B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te /b ra i n res

Research Report

In vivo glutamate decline associated with kainic acid-inducedstatus epilepticus

Natalie M. Zahra,b, Elena L. Fasano Crawfordc, Oliver Hsua,b, Shara Vincob, Dirk Mayerb,d,Torsten Rohlfingb, Edith V. Sullivana,⁎, Adolf Pfefferbaumb

aPsychiatry and Behavioral Sciences, Stanford University School of Medicine, 401 Quarry Rd., Stanford, CA 94305, USAbNeuroscience Program, SRI International, 333 Ravenswood Ave., Menlo Park, CA 94025, USAcThe Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USAdRadiology Department, Lucas MRS/I Center, Stanford University, 1201 Welsh Road, P-273, Stanford, CA 94305-5488, USA

A R T I C L E I N F O

⁎ Corresponding author. Fax: +1 650 859 2743.E-mail address: edie@stanford.edu (E.V. S

0006-8993/$ – see front matter © 2009 Elsevidoi:10.1016/j.brainres.2009.08.060

A B S T R A C T

Article history:Accepted 18 August 2009Available online 26 August 2009

Neurophysiological, biochemical, and anatomical evidence implicates glutamatergicmechanisms in epileptic seizures. Until recently, however, longitudinal characterizationof in vivo glutamate dynamics was not possible. Here, we present data using in vivomagnetic resonance spectroscopy (MRS) optimized for the detection of glutamate to identifychanges that evolve following kainic acid (KA)-induced status epilepticus. Wild-type maleWistar rats underwent whole-brain MR imaging and single-voxel MRS on a clinical 3 Tscanner equipped with a high-strength insert gradient coil. Scanning took place before andthen 3 days, 28–32 days, and 42–50 days after induction of status epilepticus. Analysescompared 5 seizure (Sz), 5 no-seizure (NoSz; received KA but did not exhibit seizures), and 6control (Con) animals. This longitudinal study demonstrated reduced glutamate levels invivo in the dorsal hippocampus 3 days and 1 month following status epilepticus in Szanimals compared with Con animals. Additionally, previous results were replicated: in theSz group, computed T2was higher in the ventral hippocampus and limbic cortex 3 days afterseizure activity compared with baseline but resolved in both regions at the 1 month scan,suggesting a transient edema. Three days following seizure activity, N-acetylaspartate(NAA) declined and lactate increased in the dorsal hippocampus of the Sz group comparedwith the Con andNoSz group; bothmetabolites approached baseline levels by the third scan.Taken together, these results support the conclusion that seizure activity following KAinfusion causes loss of glutamatergic neurons.

© 2009 Elsevier B.V. All rights reserved.

Keywords:GlutamateHippocampusCerebellumLongitudinalRatStatus epilepticus

1. Introduction

Ample data support a role for glutamate in epilepsy. In vivomicrodialysis studies report elevated extracellular glutamatelevels before or during seizures in hyperexcitable brain regionsin animal models of epilepsy (Liu et al., 1997; Sierra-Paredes et

ullivan).

er B.V. All rights reserved

al., 1998; but see Tanaka et al., 1996; Wade et al., 1987) and inpatients with medial temporal lobe epilepsy (Carlson et al.,1992; During and Spencer, 1993; Wilson et al., 1996). Asustained elevation in basal release of glutamate (Zhang etal., 1991), in addition to initiating and maintaining epilepticseizures, may be excitotoxic and contribute to cell death (e.g.,

.

66 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

Magloczky and Freund, 1993; Meldrum, 1993). In the interictalstage, in vitro studies report low glutamate content inhippocampal tissue from both animals (Koyama, 1972) andhumans (Peeling and Sutherland, 1993) potentially reflectingthe loss of glutamatergic neurons of the hippocampus. Recentin vitro magnetic resonance spectroscopy (MRS) studiesdemonstrate a reduction in the amount of 13C labeling ofglutamate in the hippocampal formation of systemicallytreated kainic acid (KA) rats (Alvestad et al., 2008) and biopsiedhippocampal tissue from patients with medial temporal lobeepilepsy (Petroff et al., 2002). However, our understanding ofthe role of altered glutamate homeostasis in epilepsy isincomplete, especially in vivo, and techniques such as invivo microdialysis have limited applicability to longitudinalstudies (Plock and Kloft, 2005).

Temporal lobe epilepsy can be modeled in rodents byadministration of KA (Williams et al., 2007). As an excitatoryamino acid agonist (Ferkany et al., 1982), KA can provokeprolonged seizures (i.e., status epilepticus), characterized bysalivation, rearing, bilateral upper extremity clonus, andfalling (Lothman and Collins, 1981), and set in motion aprocess of epileptogenesis that may lead to spontaneousseizures (Williams et al., 2007). Postmortem histologicalevaluation demonstrates a specific pattern of neuronal losslargely restricted to cells of the hippocampus, amygdala, andrelated parts of the thalamus and cortex (e.g., piriform andentorhinal cortices; Nadler and Cuthbertson, 1980; Schwob etal., 1980; Sperk et al., 1983; Sperk, 1994). This specific regionalprofile of brain damage in rodents following KA-inducedstatus epilepticus resembles Ammon's horn sclerosis (Sagarand Oxbury, 1987). Neuronal loss as a consequence of KA-induced status epilepticus involves, at last in part, anapoptotic mechanism dependent on activation of caspase-3protease activity (Faherty et al., 1999; Kondratyev and Gale,2000, 2004; Puig and Ferrer, 2002; Tokuhara et al., 2007).

In vivo studies using magnetic resonance imaging (MRI)protocols sensitive to mobile water protons (e.g., T2-weightedimages or computed T2 maps) have enabled tracking of theevolution of brain changes associated with KA-inducedseizures. T2-weighted images typically detect hyperintensitiesat 24 h (e.g., TE=272 ms; Ebisu et al., 1996) that persist for3 days after systemic KA injection (Ebisu et al., 1994) and for10 days following local intracerebral KA injections (e.g.,Bouilleret et al., 2000; Luna-Medina et al., 2007; Tanaka et al.,1993). Postmortem analyses suggest that regions of hyperin-tense T2-weighted signals in the acute period following

Table 1 – Scan schedule and weights (g)±SE.

N Scan 1 Scan 2Baseline 3 days after

Control 6 391.8±6.8a,b 397.2±7.5a

KA-no seizure 5 383.4±6.3a,b 374.6±7.4a

KA-seizure 5 396.4±6.7a,b, c 396.6±19.6

NA=not required.a T2 data analyzed.b MRS data analyzed.c Ventricular volume analyzed.

seizure activity reflect edema (e.g., Hantraye et al., 1992;Tanaka et al., 1993).

Along with gross morphology and tissue quality measure-ments from structural MRI, chemical constituents of tissue aredetectable with MRS (cf., Adalsteinsson et al., 2002; Zahr et al.,2009). N-Acetylaspartate (NAA), a presumed neuronal marker,is lower in the rat hippocampus (Najm et al., 1997, 1998) andpiriform cortex (Ebisu et al., 1996) 24 h to at least 7 days (Najmet al., 1997) after systemically administered KA; NAA reduc-tions persist for up to 84 days following local intracerebralstereotaxic injections of KA (Luna-Medina et al., 2007;Tokumitsu et al., 1997). KA administration is also associatedwith an increase in hippocampal lactate that occurs within 3 hof seizure induction (Meric et al., 1994) and persists for 7 daysafter systemic injection (Najm et al., 1997, 1998) and upwardsof 9 days after local hippocampal injection of KA (Luna-Medina et al., 2007).

