Antioxidant mechanisms in apolipoprotein E de¢cient mice prior to andfollowing closed head injury

L. Lomnitski 1;a, S. Chapman 1;a, A. Hochman b, R. Kohen c, E. Shohami d,Y. Chen d, V. Trembovler d, D.M. Michaelson a;*

a Department of Neurobiochemistry, George S. Wise Faculty of Life Sciences Tel Aviv University, Tel Aviv, Israelb Department of Biochemistry, George S. Wise Faculty of Life Sciences Tel Aviv University, Tel Aviv, Israel

c Department of Pharmacy, The Hebrew University, Jerusalem, Israeld Department of Pharmacology, The Hebrew University, Jerusalem, Israel

Received 12 August 1998; received in revised form 27 January 1999; accepted 5 February 1999

Abstract

Apolipoprotein E deficient mice have distinct memory deficits and neurochemical derangements and their recovery fromclosed head injury is impaired. In the present study, we examined the possibility that the neuronal derangements ofapolipoprotein E deficient mice are associated with oxidative stress, which in turn affects their ability to recover from closehead injury. It was found that brain phospholipid levels in apolipoprotein E deficient mice are lower than those of thecontrols (55 þ 15% of control, P6 0.01), that the cholesterol levels of the two mice groups are similar and that the levels ofconjugated dienes of the apolipoprotein E deficient mice are higher than those of control mice (132 þ 15% of P6 0.01).Brains of apolipoprotein E deficient mice had higher Mn-superoxide dismutase (134 þ 7%), catalase (122 þ 8%) andglutathione reductase (167 þ 7%) activities than control (P6 0.01), whereas glutathione peroxidase activity and the levels ofreduced glutathione and ascorbic acid were similar in the two mouse groups. Closed head injury increased catalase andglutathione peroxidase activities in both mouse groups, whereas glutathione reductase increased only in control mice. Thesuperoxide dismutase activity was unaffected in both groups. These findings suggest that the antioxidative metabolism ofapolipoprotein E deficient mice is altered both prior to and following head injury and that antioxidative mechanisms mayplay a role in mediating the neuronal maintenance and repair derangements of the apolipoprotein E deficient mice. ß 1999Elsevier Science B.V. All rights reserved.

Keywords: Alzheimer's disease; Apolipoprotein E; Oxidation stress; Antioxidant; Mouse; Brain

1. Introduction

Genetic studies suggest that the apolipoprotein E(apoE) genotype is a major risk factor for Alzheim-er's disease (AD) and that the allele apoE4, which isthe AD risk factor, exerts its e¡ects by markedlylowering the age of onset of this disease [1].

0925-4439 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved.PII: S 0 9 2 5 - 4 4 3 9 ( 9 9 ) 0 0 0 1 0 - 1

Abbreviations: AD, alzheimer's disease; ApoE, apolipopro-tein E; G-6-PD, glucose-6-phosphate dehydrogenase; GPX, glu-tathione peroxidase; GST, glutathione transferase; GR, gluta-thione reductase; SOD, superoxide dismutase

* Corresponding author. Fax: +972 (3) 6407643;E-mail : dmichael@ccsg.tau.ac.il

1 These authors contributed equally to this work.

BBADIS 61819 22-3-99

Biochimica et Biophysica Acta 1453 (1999) 359^368

ApoE de¢cient (knockout) mice [2] are a usefulmodel for studying the role of apoE in neuronalfunction, maintenance and repair. Indeed, age-de-pendent decreases of synaptic density in distinctbrain areas of apoE de¢cient mice have been demon-strated [3].It was also shown that apoE de¢cient micehave memory impairments and speci¢c derangementsin basal forebrain cholinergic neurons, which are as-sociated with decreased choline acetyltransferaselevels in the hippocampus and cortex [4] and withhyperphosphorylation of the cytoskeletal proteintau [5]. Closed head injury experiments revealedthat apoE de¢cient mice are particularly susceptibleto head trauma and that their ability to recover neu-rological and cognitive functions following this in-jury is impaired. Furthermore, histopathological ex-amination of the injured mice revealed morepronounced neuronal death in the hippocampus ofthe injured apoE de¢cient mice as compared to thecontrols [6].

Oxidative stress, namely an imbalance betweenproduction of reactive oxygen species (ROS) andtheir scavenging by antioxidants, plays a major rolein neuronal damage elicited by various head injuries[7,8]. Indeed, it was recently shown [9] that apoEde¢cient mice are impaired in their ability to replen-ish the reducing low molecular weight antioxidantswhich are oxidized following induction of closedhead injury, suggesting that the increased susceptibil-ity of the apoE de¢cient mice to closed head injurymay be related to alterations and imbalances in theirantioxidant system.

The working hypothesis of this study is that theneuronal derangements of apoE de¢cient mice areassociated with altered antioxidant levels which inturn a¡ect the ability of the mice to recover fromclose head injury. This hypothesis was presently as-sessed by measurements of the levels of distinct anti-oxidant enzymes and of low molecular weight anti-oxidants in brains of apoE de¢cient mice, both priorto and following closed head injury. The parametersmeasured include superoxide dismutase (SOD) andcatalase, the glutathione metabolized enzymes, gluta-thione peroxidase (GPX), glutathione reductase(GR) and glutathione transferase (GST), as well asthe low molecular weight antioxidants, ascorbic acidand reduced glutathione (GSH).

