Ž .Brain Research 816 1999 337–349

Research report

Behavioural analysis and susceptibility to CNS injury of four inbred strainsof mice

Stephen J. Royle, Fiona C. Collins, H. Tom Rupniak, Julie C. Barnes, Richard Anderson )

Neuroscience Unit, Glaxo Wellcome Research and DeÕelopment, Medicines Research Centre, Gunnels Wood Road, SteÕenage, Herts, SG1 2NY, UK

Accepted 20 October 1998

Abstract

Interpretation of data from gene targeting studies can be confounded by the inherent traits of the background inbred strains used in thegeneration of transgenic and null mutant mice. We have therefore compared the behaviour and response to CNS injury of four inbredstrains commonly used in molecular genetic studies to produce models of neurological disease. Adult, male 129rOla, BALBrc,

Ž .C57BLr6 and FVBrN mice 2–4 months were initially subjected to behavioural tests that comprised a neurological examination,determination of motor function and cognitive testing in the Morris water maze. Also the response to CNS injury following an acute

Ž . Ž y1 .kainic acid KA challenge 30 mg kg , i.p. was determined. The 129rOla and BALBrc strains showed significant motor deficitswhen compared with the C57BLr6 and FVBrN strains. In contrast, only the FVBrN strain showed evidence of apparent cognitiveimpairments in the water maze as evidenced by increased pathlengths to locate the escape platforms and impaired performance in a probe

Ž .trial. In addition, the FVBrN strain showed the most severe seizure response and mortality rate 62% following administration of KAŽ y1 .30 mg kg , i.p. . These behavioural changes were also associated with a greater degree of cell body and synaptophysin loss in thepyramidal CA3 hippocampal cell layer and astrogliosis 72-h post-dose. These data suggest that the FVBrN strain may not be the mostsuitable background strain for the development of new transgenic mice for the study of genes implicated in the learning and memoryprocess. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: Mouse; Inbred strain; Cognition; Kainic acid; Excitotoxicity

1. Introduction

Genetic targeting technologies provide a powerful meanswith which to study genes associated with disease pro-

w xcesses 1,2 . For example, these techniques have beenemployed to create transgenic and null mutant mice inorder to facilitate study of the roles of amyloid precursor

w x w xprotein 16,44 presenilin-1 20 and apolipoprotein Ew x36,51 in the pathogenesis of neurodegenerative disease.In addition, transgenic mice expressing mutant forms of

Ž .the CurZn superoxide dismutase SOD-1 gene have beenused to provide an animal model of peripheral motorneurone degeneration characteristic of amyotrophic lateral

Ž . w xsclerosis ALS 11,38 . The majority of gene targeting isŽ .undertaken in embryonic stem ES cells derived from the

129 strain and the modified cells injected into a distinctstrain of host blastocyst. To aid the identification of ES-de-rived offspring, the resulting chimeras are subsequently

) Corresponding author. Fax: q44-01438-764898; E-mail:ra28276@glaxowellcome.co.uk

backcrossed to an inbred strain carrying an alternative coatcolour marker. Problems can arise when the offspring from

Ž .such matings F1 generation are intercrossed, since theresultant F2 mice have recombinant genotypes derived

w xfrom the two parental mouse strains 17 and several recentreports have discussed problems associated with this tech-

w xnique 10,17,18 . The main difficulty arises in the interpre-tation of many transgenic phenotypes due to backgroundgenetic polymorphism in the inbred strains used to gener-

w xate the mutants 10,17 . For example, behavioural deficitsw x7 and susceptibility to neurodegeneration following both

Ž . w xa kainic acid KA challenge 32 and middle cerebralw xartery occlusion 8,15 have been reported in transgenic

mice but a contribution from background parental genesmay, in part, explain these phenotypes. Strategies adoptedto facilitate definition of transgene effects over geneticbackground include the use of wild type littermate controlsw x52 , multiple backcrossings into a preferred host strain andmore recently, in order to accelerate this process, thedevelopment of marker assisted ‘speed congenic’ breeding

w xmethods 3,27 .

0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0006-8993 98 01122-6

( )S.J. Royle et al.rBrain Research 816 1999 337–349338

Despite these measures, if inappropriate host strains areselected, a transgenic model may be of limited value sincemutations can have markedly different phenotypes in dif-

w xferent genetic backgrounds 41,47 . For example, overex-pression of APP on an inbred FVBrN or C57BLr6background proves lethal whereas the mutation is welltolerated on outbred lines which facilitate the development

w xof amyloid plaques 7 . The development of new trans-genic mouse models of CNS disorders will therefore criti-cally depend on the correct selection of the most suitablebackground parental strains and an informed choice willrequire a thorough understanding of the endogenous phe-notypes of the inbred strains commonly used in molecular

w xgenetics 10 . Accordingly, we have characterised the phe-notype and response to CNS injury of four common

Žlaboratory mouse strains 129rOla, BALBrc, C57BLr6.and FVBrN in order to investigate the impact of inbred

parental genotypes on behaviour and response to CNSinjury.

2. Materials and methods

2.1. Animals

Ž .Test groups comprised of male 129rOla Harlan, UKŽBALBrc, C57BLr6 and FVBrN mice all originally

.from Charles River, USA at 2–4 months of age. Thestrains were incorporated into our breeding facility in theUK and bred in specified pathogen-free conditions. Fol-lowing transfer of the experimental groups to Stevenage,the mice were allowed 2 weeks to acclimatise prior to thestudy. All animals were housed individually in a tempera-

Ž .ture-controlled environment 21 " 28C under a 12-hlightrdark cycle and allowed free access to standard labo-ratory food and water. The research complied with nationallegislation and with company policy on the Care and Useof animals.

2.2. BehaÕiour

2.2.1. Neurological assessmentsAll mice were subjected to a battery of tests to probe

Ž . Ž . Žspinal cord flexion , medulla corneal reflex , pons right-. Ž .ing reflex , and autonomic salivation function as de-

scribed in Table 1. Motor coordination was assessed on aŽ .mouse rota-rod Hugo-Basile and forelimb grip strength

determined with a grid floor and pressure transducer.

2.2.2. Locomotor actiÕityMice were individually placed in activity monitor cages

Ž .35 cm long, 21 cm wide and 18 cm deep and locomotoractivity was recorded for 1 h. There was no prior habitua-tion. Locomotor activity was quantified by measuring thenumber of beam breaks.

