine input. Here, thereduction in spike potentiation on the timolol pipette is consistent with other evidence thatnorepinephrine-mediated potentiation of the perforant path-evoked potential is dependent onlocalβ-adrenoceptor activation. The input–output data suggest a decrease in feedback facilitationafter idazoxan is likely to account for the apparent increase in feedback inhibition previouslyreported. Decreased EPSP slope ratios with similar paired pulse intervals have been reported innovel environments. Since exposure to novel environments activates locus coeruleus neurons,norepinephrine may mediate the change in EPSP slope inhibition reported in awake rats.Insummary, these results are consistent with the hypothesis that idazoxan potentiates granulecell responses to perforant path input in the dentate gyrus via increases in norepinephrine thatlead toβ-adrenoceptor activation, and, further, that idazoxan reducespairedpulse feedback spikefacilitation and enhances EPSP slope, but not spike, feedback inhibition.
ology, Memorial University of Newfoundland, St. John's, Newfoundland, Canada A1B 3X9.
rley).
er B.V. All rights reserved.
37B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
1. IntroductionLocus coeruleus innervation of the hippocampus has its dens-est terminal field in the subgranular/hilar region of the den-tate gyrus (Loy et al., 1980; Oleskevich et al., 1989). Themodulation of perforant path input to the dentate gyrus byactivation of the locus coeruleus (Harley and Sara, 1992; Har-ley and Milway, 1986; Harley et al., 1989; Klukowski and Har-ley, 1994; Walling and Harley, 2004; Walling et al., 2004), or bydirect application of norepinephrine (NE) (Neuman and Har-ley, 1983), which is secreted by locus coeruleus terminals,induces an increase in the perforant path-evoked populationspike amplitude in vivo. This NE-induced perforant path spikeamplitude potentiation is mediated by β-adrenoceptors (Har-ley et al., 1989; Walling and Harley, 2004; Walling et al., 2004;Washburn and Moises, 1989) and may relate to a direct actionof NE (Lacaille and Harley, 1985) on granule cells.
Relatively few studies, however, have addressed the role ofNE, or locus coeruleus activation, on local inhibitory circuitmodulation in dentate gyrus, despite the anatomical evidencethat these elements are heavily innervated by locus coeruleusfibers. In one earlier in vivo study, Sara and Bergis (1991) usedidazoxan (2 mg/kg) to increase NE in the dentate gyrus andstudied perforant path-evoked potentials using a paired pulseparadigm. The paired pulse paradigm, using short interstim-ulus intervals, permits an investigator to probe both modula-tion of the initial conditioning stimulus-evoked potential andthe subsequent test stimulus-evoked potential, which reflectsthe activation of feedback inhibitory circuits, as well as directperforant path input.
In their study, Sara and Bergis found increased feedbackinhibition at an interstimulus interval of 25 ms and sug-gested that NE acted to enhance feedback inhibition, aswell as to enhance the initial population spike response toglutamatergic perforant path activation (Sara and Bergis,1991). Previous work had shown that the same dose of ida-zoxan does not enhance the initial perforant path populationspike, if the locus coeruleus is lesioned using the DSP-4neurotoxin, implicating NE in the initial potentiation of thepopulation spike (Richter-Levin et al., 1991) as reported byothers. Sara and Bergis reported a single subject in whichperforant path spike amplitude of the conditioning pulsewas not enhanced. In that subject, feedback inhibition ofthe test pulse still increased. Thus, they suggest the in-creased inhibition could not be explained as an increaseddrive of principal cells engaging inhibitory circuitry morestrongly. The enhancement of feedback inhibition was re-stricted to the 25-ms interstimulus test interval and notseen at shorter or longer intervals, suggesting some specificcondition predominated at that interval.
In a study of functionally identified interneurons in thedentate gyrus (Brown et al., 2005), locus coeruleus activationwas shown to inhibit firing in all feedforward interneuronswhile either enhancing or inhibiting subpopulations of feed-back interneurons. The net effect of noradrenergic activity onfeedback inhibitory tone could not be determined from thesesingle unit data, but they reinforce the view that interneuronsare an important target of synaptically released NE.
Here, we follow the Sara and Bergis study by using ida-zoxan as a tool to elevate synaptically released NE in dentate
gyrus. We assess the prediction, based on other investigationsof NE-induced spike potentiation in dentate gyrus, that β-adrenoceptors will be involved in the potentiating effect ofidazoxan by recording with, and without, a locally diffusedβ-adrenoceptor antagonist. To evaluate the effects of ida-zoxan on feedback inhibition separately from its effects oninitial spike size, we equate for initial spike size in the analy-ses of pre- and post-idazoxan samples. In a subset of experi-ments, we use input–output data for spikes and feedbackinhibition to systematically probe spike size and feedbackinhibition before and after idazoxan.
2. Results
2.1. Histology
Consistent with the electrophysiological signature of the per-forant path-evoked potentials, recording electrode placementswere in, or near, the granule cell layer of the dentate gyrus or inthe hilus. The salinemicropipetteswere slightly anterior to thetarget site of 3.5 mm posterior to bregma, while the timololmicropipettes were posterior to that target. While the timololpipetteswere lateral to saline pipettes in the holder, theirmoreposterior location in the dentate gyrus resulted in timolol pip-ettes often being structurally medial to saline pipettes. Thus,seven timolol pipettes were placed in the medial hilar apex,while the remaining 7 timolol pipettes, and all 14 saline pip-ettes, were in the laterodorsal cell blade of dentate gyrus.