2. Materials and methods

2.1. Closed head injury

4 months old apoE de¢cient male mice (W23 g)and a matched group of control mice from the sameoriginal parent litter [2] were used in this study.Closed head injury to the indicated number ofapoE de¢cient and control mice was produced underanesthesia according to Shapira et al. [10], as modi-¢ed by Chen et al. [11]. Accordingly, a weight dropdevice was employed in which a calibrated weight(333 g) was allowed to fall from a height of 3 cmonto the exposed skull, over the left cervical hemi-sphere, 1^2 mm lateral to the midline of the midcoro-nal plane. Following recovery from anesthesia, themice were returned to their home cages with freeaccess to food and water. In addition, parallel groupsof apoE de¢cient and control mice were anaesthe-tized, their skulls were exposed but no trauma wasinduced. These mice served as sham controls. Histo-logical examinations of the brains of the 4 monthsold apoE de¢cient mice prior to closed head injuryrevealed,in accordance with previous observations,no apparent cellular or vascular pathology.

2.2. Preparation of brain homogenates

Brains of closed head-injured and sham-treatedmice were removed by decapitation prior to and5 min, 4 h and 24 h after head trauma (n = 6 foreach group of mice at each time point). The leftcortical hemisphere was then rapidly excized,promptly frozen in liquid nitrogen and stored at370³C until used. For preparation of homogenates,the brains were thawed and homogenized by anOmni EZ connect homogenizer (Omni International)at 4³C in 50 mM potassium phosphate pH 7.4, whichcontained 1 mM PMSF, 1 Wg/ml leupeptin and 1 mMEDTA. 20% of the extract was used immediately formeasurements of low molecular weight antioxidants(see below), whereas the remaining extracts were cen-trifuged at 4³C at 150 000Ug for 1 h. The resultingsupernatant was used for the enzymatic assayswhereas the pellet was used for determination of con-jugated dienes and for lipid content analysis. The GRand GPX assays were performed within 2^3 h fol-

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368360

lowing homogenization. The remaining supernatantwas stored overnight at 4³C and than used for assay-ing catalase, G-6-PD, SOD and GST.

2.3. Measurement of low molecular weightantioxidants

Freshly prepared brain homogenates were treatedwith 10% metaphosphoric acid by adding 200 Wl of15% metaphosphoric to 100 Wl homogenate and werethen centrifuged for 5 min at 10 000Ug. The result-ing supernatant was used for measurements of GSH,total glutathione and ascorbic acid as described bel-low.

2.4. GSH and ascorbic acid measurements

GSH was assayed by the glutathione assay kit(Calbio-chem-Novabiochem, San-Diego, CA, USA(CATALASE)). Total glutathione (GSH+oxidizedglutathione (GSSG)) was determined according toAdams et al. [12], using 5,5-dithiobis-2-nitrobenzoicacid (DTNB) and glutathione reductase. GSSG levelswere calculated by subtracting the measured GSHlevels from those of the total glutathione.

Ascorbic acid was assayed according to Motchnicet al. [13] by HPLC equipped with a C-18 column(Merck) and an electrochemical detector (BAS CC-5model, West Lafayet, IN, USA). The solvent systemwas acetate bu¡er 40 mM, pH = 4.78.

2.5. Enzyme assays

Glutathione reductase was assayed spectrophoto-metrically by monitoring the NADPH oxidation at340 nm according to Carlberg and Mannervik [14].

One unit of enzyme activity corresponds to the re-duction of 1 Wmol GSSG/min.

GPX was determined spectrophotometrically ac-cording to Floh and Gunzler [15] with t-butyl hydro-peroxide as substrate. One unit of enzyme activitycorresponds to the oxidation of 1 Wmol GSH/min.

GST was determined spectrophotometrically ac-cording to Habig et al. [16] using 10-chloro-2,4-dini-trophenol as substrate. One unit of enzyme activitycorresponds to the binding of 1 Wmol GSH/min.

Catalase was determined polarographically accord-ing to Hochman and Shemesh [17] using an oxygenelectrode (YSI, Yellow Springs, OH, USA) with20 mM hydrogen peroxide as substrate. One unitof enzyme activity corresponds to the release of1 Wmol O2/min.

G-6-PD was determined spectrophotometrically byfollowing the NADPH formation at 340 nm, accord-ing to Ben-Bassat and Goldberg [18]. One unit ofenzyme activity corresponds to the production of1 Wmol NADPH/min.