Table 1Neurological examination methods

Title Description

Flexion reflex Animal lifted by nape and toes lightly pinchedwith forceps. A response is registered by a

Ž .rapid foot withdrawal presentrabsent .Righting reflex Animal placed on back will resume normal

position. In addition, the animal is droppedupside down from 30 cm above a soft surface.

ŽThe animal should land on all four legs pre-.sentrabsent .

Corneal reflex The cornea of the restrained animal is lightlytouched with a hair. The animal should respond

Ž .with a blink reflex presentrabsent .Secretary signs The animals were examined for secretary signs

Ž . Žaround the mouth salivation and eyes lacri-. Ž .mation , presentrabsent .

In addition, measures of body weight and body temperature were taken.

2.2.3. Morris water mazeThe water maze consisted of a white Perspex circularŽ .pool 100 cm diameter, 30 cm deep . The pool was filled

Ž .with water 21"18C to a depth of 25 cm and renderedŽopaque by the addition of a liquid latex solution Warner

.Jenkinson . The cued escape platform was painted black,flagged and protruded 1 cm above the water surface. The

Ž .hidden escape platform 6.5 cm diameter was paintedwhite and positioned just below the water surface. Thecued task trains the mice to associate the platforms withescape from the pool and was therefore conducted prior tothe hidden platform task. The maze was isolated withscreens and distal cues placed prominently around thepool.

ŽAll mice were initially trained over four sessions twoper day, three trials per session with a 10-min intertrial

Ž .. Ž .interval ITI using the cued escape platform cued task .In this case both the starting position of the mouse and thecued platform position were both varied to avoid habitua-tion to a particular quadrant. Each trial was commenced byplacing the mouse gently in the pool facing the perimeterand activating an overhead camera connected to a comput-

Ž .erised tracking system HVS Image; Hampton, UK thatrecorded the swim paths of the mice to locate the platform.If a mouse failed to locate the platform within the maxi-mum trial period of 60 s it was led to and placed on theplatform for 10 s.

Following the cued task, the mice received eight ses-Ž .sions three trials per session, 10-min ITI over 4 consecu-

tive days with the hidden platform, the position of whichŽ .was fixed for each individual mouse hidden task . After

the eighth hidden session the mice underwent a probe trial,which consisted of a 60-s trial without the platform pre-sent, during which the search pattern was analysed. Thetime spent searching the quadrant that formerly containedthe hidden platform was recorded as a percentage andcompared with that spent searching in the other quadrants.

( )S.J. Royle et al.rBrain Research 816 1999 337–349 339

Table 2Neurological assessment of four inbred mouse strains

Inbred n Mean Mean Righting Corneal Salivation Grip Flexion Rota-rodŽ . Ž . Ž .strain weight g temperature 8C reflex reflex strength g

Ž . Ž . Ž . Ž .129rOla 12 32.3"0.5 36.4"0.3 100% 12r12 100% 12r12 0% 0r12 77.6"3.7 58.3% 7r12) 57.8"16.3)

Ž . Ž . Ž . Ž .BALBrc 12 30.9"0.7) 37.4"0.2 100% 12r12 100% 12r12 8% 1r12 73.4"5.2 50% 6r12) 47.3"4.5)

Ž . Ž . Ž . Ž .C57BLr6 12 33.0"1 36.5"0.3 100% 12r12 100% 12r12 0% 0r12 73.1"3.8 100% 12r12 96.5"10.9Ž . Ž . Ž . Ž .FVBrN 12 34.5"1 37.5"0.2 100% 12r12 100% 12r12 0% 0r12 62.9"3.6 100% 12r12 89.3"16.3

) p-0.05 vs. C57BLr6 and FVBrN strains.

2.3. Drug

Ž .Kainic acid KA, Sigma, Lot 94H0230 was dissolvedŽ . Ž .in 0.9% wrv phosphate buffered saline PBS . The dose

Ž y1 .30 mg kg is expressed in terms of free base and wasŽ .administered as an intraperitoneal i.p. injection in a dose

volume of 10 ml kgy1. Vehicle-treated mice were injectedwith PBS.

2.4. BehaÕioural response to KA

Following administration of KA, seizure severity wasŽquantified using a 5-point observer rating scale 0snormal

Ž .behaviour, 1s rare wet dog shakes WDS and rare type Aconvulsions; 2s frequent WDS and type A convulsionsŽ .no rearingrsalivation ; 3s frequent type A convulsionswith appearance of type B convulsions with rearingrsali-vation; 4s frequent type B convulsions with falling andsalivation; 5scontinuous, generalised limbic seizures and

.death within 3 h . Type A convulsions comprised focalconvulsions affecting the head and extremities, typically

starting within 30 min of KA administration. Type Bcomprised generalised tonic-clonic seizures usually start-ing between 60–90 min post-KA administration. In addi-tion, mean latency to onset of seizures, total seizure dura-tion and mortality were recorded for each treatment group.Mice were returned to their cages after 6 h or when seizureactivity had ceased.

2.5. Histology and immunohistochemistry

Seventy-two hours following administration of KA, themice were sacrificed by cervical dislocation and the brainhemisected. The left hemisphere was fixed in 2% para-formaldehyde for 18 h, paraffin embedded and serial 8 mmcoronal sections obtained at the level of the anterior hip-pocampus for histochemical analysis. In order to studygross hippocampal damage sections were stained withCresyl violet for 10 min. For glial fibrillary acidic protein,Ž .GFAP , staining slides were dewaxed in xylene prior to a30 min incubation with 3% H O in PBS. Following a 102 2

min wash in PBS, the sections were incubated for 1 h with

Ž . Ž . Ž . Ž . Ž .Fig. 1. Time course of locomotor activity LMA in 129rOla v , BALBrc I , C57BLr6 B and FVBrN ` mice. Data points are expressed as themean"S.E.M. of 12 mice per strain. Data were analysed by two-way ANOVA with strain as between-subjects factor and time as within-subjects factor.Post-hoc comparisons were made by Dunnett’s test assuming the C57BLr6 group as control. The 129rOla and BALBrc strains exhibited reduced LMA,with the 129 strain further impaired in comparison with the BALBrc mice.