2.2. Correlations of first population spike amplitude withthe P2/P1 ratio
Significant (P b 0.01) negative P2/P1 to P1 correlations, rangingfrom −0.11 to −0.78, were found for all but 1 animal, support-ing earlier evidence (Austin et al., 1989; Joy and Albertson,1987) that larger spikes are associated with stronger earlyfeedback inhibition and justifying the use of the P1-matchingtechnique for analysis of feedback inhibition.
2.3. Effect of idazoxan on the perforant path-evokedpotential to the first pulse (P1)
2.3.1. Fifteen-minute recording samples pre- andpost-idazoxanThemean baseline spike (Fig. 1A) for the saline pipette (n = 14)was 4.4mV;mean baseline spike for the timolol pipettewas 5.2mV. This difference was significant (paired t test; t13 = 3.36;P b 0.01). Following idazoxan, the spike on the saline pipetteincreased significantly to 116% of baseline (paired t test;t13 = 5.89; P b 0.01). On the timolol pipette, the increase (108%)was not significant (paired t test; t13 = 1.11). See Fig. 1A.
Mean baseline EPSP slope (Fig. 1B) for the saline pipette(n = 14) was 5.1 mV/ms; mean baseline EPSP slope for thetimolol pipette was 6.0 mV/ms. This difference was not sig-nificant (paired t test; t13 = 1.72). Post-idazoxan EPSP slope sizeaveraged 98% of baseline on the saline pipette and 93% ofbaseline on the timolol pipette.
Mean baseline P1 spike latency (Fig. 1C) for the saline pi-pette (n = 14) was 3.86 ms; mean baseline spike latency for the
Fig. 2 – The effects of idazoxan on initial population spikeamplitude for 25 different current steps beginning at 150 μAand incrementing in 25 μA steps (n = 8). (A) Mean populationspike amplitude on the saline micropipette. (B) Meanpopulation spike amplitude on the timolol micropipette.*Denotes a significant pre- versus post-idazoxan difference,P b 0.05. Error bars are standard errors of the mean.
Fig. 1 – The effects of idazoxan on initial perforantpath-evoked potentials recorded on both a saline and atimolol micropipette for a 15-min block prior to and 30 minafter injection in 14 urethane-anesthetized rats. (A) Meanpopulation spike amplitude on the saline and timololmicropipettes. (B) Mean EPSP slope on the saline and timololmicropipettes. (C) Mean population spike latency on thesaline and timolol pipettes. *Denotes a significant pre-versus post-idazoxan difference, P b 0.01. Error bars arestandard errors of the mean.
38 B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
timolol pipette was 3.91 ms. Following idazoxan, there was nosignificant change in spike latency on either pipette (3.86 ms,saline; 3.69 ms, timolol).
2.3.2. I/O curves pre- and post-idazoxanFig. 2 shows spike means obtained for the pre- and post-idazoxan I/O curves (n = 8) for each current for the saline(Fig. 2A) and timolol (Fig. 2B) pipettes separately. Pipette ×
Idazoxan × Current (2 × 2 × 25) repeated measures ANOVArevealed a significant main effect of current (F24,168 = 8.99;P b 0.0001) on spike amplitude as well as a significantIdazoxan × Current interaction (F24,168 = 2.04; P b 0.01). Neu-man–Keuls post hoc comparisons (α = 0.05) showed idazoxansignificantly increased spikes at currents ranging from 175–475 μA. See Figs. 2A and B.
Fig. 3 shows EPSP slope means obtained using pre- andpost-idazoxan I/O curves (n = 8) for each current for thesaline (Fig. 3A) and timolol (Fig. 3B) pipettes separately.Pipette × Idazoxan × Current (2 × 2 × 25) repeated measuresANOVA showed significant effects of Current (F24,168 = 5.52;P b 0.05) and Idazoxan (F1,7 = 5.70; P b 0.001) as well as asignificant Idazoxan × Current interaction (F24,168 = 2.16,P b 0.01). Newman–Keuls post hoc comparisons indicatedthat EPSP slope at each current was decreased after ida-zoxan. However, the slope decrease interacted with currentsuch that for current pulses above 275 μA, the pre-idazoxanslopes were higher than any slopes after idazoxan, while at
Fig. 3 – The effects of idazoxan on initial average EPSP slopefor 25 different current steps beginning at 150 μA andincrementing in 25 μA steps (n = 8). (A) Mean EPSP slope onthe saline micropipette. (B) Mean EPSP slope on the timololmicropipette. *Denotes a significant pre- versuspost-idazoxan difference, P b 0.05. Error bars are standarderrors of the mean.
Fig. 4 – The effects of idazoxan on the ratio of the 2ndperforant path-evoked potential measure divided by the 1stperforant path-evoked potential measure for (A) populationspike amplitude and (B) EPSP slope (n = 14) on both thesaline and timolol micropipettes. **Denotes a significantpre- versus post-idazoxan difference, P b 0.01; *denotesP b 0.05. Error bars are standard errors of the mean.
39B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
the lower currents, the slope size before idazoxan did notdiffer from the slope size at higher currents after idazoxan.
A Pipette × Idazoxan × Current repeated measures ANOVAshowed only a significant effect of current on populationspike latency (F24,168 = 8.07; P b 0.0001). Stronger currents re-duced spike latency (data not shown).