SOD. Brain homogenates were sonicated (3U15 s,Ultrasonic processor) in the presence of 1% TritonX-100 to disrupt the mitochondria and release Mn-SOD, after which the treated homogenate was cen-trifuged at 150 000Ug for 60 min. The total SODactivity in the resulting supernatant was then meas-ured spectrophotometrically, by monitoring the in-hibition of nitrite formation from hydroxylammo-nium chloride at 550 nm, as described by Pattichiset al. [19] The Mn-SOD activity was determined bythe same method, but in the presence of potassiumcyanide (10 mM) which completely inhibits Cu/Zn-SOD. The Cu/Zn-SOD activities were then calculatedby subtracting that of the Mn-SOD from the totalSOD activity. Puri¢ed SOD (5500 U/mg) obtained

Table 1Phospholipids, cholesterol and conjugated dienes levels in apoE de¢cient and control mice

Measured parameter Control ApoE de¢cient mice

Total phospholipids (Wg/mg protein) 262 þ 46 144.3 þ 29 **Conjugated dienes/phospholipids (nmol/mmol) 10.3 þ 1.2 13.6 þ 2.0 *Free cholesterol (Wg/mg protein) 160 þ 30 132 þ 20Cholesterol/phospholipids (Wg/Wg) 0.62 þ 0.05 0.98 þ 0.14**

Total phospholipids, conjugated dienes and non-esteri¢ed cholesterol in brains of apoE de¢cient and control mice were measured, asdescribed in Section 2. Results presented are mean þ S.D. of six mice in each group.*Signi¢cantly di¡erent from control mice, P6 0.05. **Signi¢cantly di¡erent from control mice, P6 0.01.

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368 361

from Biotechnology Rehovot (Israel) was used asstandard.

Occasionally, the brain tissues were slightly hemo-lytic and the resulting tissue homogenates thus con-tained some enzymes that originated from red bloodcells. In order to account for this possibility, thefollowing control assays were performed. Enzymeactivities were assayed in red blood cells hemolysatesaccording to a standard method and expressed permg hemoglobin. Quanti¢cation of the extent in¢ltrat-ing red blood cells to the brain homogenates wasperformed by assaying hemoglobin in the tissue ho-mogenates. The enzyme activities which contributedby erythrocytes were than calculated and were sub-tracted from the total observed enzyme activity. Inmost cases this correction amounted to less than 10%of the total measured activities.

2.6. Lipid analysis

The 150 000Ug pellet of the crude homogenatewas extracted by a mixture of chloroform:methanol(2:1, v/v), followed by the addition of water (1 vol-ume), as originally described by Bligh and Dyer [20].

The chloroform fraction was then evaporated andthe remaining residue was re-suspended in 1 ml hex-ane. The levels of conjugated dienes in the resultinghexane solution were than measured spectrophoto-metrically at 234 nm, utilizing a molar extinctioncoe¤cient of O= 24 600 [21]. The hexane was thenevaporated under nitrogen after which the resultingresidue was dissolved in 200 Wl ethanol and used forthe determination of non-esteri¢ed cholesterol andtotal phospholipid levels.

Non-esteri¢ed cholesterol (free cholesterol) levelswere determined utilizing the F-chol kit (BoeringerManheim, Germany) with free cholesterol (Sigma) asa standard.

Total cholesterol (esteri¢ed cholesterol) levels weredetermined by utilizing the T-chol kit (BoeringerManheim, Germany) with cholesterol ester as astandard.

Total phospholipid levels were determined accord-ing to Fiske and Subbarow [22], with sodium phos-phate as a standard.

The protein concentration was determined usingthe bicinchoninic acid protein assay kit, from Pierce(Rockford, IL, USA).

Fig. 2. Comparison of the levels of glutathione metabolizing en-zymes in brains of apoE de¢cient and control mice. The levelsof GPX, GR, GST in freshly prepared crude homogenates ofcontrol (¢lled bars) and apoE de¢cient mice (hatched bars)were determined, as described in Section 2. Results(mean þ S.D. of six mice per each group) are presented as per-centage of the mean control values. 100% GPX = 34 U/mg,100% GR = 72 U/mg, 100% GST = 1.6 U/mg. *Signi¢cantly dif-ferent from control, P6 0.01. **Signi¢cantly di¡erent from con-trol, P6 0.005.

Fig. 1. Comparison of brain Mn- and Cu/Zn-SOD and of cata-lase (CAT) levels of apoE de¢cient and control mice. Enzymaticactivities in freshly prepared cortical homogenates of control(¢lled bars) and apoE de¢cient mice (hatched bars) were deter-mined as described in Section 2. Results (mean þ S.D. of sixmice per each group) are presented as percentage of the meancontrol values. 100% Mn-SOD = 7.97 U/mg, 100% Cu/Zn-SOD = 89 U/mg, 100% catalase = 3.4 U/mg. *Signi¢cantly di¡er-ent from control, P6 0.05. **Signi¢cantly di¡erent from con-trol, P6 0.01.

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368362

Statistical analysis. All biochemical data were sub-jected to statistical analysis using two way ANOVAfollowed by a multiple comparisons test comparingthe individual enzyme levels. Data are expressed asmean þ S.D. and P values of 0.05 or less were con-sidered as signi¢cant.