()

S.J.Royle

etal.r

Brain

Research

8161999

337–

349340

Ž . Ž . Ž . Ž . Ž . Ž . Ž .Fig. 2. a Performance of 129rOla v , BALBrc I , C57BLr6 B and FVBrN ` strains in water maze learning in both cued sessions 1–4 and hidden platform sessions 5–12 tasks. Data areŽ . Ž .expressed as mean pathlengths ns12 per group . Statistical significance was determined by two-way ANOVA followed by Dunnett’s test assuming the C57BLr6 group as control. b Performance of the

Žsame mice in a 60-s probe trial conducted 10 min following the 12th session. Each strain is represented by four histobars corresponding to each quadrant of the water maze adjacent left, island quadrant,. Ž .adjacent right and opposite . Note that the 129rOla, C57BLr6 and BALBrc strains show good evidence of spatial learning whereas the FVBrN mice fail to show a quadrant preference. c Median probe

data path plots of the four inbred strains.

( )S.J. Royle et al.rBrain Research 816 1999 337–349 341

Table 3Ž y1 .Seizure response to kainic acid 30 mg kg , i.p. of four inbred strains of mice

Ž .Inbred strain n Treatment Median peak Mean latency Total seize Mortality %Ž . Ž .seizure score to seize min time min

y1 Ž .129rOla 8 KA 30 mg kg 0.5 0–1 50"4 0.3"0.2 0y1 Ž .BALBrc 7 KA 30 mg kg 0 0–1 16"3)) 7"5 29y1 Ž .C57BLr6 8 KA 30 mg kg 4 3.8–4.3 28"2 89"44)) 12.5y1 Ž .FVBrN 8 KA 30 mg kg 4 3.8–4.3 24"3)) 97"44)) 62)

Ž .Seizure severity was quantified on a 5-point observer rating scale see text . Note the increased susceptibility to seizures observed in FVBrN strain asw Ž . x w Ž . xevidenced by a significantly reduced latency to onset F 3,28 s5.6, p-0.05 and increased duration of seizures F 3,28 s4, ps0.01 and mortality

Ž .rate ps0.04 .

5% normal horse serum before overnight application ofŽ .1:500 diluted GFAP antibody Sigma, G-9269 at 58C. The

following day sections were washed in PBS before incuba-tion for 30 min with biotinylated anti-mouse antibodyŽ .Vectastain elite ABC-peroxidase kit . Following a furtherwash, the slides were then incubated for 30 min with ABCreagent before a final wash and a 4.5 min incubation with

Ž .DAB peroxidase substrate kit SK-4100, Vector labs .In order to determine synapse loss, sections were de-

waxed in xylene prior to overnight incubation with undi-Žluted primary synaptophysin antibody MCA860, Serotec,

.UK at 58C. The following day slides were washed in PBSbefore incubation for 30 min with biotinylated anti-mouse

Ž .antibody Vectastain elite ABC-mouse kit . Following afurther wash, the slides were then incubated for 30 min

Ž .with streptavidin-fluorescein 2% in PBS, Vector labsbefore a final wash and mounting with vectorshield mount-

Ž .ing medium Vector labs . All sections stained with eitherthe GFAP or synaptophysin antibody were processed at thesame time and in a random fashion.

2.6. Image analysis

Gross hippocampal cell damage was assessed by lightmicroscopy using a Leica Q600 Image Analyser and DM

Ž .microscope Leica UK, Milton Keynes and quantified byŽcounting viable neurones, within a counting frame 1500

2 .pp , where both nucleus and cell membrane were intact.For GFAP staining, background levels were determinedfrom negative controls and all sections normalised. GFAPimmunoreactivity was quantified using an image analysisprogram. Positively stained structures were counted within

Ž 2 .a 0.21-mm=0.17-mm field 0.036 mm .Images of synaptophysin stained sections were taken

Žwith a Leica TCS4D confocal microscope Leica.Lasertechnic, Heidelberg, Germany with 488 nm fluores-

cence excitation and 530 nm emission. A 10= objectivelens was used and the height of the microscope stage wasadjusted to give the brightest image from each section, andthe gain of the microscope was set for the section with thebrightest fluorescence. Images were stored on optical diskŽ .Maxell, MA-M128 and transferred to the image analyserfor measurement.

Analysis of confocal images was made with a LeicaŽ .Q600MC image analyser Leica UK, Cambridge, UK .

The analysis was automated using a QUIPS macro thatperforms a sequence of functions as follows: For eachimage, the image analyser was first used to threshold thefluorescent staining from the background, and then todefine sample areas within the hippocampal subfields.Binary image arithmetic was used to display the threeareas in colour overlying the grey image of fluorescence,and to avoid overlapping between the areas. The imageanalyser then measured the mean grey level for each of the

Žareas grey level is proportional to intensity of fluores-.cence . Mean grey levels from negatively stained sections

were then subtracted from each result to obtain intensity offluorescence without background staining.

2.7. Lipid peroxidation

The left cortical hemisphere of each mouse was ho-Žmogenised in Ripa Lysis buffer 50 mM Tris–HCl, 150

mM NaCl, 1% deoxycholic acid, 0.1% SDS, 1% Triton.X-100 adjusted to pH 7.5 , aliquoted, and stored at y808C.

Basal lipid peroxidation was determined by measuringŽ . Ž .malondialdehyde MDA and 4-hydroxynonenal 4-HNE

Ž .concentrations using a standard kit R&D Systems .

Ž . Ž .Fig. 3. Time course of seizure severity in 129rOla v , BALBrc I ,Ž . Ž . Ž y1 .C57BLr6 B and FVBrN ` strains following KA 30 mg kg , i.p .

Ž .Data are expressed as mean"S.E.M. seizure severity scores see text .

( )S.J. Royle et al.rBrain Research 816 1999 337–349342

( )S.J. Royle et al.rBrain Research 816 1999 337–349 343

Ž . Ž y1 . Ž .Fig. 5. Normal pyramidal cell counts in sections following vehicle I or 72 h post-KA 30 mg kg B . Note the significantly decreased number ofŽ . Ž . Žcells with intact nuclei and plasma membrane in the CA1 69% and CA3 91% subfields of the FVBrN sections )) p-0.01 vs. vehicle unpaired

.t-test .