2.4. Effect of idazoxan on paired pulse inhibition usingP2/P1 values associated with overlapping P1 values that werenot significantly different pre- and post-idazoxan
Fig. 4A shows mean P2/P1 spike ratios pre- and post-idazoxanfor saline and timolol pipettes (n = 14). The mean baseline P2/P1 spike ratio on the saline pipette was 0.481; mean baselineP2/P1 spike ratio on the timolol pipette was 0.614. This differ-ence was not significant (paired t test; t13 = 1.72). The P1 meanof the selected spikes on the saline pipette was 4.72 mV (meannumber of records used for each rat in the baseline peri-od = 233), while the P1 mean on the timolol pipette was 5.31
mV (mean number of records used for each rat in the baselineinterval = 242). Following idazoxan, ratios decreased to 0.376on the saline pipette (paired t test; t13 = 3.71, P b 0.01) and to0.471 on the timolol pipette (paired t test; t13 = 3.17; P b 0.01)indicating increases in inhibition on both pipettes, althoughthe degree of inhibition did not differ significantly. Thematched post-idazoxan P1 mean on the saline pipette was4.67 mV for the 15-min interval analyzed (mean number ofrecords in each rat's average = 328), while the matched post-idazoxan P1 mean on the timolol pipette was 5.43 mV (meannumber of records in each rat's average = 315).
Fig. 4B shows the mean P2/P1 EPSP slope ratios for pre- andpost-idazoxan for both pipettes (n = 14). The mean baseline P2/P1 EPSP slope ratio for the saline pipette was 0.730; the meanbaseline P2/P1 EPSP slope ratio for the timolol pipettewas 0.721.This difference was not significant (paired t test; t13 = 0.45). P1slopemeanon the saline pipettewas 5.4mV/ms,while P1 slopemean on the timolol pipette was 6.1 mV/ms. Following ida-zoxan, the P2/P1 EPSP slope ratio on the timolol pipette wassignificantly decreased to 0.653 (paired t test; t13 = 2.59, P b 0.05),indicating an increase in EPSP slope inhibition. There was nocorresponding change on the saline pipette; post-idazoxan, theP2/P1 EPSP slope ratio was 0.718. Post-idazoxan, the P1 EPSP
40 B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
slope mean on the saline pipette was 4.7 mV/ms, while the P1mean on the timolol pipette was 5.2 mV/ms.
2.5. Effect of idazoxan and timolol on paired pulseinhibition at variable current levels
Fig. 5 compares mean P2/P1 spike amplitude ratios obtainedusing pre- and post-idazoxan I/O curves (n = 8) for each cur-rent for the saline (Fig. 5A) and timolol (Fig. 5B) pipettes sep-arately. Paired pulse facilitation can be seen at lower currents.Although there was variation in responses at lower currentsprior to idazoxan, all rats showed some degree of paired pulsefacilitation prior to idazoxan at the lower currents.
Repeated measures ANOVA showed that there was a sig-nificant Idazoxan × Current interaction (F24,168 = 4.19;P b 0.0001). Neuman–Keuls post hoc comparisons showedidazoxan caused a reduction of a facilitation (rather than an
Fig. 5 – The effects of idazoxan on the ratio of the 2ndperforant path-evoked population spike divided by the 1stperforant path-evoked population spike at 25 current levelsbeginning at 150 μA and incrementing in 25 μA steps for(A) the saline micropipette and (B) the timolol micropipette.*Denotes a significant pre- versus post-idazoxan difference,P b 0.05. Error bars are standard errors of the mean.
Fig. 6 – The effects of idazoxan on the ratio of the 2ndperforant path-evoked EPSP slope divided by the 1stperforant path-evoked EPSP slope at 25 current levelsbeginning at 150 μA and incrementing in 25 μA steps for(A) the saline micropipette and (B) the timolol micropipette.*Denotes a significant pre- versus post-idazoxan difference,P b 0.05. Error bars are standard errors of the mean.
increase in inhibition) of spike P2/P1 ratios at the lower cur-rents of 150 and 175 μA.
Fig. 6 compares mean P2/P1 EPSP slope ratios obtainedusing pre- and post-idazoxan I/O curves (n = 8) for each currentfor the saline (Fig. 6A) and timolol (Fig. 6B) pipettes separately.
Repeated measures Pipette × Current × Idazoxan ANOVAshowed that there was a significant main effect of idazoxan indecreasing the EPSP slope P2/P1 ratio (F1,7 = 17.4; P b 0.01) aswell as a significant main effect of current (F1,7 = 1.63; P b 0.05)whichwas reflected in increasing inhibition at higher currents.
2.6. EPSP slope population spike amplitude correlations
On both the saline and timolol pipettes, there were significantpositive correlations between EPSP slope and population spikeamplitude in each animal (n = 14) pre-idazoxan (P b 0.01) and,
41B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
in all but 1 animal, post-idazoxan (P b 0.01). The slope–spikecorrelation was significantly smaller after idazoxan adminis-tration on both pipettes in all animals. Overall, mean correla-tions were 0.78 pre-idazoxan and 0.43 post-idazoxan for thesaline pipette (paired t test; t = 5.69; df = 13; P b 0.0001) and 0.79pre-idazoxan and 0.50 post-idazoxan for the timolol pipette(paired t test; t = 3.67; df = 13; P = 0.01).
2.7. EPSP slope and population spike P2/P1 ratiocorrelations
There were no significant correlations between the spike P2/P1 ratios and the EPSP slope P2/P1 ratio pre- or post-idazoxanon either pipette (P N 0.01) in any animal (n = 14).