3. Results

The possibility that apoE de¢cient mice are at achronic state of oxidative stress was examined by theassessment of the levels of distinct brain antioxidantenzymes and oxidation products. The brain levels ofMn-SOD, Cu/Zn-SOD and catalase activities of con-trol and apoE de¢cient mice are depicted in Fig. 1.As can be seen, the activity of Mn-SOD, which is the

Fig. 3. The e¡ects of closed head injury on brain GPX and GR activities of apoE de¢cient and control mice. GPX and GR activitiesof brain homogenates of sham (¢lled bars) and closed head injury treated (hatched bars) control (A) and apoE de¢cient (B) micewere measured 24 h following head injury, as described in Section 2. Results are mean þ S.D. of six mice per group. **Signi¢cantlydi¡erent from control, P6 0.01.

Fig. 4. The e¡ects of closed head injury on brain catalase (CAT) and SOD activities of apoE de¢cient and control mice. Catalase andCu/Zn-SOD activities of brain homogenates of sham (¢lled bars) and closed head injury treated (hatched bars) control (A) and apoEde¢cient (B) mice were measured 24 h following head injury, as described in Section 2. Results (mean þ S.D. of six mice per eachgroup) are presented as percentage of the mean of the untreated control group. 100% Catalase = 3.4 U/mg, 100% Cu/Zn-SOD =89 U/mg. ***Signi¢cantly di¡erent from untreated control group, P6 0.005. **Signi¢cantly di¡erent from untreated control group,P6 0.01.

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368 363

mitochondrial isoenzyme, was signi¢cantly higher inbrains of the apoE de¢cient mice group than in thoseof the controls (134 þ 7% of control, P6 0.01). Sim-ilarly, the activity of Cu/Zn-SOD, which is a cyto-solic enzyme, was also higher in brains of apoEde¢cient mice (122 þ 8% of control, P6 0.05). Fur-thermore, catalase, whose function is to remove hy-drogen peroxide, was also more active in the brainsof the apoE de¢cient mice (134 þ 6% of control,P6 0.01).

The levels of glutathione metabolizing enzymes inthe brains of the two mice groups are presented inFig. 2. As can be seen, of the two peroxide scaveng-ing enzymes, the level of GST was lower in the apoEde¢cient mice (62 þ 4% of control, P6 0.01), whereasthe GPX activity of these mice was slightly, thoughnot signi¢cantly, higher than that of the controls(114 þ 15% of control). In contrast, the GR activity,whose function is to replenish GSH, was markedlyhigher in the apoE de¢cient mice (167 þ 7% of con-trol, P6 0.005). The activity of brain G-6-PD, theenzyme which produces the reducing power forGR, was also higher in the apoE de¢cient micethan in the controls (respectively, 75.8 þ 8.7 and60.4 þ 1.7 U/mg).

In order to assess whether the observed di¡erencesare related to lack of apoE and are not due to acci-

dental strain di¡erences, we measured the levels ofantioxidative enzymes in the parental strains. AsapoE de¢cient and control mice di¡ered mainly intheir SOD and GR activities (see Figs. 1 and 2),the experiments focused on these enzymes. The re-sults obtained revealed that the parental strainsC67BL/6 and OLA129 contain respectively103 þ 9.5 and 99 þ 11% of the control total SOD lev-els and that the corresponding GR activities were89 þ 12 and 111 þ 10% of the control (n = 5 in eachmouse strain). There were no signi¢cant di¡erences(Ps 0.01) between the enzymes activities of the pa-rental strains and those of the control mice.

Measurements of brain levels of low molecularweight antioxidants revealed no di¡erences betweenthe apoE de¢cient and control mice in the levels ofeither GSH (respectively, 14.8 þ 1.4 and 15 þ 1.1nmol/mg protein) or ascorbic acid (respectively,51 þ 4 and 47.8 þ 7 nmol/mg protein).

The possibility that the observed alterations inantioxidant levels in the apoE de¢cient mice are as-sociated with chronic oxidative stress was investi-gated by measurements of the relative levels of con-jugated dienes in brain membranes of the two micegroups. As shown in Table 1, the total brain mem-branes of apoE de¢cient mice contained 45% lessbrain phospholipids per protein than did those of

Fig. 5. The e¡ect of closed head injury on ascorbic acid levels in brains of apoE de¢cient and control mice. Control (¢lled bars, panelA) and apoE de¢cient (hatched bars, panel B) mice were either subjected to closed head injury or sham-treated as described in Section2. Results presented are the mean þ S.D. of six mice per group. *Signi¢cantly di¡erent from corresponding sham group, P6 0.05.**Signi¢cantly di¡erent from corresponding sham group, P6 0.01.

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368364

the control mice and 30% higher levels of conjugateddienes per phospholipids. Interestingly, the levels ofbrain membrane cholesterol of the two mice groupswere similar, which resulted in a higher cholesterol/phospholipids molar ratio in the apoE de¢cient mice,as compared to that of the controls (Table 1).

The e¡ects of head injury on the antioxidant con-tent and oxidation products of the injured corticalhemisphere of apoE de¢cient and control mice wereinvestigated 24 h after closed head injury. Accord-ingly, the levels of oxidation products were assessedby measurements of the levels of brain-conjugateddienes, prior to and following injury. This revealedthat closed head injury elevated brain-conjugated di-enes levels in both apoE de¢cient mice (from13.6 þ 2.0 to 16.7 þ 0.7 nmol/mmol) and controlmice (from 10.3 þ 1.2 to 15.5 þ 0.6 nmol/mmol) andthat the resulting levels of conjugated dienes of thetwo mice groups were similar.