2.8. Data analysis and statistics

The four inbred strains were analysed for strain differ-Žences using either parametric one-way ANOVA for body

weight and temperature, grip strength, rota-rod and water. Žmaze probe data or non-parametric all other neurological

.data tests. The water maze acquisition and LMA studieswere analysed by two-way ANOVA with sessionrtime bin

Ž .as the within-subjects repeated measures variable andstrain as the between-subjects variable. In all cases when asignificant main effect andror significant interaction termswere found, subsequent post-hoc comparisons were madeby Dunnett’s test comparing all strains to C57BLr6. Lipidperoxidation and seizure measures were analysed by one-way ANOVA except mortality, which was determined byx 2 methods followed by Fischer’s exact test. Vehicle andkainate-treated image analysis comparisons were made byunpaired t-test.

3. Results

3.1. BehaÕiour

3.1.1. Neurological assessmentTable 2 shows a neurological assessment of the four

inbred mouse strains. No obvious differences were seenbetween strains in body temperature, righting and cornealreflexes, secretary signs or grip strength. However, theBALBrc strain showed a significantly reduced body mass

w Ž . xcompared with the other strains F 3,44 s4, ps0.02

and in addition to the 129rOla mice, impaired flexionw Ž . Ž .xreflex BALBrc ps0.01 and 129rOla ps0.04 and

reduced rota-rod latency scores in comparison withw Ž . xC57BLr6 and FVBrN mice F 3,44 s3, ps0.02 .

Total locomotor activity scores clearly revealed amarked hypoactivity in the 129rOla and BALBrc strainscompared with the C57BLr6 and FVBrN strains, i.e.,total activity scores for 129rOla, BALBrc, C57BLr6 andFVBrN mice were 426"63)), 788"109), 1221"150

w Ž .and 2191 " 681, respectively F 3,44 s 26, )) p -x. Ž .0.001 . Time course analysis Fig. 1 showed a significant

w Ž . xmain effect of strain F 3,44 s24, p-0.001 and strainw Ž . x= time interaction F 3,44 s91, p-0.001 with the

129rOla and BALBrc strains exhibiting a lower rate ofinitial exploratory activity.

3.1.2. Morris water mazeŽ .Fig. 2a sessions 1–4 shows pathlength to locate the

Žvisible platform in the cued task and Fig. 2a sessions.5–12 pathlength to locate the hidden platform. A main

w Ž .effect of strain was found in the cued F 3,44 s74,x Ž Ž . xp-0.01 and hidden F 3,44 s50, p-0.01 tasks with

the FVBrN strain showing significantly increased path-lengths to locate the escape platforms in comparison withthe other three strains. In addition, the FVBrN strainexhibited significantly elevated swim speeds at each ses-

w Ž . xsion F 3,44 s9, p-0.001 . For example, swim speedsat the 12th session were 13 " 0.6, 13 " 1, 16 " 1,26"0.4)) cm sy1 for 129rOla, BALBrc C57BLr6and FVBrN mice, respectively. Analysis of probe datafollowing the 12th session showed significant spatial learn-

Ž . y1 Ž . Ž .Fig. 4. Cresyl violet staining of the hippocampal CA3 pyramidal layer 72 hours following vehicle left panels or KA 30 mg kg , i.p. right panels in: aŽ . Ž . Ž . w x129rOla, b BALBrc, c C57BLr6, d FVBrN sections =200 . Note the densely stained, shrunken cells in the pyramidal layers of KA-treated tissues

with extensive cell loss in the FVBrN strain.

( )S.J. Royle et al.rBrain Research 816 1999 337–349344

( )S.J. Royle et al.rBrain Research 816 1999 337–349 345

Ž . Ž y1 . Ž .Fig. 7. KA-induced increases in hippocampal GFAP immunoreactivity following vehicle I or 72 h post-KA 30 mg kg , i.p. B . Histobars showŽ .mean"S.E.M. positively stained structures per area see text . Note the significantly increased number of positively stained structures in the FVBrN

Ž .sections. ) p-0.05 vs. vehicle unpaired t-test .

Žing in the 129rOla, BALBrc, and C57BLr6 mice all.p-0.001 as evidenced by a significantly greater time

spent in the former platform quadrant. In contrast theFVBrN mice failed to show a significant quadrant prefer-

Ž . w Ž . xence Fig. 2b F 3,44 s0.6, ps0.6 . Median probepath plots for each of the strains following the 12th sessionare shown in Fig. 2c.

3.2. BehaÕioural response to KA

Ž y1 .Administration of KA 30 mg kg , i.p. produced amarked strain-dependent seizure response with the follow-ing rank order: C57BLr6 s FVBrN ) BALBrc )

Ž .129rOla Table 3 . The behavioural changes initially man-ifest as reduced LMA, catatonia, head bobbing and facewashing progressing to wet dog shakes, tremors and finallyrearing, generalised seizures and in some instances death.Whilst all strains developed seizures, the C57BLr6 andFVBrN strains exhibited a greater seizure severity score

Ž .and seizure duration Fig. 3 . In addition, the FVBrNŽ .mice showed the highest mortality rate 62% .

3.3. Immunohistochemistry

Cresyl violet staining of hippocampal sections 72 hfollowing KA treatment revealed marked pyramidal cellloss in the CA3 subfield when compared with vehicle-

Ž .treated sections Fig. 4 Damaged neurones appeared

shrunken with dense pyknotic nuclei. There was markeddisruption of the cytoarchitecture and increased spacing ofthe pyramidal cell layer. Sections from the 129rOla,BALBrc and C57BLr6 mice showed a similar degree ofdamage but those from the FVBrN mice showed a morepronounced cell loss with almost complete destruction of

Ž .the CA3 cell layer Fig. 5 .

3.3.1. GFAP immunoreactiÕityVehicle-treated sections showed a low level of GFAP

immunoreactivity distributed evenly throughout the hip-pocampus. In contrast, immunostaining of KA-treated tis-sues revealed a marked astrogliosis in all strains as evi-denced by increases in the size, density and ramification of

Ž .GFAP expressing cells Fig. 6 . Reactive astrocytes werefound throughout the hippocampal area but none wereobserved in the pyramidal cell layer. Image analysis tech-niques revealed that the FVBrN strain displayed a moreintense astrogliotic reaction following KA administration

Ž .when compared with the other three strains Fig. 7 .