3. Discussion
3.1. β-Adrenoceptor-dependent potentiation of populationspike amplitude
Perforant path-evoked population spike amplitude was po-tentiated 30 min after idazoxan administration. This effectwas stronger on the saline than on the timolol micropipettein the overall group data set (n = 14). Potentiation of perforantpath population spike amplitude in vivo by idazoxan (Saraand Bergis, 1991; Richter-Levin et al., 1991) and by NE (Neu-man and Harley, 1983) has been observed previously. As in theearlier idazoxan studies, the effect was most evident at lowercurrent levels in the I/O curve. This is the first study to dem-onstrate that idazoxan potentiation of perforant path-evokedpopulation spike amplitude is attenuated by a β-adrenoceptorantagonist, consistent with other evidence that idazoxan'seffects are mediated through increased levels of NE (Richter-Levin et al., 1991). Although a larger baseline population spikeon the timolol pipette suggests ceiling effects are a possibleexplanation for the weaker potentiation, an examination ofthe I/O curves indicates smaller potentiation on the timololpipette even when spike amplitudes were small.
3.2. Increase in apparent feedback inhibition of populationspike amplitude is a function of loss of facilitation
As previously reported (Austin et al., 1989; Joy and Albertson,1987), the strength of feedback inhibition was related to thesize of the first population spike in the dentate gyrus. In thepresent study, a set of paired pulse pairs matched for the sizeof the first population spike from both before, and after, ida-zoxan administration was compared. The results corroborat-ed the report that feedback inhibition was apparentlyincreased after idazoxan administration (Sara and Bergis,1991), even when the initial conditioning spikes were notsignificantly different.
An examination of I/O current data provided an alternateinterpretation of these results however. In the I/O data, therewas no change in feedback inhibition, but a marked loss offacilitation that was readily observable at low currents. Thisloss of facilitation was the only change observable in the I/Odata, and this suggests that the apparent increase in feedbackinhibition at an average stimulus current of 275 μA (the larger
data set) is due to the loss of facilitation. The change in feed-back effects on the population spike was unaffected by β-adrenoceptor blockade suggesting mediation by otherreceptors.
Other studies of short paired pulse intervals in dentategyrus have reported facilitation, as well as inhibition, atshort intervals (Austin et al., 1989; Joy and Albertson, 1987;Sloviter, 1991; Oliver and Miller, 1985; Racine and Milgram,1983). Joy and Albertson suggest early facilitation dependson feedforward interneurons, although its maximal effectoccurs later than initial granule cell spiking. Feedforward fa-cilitation can be revealed to contribute at short paired pulseintervals (e.g., by reducing GABA-A or Cl− (Oliver and Miller,1985)) Austin et al. (1989) report that paired pulse facilitationdominates at short intervals in active waking rats, whilepaired pulse inhibition dominates in immobile waking andsleep. Facilitation requires orthodromic stimulation, unlikeearly inhibition that can be produced by antidromic stimula-tion (Oliver and Miller, 1985), and facilitation is seen even inthe absence of an initial population spike (Sloviter, 1991;Racine and Milgram, 1983), consistent with its proposed me-diation by feedforward interneurons.
Locus coeruleus activation inhibits functionally identifiedfeedforward interneurons (Brown et al., 2005). This actionprovides a single mechanism for both the increase in initialpopulation spike (loss of feedforward inhibition) and the ap-parent increase in feedback inhibition (loss of feedforwardfacilitation). The latter hypothesis could be further exploredby looking at facilitation at longer intervals.
3.3. EPSP slope depression
Idazoxan did not significantly affect EPSP slope when meandata taken prior to, and 30 min following, idazoxan werecompared, but when I/O curve data were examined, EPSPslope was consistently depressed relative to pre-idazoxanmeasures. This effect was exaggerated on the timolol pipette,suggesting a contribution of α-adrenoceptors, but the pipettedifference was not significant. The failure to see a depressioneffect in the larger data set may relate to the lower currentused there (mean = 275 μA), since depression was more pro-nounced at higher current levels in the I/O curve data. Depres-sion of EPSP slope with idazoxan has been reported previously(Richter-Levin et al., 1991). EPSP slope depression was notobserved after a neurotoxic lesion of the locus coeruleus,suggesting that EPSP slope depression requires NE.
In vitro NE has consistent EPSP slope effects that varydepending on the perforant path component stimulated. NEdepresses the lateral perforant path EPSP slope but enhancesthe medial perforant path EPSP slope (Dahl and Sarvey, 1989).In vivo NE effects on EPSP slope are inconsistent, possiblydue to a mixture of lateral and medial perforant path con-tributions. In the present study, spike latencies argue formedial perforant path activation, but lateral perforant pathfibers were likely co-activated, especially at higher currents,and might then be responsible for the EPSP slope depressioneffect observed. Alternatively, another action of idazoxan,such as its effects on imidazoline1 receptors (Regunathan etal., 1993), or its blockade of heteroreceptor α2 effects (Boehm,1999), may account for the consistent EPSP slope depression
42 B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
seen in the idazoxan in vivo studies, but not with other NEmanipulations.
With NE, spike amplitude is often increased in the absenceof increases in EPSP slope suggesting an increase in EPSPslope/spike amplitude coupling. A similar effect is seen inthe present study. In addition, EPSP slope is less predictive ofspike size after idazoxan, an effect that has not been exam-ined yet with other NE manipulations.