Head trauma induced signi¢cant (P6 0.01) in-creases in the GPX and GR activities of the controlmice (respectively, from 33.9 þ 3.6 to 73 þ 11.4 U/mgand from 72 þ 7.6 to 100 þ 3.4 U/mg, P6 0.01) (Fig.3). However, whereas the GPX activity of the apoEde¢cient mice also increased following head injury(from 38.5 þ 6.9 to 75 þ 16 U/mg), their GR activitydecreased slightly (Fig. 3). Further enzymatic meas-urements revealed that the GST activity of the con-trol group decreased signi¢cantly by 39 þ 11%, from1.6 þ 0.18 to 0.97 þ 0.1 U/mg (P6 0.05), whereas thatof the apoE de¢cient mice decreased only slightlyand not signi¢cantly, from 1.25 þ 0.07 to 1.07 þ0.01 U/mg. The brain G-6-PD activity of the controlmice was not a¡ected by head trauma, whereas thatof the apoE de¢cient mice decreased (from 75.8 þ 8.8to 58.9 þ 3.8 mU/mg (P6 0.05). Fig. 4 depicts thee¡ects of closed head injury on brain catalase andCu/Zn-SOD levels of the two mice groups. As can beseen, catalase activities of the control and apoE de-¢cient groups increased following head injury to200 þ 40 and 170 þ 34% of the respective sham-treated mice, whereas their Cu/Zn-SOD activities de-creased to respectively, 75 þ 3.0 and 70 þ 17% of thesham controls. Thus the catalase and Cu/Zn-SODactivities of the two mice groups were similarly af-fected by head injury.

The possibility that closed head injury has a di¡er-ential e¡ect on the levels of low molecular weight

antioxidants was examined by measurements of thelevels of ascorbic acid and glutathione in the twomice groups at varying time intervals following in-jury. As shown in Fig. 5, the levels of brain ascorbicacid of the two mice groups were both markedlyreduced at 5 min post injury and were not signi¢-cantly changed up to 24 h post injury (Fig. 5). Thelevels of reduced glutathione of the apoE de¢cientand control mice, measured 24 h post injury, weresimilarly and signi¢cantly reduced to 75 þ 5% of theuntreated groups. By contrast, the levels of GSSG,which comprises 1^2% of the total GSH, were notsigni¢cantly changed, suggesting that there is no dif-ference in either total or reduced glutathione levels ofthe two injured mice groups.

4. Discussion

The present study examined the possibility that thederangements in neuronal maintenance and repair ofapoE de¢cient mice are related to alterations in anti-oxidant mechanisms. This was performed by meas-urements, prior to and following closed head injury,of the levels of antioxidant enzymes, of low molec-ular weight antioxidants and of parameters of oxida-tive damages, in brains of apoE de¢cient mice andcontrol mice. The results obtained revealed consis-tent di¡erences prior to head injury between brainactivities of antioxidant enzymes of apoE de¢cientmice and controls, but that the two mice groupscontain similar levels of low molecular weight anti-oxidants. Closed head injury induced similar changesin the levels of antioxidant enzymes and of ascorbicacid and glutathione of the two mice group.

The observed increases in basal levels of Mn-SOD,GR and catalase of the apoE de¢cient mice (Figs. 1and 2) are similar to those previously observed inanimal models under chronic stress situations, suchas during aging [23] and in Bcl-2 de¢cient mice [24],as well as in human diseases such as AD, Parkinson'sdisease and other neurodegenerative diseases [25^28].The important result that the brain levels of antiox-idative enzymes in the parental mice strains, OLA129and C57BL/6, were the same and similar to those ofthe control mice suggests that the presently observedbrain enzymatic changes and chronic oxidative stressof the apoE de¢cient mice are indeed related to the

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368 365

lack of apoE itself and are not a result of accidentalstrain di¡erences. The assertion that apoE de¢cientmice are under chronic oxidative stress, is also sup-ported by the observed elevated levels of conjugateddienes and by the recent report that these mice con-tain elevated levels of 3-nitrotyrosin [29].

ApoE, the main lipoprotein and lipid transporterin the brain [30], plays an important role in mem-brane biosynthesis and in neurite outgrowth [31,32].Thus, it is possible that apoE de¢ciency results incompositional and structural alterations in brainmembranes, which in turn a¡ect the stability of themitochondria and result in electron leakage and inthe release of reactive oxygen species. The present¢nding that the ratio of cholesterol to phospholipidsof brain membranes in the apoE de¢cient mice iselevated (Table 1) is consistent with the above inter-pretation, as such changes are known to increase themembrane sensitivity to reactive oxygen species [33^36]. An alternative mechanism for explaining the in-creased oxidative susceptibility of the apoE de¢cientmice is based on the recent ¢ndings that apoE canact as an antioxidant [9,37,38] possibly by function-ing as an iron chelator [38]. Accordingly, it is possi-ble that in the absence of apoE, there is an increasein the levels of free iron, which in turn triggers theproduction of reactive oxygen species [8,39].