The present in vivo longitudinal study was undertaken toreplicate previous MRI and MRS results and to extend earlierwork by quantifying glutamate levels at baseline and as longas 1 month following KA-induced status epilepticus. Wehypothesized that MRI would detect a transient increase incomputed T2, indicative of edema in response to systemic KA-induced status epilepticus, and that structural evidence forneuronal loss in limbic structures would be paralleled byventricular enlargement. We also predicted that NAA levelswould decrease and lactate levels would increase in the dorsalhippocampus as a result of KA-induced seizure activity.Notably, based on recent in vitro studies using 13C labeledglucose, we expected levels of glutamate to decline in thedorsal hippocampus in the interictal period following statusepilepticus. Confirmation of previous results (reduced NAAand increased T2 and lactate) supports the novel in vivofinding of decreased glutamate following KA-induced statusepilepticus. Radiological results were verified with postmor-tem immunohistochemical staining for caspase-3.

2. Results

Between-group analyses were conducted on computed T2data from scans 1 and 2 and on MRS data from scans 1, 2, and3. Only the seizure (Sz) group was scanned 4 times and formedthe basis for within-subject analyses of ventricular volume(Table 1). Bodyweight data for the control (Con) and no seizure(NoSz) groups is absent at the 42- to 50-day time point since

Scan 3 Scan 4KA 28–32 days after KA 42–50 days after KA

,b 456.7±9.4b NA,b 441.6±6.0b NAa,b,c 452.8±17.5b,c 473.8±18.4b,c

67B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

these groups did not receive a 4th scan. For the first 3 scans,there were no group differences in body weights; within eachgroup, however, rats were significantly larger at scan 3compared with scans 1 and 2.

2.1. Changes in computed tissue T2

Regions of interest (ROI) were manually drawn on late-echoimages in the dorsal and ventral hippocampus, and in corticalregions (starting at∼−7.00 Bregma) lateral to the hippocampusextending from the entorhinal cortex to the piriform cortex(referred to as the limbic cortex, Fig. 1). For statistical analysis,tissue T2 values were computed using the average ROI valuesfor the early- and late-echo images. A 3-group (Con, Sz, NoSz)by 2-time (baseline and 3 days after KA) analysis of variance(ANOVA) examining computed T2 changes in the dorsalhippocampus did not reveal any group, time, or interactioneffects (Fig. 1). In the ventral hippocampus, a similar ANOVAyielded a group-by-time interaction (F(2,13)=6.01, p=.0142,GG= .0142; nonparametric Kruskal–Wallis (KW) results:H=6.38, p=.041), indicating that the Sz group had a significantcomputed T2 increase between scan 1 and scan 2 relative tothe Con and NoSz groups, neither of which changed betweenscans. Follow-up t-tests confirmed a greater computed T2change (percent increase from baseline) in the ventral

Fig. 1 – Top:Manually drawn regions of interest (ROIs) on late-echoT2 increase in the 3 regions, 3 days after kainic acid-induced statuSz and NoSz groups.

hippocampus 3 days after KA-induced status epilepticus inthe Sz group relative to the Con group (t(9)=3.09, p=.0129) andthe NoSz group (t(8)=2.43, p=.0412, Fig. 1). The limbic cortexshowed a similar pattern, where an ANOVA yielded a group-by-time interaction (F(2,13)=11.85, p=.0012, GG=.0012; KWresults: H=9.97, p=.0069). The Sz group had increased com-puted T2 in the limbic cortex 3 days after KA-induced statusepilepticus compared with the Con (t(9)=4.26, p=.0021) andNoSz groups (t(8)=3.38, p=.0097) (Fig. 1). One month (28–32days) following status epilepticus, the computedT2 increasein the limbic cortex had resolved; in the ventral hippocampus,fluid-filled cysts were clearly evident precluding meaningfulcomputation of T2 values at the third scan (Fig. 2).

2.2. Ventricular enlargement

Within the Sz group, a repeated-measures ANOVA examiningventricular volume across the 4 MRI sessions revealed asignificant effect of time (F(3,12)=10.68, p=.001, nonparamet-ric Friedman test results: χ2=10.68, p=.0136), indicating anincrease in ventricular volume (Fig. 3). Three days after KA-induced status epilepticus, the Sz group showed no significantincrease in ventricular volume (p=.8573). One month afterstatus epilepticus, however, ventricular volume in the 5 Sz ratswas 7.6 fold higher than baseline (p=.0195) and 7.3 fold higher

axial slices for computation of T2. Bottom: Percent computeds epilepticus. *p≤ .05 between Sz and Con groups and between

Fig. 2 – Late-echo (top panel) and computed T2 (bottom panel) MR images of an axial brain slice of a single rat. Closed arrowspoint to the ventral hippocampus, open arrows to the limbic cortex. Note the regions of hyperintensity on computed T2 imagesat 3 days, and the presence of fluid-filled cysts in the ventral hippocampus at 28 and 42 days following status epilepticus.

Fig. 3 – Top: Early-echo MR images of an axial brain slice of a single seizure rat demonstrating enlargement of ventricularvolume. Bottom: Thresholded ventricles on T2 computed images from the same animal.

68 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

Fig. 4 – Quantified ventricular volume of each seizure animal at the 3 time points following kainic acid-induced statusepilepticus.

Fig. 5 – Voxel placement in the dorsal hippocampus andcerebellum including reproducibility of voxel placementacross the first 3 scan time points.

69B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

than it was 3 days after KA administration (p=.027). Between1month (28–32 days) and 42–50 days after status epilepticus, 2of the 5 seizure animals showed ventricular shrinkage,although individual variability among seizure animals waslarge (Fig. 4).

2.3. Metabolite changes

MRS data were collected in the dorsal hippocampus toevaluate the effects of the KA-induced status epilepticusand in the cerebellum as a control brain structure thought tobe spared by seizure activity (Fig. 5). In the dorsal hippocam-pus, a 3-group (Con, NoSz, Sz) repeated-measures (3 scansessions, 6 metabolite signals) ANOVA revealed group-by-metabolite (F(10,65)=2.84, p=.006, GG=.02) and group-by-metabolite-by-time (F(20,130)=3.5, p=.0001, GG=.005) interac-tions, indicating a significant effect of status epilepticus onmetabolite levels. Three days after seizure induction, NAA inthe Sz group was significantly lower than in the Con (t(9)=2.62, p=.014) and NoSz groups (t(8)=2.94, p=.0093) (Fig. 6). Atthe 1 month scan there was 50% recovery of NAA, with levelswhich had dropped to 69% of baseline up to 83% of baseline;however, NAA levels in the Sz group were still significantlylower than the Con group (t(9)=2.66, p=.0129; Fig. 7). Withinthe Sz group, the decrease in NAA between baseline and3 days following seizure activity was significant (t(4)=2.99,p=.0401), as was the increase in NAA between 3 and 28–32 days following seizures (t(4)=3.66, p=.0215).

Lactate increased 3 days after status epilepticus in thedorsal hippocampus in the Sz group compared with both theCon (t(9)=3.55, p=.0031) and NoSz groups (t(8)=2.96, p=.0091).By the third scan, 1 month after status epilepticus, lactate

levels in the dorsal hippocampus were indistinguishableamong the 3 groups (Fig. 7). Within the Sz group, thesignificant increase in lactate 3 days following status epilepti-cus (t(4)=3.93, p=.0171) was followed by a significant decrease

Fig. 6 – Sample spectra froma single seizure animal (dashed, red) and a single control animal (solid, blue) for voxels in the dorsalhippocampus (top) and cerebellum (bottom) across the first 3 scan sessions.

70 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

in lactate at the scan performed 1 month following seizureactivity (t(4)=3.01, p=.0397).

MRS-detectable glutamate in the dorsal hippocampusshowed modest but predicted (1-tailed t-tests) changes.Specifically, glutamate tended to be lower in the Sz groupcompared with the Con group at 3 days and 1 month (3 days:t(9)=1.93, p=.0427, 1 month: t(9)=1.80, p=.0528) and lowerthan the NoSz group at 3 days (t(8)=1.84, p=.0515) but not at1 month following seizure activity. Within the Sz group, thedecrease in glutamate observed between baseline and 3 daysfollowing seizures was significant (t(4)=3.07, p=.0186); at thethird scan, 1 month following seizures, glutamate levels in theSz group remained lower than at baseline (t(4)=1.97, p=.0603).Status epilepticus did not significantly affect any of the otherMRS-measured metabolites in the dorsal hippocampus.