3.3.2. Synaptophysin immunoreactiÕityImmunocytochemical analysis revealed characteristic

punctate immunostaining of presynaptic terminals using anantibody against the presynaptic vesicle protein, synapto-physin. Quantitative image analysis of synaptophysin im-munolabelled sections failed to reveal a significant reduc-

Ž . Ž . Ž . Ž . w xFig. 6. GFAP staining of adjacent sections revealing astrogliosis in: a 129rOla, b BALBrc, c C57BLr6, d FVBrN =200 72 h followingy1 Ž .injection of KA 30 mg kg , i.p. right panels . Note the more intense GFAP labelling, thicker, shorter processes and swollen soma of astrocytes in the

Ž .KA-treated sections compared with respective vehicle-treated sections left panels .

( )S.J. Royle et al.rBrain Research 816 1999 337–349346

Fig. 8. Changes in synaptophysin immunoreactivity in the inner, middleŽ . Ž y1 .and outer molecular layers in vehicle I and KA 30 mg kg , i.p.

Ž .B treated FVBrN sections as determined by laser scanning confocalmicroscopy and subsequent image analysis. Data are presented as mean"

Ž .S.E.M. ) p-0.05 vs. vehicle unpaired t-test .

tion in the level of immunoreactivity in the inner, middleand outer molecular layers of the hippocampus in KA-treated sections from the 129rOla, BALBrc and C57BLr6

Ž .mice data not shown . In contrast, the FVBrN miceshowed a 50.4% and 49.8% loss of synaptophysin im-munoreactive presynaptic terminals in the middle and outerlayers respectively compared with vehicle-treated miceŽ .Fig. 8 . Synaptophysin levels in the dentate gyrus re-mained unchanged by KA treatment in all strains. Omis-sion of the GFAP and synaptophysin antibodies resulted in

Ž .a complete absence of signal data not shown .

3.4. Lipid peroxidation

The effect of KA on lipid peroxidation in corticalhomogenates was assessed by measuring the change in

Ž . Ž .malondialdehyde MDA and 4-hydroxynonenal 4-HNEconcentrations 72 h post-dose. The degree of lipid per-oxidation measured in vehicle-treated tissues of the129rOla, C57BLr6 and BALBrc strains did not appear

Ž .significantly elevated in the KA-treated samples Table 4 .However, MDA and 4-HNE levels in KA-treated samplesprepared from the FVBrN mice were significantly in-

Ž .creased ps0.02 by 36% in comparison with vehicle.

4. Discussion

In order to avoid over-interpretation of a mutant pheno-type a comprehensive understanding of the traits of thehost strains in which the phenotype will be studied is

w xessential 10,17 . The aim of the present study was there-fore to investigate further the endogenous characteristics offour inbred mouse strains commonly used in moleculargenetics.

A comparison of neurological function in the four in-bred strains revealed an absence of the flexion reflex inapproximately 50% of the 129rOla and BALBrc micewhich suggested impaired cerebellarrspinal cord functionw x48 . Further evidence to support motoric deficits in these

strains was obtained from analysis of LMA and rota-rodstudies, which showed clear reductions in exploratory be-haviour and coordination respectively. Low exploratorybehaviour has previously been reported in BALBrc micew x25 and inter-strain variation in the ‘emotional’ status ofthe mice may play a role in this phenotype since anxietyand stress have been shown to depress LMA in micew x9,28 . The marked hyperactivity observed in the FVBrNstrain in this study may provide a favourable backgroundupon which to produce mutations designed to mimic themotor impairments that characterise ALS, since a highlevel of activity would be desirable to define clear trans-gene-induced deficits.

Analysis of acquisition and probe data from the Morriswater maze study revealed an apparent cognitive deficit inthe FVBrN strain. However, since pathlengths to locatethe cued platform were significantly elevated in addition tothose in the hidden task, these data suggest that the deficitsin FVBrN performance may not be due to learning andmemory impairments alone. Rather, since the retinal de-

Ž .generation gene rd has been identified on chromosome 5w xof the FVBrN strain 46 deficits in visual acuity are

likely to underlie this phenotype. Perhaps surprisingly, thebehaviour of the 129rOla mice was indistinguishable fromthat of the BALBrc and C57BLr6 strains in that allshowed evidence of good spatial learning and memory.The 129 strain is widely used in genetic targeting researchas it is one of the most valuable donor strains of ES celllines but the majority of cognitive studies have charac-

w xterised the 129 strain as a poor learner 17,35,50 . How-w xever, there are many substrains within the 129 family 40

and evidence is accumulating to support behavioural het-erogeneity within the substrains. For example, the 129

ŽSvHsd substrain also known as the 129 Sv-ter differsw xfrom other 129 substrains at many genetic loci 12,40 and

this may account for the variation in water maze perfor-w xmance. Indeed, Montkowski et al. 31 recently demon-

strated heterogeneity in water maze learning in four 129substrains with the 129rOla substrain showing good evi-

Table 4Ž y1 .The effect of kainic acid 30 mg kg , i.p. on lipid peroxidation in

cortical homogenates prepared 72 h following dosing with either vehicleŽ .PBS or kainic acid

Inbred strains n Treatment MDAq4HNE concentrationy1Ž w x .nmoles mg protein

129rOla 4 Vehicle 6.2"0.5y18 KA 30 mg kg 7"0.8

BALBrc 4 Vehicle 6.7"0.4y15 KA 30 mg kg 8.5"0.9

C57BLr6 4 Vehicle 5.6"0.3y16 KA 30 mg kg 5.5"0.3

FVBrN 4 Vehicle 4.5"0.4y13 KA 30 mg kg 6.1"0.1)

Data are expressed as mean"S.E.M. MDAq4-HNE concentrationsŽ w xy1 .nmoles mg protein .

Ž .) p-0.05 vs. vehicle unpaired t-test .

( )S.J. Royle et al.rBrain Research 816 1999 337–349 347

dence of learning and memory in agreement with thepresent study.