Closed head injury increased the levels of GPX,GR and catalase in control mice. By contrast, inthe apoE de¢cient mice only GPX and catalase ac-tivities were so increased (Figs. 3 and 4), whereas theGR activity was una¡ected (compare Fig. 3A to B).Since the GR activity of the apoE de¢cient mice isalready elevated prior to head injury (Fig. 2) it ispossible that it is already maximally elevated andthus cannot be further activated by the head injury.Increases in antioxidant enzyme activities have alsobeen observed in other experimental models whichinvolve oxidative stress [40]. These changes may bedue to phagocytic activity, as well as to intrinsicneuronal responses, and are believed to represent aprotection response, whereby they accelerate removalof free radicals. Accordingly, the antioxidant protect-ing responses of the apoE de¢cient and control micefollowing head injury are similar (see Figs. 3 and 4)and override the more subtle enzymatic di¡erencesbetween the two mice groups which were existed pri-or to head injury (see Figs. 1 and 2).

It was previously shown, utilizing cyclic voltamme-try measurements, that the two mice groups containsimilar levels of total reducing compounds and thatclosed head injury results in the depletion of thesecompounds, which subsequently recover in controlbut not in apoE de¢cient mice [9]. The oxidationpotential peak of ascorbic acid is about 400 mV.However, since the 400 mV oxidation peak of thetwo mice groups recovered more rapidly and exten-sively following closed head injury [9] than did theascorbic acid levels (Fig. 5), it is likely that the 400mV cyclic voltammetry peaks also detect changes inbrain low molecular weight antioxidants other thanascorbic acid [41,42]. The impaired ability of theapoE de¢cient mice to regenerate antioxidants fol-lowing head injury is probably not due to di¡erencesin antioxidative enzyme activities, whose levels in thetwo mice groups are similar following trauma (Figs.3 and 4). It may, however, be due to impairments inrecycling of unidenti¢ed reducing agents, other thanascorbic acid and glutathione. Alternatively, it maybe directly linked to apoE, which was recently shownto be an antioxidant [38] and whose de¢ciency istherefore expected to result in a diminished antiox-idant capacity.

The pathophysiological rami¢cations of the ob-served alterations in the antioxidative system ofapoE de¢cient at rest and following head traumaare not fully known. Recent ¢ndings suggest thatapoE de¢cient mice are more susceptible to ischemictreatment than control mice [43] and that humansubjects with the apoE4 genotype recover less wellfrom head injury than subjects who lack this allele[44]. It is possible that these abnormalities are relatedto antioxidative/oxidative alterations, similar to thosewhich are reported in the current study. The basalneurochemical derangements of the apoE de¢cientmice prior to head injury are speci¢c in that, forexample, cholinergic but not dopaminergic neuronsare a¡ected in these mice [45]. In view of the recent¢ndings that the cholinergic system is particularlysusceptible to apoE de¢ciency [45], it will be of in-terest to determine the antioxidative/oxidative bal-ance in distinct neuronal populations of the apoEde¢cient mice and to examine the extent to whichthey correlate with the neuropathology of these neu-rons. In addition, the extent to which distinct anti-oxidants [46,47] and iron chelators [39] can protect

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368366

the mice from the neuropathological consequences ofapoE de¢ciency remains to be determined.

In conclusion, the present study revealed thatapoE de¢cient mice are in a chronic state of oxida-tive stress and suggest that the brain antioxidativemechanism may play a role in the cognitive and neu-ronal impairments of the apoE de¢cient mice and intheir increased susceptibility to head injury.

Acknowledgements

We thank Prof. A. Devir from the Department ofNeurobiochemistry, Tel-Aviv University, for her ad-vise and generous help in the lipid analysis. Thiswork was supported in part by grants to DMMfrom the Fund for Basic Research of the Israel Acad-emy of Sciences (Grant number 670/96), from theUnited States-Israel Binational Science Foundation(Grant number 95/16), from the Harry Stern Nation-al Center for Alzheimer's Disease, from the JointGerman-Israeli Research Programm (Grant 1626)and from the Joseph K. and Inez Eichenbaum Foun-dation and to LL from the Rubinowicz Fund.

References

[1] D. Roses, Apolipoprotein E a¡ects the rate of Alzheimer'sdisease expression: beta-amyloid burden is a secondary con-sequence dependent on ApoE genotype and duration of dis-ease, J. Neuropathol. Exp. Neurol. 53 (1994) 429^431.

[2] A.S. Plump, J.D. Smith, T. Hayek, K. Alto-Stala, A. Walsh,J.G. Verstugygg, M. Rubin, J.L. Breslow, Severe hypercho-lesterolemia and atherosclerosis in apolipoprotein E-de¢cientmice created by homologous recombination in ES cells, Cell16 (1992) 343^353.

[3] M. Masliah, N. Mallory, M. Ge, I. Alford, A.D. Veinberg,D. Roses, Neurodegeneration in the central nervous systemof apoE-de¢cient mice, Exp. Neurol. 136 (1995) 107^122.

[4] E. Gordon, I. Grauer, E. Genis, D.M. Sehayek, D.M. Mi-chaelson, Memory de¢cits and cholinergic impairments inapolipoprotein E-de¢cient mice, Neurosci. Lett. 199 (1995)1^4.