In the cerebellum, a similar 3-group repeated-measuresANOVA yielded a group-by-metabolite-by-time interaction (F(20,130)=2.36, p=.002, GG=.03), likewise indicating a signifi-cant seizure effect on metabolite levels. Three days afterseizure induction, therewas a significant increase inCho in thecerebellum of the Sz group compared with the Con (t(9)=2.47,p=.0353) and NoSz groups (t(8)=2.49, p=.0374) that resolved1 month later (Fig. 8). The Glx peak was also higher in thecerebellum of the Sz group 3 days after seizures comparedwith the NoSz group (t(8)=2.56, p=.0337), but this was in partdue to a drop in Glx in the NoSz group (Fig. 8). A significantdifference in Glu was observed between the Con and NoSz

groups (t(9)=2.95, p=.0163) 1 month after KA administration,but the significance was carried by an inexplicable increase inGlu in the Con group.

2.4. Immunohistochemistry

A 3 group-by-3 region (CA1, CA3, DG) repeated-measuresANOVA revealed group effects (F(2,12)=15.83, p=.0004), indicat-ing a disproportionately higher number of caspase-3 stainingcells in the Sz group than in the Con or NoSz groups. Follow-upt-tests revealed that the Sz group had a higher number ofcaspase-3 staining cells compared with the Con and NoSzgroups in the CA1 (Con t(9)=4.18, p=.0024; NoSz t(7)=5.65,p=.0008) and CA3 (Con t(9)=2.58, p=.0296; NoSz t(7)=3.23,p=.0144) regions of the hippocampus. Total caspase-3 stainingin the hippocampus was likewise higher in the Sz group thantheCon (t(9)=4.92, p=.0008) andNoSz groups (t(7)=5.38, p=.001).

2.5. Correlations

The relationships between ventricular volume, NAA, andglutamate at the third scan were explored using simpleregression analysis including only the Sz group. A significantcorrelation between ventricular volume and glutamate levels(ρ=1, p= .05) revealed that lower glutamate levels wereassociated with larger ventricular volumes (Fig. 9). Similarly,lower NAA levels were associated with larger ventricular

Fig. 7 –Water-referencedmetabolite levels in the dorsal hippocampus at baseline (MRS 1), 3 days (MRS 2), and 28–32 days (MRS3) after kainic acid (Sz+NoSz) or water (Con) injection as quantified using peak integration. *p<.05 between Sz and Con, #p<.05between Sz and NoSz.

71B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

volumes (ρ=.9, p=.07; Fig. 9). Glutamate and NAA were relatedsuch that lower glutamate correlated with lower NAA (ρ=.9,p=.07; Fig. 9).

3. Discussion

KA per se was ineffective in producing in vivo evidence of brainchanges; by contrast, status epilepticusproducedboth local andglobal changes (O'Shaughnessy and Gerber, 1986). At 3 daysfollowing status epilepticus, edematous tissue localized bilat-erally to the ventral hippocampus and limbic cortex, asevidenced by the increase in computed T2. One monthfollowing status epilepticus, theT2 increase in the limbic cortexresolved, cavitations were present in the ventral hippocampus,and signs of ex vacuo ventricular dilation were evident. Partialrecovery of ventricular volume was observed 42–50 daysfollowing seizure induction in 2 of 5 seizure rats. At the acute,3 day time point, the dorsal hippocampus revealed significantdeclines in NAA and glutamate and increases in lactateselective to rats that exhibited status epilepticus. While a fallin lactate to baseline levels was observed, NAA and glutamatelevels remained lower in the Sz comparedwith theCon group atthe third scan, 1 month following status epilepticus. At

postmortem, 50 days after status epilepticus, immunohisto-chemistry revealed a higher number of caspase-3 stained cellsin CA1 and CA3 regions of the hippocampus in the Sz groupcompared with the NoSz and Con groups, indicative of celldeath and consistent with the enduring depression of NAA andglutamate in the Sz group.

3.1. Changes in computed T2

Late-echo fast spin-echo (FSE) images revealed hyperintensi-ties quantified as increased computed T2 values in the ventralhippocampus and limbic cortex of the Sz group 3 days afterstatus epilepticus in agreement with previous MR studiescarried out 1 – 3 days following systemic KA-induced seizures(Ebisu et al., 1994, 1996; Righini et al., 1994) and upwards of9 days following local KA injections into the hippocampus(Bouilleret et al., 2000; Luna-Medina et al., 2007; Tanaka et al.,1993). Brain edema begins 2–24 h after KA injection (Lassmannet al., 1984; Schwob et al., 1980; Sperk et al., 1983), andpostmortem evidence suggests that the hyperintense T2-weighted signal observed at acute times following KA-inducedseizure activity reflects acute edema (Bouilleret et al., 2000;Norman et al., 1990; Tanaka et al., 1993). In humans withtemporal lobe epilepsy, focal regionswith high signal intensity

Fig. 8 – Water-referenced metabolite levels in the cerebellum at baseline (MRS 1), 3 days (MRS 2), and 28–32 days (MRS 3) afterkainic acid (Sz+NoSz) or water (Con) injection as quantified using peak integration. *p<.05 between Sz and Con, #p<.05between Sz and NoSz, +p<.05 between Con and NoSz.

72 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

on T2-weighted images have been noted and ascribed totransient focal edema when the MR scan is performed withinhours of a seizure (Kramer et al., 1987; Stone et al., 1986).

Fig. 9 – Correlations (non parametric Spearman rank) between veglutamate at the third scan.

At the third scan, 1 month after status epilepticus, areaswith high computed T2 in the ventral hippocampus andlimbic cortex had resolved: in the ventral hippocampus,

ntricular volume and NAA or glutamate, and NAA and

73B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

hyperintensity was replaced with fluid-filled cysts. Thiscontrasts with studies demonstrating persisting high signalintensity on T2-weighted images at 1 month (Tanaka et al.,1993) and 4 months (Bouilleret et al., 2000) following seizureactivity. These inconsistent resultsmay be due to the injectionmethod used (previous studies used local stereotaxic injec-tions of KA into limbic regions) or because in previous studies,epileptogenesis resulted in repetitive spontaneous motorseizures acutely associated with an increase in T2-weightedsignal intensity and indicative of edema.

3.2. Ventricular enlargement

One month after seizure induction, lateral ventricular volumein the Sz group was significantly larger than at baseline and3 days after KA administration, in agreement with MRI studiesperformed 9–10 days following KA treatment (Righini et al.,1994; Wolf et al., 2002). Previous postmortem histologyperformed 1–7 days after systemic KA treatment revealedwidespread neural damage in the hippocampal formation,amygdaloid complex, limbic cortex (piriform and entorhinalcortices), and midline thalamus (Bates et al., 1988; Sperk,1994). Thus, enlargement of the lateral ventricles is assumedto result from atrophy of the surrounding brain tissue, namelythe hippocampus, amygdala, piriform cortex, and thalamus(Roland and Savage, 2007). Although considered a marker forhippocampal atrophy related to Alzheimer's disease (Camaraet al., 2008), cautionmust be used when interpreting increasesin ventricular volume, i.e., dysmorphology of a non-tissue, as asurrogate, to draw conclusions about the target structure(Pfefferbaum et al., 2004b).

3.3. Metabolite changes

The decrease in NAA observed in the dorsal hippocampus3 days following status epilepticus comports with studies inwhich a reduced NAA signal was observed in the hippocam-pus 12 h to 7 days following systemic KA-induced seizures(Ebisu et al., 1994; Ebisu et al., 1996; Najm et al., 1997, 1998) and1 to 84 days following local stereotaxic KA injections (Luna-Medina et al., 2007; Mao et al., 2007; Tokumitsu et al., 1997). Inthe current study, the reduction in NAA in the dorsalhippocampus was not associated with changes in computedT2 (Ebisu et al., 1996; Tokumitsu et al., 1997) suggesting thatMRS may detect a subtler or different form of tissue damagethan is possible with T2-weighted imaging or regionallycomputed T2.