KA reliably induces seizures in both rats and micew x w x23,42 and is widely used to model epilepsy 33 . Theseizures are associated with characteristic neurodegenera-tion in the hippocampus with the CA1–CA3 subfields

w xproving particularly vulnerable 43 . Analysis of KA-in-duced seizure activity revealed a marked strain-dependentresponse with the 129rOla mice proving relatively resis-tant to the convulsive effects of KA. In contrast, intenseand protracted seizure activity was observed in the FVBrNand C57BLr6 strains with the FVBrN mice exhibiting thehighest mortality rate. The increased vulnerability of theFVBrN strain to seizures was paralleled histologically bya greater degree of cell loss observed predominantly in theCA3 pyramidal cell layer and significant loss of synapto-physin immunoreactivity throughout the hippocampalfields. In addition, this cell damage was also associatedwith an intense astrogliosis. This suggests that the FVBrNstrain is more susceptible to neurodegeneration and im-paired neuronal function following KA induced excitotoxi-city. One of the mechanisms thought to underlie KAmediated excitotoxicity is via the generation of free radi-

w xcals 4,6 which subsequently lead to lipid peroxidationw x w x45 and neuronal membrane damage 29,30,37 . In agree-ment with the histological findings measurements of corti-cal lipid peroxidation revealed a significant increase in theFVBrN strain following KA that was not observed in theother strains. Thus the behavioural, immunohistochemicaland biochemical data all suggest that the FVBrN strain ismore susceptible to KA-induced seizures and neurodegen-eration.

It is unlikely that the inter-strain variation in response toKA is due to pharmacokinetic reasons since previousstudies have shown no evidence for differential uptake of

w xthe compound into the murine CNS 13,39 . Rather, inter-w xstrain differences in kainate receptor densities 24 andror

the presence of genes that either mediate neuroprotectionor predispose to excitotoxic damage may govern the levelof injury. The contribution of background genetic factors

Ž .was explored recently in a quantitative trait loci QTLstudy that identified as many as eight chromosomal locithat are strongly linked to the differential response of

w xinbred strains to KA-induced seizures 14 . The locuspresent on chromosome 1 accounted for 17% of the ge-netic variance between strains and candidate genes at thislocus include the a- and b-subunit of the NaqrKq-ATPase

w xpump 22 . Clearly, compromised function of this iontransporter could lead to perturbed neuronal polarisationand render the FVBrN strain more susceptible to ectopicdischarge. Indeed, pharmacological inhibition of this en-zyme potentiates excitotoxic injury of adult rat hippocam-

w xpal neurones 5 and such an effect may explain thepreviously reported appearance of spontaneous seizures in

w xthis strain 19 and the increased susceptibility to KA-in-duced seizures in the present study. In addition, it is

interesting to note that the high mortality rate observed inthis strain is consistent with poor survival of transgenic

w xmice produced on this background 7,21,26 .The present data are in conflict with those of

w xSchauwecker and Steward 39 who reported comparableseizure activity in the same strains following administra-tion of KA at the same dose used in the present study.

Ž w x.Nevertheless, Table 1 p. 4104 of Ref. 39 which sum-marised seizure type in each strain, stated that 85% ofFVBrN and only 20% of 129rSvEMS mice exhibited‘severe seizures’. The authors also reported relatively littleexcitotoxic cell loss in C57BLr6 and BALBrc strains.Whilst our data is in agreement with the observation thatC57BLr6 mice are less susceptible to cell loss, we ob-served a 34.7% reduction in normal cell number in theCA1 region of BALBrc mice. However, differences in the

Žroute of administration, the substrains of mice used for.example, 129rSvEMS vs. 129rOla and time points used

to study histological changes may account for these dis-crepancies.

FVBrN mice have proven useful in transgenic workdue to their vigorous reproductive performance and thepresence of prominent pro-nuclei in fertilised eggs that

w xfacilitate microinjection of DNA 46 . However, takentogether the data from the present study suggest that thisstrain may not be the most suitable for the development ofnew transgenic mice to aid the study of genes implicatedin CNS disorders. This may be particularly relevant forlongitudinal studies where inbred traits such as sponta-neous seizure activity and retinal degeneration may beexacerbated with age. The FVBrN strain will however,remain useful in gene targeting studies designed to interro-gate genes involved in peripheral disease, for example, in

w ximmunological studies 34,49 .

Acknowledgements

We gratefully acknowledge Dr. Neil Clarke for expertadvice on transgenic technology.

References

w x1 A. Aguzzi, S. Brandner, U. Sure, D. Ruedi, S. Isenmann, Transgenic¨and knockout mice: models of neurological disease, Brain Pathol. 4Ž .1994 3–20.

w x2 J.K. Andersen, O.P. Jurma, Use of genetically engineered mice asmodels for understanding human neurodegenerative disease, J. Am.

Ž .Geriatric Soc. 44 1996 717–722.w x Ž .3 R. Behringer, Supersonic congenics?, Nat. Genet. 18 1998 108.w x4 R. Bose, C.L. Schnell, C. Pinsky, V. Zitko, Effects of excitotoxins

Ž .on free radical indices in mouse brain, Toxicol. Lett. 60 1992211–219.

( )S.J. Royle et al.rBrain Research 816 1999 337–349348

w x5 M.L. Brines, A.O. Dare, N.C. deLanerolle, The cardiac glycosideoubain potentiates excitotoxic injury of adult neurones in rat hip-

Ž .pocampus, Neurosci. Lett. 191 1995 145–148.w x6 A.J. Bruce, M. Baudry, Oxygen free radicals in rat limbic structures

Ž .after kainate-induced seizures, Free Radic. Biol. Med. 18 1995993–1002.

w x7 G.A. Carlson, D.R. Borchelt, A. Dake, S. Turner, V. Danielson, J.D.Coffin, C. Eckman, J. Meiners, S.P. Nilsen, S.G. Younkin, K.K.Hsiao, Genetic modification of the phenotypes produced by amyloidprecursor protein overexpression in transgenic mice, Hum. Mol.