[5] I. Genis, I. Gordon, E. Sehayek, D.M. Michaelson, Phos-phorylation of tau in apolipoprotein E-de¢cient mice, Neu-rosci. Lett. 199 (1995) 5^8.

[6] Y. Chen, L. Lomnitski, D.M. Michaelson, E. Shohami, Mo-tor and memory de¢cits in apoE-de¢cient mice after closedhead injury, Neuroscience 30 (1997) 1255^1262.

[7] M. Bochachio, N. Latronico, D.G. Zani, M. Mariotti, L.

Morandini, A.M. Acquarolo, A. Candiani, Free radical-in-duced lipoperoxidation and severe head injury. A clinicalstudy, Intensive Care Med. 16 (1990) 444^447.

[8] B. Halliwell, Reactive oxygen species and the central nervoussystem, J. Neurochem. 59 (1992) 1609^1623.

[9] L. Lomnitski, R. Kohen, Y. Chen, E. Shohami, V. Trem-bovler, T. Vogel, D.M. Michaelson, Reduced levels of anti-oxidants in brains of apolipoprotein E - de¢cient mice fol-lowing closed head injury, Pharmacol. Biochem. Behav. 56(1997) 669^673.

[10] Y. Shapira, E. Shohami, A. Sidi, D. So¡er, S. Freeman, S.Cotev, Experimental closed head injury in rats: mechanical,pathophysiologic, and neurologic properties, Crit. CareMed. 16 (1988) 258^265.

[11] Y. Chen, S. Constantini, V. Trembovler, M. Weinstock, E.Shohami, An experimental model of closed head injury inmice: pathophysiology, histopathology, and cognitive de¢-cits, J. Neurotrauma 13 (1996) 557^567.

[12] D. Adams Jr., B.H. Lauterburg, J.R. Mitchell, Plasma glu-tathione and glutathione disul¢de in the rat: regulation andresponse to oxidative stress, J. Pharmacol. Exp. Ther. 227(1983) 749^754.

[13] P.A. Motchnik, B. Frei, B.N. Ames, Measurement of anti-oxidants in human blood plasma, Methods Enzymol. 234(1994) 269^279.

[14] L. Carlberg, B. Mannervik, Glutathione reductase, MethodsEnzymol. 113 (1985) 484^499.

[15] L. Floh, W.A. Gunzler, Glutathione peroxidase, MethodsEnzymol. 105 (1984) 115^121.

[16] W.A. Habig, M.J. Pabst, W.B. Jacoby, Glutathione S-trans-ferase, J. Biol. Chem. 249 (1974) 7130^7139.

[17] A. Hochman, A. Shemes, Puri¢cation and characterizationof catalase-peroxidase from the photosynthetic bacteriumRhodopseudomonas capsulata, J. Biol. Chem. 262 (1987)6871^6876.

[18] A. Ben-Bassat, I. Goldberg, Puri¢cation and properties ofglucose-6-phosphate dehydrogenase (NADP�/NAD�) and 6-phosphogluconate dehydrogenase (NADP�/NAD�) frommethanol-grown Pseudomonas C, Biophys. Biochym. Acta611 (1980) 1^10.

[19] K. Pattichis, L.L. Louca, V. Glover, Quantitation of solublesuperoxide dismutase in rat striata, based on the inhibitionof nitrite formation from hydroxy-lammonium chloride,Anal. Biochem. 221 (1994) 428^431.

[20] B.G. Bligh, W.J. Dyer, A rapid method of total lipid extrac-tion and puri¢cation, Can. J. Biochem. Physiol. 37 (1959)497^509.

[21] R.O. Recknagel, E.A. Glende Jr., Spectrophotometric detec-tion of lipid conjugated dienes, Methods Enzymol. 105(1984) 331^337.

[22] C.H. Fiske, Y. Subbarow, The calorimetric determination ofphosphorus, J. Biol. Chem. 66 (1925) 375^400.

[23] S. Hussain, W. Slikker Jr., S.F. Ali, Age-related changes inantioxidant enzymes, superoxide dismutase, catalase, gluta-thione peroxidase and glutathione in di¡erent regions ofmouse brain, Int. J. Dev. Neurosci. 13 (1995) 811^817.

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368 367

[24] A. Hochman, H. Sterninl, S. Gorodin, S. Korsmeyer, I. Ziv,E. Melamed, D. O¡en, Enhanced oxidative stress and alteredantioxidants in brains of Bcl-2 de¢cient mice, J. Neurochem.71 (1998) 741^748.

[25] M.A. Pappolla, R.A. Omar, K.S. Kim, N.K. Robakis, Im-munohistochemical evidence of antioxidant stress in Alz-heimer's disease, Am. J. Pathol. 140 (1992) 621^628.

[26] L. Chen, J.S. Richardson, J.E. Caldwell, L.C. Ang, Regionalbrain activity of free radical defense enzymes in autopsysamples from patients with Alzheimer's disease and fromnon-demented controls, Int. J. Neurosci. 75 (1994) 83^90.