A 50% recovery of NAA levels was observed in the Sz groupat 1 month compared with levels at 3 days following statusepilepticus. Because it is unlikely that neurogenesis accountsfor the recovery in NAA levels (Nixon, 2006), and becauseprevious studies suggest that NAA levels can decline inresponse to KA-induced seizures even with little histologicalevidence for neuronal loss (cf., Ebisu et al., 1996), it ispresumed that the interpretation of NAA as solely a markerof neuronal viability is insufficient (Moffett et al., 2007). Theprofound but transient decrease in NAA observed at 3 daysfollowing seizure activity may reflect mitochondrial dysfunc-tion or compromised ATP production at the acute time point(Moffett et al., 2007; Retz and Coyle, 1982). By contrast, the

fraction of NAA that remained attenuated in the Sz group atthe 1 month scan may indeed represent a loss of neurons, aninterpretation underscored by the relationship between NAAand ventricular volume (Fig. 9).

Levels of lactate were increased 3 days following KA-induced seizures (Luna-Medina et al., 2007; Najm et al., 1998)but returned to baseline levels 1 month later (Najm et al.,1997). KA administration is associated with energy failure: adramatic increase in glucose metabolism (Lothman andCollins, 1981) and a rapid and significant reduction in ATPand glucose (Retz and Coyle, 1982) may lead to accumulationof lactate. Concurrently, mitochondrial dysfunction related toimpending neuronal lossmay also increase lactate production(Ebisu et al., 1996). Evidence for an increase in lactate followingKA-induced seizure activity has been provided by severalstudies (e.g., Luna-Medina et al., 2007; Meric et al., 1994; Najmet al., 1997) and interpreted as reflecting an increase in cellularactivity and metabolism leading to energy failure; themismatch between oxidative metabolism and glycolysis mayproduce lactate.

Comporting with in vitro results (Alvestad et al., 2008;Petroff et al., 2002), the current in vivo study revealed astatistically significant reduction in the levels of glutamate inthe dorsal hippocampus at 3 days following status epilepticusin the Sz group compared with the NoSz and Con groups. Atthe 1-month scan, the levels of glutamate tended to be lowerin the Sz group compared with the Con group. In humanstudies, levels of glutamate were described as lowest in glioticregions of the epileptic hippocampus (Peeling and Sutherland,1993; Petroff et al., 2002) suggesting that hippocampalglutamate levels in epilepsy may reflect severity of neuronalloss and the associated gliosis. The hippocampal formation ofrats following KA-induced status epilepticus demonstratesreduced 13C-labeling of glutamate, implying reduced gluta-mate content and variously interpreted as reflecting loss ofglutamatergic neurons of the hippocampus or mitochondrialdysfunction of the remaining neurons (Alvestad et al., 2008).The correlation observed in the current study betweenreduced glutamate levels and enlarged ventricular volume(suggestive of neuronal loss), however, supports the formerinterpretation (Fig. 9).

For the most part, metabolite levels in the cerebellum didnot change as a function of KA administration or statusepilepticus. An exception was an increase in Cho 3 daysfollowing KA-induced status epilepticus. This increase in Chois a novel finding that, if not a chance occurrence, may reflectgeneralized membrane breakdown, demyelination, or inflam-mation (Govindaraju et al., 2000) accompanying status epi-lepticus. Because the CT-PRESS data were acquired at TEsranging from 36.6 ms to 241.4 ms (average TE= 139 ms), thesignal intensities in the spectrum could be significantlyaffected by T2. However, if the metabolites showed a similartrend to longer T2 as water, this would lead to higher signalsfrom all metabolites and, hence, could not explain theobserved signal decreases from NAA and glutamate. Althougha longer T2 could plausibly explain the increase in the lactatesignal, we believe a change in lactate concentration was thedominant cause, in agreement with previous findings andinterpretations (e.g., Luna-Medina et al., 2007; Meric et al.,1994; Najm et al., 1997, 1998).

74 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

3.4. Immunohistochemistry

Some histological studies implicate apoptotic processes in theneuronal death that occurs in brain regions targeted by KA-induced status epilepticus (Faherty et al., 1999; Kondratyevand Gale, 2000). As caspase-3 proteases play a crucial role inapoptosis, immunohistochemistry for the caspase-3 proteinwas carried out to evaluate apoptotic cells in the CA1, CA3, andDG regions of the entire hippocampus (both dorsal and ventralregions). A significantly higher number of caspase-3 stainingcells were observed in the Sz group than in the Con and NoSzgroups, consistent with previous studies (Faherty et al., 1999;Kondratyev and Gale, 2000). Status epilepticus-induced braindamage is pronounced in CA1 and CA3 regions of thehippocampus (Bouilleret et al., 2000; Sperk et al., 1983),while the DG is relatively spared (Ben-Ari et al., 1981; Lothmanand Collins, 1981; Olney et al., 1986). This pattern of insult andsparing is consistent with the current findings wherecaspase-3 staining was pronounced in the CA1 and CA3regions of the Sz group but indistinguishable among thethree groups in the DG.

3.5. Limitations

Animals receiving KA and exhibiting status epilepticus areprone to the development of spontaneous seizures following alatent period. The length of the latent period, and thefrequency of seizure occurrence, is, however, a contentiousarea of research (e.g., Williams et al., 2007). In this study,although animals were regularly observed during the firstweek following KA administration when no spontaneousseizure activity was perceived, the animals were not moni-tored by video or electroencephalographic recordings 24 h aday for accurate detection of the first spontaneous motorseizure (Williams et al., 2007). Thus, it is possible that at leastsome of the animals in the Sz group had spontaneous seizuresby scan 4 that may explain the high variability in ventricularvolume enlargement and recovery.

The peak integration technique is potentially more proneto be affected by changes or differences in T2′ (i.e., the timeconstants for the signal decay due to B0 inhomogeneities)than are line-fitting routines such as the LCModel. However,this quantitation technique was used to quantify MRS databecause one-dimensional (1D) linear least squares fittingmethods such as the LCModel are in general not applicableto quantifying two-dimensional (2D) constant time pointresolved spectroscopy (CT-PRESS) data. However, efforts areunderway to develop an LCModel-type quantification methodsuitable for CT-PRESS data that will take advantage of the full2D spectral information.

4. Conclusions

Unlike postmortem analyses where observations are limited tospecific time points following insult, in vivo MR imaging andspectroscopyprovide theopportunity to track the evolutionandresolution of brain compromise. Three days following KA-induced status epilepticus, increases in computed T2 in theventral hippocampus and limbic cortex indicated edema,

whereas decreases in NAA and glutamate, and increases inlactate in the dorsal hippocampus indicated energy failure andthe onset of tissue damage. One month following statusepilepticus, the computed T2 increase had resolved, indicatingthat the edema was transient, and lactate levels normalized,indicating a return to normal glucose utilization or mitochon-drial function. Some cavitation of ventral hippocampal tissuelimited tissue volume recovery. NAA and glutamate levelsremained significantly lower than baseline levels and, togetherwith correlations with ventricular volume and the significantpresence of caspase-3 staining indicate that a fraction of cells inthe dorsal hippocampus were irrecoverably damaged. At thefinal scan, 42–50 days after status epilepticus onset, shrinkageof theenlargedventricles occurred in someanimals thatmaybeindicative of tissue recovery. Detection of transient andpermanent changes in the brain highlights the utility of MRIandMRS as tools to characterize the structural and neurometa-bolic profile of disease processes and their potential forevaluating the effects of medical and surgical treatments fordiseases such as epilepsy.

5. Experimental procedures

5.1. Animals

The study group comprised 16 male, wild-type Wistar rats(Charles River Laboratories) singly housed with free access tofood (vitamin- and mineral-enriched mouse and rat diet,#7001; Teklad, Madison, WI) and water. Rats were acclimatedfor 4 weeks before the first imaging session (baseline scan), atwhich time their average weight was 390.6±15.3 g. TheInstitutional Animal Care and Use Committees at SRIInternational and Stanford University approved allprocedures.