Ž .Genet. 6 1997 1951–1959.w x8 P.H. Chan, C.J. Epstein, Y. Li, T.T. Huang, E. Carlson, H. Ki-

nouchi, G. Yang, H. Kamii, S. Mikawa, T. Kondo et al., Transgenicmice and knockout mutants in the study of oxidative stress in brain

Ž .injury, J. Neurotrauma 12 1995 815–824.w x9 J.N. Crawley, L.G. Davis, Baseline exploratory activity predicts

anxiolytic responsiveness to diazepam in five mouse strains, BrainŽ .Res. Bull. 8 1982 609–612.

w x10 J.N. Crawley, J.K. Belknap, A. Collins, J.C. Crabbe, W. Frankel, N.Henderson, R.J. Hitzeman, S.C. Maxson, L.L. Miner, A.J. Silva,J.M. Wehner, A. Wynshaw-Boris, R. Paylor, Behavioural pheno-types of inbred mouse strains: implications and recommendations for

Ž .molecular studies, Psychopharmacology 132 1997 107–124.w x11 M.C. Dal Canto, M.E. Gurney, A low expressor line of transgenic

Ž .mice carrying a mutant human Cu, Zn superoxide dismutase SOD1gene develops pathological changes that most closely resemble those

Ž .in human amyotrophic lateral sclerosis, Acta Neuropathol. 93 1997537–550.

w x12 W.F. Dietrich, J. Miller, R. Steen, M.A. Merchant, D. Damron-Boles,Z. Husain, R. Dredge, M.J. Daly, K.A. Ingalls, T.J. O’Connor, Acomprehensive genetic map of the mouse genome, Nature 380Ž .1996 149–152.

w x13 T.N. Ferraro, G.T. Golden, G.G. Smith, W.H. Berrettini, Differentialsusceptibility to seizures induced by kainic acid treatment in mature

Ž .DBAr2J and C57BLr6J mice, Epilepsia 36 1995 301–307.w x14 T.N. Ferraro, G.T. Golden, G.G. Smith, N.J. Schork, P.St. Jean, C.

Ballas, H. Choi, W.H. Berrettini, Mapping murine loci for seizureŽ .response to kainic acid, Mamm. Genome 8 1997 200–208.

w x15 R.M. Friedlander, V. Gagliardini, H. Hara, K.B. Fink, W. Li, G.MacDonald, M.C. Fishman, A.H. Greenberg, M.A. Moskowitz, J.Yuan, Expression of a dominant negative mutant of interleukin-1 b

converting enzyme in transgenic mice prevents neuronal cell deathinduced by trophic factor withdrawal and ischemic brain injury, J.

Ž .Exp. Med. 185 1997 933–940.w x16 D. Games, D. Adams, R. Alessandrini, R. Barbour, P. Berthelette,

C. Blackwell, T. Carr, J. Clemens, T. Donaldson, F. Gillespie,Alzheimer-type neuropathology in transgenic mice overexpressing

Ž .V717F beta-amyloid precursor protein, Nature 373 1995 523–527.w x17 R. Gerlai, Gene-targeting studies of mammalian behavior: is it the

Ž .mutation or the background genotype?, Trends Neurosci. 19 1996177–181.

w x18 R. Gerlai, K.J. Millen, K. Herrup, K. Fabien, A.L. Joyner, J. Roder,Impaired motor learning performance in cerebellar En-2 mutant

Ž .mice, Behav. Neurosci. 110 1996 126–133.w x19 M.F. Goelz, J. Mahler, J. Harry, P. Myers, J. Clark, J.E. Thigpen,

D.B. Forsythe, Neuropathologic findings associated with seizures inŽ .FVB mice, Lab. Anim. Sci. 48 1998 34–37.

w x20 L. Holcomb, M.N. Gordon, E. McGowan, X. Yu, S. Benkovic, P.Jantzen, K. Wright, I. Saad, R. Mueller, D. Morgan, S. Sanders, C.Zehr, K. O’Campo, J. Hardy, C.M. Prada, C. Eckman, S. Younkin,K. Hsiao, K. Duff, Accelerated Alzheimer-type phenotype in trans-genic mice carrying both mutant amyloid precursor protein and

Ž .presenilin 1 transgenes, Nat. Med. 4 1998 97–100.w x21 K.K. Hsiao, D.R. Borchelt, K. Olson, R. Johannsdottir, C. Kitt, W.

Yunis, S. Xu, C. Eckman, S. Younkin, D. Price, C. Iadecola, H.B.Clark, G. Carlson, Age-related CNS disorder and early death in

transgenic FVBrN mice overexpressing Alzheimer amyloid precur-Ž .sor proteins, Neuron 15 1995 1203–1218.

w x22 R.B. Kent, D.A. Fallows, E. Geissler, T. Glaser, J. Emanuel, P.A.Lalley, R. Levenson, D.E. Houseman, Genes encoding a and b

subunits of Na, K ATPase are located on three different chromo-Ž .somes in the mouse, Proc. Natl. Acad. Sci. USA 84 1987 5369–

5373.w x23 T.G. Kokate, A.L. Cohen, E. Karp, M.A. Rogawski, Neuroactive

steroids protect against pilocarpine- and kainic acid-induced limbicseizures and status epilepticus in mice, Neuropharmacology 35Ž .1996 1049–1056.

w x24 V.C. Kurschner, R.L. Petruzzi, G.T. Golden, W.H. Berrettini, T.N.¨Ferraro, Kainate and AMPA receptor binding in seizure-prone and

Ž .seizure-resistant inbred mouse strains, Brain Res. 780 1998 1–8.w x25 S.F. Logue, E.H. Owen, D.L. Rasmussen, J.M. Wehner, Assessment

of locomotor activity, acoustic and tactile startle, and prepulseinhibition of startle in inbred mouse strains and F1 hybrids: implica-tions of genetic background for single gene and quantitative trait loci

Ž .analyses, Neuroscience 80 1997 1075–1086.w x26 J.F. Mahler, W. Stokes, P.C. Mann, M. Takaoka, R.R. Maronpot,

Spontaneous lesions in aging FVBrN mice, Toxicol. Pathol. 24Ž .1996 710–716.

w x27 P. Markel, P. Shu, C. Ebeling, G.A. Carlson, D.L. Nagle, J.S.Smutko, K.J. Moore, Theoretical and empirical issues for marker-as-

Ž .sisted breeding of congenic mouse strains, Nat. Genet. 17 1997280–284.

w x28 C. Mathis, P.E. Neumann, H. Gershenfeld, S.M. Paul, J.N. Crawley,Genetic analysis of anxiety-related behaviors and responses to ben-zodiazepine-related drugs in AXB and BXA recombinant inbred

Ž .mouse strains, Behav. Genet. 25 1995 557–568.w x29 M. Miyamoto, J.T. Coyle, Idebenone attenuates neuronal degenera-

tion induced by intrastriatal injection of excitotoxins, Exp. Neurol.Ž .108 1990 38–45.

w x30 M. Miyamoto, T.H. Murphy, R.L. Schnaar, J.T. Coyle, Antioxidantsprotect against glutamate-induced cytotoxicity in a neuronal cell

Ž .line, J. Pharmacol. Exp. Ther. 250 1989 1132–1140.w x31 A. Montkowski, M. Poettig, A. Mederer, F. Holsboer, Behavioural

performance in three substrains of mouse strain 129, Brain Res. 762Ž .1997 12–18.

w x32 R.S. Morrison, H.J. Wenzel, Y. Kinoshita, C.A. Robbins, L.A.Donehower, P.A. Schwartzkroin, Loss of the p53 tumor suppressorgene protects neurons from kainate-induced cell death, J. Neurosci.