3.4. The increase in feedback inhibition of EPSP slopeparallels the effects of novelty
This is the first study to examine feedback inhibition of EPSPslope with idazoxan. Data from the matched evoked poten-tials did not reveal a significant change on the saline pipette,but the EPSP slope ratio on the timolol micropipette was sig-nificantly smaller after idazoxan. The I/O curve data, however,reveal a reduction in the P2/P1 ratio of EPSP slope, at allcurrent levels, across both micropipettes, although consistentwith the larger group data, mean differences are smaller atlow current levels on the saline pipette. The greater strengthof the effect on the timolol pipette again suggests that α-adrenoceptors may selectively contribute to the EPSP sloperatio decrease.
A significant increase in EPSP slope feedback inhibition inthe dentate gyrus in awake rats exploring a novel environ-ment has been reported for EPSP slopes recorded at the den-drites (Moser, 1996). The increase in EPSP slope feedbackinhibition was not related to temperature changes. EPSPslope measured at the soma level, as here, is a less directmeasure of dendritic current than EPSP slope measured inthe dendrites. However, in a previous study, EPSP slopechanges measured at both somatic and dendritic levels,using the two micropipette methodology with timolol, gaveparallel results in terms of magnitude and direction of EPSPslope change (Munro et al., 2001).
Naturally released NE could account for the increase inEPSP slope inhibition described in awake rats, since thelocus coeruleus, the source of dentate gyrus NE, is consis-tently activated by novelty (Sara et al., 1994; Vankov et al.,1995). Using novel stimuli in a holeboard, as well as place-ment in a novel holeboard, we have demonstrated a poten-tiation of perforant path population spike amplitude, whichis blocked by a β-adrenoceptor antagonist (Kitchigina et al.,1997). This effect is longer lasting in a novel environment,similar to the time course described for the increase inpaired pulse EPSP slope inhibition in novel environments(Moser, 1996). However, the actions of idazoxan on α2 het-eroreceptors on cholinergic, glutamatergic, or gabaergicinputs or other systemic autonomic effects of idazoxan can-not be ruled out as being critical for the feedback effectsobserved here.
Since the increase in EPSP slope inhibition was only ob-served when the first perforant path-evoked potential eliciteda population spike in the awake rat study, it suggests feedbackinterneuron activation is responsible for the effect. Locuscoeruleus activation, as noted earlier, excites a feedback in-terneuron subpopulation, which might be related to the den-dritic EPSP inhibition effect reported in awake rats (Brown etal., 2005).
3.5. Summary
Consistent with the relatively dense noradrenergic innerva-tion of interneuron regions of the dentate gyrus, the presentstudy finds that idazoxan induces a β-adrenoceptor-sensitivepotentiation of the perforant path population spike and mod-ulates feedforward and feedback interneuron circuits. An ap-parent increase in population spike feedback inhibition isrevealed as a reduction in feedforward facilitation. An in-crease in EPSP slope feedback inhibition was also discovered.The increased EPSP slope feedback inhibition resembles theeffects of novelty on paired pulse EPSP slope feedback inhibi-tion in awake rats and is consistent with increased feedbackinterneuron activity following locus coeruleus activation.However, other effects of idazoxan on non-noradrenergic cir-cuits cannot be ruled out. Overall, the patterns observed areconsistent with the hypothesis that locus coeruleus activationand NE modulate dentate gyrus interneuron circuitry as wellas influencing principal cell excitability through a β-adreno-ceptor mechanism.
4. Experimental procedures
4.1. Subjects
Fourteen male Sprague–Dawley rats (250–350 g) obtained from theMemorial University Vivarium served as subjects. The experimen-tal protocol conformed to the Canadian Council of Animal Careguidelines and was approved by the Institutional Committee onAnimal Care. Urethane-anesthetized rats (1.5 g/kg, i.p.) wereplaced stereotaxically with lambda and bregma in the horizontalplane. Temperature was maintained at 37 ± 1 °C using a rectalprobe coupled to a heating pad and temperature control unit.
4.2. Electrode placements
A bipolar stimulating electrode (NE100) was directed at the per-forant path (7.2 mm posterior, 4.1 mm lateral to bregma, and 3.0–3.5 mm ventral to brain surface). Two recording glass micropip-ettes (40–50 μm tips) were directed at the cell body layer or hilus ofthe dentate gyrus (3.5 mm posterior, 2.0 mm lateral to bregma,and 3.0–3.2 mm ventral to brain surface). One recording pipettecontained 0.9% physiological saline, while the second contained100 mM timolol hydrochloride (Sigma) in 0.9% physiological sa-line. Pipettes were separated by 0.5 to 1 mm with the timololpipette posterolateral to the saline pipette. Pipette impedancevaried from 1 to 3 MΩ.
The choice of pipette impedances and separations for dentategyrus recording was based on a methodological innovation intro-duced by Steward et al. (1990) and on previous work in our ownlaboratory (Munro et al., 2001). Steward and Tomasulo showedthat 8 mM biculluline diffusing from a 1–3 MΩ recording micropi-pette placed ∼1 mm from a 0.9% saline micropipette continuouslyaffects its own evoked potential (multiple spikes) without causingdetectable changes in the responses or occurrence of feedbackinhibition at the nearby saline micropipette.
We have previously substituted timolol for bicuculline (Munroet al., 2001) in this protocol. We had shown earlier that timololinfused into dentate gyrus prevents glutamatergic locus coeruleusactivation from potentiating the perforant path-evoked popula-tion spike (Harley and Evans, 1988). In the micropipette diffusionprotocol, we demonstrated blockade of paragigantocellularis po-tentiation of perforant path-evoked spike amplitude on the timo-lol, but not the saline, micropipette. The paragigantocellularis
43B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
nucleus provides glutamatergic excitatory drive to locus coeru-leus (Aston-Jones et al., 1991). Late long-term potentiation wasalso prevented on the timolol pipette as reported for timolol invitro (Bramham et al., 1997).