5.2. Kainic acid-induced status epilepticus

Kainic acid (Cayman Chemical, Ann Arbor, MI), purchased as acrystalline solid, was diluted in phosphate-buffered saline(PBS) to a final concentration of 2 mg/ml (pH adjusted to 7.2).After the baseline scan, 10 of 16 rats were administeredintraperitoneally (i.p.) a starting dose of 10 mg/kg KA, then10 mg/kg KA i.p. each hour for 4 h (Hellier et al., 1998; Hellierand Dudek, 2006) and monitored for seizure activity accordingto a modified Racine scale: 0=no changes, .5=wet dog shakes,1=mouth and facial movements, 2=head nodding, 3=fore-limb clonus, 4=rearing, 5=rearing and falling (Racine, 1972).Five of the ten animals that received KA exhibited seizureactivity with a minimum score of 3 for at least 3 h (total dose:10–50 mg/kg); this behavioral pattern was classified as statusepilepticus and these animalswere included in the seizure (Sz)group. The other 5 of 10 rats that received KA (total KA dose=50 mg/kg) did not exhibit seizure activity and were classifiedas the no-seizure (NoSz) animals. The remaining 6 animalswere injected i.p. with equivalent volumes of saline and weredesignated the control (Con) group. Seizure animals weremonitored daily until they consumed food and water,groomed, and defecated as normal (∼1 week). Until thispoint of recovery, animals were kept hydrated with

75B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

subcutaneous injections of saline (10 ml) and fed Critical Care(Oxproline, Murdock, NE) intragastrically.

5.3. MR scanning procedures

5.3.1. ScheduleAll animals were scanned at baseline (scan 1), and then 3 days(scan 2) and 1 month (28–32 days, scan 3) after KA adminis-tration (Table 1). To track local brain damage progress orrecovery, the Sz group had an additional scanning session 42–50 days (scan 4) after status epilepticus induction.

5.3.2. Anesthesia and monitoringAnimals were held in an MR-invisible structure providingsupport for a radiofrequency (RF) coil and a nose cone fordelivery of isoflurane anesthesia (1.5% to 3.5%) and oxygen(1.5 l/min) (Adalsteinsson et al., 2004; Pfefferbaum et al.,2004a). Blood oxygen saturation, pulse rate, rectal tempera-ture, and respiration were monitored throughout the exper-iment. Heating was provided by pre-warmed bags of salinesolution placed under the animals. Scanning sessions were∼2 h for each rat.

5.3.3. MRI acquisitionThe experiments were conducted on a clinical 3T GE Signa MRscanner equipped with a high-strength insert gradient coil(peak strength=600 mT/m, peak slew rate=3200 T/m/s;Chronik et al., 2000a,b). The gradient system was operated atan amplitude of 500mT/mwith a slew rate of 1800mT/m/ms.Acustom-made rat brain quadrature head coil (∅=44 mm) wasused for both RF excitation and signal reception. A gradient-recalled echo (GRE) localizer scan (echo time (TE)/repetitiontime (TR)=1.9/7.7ms, field of view (FOV)=80×80mm2, 256×128matrix, 5 mm slice per plane) was used to position theanimals in the scanner and for graphical prescription of thesubsequent scans. High-resolution, dual-echo, FSE imageswere acquired in the coronal plane, transaxial to the magnetsystem bore (TE1/TE2/TR=12/60/5000 ms, FOV=64×48 mm2,256×192 matrix, echo train length=8, 50 slices, .3 mm thick,0 mm skip (no slice gap), 4 acquisitions, NEX=3). Nominalacquisition resolution was .25 mm×.25 mm in plane. Herein,“early echo” refers to TE1=12 ms and “late echo” refers toTE2=60 ms.

5.3.4. Image postprocessingMRI data from scans 1, 2, and 3 were quantified afterregistration to the baseline images of one untreated, referenceanimal (Pfefferbaum et al., 2008). For registration, to supportmore finely sampled deformation fields, the data were up-sampled to near-isotropic voxels (.117×.117× .125 mm) usingcubic spline interpolation. Each animal's scan 1 image wasregistered to the reference animal's scan 1 data with rigidfollowed by affine followed by nonrigid registration (Rohlfingand Maurer, 2003; Rohlfing et al., 2006, 2008). For longitudinalalignment, each animal's scan 2 and scan 3 images weresimilarly registered to its own scan 1 data. For the Sz group,each animal's scan 4 images were also registered to its ownscan 1 data.

By concatenating inter-animal and within-animal trans-formations for scans 2 and 3, rather than registering scan 2

and scan 3 images to the reference animal independently, itwas ensured that the unavoidable residual misalignmentsbetween animals were consistent for all three scans withineach animal. Rigid registrations, preserving relative within-animal size, were retained for individual and group averagedisplay. T2 maps (computed voxel-by-voxel: T2=[TE2−TE1]/[log(IntensityTE1)− log(IntensityTE2)]) were constructed for dis-play and are herein referred to as computed T2. The threeregions of interest for tissue T2 computation were drawn onthe time 2 data for each rat. After rigid alignment (preservingall size information) of all sessions to the same image space,the ROIs were projected onto times 1 and 3 images andmanually edited to ensure that no ventricular enlargement orcavitation was included in the ROI. A 5 slice ROI encompassingthe lateral ventricle was defined on a group average time 1brain. Rigid inter-subject alignment for all animals for alltimes preserved ventricular size and allowed for projection ofthe ventricular ROI to each animal for each observation time.Thresholding for ventricular volume was accomplished byconstructing histograms for all computed T2 voxels in the ROI,normalizing the histograms to time 1, and defining all voxelsgreater than 110 ms as cerebral spinal fluid (Fig. 3).

5.3.5. MRS acquisitionFSE images were used to prescribe voxels in the dorsalhippocampus to determine seizure effects and in the cerebel-lum to provide a control structure. Optimization of voxel sizefor high signal-to-noise ratio (SNR) precluded MRS measure-ments in the smaller ventral hippocampus. The dorsalhippocampal voxel (∼ .16 cm3=9.8×4×4 mm3) was centeredat approximately −4.00 mm Bregma (and extended 2 mmanterior and posterior to this central point), extended 4.9 mmto the right and left of midline, and 4 mm inferior to −3.10 mmBregma, according to the atlas of Paxinos and Watson (2005).The cerebellar voxel (∼ .19 cm3=8.5×5×4.6 mm3) extended2.5 mm anterior and posterior to a central position atapproximately-11.80 mm Bregma, 4.25 mm to the right andleft of midline, and 4.6 mm inferior to −3.60 mm Bregma,according to the atlas of Paxinos and Watson (2005).

Spectroscopy was performed with a constant time pointresolved spectroscopy (CT-PRESS) sequence (Dreher andLeibfritz, 1999; Mayer and Spielman, 2005; Mayer et al.,2007), which allows for the detection of J-coupled resonances(e.g., glutamate) with increased SNR and spectral resolutionby using effective homonuclear decoupling. CT-PRESS con-sists of a modified PRESSmodule in which the position of thelast refocusing pulse is shifted to encode the chemical shift(CS) in the second time dimension (t1). The pulse was shiftedin 129 steps with increment Δt1 /2= .8 ms corresponding to aspectral width (SW1) of 625 Hz in f1. The average TE of thesequence was 139 ms, optimized for the detection ofglutamate. To increase SNR, data acquisition (2048 complexpoints at SW2= 5000 Hz) started immediately after the lastcrusher gradient of the second refocusing pulse. A three-pulse CS-selective sequence for water suppression and anouter volume suppressionmodule using selective saturationpulses preceded CT-PRESS. TR was 2 s, and 4 excitationswere performed without data acquisition to establish asteady state. With 6 averages, the acquisition time perspectrum was 26:36 min. An additional acquisition without

76 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

water suppression was carried out (17 CS encoding steps,Δt1 /2= 6.4 ms, 2 averages, Tacq=1:16 min) to measure tissuewater content used to normalize the metabolite signalintensities.