[27] P. Jenner, Oxidative damage in neurodegenerative disease,Lancet 344 (1994) 796^798.

[28] G. Benzi, A. Moretti, Are reactive oxygen species involvedin Alzheimer's disease?, Neurobiol. Aging 16 (1995) 661^674.

[29] R.T. Matthews, M.F. Beal, Increased 3-nitrotyrosine inbrains of apoE-de¢cient mice, Brain Res. 718 (1996) 181^184.

[30] R.W. Mahley, B.P. Nathan, R.E. Pitas, Apolipoprotein Estructure, function and possible roles in Alzheimer's disease,in: R.J. Wurtman, S. Corkin, J.H. Growdon, R.M. Nitsch(Eds.), The Neurobiology of Alzheimer's Disease, Vol. 777,Academy of Sciences, New York, 1996, pp. 139^145.

[31] B.P. Nathan, S. Bellosta, D.A. Sanan, K.H. Weisgraber,R.W. Mahley, R.E. Pitas, Di¡erential e¡ects of apolipopro-tein E3 and E4 on neuronal growth in vitro, Science 264(1994) 850^852.

[32] J. Poirier, Apolipoprotein E in animal models of CNS injuryand in Alzheimer's disease, Trends Neurosci. 17 (1995) 525^530.

[33] P. Viani, G. Cervato, A. Fiorilli, B. Cestaro, Age relateddi¡erences in synaptosomal peroxidative damage and mem-brane properties, J. Neurochem. 56 (1991) 253^258.

[34] I. Hajimohammadreza, M. Brammer, Brain membrane £uid-ity and lipid peroxidation in Alzheimer's disease, Neurosci.Lett. 112 (1992) 333^337.

[35] L. Ginsberg, J.R. Atack, S.I. Rapoport, N.L. Gershfeld,Evidence for a membrane lipid defect in Alzheimer's disease,Mol. Chem. Neuropathol. 19 (1993) 37^46.

[36] K. Wells, A.A. Farooqui, L. Liss, L.A. Horrocks, Neuralmembrane phospholipids in Alzheimer's disease, Neuro-chem. Res. 20 (1995) 1329^1333.

[37] M. Aviram, Interaction of oxidized low density lipoproteinwith macrophages in atherosclerosis, and the antiatheroge-nicity of antioxidants, Eur. J. Clin. Chem. Clin. Biochem. 34(1996) 599^608.

[38] N. Miyata, J.D. Smith, Apolipoprotein E allele-speci¢c anti-oxidant activity and e¡ects on cytotoxicity by oxidative in-sults and beta-amyloid peptides, Nat. Genet. 14 (1996) 55^61.

[39] S.S. Panter, J.M. Braughler, E.D. Hall, Dextran-coupled de-feroxamine improves outcome in a murine model of headinjury, J. Neurotrauma 9 (1992) 47^53.

[40] M.A. Otero, D. Avila, L.B. Em¢etzoglou, D. Clerch, V.Massaro, Increased manganese superoxide dismutase activ-ity, protein and mRNA levels and concurrent induction oftumor necrosis factor alpha in radiation-initiated Syrianhamster cells, Mol. Carcinog. 17 (1996) 175^180.

[41] E.C. Tayag, S.N. Nair, S. Wahhab, C.D. Katsetos, J.W.Lighthall, J.C. Lehmann, Cerebral uric acid increases follow-ing experimental traumatic brain injury in rat, Brain Res.733 (1996) 287^291.

[42] E. Beit-Yannai, R. Kohen, M. Horowitz, V. Trembovler, E.Shohami, Changes of biological reducing activity in rat brainfollowing closed head injury: A cyclic voltammetry study innormal and heat-acclimated rats, J. Cereb. Blood FlowMetab. 17 (1997) 273^279.

[43] H. Sheng, D.T. Laskowitz, E. Bennetr, D.E. Schmechel,R.D. Bart, A.M. Saunders, R.D. Pearlstein, A.D. Ropses,D.S. Warner, Apolipoprotein E isoform-speci¢c di¡erencesin outcome from focal ischemia in transgenic mice, J. Cereb.Blood Flow Metab. 18 (1998) 361^366.

[44] K. Johansson, N. Bogdanovic, H. Kalimo, B. Winblad, M.Viitanen, Alzheimer's disease and apolipoprotein E4 allele inolder drivers who died in automobile accidents, Lancet 349(1997) 1143^1144.

[45] S. Chapman, D. Michaelson, Speci¢c neurochemical de-rangements of brain projecting neurons in apolipoproteinE-de¢cient mice, J. Neurochem. 70 (1998) 708^714.

[46] S. Sen, H. Goldman, M. Morehead, S. Murphy, J.W. Phillis,Alpha-phenyl-tert-butyl-nitrone inhibits free radical releasein brain concussion, Free Rad. Biol. med. 16 (1994) 685^691.

[47] L. Packer, H.J. Tritschler, K. Wessel, Neuroprotection bythe metabolic antioxidant alpha-lipoic acid, Free Radic.Biol. Med. 22 (1997) 359^378.

BBADIS 61819 22-3-99

L. Lomnitski et al. / Biochimica et Biophysica Acta 1453 (1999) 359^368368