This evidence of antagonism of β-adrenoceptor-mediated NEeffects in dentate gyrus by timolol supports our present use oftimolol diffusion to address the question of local β-adrenoceptorinvolvement in idazoxan-induced effects. The depths of stimulat-ing and recording electrodes were determined by monitoringresponses (see Experimental protocol). Concomitant traces froma saline and timolol filled pipette in the same experiment areshown in Fig. 7.
4.3. Stimulation and recording of evoked potentials
The ‘Workbench’ program in ‘Brainwave’ (Datawave Technolo-gies) was used to control stimulation and recording. Two mono-phasic 0.2-ms pulses with an interstimulus interval (ISI) of 15–30ms were generated by the Datawave A/D board and deliveredthrough a constant current unit at a frequency of 0.1 Hz. Signalswere amplified on a Grass pre-amplifier at a bandwidth of 0.1–3.0kHz and displayed on a dual-channel analogue oscilloscope.Waveforms were digitized on-line at a rate of 10,000 points/s,displayed on a computer monitor, and stored using Brainwave.
4.4. Experimental protocol
For each subject, stimulating electrodes were lowered first. Thenrats received paired pulse stimulation every 10 s, typically at 400–600 μA, as recording electrodes were lowered, until a maximalnegative-going population spike was identified. Position of therecording and stimulating electrodes was adjusted untilresponses to the first pulse (P1) on the two pipettes matchedclosely with respect to amplitude and spike latencies were 4 msor less. The interstimulus interval was adjusted until paired pulseinhibition was observed on both pipettes, with the constraint thatspike amplitude for the second pulse was large enough to bemeasured.
In the initial step, an input–output (I/O) curve was obtained,beginning at a current of 50 μA, and proceeding in 25 μA incre-ments, at intervals of 1 min, with 6 stimulations per current value,to the current yielding apparent maximal spike amplitude. I/Ocurves for each animal consisted of 25–35 different currents.
In the second step, stimulation was adjusted to the lowestcurrent at which a measurable spike amplitude was observed inthe second response on both pipettes and a 15-min pre-idazoxanbaseline was taken. Sara and Bergis (1991) had reported a signif-icant spike enhancing effect of idazoxan only at low currentpulses. As noted earlier, the interstimulus interval was selected
Fig. 7 – Paired pulse waveforms recorded concomitantly on a salmicropipette spaced ~1 mm apart.
as an interval that produced feedback inhibition and provided ameasurable second spike. Interstimulus intervals of 22–25 mswere employed for 8 rats; 15–18 ms intervals were used for 4rats, and a 30-ms interstimulus interval was used for 2 rats.While Sara and Bergis had only seen idazoxan-induced increasesin inhibition with a 25-ms interstimulus interval, rats tested hereusing the shorter and longer intervals all exhibited a decrease inpaired pulse spike ratio after idazoxan indicative of increasedfeedback inhibition.
In the third step, the subject was injected with idazoxan hy-drochloride (5 mg/kg, i.p.), after which evoked potentials wererecorded for approximately 1.5 h post-drug. The dose of 5 mg/kgidazoxan was chosen based on microdialysis data in anesthetizedrats. Microdialysismeasurements demonstrate that 1mg/kg dosesare unreliable in elevating NE in forebrain (Thomas and Holman,1991; Sacchetti et al., 1999), while 3 mg/kg gives consistent, butsignificantly smaller NE elevation than 5 mg/kg (Thomas and Hol-man, 1991). Idazoxan at a 5 mg/kg dose approximately doubles NElevels, and higher doses do not increase levels further. Idazoxanantagonizes α2 receptors, reducing autoinhibition on NE terminalsin the forebrain (increasing synaptic NE) and on locus coeruleusneurons (increasing NE neuronal firing) (Linner et al., 1999). Whileidazoxan also interacts with imidazoline1 receptors, the majorityof its binding occurs to α2 receptors (Raymon et al., 1992).
For 8 animals, a fourth step was included in the protocol inwhich a post-idazoxan I/O curve was obtained at the end of the1.5-h recording period using the same current values and timeintervals used in the pre-drug I/O curve.
4.5. Histology
After recording, the brain was removed and frozen in 2-methylbutane that had been previously cooled in a −80 °C freezer. Brainswere stored in the same freezer until sectioning.
For verification of recording electrode position, brains were cutat 30 μm on a Jung Frigocut cryostat microtome, and alternateconsecutive sections were taken. One set was subjected to anacetylcholinesterase-metachromatic Nissl staining procedure(Paxinos, 1998), and the other set was stained for glycogen phos-phorylase a (Harley and Bielajew, 1992). The latter stain facilitateddetermination of the pipette tip placements.
4.6. Evoked potential parameter extraction
For each animal, the parameter extraction option in the ‘Work-bench’ program was used to measure spike amplitude and EPSPslope. Measurement #25 (Peak to valley) from the ‘Electrophysiol-ogy’ function set was used to obtain population spike amplitude,while measurement #1 was used to determine EPSP slope.
ine-filled (upper trace) and a timolol-filled (lower trace)
44 B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
4.7. Statistics
4.7.1. The effect of idazoxan on the response to the first pulseOver all 14 animals, a paired t test was carried out in whichpopulation spike and EPSP slope measurements from the re-sponse to the first pulse obtained during the 15-min pre-idazoxanbaseline (90 records) in step 2 of the protocol were compared tomeasures in the 15-min period from 30 min post-idazoxan injec-tion to 45 min post-injection.