The amount of cerebral spinal fluid and tissue water in thespectroscopic voxel was estimated by fitting the data acquiredwithout water suppression across the 17 TEs to a bi-exponential model, as described in Mayer et al. (2007).Apodization of the water-suppressed data involved multipli-cation with sine-bell functions in both time dimensions andzero filling up to 4096×1024 data points. A t1-dependent shiftwas applied in t2 to correct for the different start of dataacquisition. After performing a two-dimensional fast Fouriertransform, effectively decoupled one-dimensional CT-PRESSspectra were obtained by integrating the signal along f2 withina ±13 Hz interval around the spectral diagonal. Metabolitesignals in the one-dimensional spectra were evaluated bypeak integration with an interval of ±6 Hz. The quality of thespectra allowed evaluation of signals from N-acteylaspartate(NAA, 2.02 ppm), total creatine (tCr, 3.03 ppm and 3.93 ppm),choline-containing compounds (Cho, 3.24 ppm), glutamate(Glu, 2.36 ppm), the combined resonances of glutamate andglutamine (Glx, 3.78 ppm), and lactate (1.31 ppm) (cf.,Govindaraju et al., 2000).

To investigate spatial consistency of voxel prescriptions(i.e., their reproducibility), the voxel ROIs were generated asbinary masks for each acquisition, i.e., for each animal andeach scan time. For each animal, the ROIs from the follow-upscans were reformatted into the baseline scan coordinates ofthe same animal via the rigid coordinate transformationspreviously determined by image-to-image registration. A Diceoverlap measure (Dice, 1945) was then computed between thebaseline and each of the reformatted follow-up ROIs, again foreach animal separately. The resulting overlap scores were asfollows: for the hippocampal ROIs, .868±.011; for the cerebellarROIs, .883±.008. To put these values into perspective, we referto previous results (Rohlfing et al., 2004) that quantified thedependence of the Dice measure on the shape and size of theregion of interest. In short, the Dice measure betweendiscretely sampled regions is more sensitive for smallerstructures and for structures with larger surface areas relativeto their volumes. In particular, given the surface-to-volumeratio of an ROI, one can compute the Dice overlap betweenthat ROI and the same ROI after eroding it by one pixel. Theresult represents the Dice overlap that corresponds to under-segmentation of the ROI by one pixel everywhere along itssurface. For the ROIs under consideration here, the meansurface-to-volume ratio was .301 for the hippocampal ROIsand .277 for the cerebellumROIs. For a surface-to-volume ratioof .3, a previously derived formula (Rohlfing et al., 2004) formissing one pixel along the entire surface of the ROI yields aDice overlap of .82. The reproducibility of our ROIs is thussubstantially better.

5.4. Immunohistochemistry

After MR scanning session 4, 50 days following statusepilepticus induction, all animals were anesthetized andperfused transcardially with 4% paraformaldehyde (PF) solu-tion. Brains were removed, postfixed in 4% PF (4°C overnight),

and then transferred for storage and shipment into a 30%sucrose phosphate buffer solution (PBS). The histologist (E.L.F.C.) sectioned brains into serial coronal 40 μmsections using anAmerican Optical freezing microtome. Sections were saved instrict anatomical order, collected in a 1/8 series, and stored in.1% NaN3 in .1 M PBS at 4 °C until processing for immunohis-tochemistry (IHC).

Every 7th section of the hippocampus, Bregma −1.8 mm to−6.3 mm (Paxinos andWatson, 1997), was mounted on a glassslide (SuperFrost Plus: Fisher Scientific, Hampton, NH) andallowed to air-dry overnight. Prior to antibody incubation,sections were pretreated as follows: antigen unmasking (.01 Mcitric acid, pH 6.0, 95°C, 10 min), membrane permeabilization(.1% trypsin in .1 M Tris and .1% CaCl2, 10 min), and DNAdenaturing (2N HCl in .1 M PBS, 30 min). Sections were thenincubated in 1% hydrogen peroxide in .1 M PBS (20 min) toremove any endogenous peroxide activity. Non-specific bind-ing was blocked by incubating sections in 5% normal horseserum (NHS) and .5% Triton X-100 in PBS for 60 min. Sectionswere incubated in rabbit polyclonal cleaved caspase-3 (Asp175) antibody (Cell Signaling Technologies, Danvers, MA)diluted 1:500 in 5% NHS, .5% Tween 20 in PBS (24 h), andthen in undiluted Rabbit ImmPRESS (Vector Labs, Burlingame,CA; 60 min). DABIMPACT (Vector Labs) was used to visualizethe staining reaction. Sections were counterstained with FastRed (Vector Labs), dehydrated through a graded series ofethanols, and coverslipped with Permount. Omission of theprimary antibody resulted in a lack of specific staining.

Cell counts were performed in the neuronal cell layers ofthe hippocampus using 200× (20× objective, .40 NA) and 400×magnifications (40× objective, .65 NA) under bright fieldillumination on a Nikon SBR-KT microscope. All caspase-3positive pyramidal cells in CA1, CA3, and the granule cell layerof the dentate gyrus were counted in the right hemisphere ofevery 7th hippocampal section throughout the anterior toposterior extent of the hippocampal formation.

5.5. Statistical analysis

Group differences were analyzed using repeated-measuresANOVA and confirmed, in certain instances, by the Kruskal–Wallis (KW) one-way ANOVA by ranks. Only group effects andtheir interactions were of interest to this analysis. Follow-upbetween group differences were conducted with t-tests (1-tailed for predicted directions).

Acknowledgments

This work was supported by the National Institute on AlcoholAbuse and Alcoholism (AA005965, AA013521-INIA).

R E F E R E N C E S

Adalsteinsson, E., Sullivan, E.V., Pfefferbaum, A., 2002.Biochemical, functional and microstructural magnetic reso-nance imaging (MRI). In: Liu, Y., Lovinger, D.M. (Eds.), Methodsin Alcohol-Related Neuroscience Research. Vol. CRC Press,Boca Raton, FL, pp. 345–372.

77B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

Adalsteinsson, E., Hurd, R.E., Mayer, D., Sailasuta, N., Sullivan, E.V.,Pfefferbaum, A., 2004. In vivo 2D J-resolved magneticresonance spectroscopy of rat brain with a 3-T clinical humanscanner. NeuroImage 22, 381–386.

Alvestad, S., Hammer, J., Eyjolfsson, E., Qu, H., Ottersen, O.P.,Sonnewald, U., 2008. Limbic structures show alteredglial–neuronal metabolism in the chronic phase of kainateinduced epilepsy. Neurochem. Res. 33, 257–266.

Bates, J.F., Peake, L., Swearengen, E.S., Hall, T.W., Standish, L.J.,1988. CL 218-872 pretreatment and intervention blockkainate-induced convulsions and neuropathology. Behav.Neurosci. 102, 84–92.

Ben-Ari, Y., Tremblay, E., Riche, D., Ghilini, G., Naquet, R., 1981.Electrographic, clinical and pathological alterations followingsystemic administration of kainic acid, bicuculline orpentetrazole: metabolic mapping using the deoxyglucosemethod with special reference to the pathology of epilepsy.Neuroscience 6, 1361–1391.

Bouilleret, V., Nehlig, A., Marescaux, C., Namer, I.J., 2000. Magneticresonance imaging follow-up of progressive hippocampalchanges in a mouse model of mesial temporal lobe epilepsy.Epilepsia 41, 642–650.

Camara, O., Schnabel, J.A., Ridgway, G.R., Crum, W.R., Douiri, A.,Scahill, R.I., Hill, D.L., Fox, N.C., 2008. Accuracy assessment ofglobal and local atrophy measurement techniques withrealistic simulated longitudinal Alzheimer's disease images.NeuroImage 42, 696–709.