Ž .16 1996 1337–1345.w x33 J.V. Nadler, Mini-review: kainic acid as a tool for the study of

Ž .temporal lobe epilepsy, Life Sci. 29 1981 2031–2042.w x34 A. Nii, D.A. Reynolds, H.A. Young, J.M. Ward, Osteochondrodys-

plasia occurring in transgenic mice expressing interferon-gamma,Ž .Vet. Pathol. 34 1997 431–441.

w x35 E.H. Owen, S.F. Logue, D.L. Rasmussen, J.M. Wehner, Assessmentof learning by the Morris water task and fear conditioning in inbredmouse strains and F hybrids: implications of genetic background for1

single gene mutations and quantitative trait loci analyses, Neuro-Ž .science 80 1997 1087–1099.

w x36 J.A. Piedrahita, S.H. Zhang, J.R. Hagaman, P.M. Oliver, N. Maeda,Generation of mice carrying a mutant apolipoprotein E gene inacti-vated by gene targeting in embryonic stem cells, Proc. Natl. Acad.

Ž .Sci. USA 89 1992 4471–4475.w x37 P.S. Puttfarcken, R.L. Getz, J.T. Coyle, Kainic acid-induced lipid

peroxidation: protection with butylated hydroxytoluene and U78517FŽ .in primary cultures of cerebellar granule cells, Brain Res. 624 1993

223–232.w x38 M.E. Ripps, G.W. Huntley, P.R. Hof, J.H. Morrison, J.W. Gordon,

Transgenic mice expressing an altered murine superoxide dismutasegene provide an animal model of amyotrophic lateral sclerosis, Proc.

Ž .Natl. Acad. Sci. USA 92 1995 689–693.w x39 P.E. Schauwecker, O. Steward, Genetic determinants of susceptibil-

( )S.J. Royle et al.rBrain Research 816 1999 337–349 349

ity to excitotoxic cell death: implications for gene targeting ap-Ž .proaches, Proc. Natl. Acad. Sci. USA 94 1997 4103–4108.

w x40 E.M. Simpson, C.C. Linder, E.E. Sargent, M.T. Davisson, L.E.Mobraaten, J.J. Sharp, Genetic variation among 129 substrains andits importance for targeted mutagenesis in mice, Nat. Genet. 16Ž .1997 19–27.

w x41 O. Smithies, N. Maeda, Gene targeting approaches to complexgenetic diseases: atherosclerosis and essential hypertension, Proc.

Ž .Natl. Acad. Sci. USA 92 1995 5266–5272.w x42 G. Sperk, Kainic acid seizures in the rat, Progress Neurobiol. 42

Ž .1994 1–32.w x43 S.M. Strain, R.A. Tasker, Hippocampal damage produced by sys-

Ž .temic injections of domoic acid in mice, Neuroscience 44 1991343–352.

w x44 C. Sturchler-Pierrat, D. Abramowski, M. Duke, K.H. Wiederhold, C.Mistl, S. Rothacher, B. Ledermann, K. Burki, P. Frey, P.A. Pa-ganetti, C. Waridel, M.E. Calhoun, M. Jucker, A. Probst, M.Staufenbiel, B. Sommer, Two amyloid precursor protein transgenicmouse models with Alzheimer disease-like pathology, Proc. Natl.

Ž .Acad. Sci. USA 94 1997 13287–13292.w x45 L. Sztriha, Increased lipid peroxide formation in the rat forebrain

Ž .during kainic acid seizures, Biomedica Biochimica Acta 45 1986491–494.

w x46 M. Taketo, A.C. Schroeder, L.E. Mobraaten, K.B. Gunning, G.Hanten, R.R. Fox, T.H. Roderick, C.L. Stewart, F. Lilly, C.T.

Hansen, P.A. Overbeek, FVBrN: an inbred mouse strain preferableŽ .for transgenic analyses, Proc. Natl. Acad. Sci. USA 88 1991

2065–2069.w x47 D.W. Threadgill, A.A. Dlugosz, L.A. Hansen, T. Tennenbaum, U.

Lichti, D. Yee, C. LaMantia, T. Mourton, K. Herrup, R.C. Harris,Targeted disruption of mouse EGF receptor: effect of genetic back-

Ž .ground on mutant phenotype, Science 269 1995 230–234.w x48 D.E. Tupper, R.B. Wallace, Utility of the neurological examination

Ž .in rats, Acta Neurobiol. Exp. 40 1980 999–1003.w x49 J. Wang, W.K. Yang, Y. Yang, S.J. Wei, D.M. Yang, J. Whang-Peng,

Y.M. Chen, K.L. Ting, C.C. Ting, Paradoxical effect of GM-CSFgene transfer on the tumorigenicity and immunogenicity of murine

Ž .tumors, Int. J. Cancer 75 1998 459–466.w x50 D.P. Wolfer, U. Muller, M. Stagliar, H.P. Lipp, Assessing the

effects of the 129rSv genetic background on swimming navigationlearning in transgenic mutants: a study using mice with a modified

Ž .beta-amyloid precursor protein gene, Brain Res. 771 1997 1–13.w x51 P.T. Xu, D. Schmechel, T. Rothrock-Cristian, D.S. Burkhart, H.L.

Qui, B. Popko, P. Sullivan, N. Maeda, A.M. Saunders, A.D. Roses,J.R. Gilbert, Human apolipoprotein E2, E3 and E4 isoform-specifictransgenic mice: human-like pattern of glial and neuronal immuno-reactivity in central nervous system not observed in wild-type mice,

Ž .Neurobiol. Dis. 3 1996 229–245.w x52 A. Zimmer, Gene targeting and behaviour: a genetic problem re-

Ž .quires a genetic solution, Trends Neurosci. 19 1996 470.