In the 8 animals with post-idazoxan I/O curves, repeated mea-sures ANOVAs for mean spike and EPSP slope measures for thefirst pulse were carried out for currents that all 8 rats had received(150–750 μA) between pre- and post-drug I/O curves.
4.7.2. The effect of idazoxan and timolol on paired pulse inhibition (P1overlap pre- and post-idazoxan)For each pair of recorded responses to paired pulse stimulation oneach pipette, an inhibition ratio (P2/P1) value was calculated bydividing the spike amplitude of the response to the second pulseby that to the first. Since the P2/P1 population spike ratio varieswith magnitude of P1, spike amplitudes were ordered on an Excelspreadsheet, and the area of P1 overlap for pre-idazoxan and 30-min post-idazoxan spikes was determined. As necessary, an ap-proximately equal number of values were removed from theextremes of each of the pre- and post-idazoxan distributionsuntil P1 values were not significantly different using an unpairedt test for unequal variances (until P N 0.05). One t test was done foreach of the recording pipettes. The P2/P1 spike and EPSP sloperatios associated with the P1 values remaining after removal ofextremes were used to calculate pre- and post-idazoxan P2/P1means which were subjected to statistical analysis of groupeddata in the form of paired t tests in order to determine if pairedpulse ratios before idazoxan were significantly different fromthose after the drug.
4.7.3. The effect of idazoxan and timolol on paired pulse inhibition atconstant current levels (I/O curves pre- and post-idazoxan)In the 8 animals that had post-idazoxan I/O curves, comparisonsof spike and EPSP slope inhibition measures were carried outusing a Pipette × Drug × Current ANOVA to analyze the differencesbetween the mean P2/P1 values associated with each current.
4.8. Correlations
To confirm the reported relationship between slope and spike,and between P1 spike and P2/P1 spike ratios, Pearson correlationcoefficients were determined for each pipette, using all recordspre-idazoxan, and 30 min post-idazoxan and later for eachanimal.
Acknowledgment
This research was supported by NSERC grant 9791 to C.W.Harley.
R E F E R E N C E S
Aston-Jones, G., Shipley, M.T., Chouvet, G., Ennis, M., Van, B.E.,Pieribone, V., Shiekhattar, R., Akaoka, H., Drolet, G., Astier, B.,1991. Afferent regulation of locus coeruleus neurons: anatomy,physiology and pharmacology. Prog. Brain Res. 88, 47–75.
Austin, K.B., Bronzino, J.D., Morgane, P.J., 1989. Paired-pulsefacilitation and inhibition in the dentate gyrus is dependent onbehavioral state. Exp. Brain Res. 77, 594–604.
Boehm, S., 1999. Presynaptic alpha2-adrenoceptors controlexcitatory, but not inhibitory, transmission at rat hippocampalsynapses. J. Physiol. 519 (Pt. 2), 439–449.
Bramham, C.R., Bacher, S.K., Sarvey, J.M., 1997. LTP in the lateralperforant path is beta-adrenergic receptor-dependent.NeuroReport 8, 719–724.
Brown, R.A., Walling, S.G., Milway, J.S., Harley, C.W., 2005. Locusceruleus activation suppresses feedforward interneurons andreduces beta-gamma electroencephalogram frequencies whileit enhances theta frequencies in rat dentate gyrus. J. Neurosci.25, 1985–1991.
Dahl, D., Sarvey, J.M., 1989. Norepinephrine inducespathway-specific long-lasting potentiation and depression inthe hippocampal dentate gyrus. Proc. Natl. Acad. Sci. U. S. A. 86,4776–4780.
Harley, C.A., Bielajew, C.H., 1992. A comparison of glycogenphosphorylase a and cytochrome oxidase histochemicalstaining in rat brain. J. Comp. Neurol. 322, 377–389.
Harley, C.W., Evans, S., 1988. Locus coeruleus-inducedenhancement of the perforant path-evoked potential: effects ofintradentate beta blockers. In: Woody, C.D., Alkon, D.L.,McGaugh, J.L. (Eds.), Cellular Mechanisms of Conditioning andBehavioral Plasticity. Plenum Publishing, pp. 415–423.
Harley, C.W., Milway, J.S., 1986. Glutamate ejection in the locuscoeruleus enhances the perforant path-evoked populationspike in the dentate gyrus. Exp. Brain Res. 63, 143–150.
Harley, C.W., Sara, S.J., 1992. Locus coeruleus bursts induced byglutamate trigger delayed perforant path spike amplitudepotentiation in the dentate gyrus. Exp. Brain Res. 89,581–587.
Harley, C.W., Milway, J.S., Lacaille, J.-C., 1989. Locus coeruleuspotentiation of dentate gyrus responses: evidence for twosystems. Brain Res. Bull. 22, 643–650.
Joy, R.M., Albertson, T.E., 1987. Interactions of lindane withsynaptically mediated inhibition and facilitation in the dentategyrus. Neurotoxicology [OAP] 8, 529–542.
Kitchigina, V., Vankov, A., Harley, C., Sara, S.J., 1997. Novelty-elicited, noradrenaline-dependent enhancement of excitabilityin the dentate gyrus. Eur. J. Neurosci. 9, 41–47.