Carlson, H., Ronne-Engstrom, E., Ungerstedt, U., Hillered, L., 1992.Seizure related elevations of extracellular amino acids inhuman focal epilepsy. Neurosci Lett 140, 30–32.

Chronik, B., Alejski, A., Rutt, B.K., 2000a. Design and fabrication of athree-axis multilayer gradient coil for magnetic resonancemicroscopy of mice. Magma 10, 131–146.

Chronik, B., Alejski, A., Rutt, B.K., 2000b. Design and fabrication ofa three-axis multilayer gradient coil for magnetic resonancemicroscopy of mice. Magma 10, 131–146.

Dice, L., 1945. Measures of the amount of ecologic associationbetween species. Ecology 26, 297–302.

Dreher, W., Leibfritz, D., 1999. Detection of homonucleardecoupled in vivo proton NMR spectra using constant timechemical shift encoding: CT-PRESS. Magn. Reson. Imaging 17,141–150.

During, M.J., Spencer, D.D., 1993. Extracellular hippocampalglutamate and spontaneous seizure in the conscious humanbrain. Lancet 341, 1607–1610.

Ebisu, T., Rooney, W.D., Graham, S.H., Weiner, M.W., Maudsley,A.A., 1994. N-Acetylaspartate as an in vivo marker of neuronalviability on kainate-induced status epilepticus: 1H magneticresonance spectroscopic imaging. J. Cereb. Blood Flow Metab.14, 373–382.

Ebisu, T., Rooney, W.D., Graham, S.H., Mancuso, A., Weiner, M.W.,Maudsley, A.A., 1996. MR spectroscopic imaging anddiffusion-weighted MRI for early detection of kainate-induced status epilepticus in the rat. Magn. Reson. Med. 36,821–828.

Faherty, C.J., Xanthoudakis, S., Smeyne, R.J., 1999. Caspase-3-dependent neuronal death in the hippocampus followingkainic acid treatment. Brain Res. Mol. Brain Res. 70, 159–163.

Ferkany, J.W., Zaczek, R., Coyle, J.T., 1982. Kainic acid stimulatesexcitatory amino acid neurotransmitter release at presynapticreceptors. Nature 298, 757–759.

Govindaraju, V., Young, K., Maudsley, A.A., 2000. Proton NMRchemical shifts and coupling constants for brain metabolites.NMR Biomed. 13, 129–153.

Hantraye, P., Leroy-Willig, A., Denys, A., Riche, D., Isacson, O.,Maziere, M., Syrota, A., 1992. Magnetic resonance imaging tomonitor pathology of caudate–putamen after excitotoxin-induced neuronal loss in the nonhuman primate brain. Exp.Neurol. 118, 18–23.

Hellier, J.L., Dudek, F.E., 2006. Chemoconvulsant model of chronicspontaneous seizures. In: Skolnick, P. (Ed.), Short Protocols inNeuroscience: Systems and Behavioral Method Vol. JohnWileyand Sons Ltd.

Hellier, J.L., Patrylo, P.R., Buckmaster, P.S., Dudek, F.E., 1998.Recurrent spontaneousmotor seizures after repeated low-dosesystemic treatment with kainate: assessment of a rat model oftemporal lobe epilepsy. Epilepsy Res. 31, 73–84.

Kondratyev, A., Gale, K., 2000. Intracerebral injection of caspase-3inhibitor prevents neuronal apoptosis after kainic acid-evokedstatus epilepticus. Brain Res. Mol. Brain Res. 75, 216–224.

Kondratyev, A., Gale, K., 2004. Latency to onset of statusepilepticus determines molecular mechanisms of seizure-induced cell death. Brain Res. Mol. Brain Res. 121, 86–94.

Koyama, I., 1972. Amino acids in the cobalt-induced epileptogenicand nonepileptogenic cat's cortex. Can. J. Physiol. Pharmacol.50, 740–752.

Kramer, R.E., Luders, H., Lesser, R.P., Weinstein, M.R., Dinner, D.S.,Morris, H.H., Wyllie, E., 1987. Transient focal abnormalities ofneuroimaging studies during focal status epilepticus. Epilepsia28, 528–532.

Lassmann, H., Petsche, U., Kitz, K., Baran, H., Sperk, G.,Seitelberger, F., Hornykiewicz, O., 1984. The role of brain edemain epileptic brain damage induced by systemic kainic acidinjection. Neuroscience 13, 691–704.

Liu, Z., Stafstrom, C.E., Sarkisian, M.R., Yang, Y., Hori, A., Tandon,P., Holmes, G.L., 1997. Seizure-induced glutamate release inmature and immature animals: an in vivo microdialysis study.NeuroReport 8, 2019–2023.

Lothman, E.W., Collins, R.C., 1981. Kainic acid induced limbicseizures: metabolic, behavioral, electroencephalographic andneuropathological correlates. Brain Res. 218, 299–318.

Luna-Medina, R., Cortes-Canteli, M., Sanchez-Galiano, S.,Morales-Garcia, J.A., Martinez, A., Santos, A., Perez-Castillo, A.,2007. NP031112, a thiadiazolidinone compound, preventsinflammation and neurodegeneration under excitotoxicconditions: potential therapeutic role in brain disorders.J. Neurosci. 27, 5766–5776.

Magloczky, Z., Freund, T.F., 1993. Selective neuronal death in thecontralateral hippocampus following unilateral kainateinjections into the CA3 subfield. Neuroscience 56, 317–335.

Mao, H., Toufexis, D., Wang, X., Lacreuse, A., Wu, S., 2007. Changesofmetabolite profile in kainic acid induced hippocampal injuryin rats measured by HRMAS NMR. Exp. Brain Res. 183, 477–485.

Mayer, D., Spielman, D.M., 2005. Detection of glutamate in thehuman brain at 3 T using optimized constant time pointresolved spectroscopy. Magn. Reson. Med. 54, 439–442.

Mayer, D., Zahr, N.M., Sullivan, E.V., Pfefferbaum, A., 2007. In vivometabolite differences between the basal ganglia andcerebellum of the rat brain detected with proton MRS at 3T.Psychiatry Res 154, 267–273.

Meldrum, B.S., 1993. Excitotoxicity and selective neuronal loss inepilepsy. Brain Pathol. 3, 405–412.

Meric, P., Barrere, B., Peres, M., Gillet, B., Berenger, G., Beloeil, J.C.,Seylaz, J., 1994. Effects of kainate-induced seizures on cerebralmetabolism: a combined H-1 and P-31 NMR study in rat. BrainRes. 638, 53–60.

Moffett, J.R., Ross, B., Arun, P., Madhavarao, C.N., Namboodiri,A.M., 2007.N-Acetylaspartate in theCNS: fromneurodiagnosticsto neurobiology. Prog. Neurobiol. 81, 89–131.

Nadler, J.V., Cuthbertson, G.J., 1980. Kainic acid neurotoxicitytoward hippocampal formation: dependence on specificexcitatory pathways. Brain Res. 195, 47–56.

Najm, I.M., Wang, Y., Hong, S.C., Luders, H.O., Ng, T.C., Comair,Y.G., 1997. Temporal changes in proton MRS metabolitesafter kainic acid-induced seizures in rat brain. Epilepsia 38, 87–94.

Najm, I.M., Wang, Y., Shedid, D., Luders, H.O., Ng, T.C., Comair,Y.G., 1998. MRS metabolic markers of seizures andseizure-induced neuronal damage. Epilepsia 39, 244–250.

78 B R A I N R E S E A R C H 1 3 0 0 ( 2 0 0 9 ) 6 5 – 7 8

Nixon, K., 2006. Alcohol and adult neurogenesis: roles inneurodegeneration and recovery in chronic alcoholism.Hippocampus 16, 287–295.

Norman, A.B., Thomas, S.R., Pratt, R.G., Samaratunga, R.C.,Sanberg, P.R., 1990. T1 and T2 weighted magnetic resonanceimaging of excitotoxin lesions and neural transplants in ratbrain in vivo. Exp. Neurol. 109, 164–170.