Klukowski, G., Harley, C.W., 1994. Locus coeruleus activationinduces perforant path-evoked population spike potentiationin the dentate gyrus of awake rat. Exp. Brain Res. 102,165–170.
Lacaille, J.-C., Harley, C.W., 1985. The action of norepinephrine inthe dentate gyrus: beta-mediated facilitation of evokedpotentials in vitro. Brain Res. 358, 210–220.
Linner, L., Arborelius, L., Nomikos, G.G., Bertilsson, L., Svensson,T.H., 1999. Locus coeruleus neuronal activity andnoradrenaline availability in the frontal cortex of ratschronically treated with imipramine: effect of alpha2-adrenoceptor blockade. Biol. Psychiatry 46, 766–774.
Loy, R., Koziell, D.A., Lindsey, J.D., Moore, R.Y., 1980. Noradrenergicinnervation of the adult rat hippocampal formation. J. Comp.Neurol. 189, 699–710.
Moser, E.I., 1996. Altered inhibition of dentate granule cells duringspatial learning in an exploration task. J. Neurosci. 16,1247–1259.
Munro, C.A., Walling, S.G., Evans, J.H., Harley, C.W., 2001.Beta-adrenergic blockade in the dentate gyrus in vivo preventshigh frequency-induced long-term potentiation of EPSP slope,but not long-term potentiation of population spike amplitude.Hippocampus 11, 322–328.
Neuman, R.S., Harley, C.W., 1983. Long-lasting potentiation of thedentate gyrus population spike by norepinephrine. Brain Res.273, 162–165.
Oleskevich, S., Descarries, L., Lacaille, J.-C., 1989. Quantifieddistribution of noradrenaline innervation in the hippocampusof adult rat. J. Neurosci. 9, 3803–3815.
Oliver, M.W., Miller, J.J., 1985. Alterations of inhibitory processes in
45B R A I N R E S E A R C H 1 0 7 2 ( 2 0 0 6 ) 3 6 – 4 5
the dentate gyrus following kindling-induced epilepsy.Exp. Brain Res. 57, 443–447.
Paxinos, G.W.C., 1998. The Rat Brain in Stereotaxic Coordinates.Academic Press, San Diego.
Racine, R.J., Milgram, N.W., 1983. Short-term potentiationphenomena in the rat limbic forebrain. Brain Res. 260, 201–216.
Raymon, H.K., Smith, T.D., Leslie, F.M., 1992. Furtherpharmacological characterization of [3H]idazoxan bindingsites in rat brain: evidence for predominant labeling of alpha2-adrenergic receptors. Brain Res. 582, 261–267.
Regunathan, S., Feinstein, D.L., Reis, D.J., 1993. Expression ofnon-adrenergic imidazoline sites in rat cerebral corticalastrocytes. J. Neurosci. Res. 34, 681–688.
Richter-Levin, G., Segal, M., Sara, S., 1991. An α2 antagonist,idazoxan, enhances EPSP-spike coupling in the rat dentategyrus. Brain Res. 540, 291–294.
Sacchetti, G., Bernini, M., Bianchetti, A., Parini, S., Invernizzi, R.W.,Samanin, R., 1999. Studies on the acute and chronic effects ofreboxetine on extracellular noradrenaline and othermonoamines in the rat brain. Br. J. Pharmacol. 128, 1332–1338.
Sara, S.J., Bergis, O., 1991. Enhancement of excitability andinhibitory processes in hippocampal dentate gyrus bynoradrenaline: a pharmacological study in awake, freelymoving rats. Neurosci. Lett. 126, 1–5.
Sara, S.J., Vankov, A., Hervé, A., 1994. Locus coeruleus-evokedresponses in behaving rats: a clue to the role of noradrenalinein memory. Brain Res. Bull. 35, 457–465.
Sloviter, R.S., 1991. Feedforward and feedback inhibition ofhippocampal principal cell activity evoked by perforant pathstimulation: GABA-mediated mechanisms that regulateexcitability in vivo. Hippocampus 1, 31–40.
Steward, O., Tomasulo, R., Levy, W.B., 1990. Blockade of inhibitionin a pathway with dual excitatory and inhibitory actionunmasks a capability for LTP that is otherwise not expressed.Brain Res. 516, 292–300.
Thomas, D.N., Holman, R.B., 1991. A microdialysis study of theregulation of endogenous noradrenaline release in the rathippocampus. J. Neurochem. 56, 1741–1746.
Vankov, A., Hervé-Minvielle, A., Sara, S.J., 1995. Response tonovelty and its rapid habituation in locus coeruleus neurons ofthe freely exploring rat. Eur. J. Neurosci. 7, 1180–1187.
Walling, S.G., Harley, C.W., 2004. Locus ceruleus activationinitiates delayed synaptic potentiation of perforant path inputto the dentate gyrus in awake rats: a novel beta-adrenergic-and protein synthesis-dependent mammalian plasticitymechanism. J. Neurosci. 24, 598–604.
Walling, S.G., Nutt, D.J., Lalies, M.D., Harley, C.W., 2004. Orexin-Ainfusion in the locus ceruleus triggers norepinephrine (NE)release and NE-induced long-term potentiation in the dentategyrus. J. Neurosci. 24, 7421–7426.
Washburn, M., Moises, H.C., 1989. Electrophysiological correlatesof presynaptic alpha 2-receptor-mediated inhibition ofnorepinephrine release at locus coeruleus synapses in dentategyrus. J. Neurosci. 9, 2131–2140.