O'Shaughnessy, D., Gerber, G.J., 1986. Damage induced by systemickainic acid in rats is dependent upon seizure activity—abehavioral andmorphological study. Neurotoxicology 7, 187–202.

Olney, J.W., Collins, R.C., Sloviter, R.S., 1986. Excitotoxicmechanisms of epileptic brain damage. Adv. Neurol. 44,857–877.

Paxinos, G., Watson, C., 2005. The Rat Brain in StereotaxicCoordinates, Vol. Elsevier Academic Press, London.

Peeling, J., Sutherland, G., 1993. 1H magnetic resonancespectroscopy of extracts of human epileptic neocortex andhippocampus. Neurology 43, 589–594.

Petroff, O.A., Errante, L.D., Rothman, D.L., Kim, J.H., Spencer, D.D.,2002. Glutamate–glutamine cycling in the epileptic humanhippocampus. Epilepsia 43, 703–710.

Pfefferbaum, A., Adalsteinsson, E., Sullivan, E.V., 2004a. In vivostructural imaging of the rat brain with a 3-T clinical humanscanner. J. Magn. Reson. Imaging 20, 779–785.

Pfefferbaum, A., Sullivan, E.V., Carmelli, D., 2004b. Morphologicalchanges in aging brain structures are differentially affected bytime-linked environmental influences despite strong geneticstability. Neurobiol. Aging 25, 175–183.

Pfefferbaum, A., Zahr, N.M., Mayer, D., Vinco, S., Orduna, J.,Rohlfing, T., Sullivan, E.V., 2008. Ventricular expansion inwild-typeWistar rats after alcohol exposure by vapor chamber.Alcoholism, Clin. Exp. Res. 32, 1459–1467.

Plock, N., Kloft, C., 2005. Microdialysis—theoretical backgroundand recent implementation in applied life-sciences. Eur. J.Pharm. Sci. 25, 1–24.

Puig, B., Ferrer, I., 2002. Caspase-3-associated apoptotic cell deathin excitotoxic necrosis of the entorhinal cortex followingintraperitoneal injection of kainic acid in the rat. Neurosci. Lett.321, 182–186.

Racine, R.J., 1972. Modification of seizure activity by electricalstimulation. II. Motor seizure. Electroencephalogr. Clin.Neurophysiol. 32, 281–294.

Retz, K.C., Coyle, J.T., 1982. Effects of kainic acid on high-energymetabolites in the mouse striatum. J. Neurochem. 38,196–203.

Righini, A., Pierpaoli, C., Alger, J.R., Di Chiro, G., 1994. Brainparenchyma apparent diffusion coefficient alterationsassociated with experimental complex partial statusepilepticus. Magn. Reson. Imaging 12, 865–871.

Rohlfing, T., Maurer Jr., C.R., 2003. Nonrigid image registration inshared-memory multiprocessor environments withapplication to brains, breasts, and bees. IEEE Trans. Inf.Technol. Biomed. 7, 16–25.

Rohlfing, T., Brandt, R., Menzel, R., Maurer Jr, C.R., 2004. Evaluationof atlas selection strategies for atlas-based image segmenta-tion with application to confocal microscopy images of beebrains. NeuroImage 21, 1428–1442.

Rohlfing, T., Sullivan, E.V., Pfefferbaum, A., 2006. Deformation-based brain morphometry to track the course of alcoholism:differences between intra-subject and inter-subject analysis.Psychiatry Res.: NeuroImaging 146, 157–170.

Rohlfing, T., Zahr, N.M., Sullivan, E.V., Pfefferbaum, A., 2008. TheSRI24multi-channel brain atlas: construction and applications.Medical Imaging 2008: Image Processing. Proc. SPIE 6914, EID691409 (12 pages).

Roland, J.J., Savage, L.M., 2007. Blunted hippocampal, but notstriatal, acetylcholine efflux parallels learning impairment indiencephalic-lesioned rats. Neurobiol. Learn. Mem. 87, 123–132.

Sagar, H.J., Oxbury, J.M., 1987. Hippocampal neuron loss intemporal lobe epilepsy: correlation with early childhoodconvulsions. Ann. Neurol. 22, 334–340.

Schwob, J.E., Fuller, T., Price, J.L., Olney, J.W., 1980. Widespreadpatterns of neuronal damage following systemic orintracerebral injections of kainic acid: a histological study.Neuroscience 5, 991–1014.

Sierra-Paredes, G., Galan-Valiente, J., Vazquez-Illanes, M.D.,Aguilar-Veiga, E., Soto-Otero, R., Mendez-Alvarez, E.,Sierra-Marcuno, G., 1998. Extracellular amino acids in therat hippocampus during picrotoxin threshold seizures inchronic microdialysis experiments. Neurosci. Lett. 248,53–56.

Sperk, G., 1994. Kainic acid seizures in the rat. Prog. Neurobiol. 42,1–32.

Sperk, G., Lassmann, H., Baran, H., Kish, S.J., Seitelberger, F.,Hornykiewicz, O., 1983. Kainic acid induced seizures:neurochemical and histopathological changes. Neuroscience10, 1301–1315.

Stone, J.L., Hughes, J.R., Barr, A., Tan, W., Russell, E., Crowell, R.M.,1986. Neuroradiological and electroencephalographic featuresin a case of temporal lobe status epilepticus. Neurosurgery 18,212–216.

Tanaka, S., Tanaka, T., Kondo, S., Hori, T., Fukuda, H., Yonemasu,Y., Tanaka, M., Shindo, K., 1993. Magnetic resonance imaging inkainic acid-induced limbic seizure status in cats. Neurol. Med.Chir. (Tokyo) 33, 285–289.

Tanaka, K., Graham, S.H., Simon, R.P., 1996. The role of excitatoryneurotransmitters in seizure-induced neuronal injury in rats.Brain Res. 737, 59–63.

Tokuhara, D., Sakuma, S., Hattori, H., Matsuoka, O., Yamano, T.,2007. Kainic acid dose affects delayed cell death mechanismafter status epilepticus. Brain Dev. 29, 2–8.

Tokumitsu, T., Mancuso, A., Weinstein, P.R., Weiner, M.W.,Naruse, S., Maudsley, A.A., 1997. Metabolic and pathologicaleffects of temporal lobe epilepsy in rat brain detected by protonspectroscopy and imaging. Brain Res. 744, 57–67.

Wade, J.V., Samson, F.E., Nelson, S.R., Pazdernik, T.L., 1987.Changes in extracellular amino acids during soman- and kainicacid-induced seizures. J. Neurochem. 49, 645–650.

Williams, P.A., Hellier, J.L., White, A.M., Staley, K.J., Dudek, F.E.,2007. Development of spontaneous seizures after experimentalstatus epilepticus: implications for understandingepileptogenesis. Epilepsia 48 (Suppl 5), 157–163.

Wilson, C.L., Maidment, N.T., Shomer, M.H., Behnke, E.J.,Ackerson, L., Fried, I., Engel Jr, J., 1996. Comparison of seizurerelated amino acid release in human epileptic hippocampusversus a chronic, kainate rat model of hippocampal epilepsy.Epilepsy Res. 26, 245–254.

Wolf, O.T., Dyakin, V., Patel, A., Vadasz, C., de Leon, M.J., McEwen,B.S., Bulloch, K., 2002. Volumetric structural magneticresonance imaging (MRI) of the rat hippocampus followingkainic acid (KA) treatment. Brain Res. 934, 87–96.

Zahr, N.M., Mayer, D., Hsu, O., Vinco, S., Orduna, J., Sullivan, E.V.,Pfefferbaum, A., 2009. In vivo neurochemical evidence foralcohol binge induced neural compromise and recovery inalcohol-preferring rats. Research Society on Alcoholism.San Diego, CA.

Zhang, W.Q., Hudson, P.M., Sobotka, T.J., Hong, J.S., Tilson, H.A.,1991. Extracellular concentrations of amino acid transmittersin ventral hippocampus during and after the development ofkindling. Brain Res. 540, 315–318.