neuroprotective effect of aminoguanidine on transient focal ischaemia in the rat brain

Brain Research 905 (2001) 91–103www.elsevier.com/ locate /bres

Research report

Neuroprotective effect of aminoguanidine on transient focal ischaemiain the rat brain

a , b c d a*Diana Cash , John S. Beech , Richard C. Rayne , Philip M.W. Bath , Brian S. Meldrum ,aSteve C.R. Williams

aInstitute of Psychiatry, King’s College, University of London, London, UKbDepartment of Medicine, Division of Anaesthesia, University of Cambridge, Cambridge, UK

cDepartment of Biology, Birkbeck College, University of London, London, UKdDivision of Stroke Medicine, University of Nottingham, Nottingham, UK

Accepted 23 April 2001

Abstract

Using serial magnetic resonance imaging we have evaluated the effectiveness of aminoguanidine (AG) as a neuroprotective agent in arat model of transient middle cerebral artery occlusion (MCAO). Because aminoguanidine’s neuroprotective properties have primarilybeen ascribed to its action as iNOS inhibitor, we also performed a biochemical analysis of nitric oxide metabolites and NOS isoforms inour model of ischaemia. Daily injections of AG (100 mg/kg) or saline, were started at 6 h after the occlusion and the effects of thistreatment on lesion progression monitored by T -weighted MRI at 6 (pre-treatment scan), 24 and 72 h. Measurements of lesion volumes2

showed that between 6 and 72 h post-MCAO, lesion growth was slower in AG-treated rats than in control rats. This difference was mostpronounced between 24 and 72 h post-MCAO when AG halted the lesion volume expansion observed in control rats. Measurements ofplasma NOx (nitrite plus nitrate) at 0, 24, 48 and 72 h after MCAO, showed that NO levels did not differ significantly between the AG-and saline-treated groups at any time-point. Moreover, NOS activity assays revealed that no iNOS activity was present in any of the brainstested and that constitutive neuronal NOS activity was similar across the two hemispheres between both groups. The absence of iNOSprotein in the ischaemic and contralateral hemispheres at 48 and 72 h after MCAO (control group only) was confirmed by Western blotanalysis. These results suggest that AG treatment reduces the rate of growth of ischaemic lesions, perhaps preserving the functioning ofperifocal neurons. Our observations contradict suggestions that high levels of NO generated by iNOS are partially responsible forexacerbating the neuronal damage in the postischaemic phase of MCAO. Although this does not rule out a role for AG as aneuroprotective agent via its ability to inhibit iNOS, these findings indicate that neuroprotective actions of AG may also be mediated viaother cellular targets. 2001 Elsevier Science B.V. All rights reserved.

Theme: Disorders of the nervous system

Topic: Ischemia

Keywords: Aminoguanidine; Ischaemia; Inducible nitric oxide synthase; Magnetic resonance imaging; Middle cerebral artery occlusion; Rat

1. Introduction

In thromboembolic stroke, the region of ischaemicAbbreviations: AG, Aminoguanidine; DWI, Diffusion-weighted imag-

penumbra surrounding the core of ischaemic brain tissue ising; eNOS, Endothelial NOS; iNOS, Inducible NOS; L-NAME, N-nitro-l-metabolically compromised but still potentially viable forarginine methyl ester; MCAO, Middle cerebral artery occlusion; nNOS,

Neuronal NOS; NO, Nitric oxide; NOS, Nitric oxide synthase; NOx, hours after the initial insult [13,20,33,35]. The rescue ofNitrates plus nitrites; PDWI, Proton density-weighted imaging; T WI,2 this perifocal, penumbral region from irreversible damageT -weighted imaging2 is a major goal of new therapies for treating stroke [25,57].

*Corresponding author. Neuroimaging Research, Institute of Psychi-Therapy that is targeted to these late events is sensibleatry, 4 Windsor Walk, London SE5 8AF, UK. Tel.: 144-207-848-0083;given that most stroke victims do not present and receivefax: 144-207-848-0055.

E-mail address: d.cash@iop.kcl.ac.uk (D. Cash). medical care until hours after the onset of the injury.

0006-8993/01/$ – see front matter 2001 Elsevier Science B.V. All rights reserved.PI I : S0006-8993( 01 )02508-2

92 D. Cash et al. / Brain Research 905 (2001) 91 –103

Additional complication arises in approximately 50% of sion, a paradigm believed to mimic closely human throm-thromboembolic strokes [38]: spontaneous reperfusion, boembolic stroke [33]. We have measured the extent ofoccurring consequent to recanalisation of the occluded cerebral tissue damage, in AG- and saline-treated rats, overblood vessel, can paradoxically exacerbate vascular and a period of 3 days. The first dose of AG was administeredparenchymal injury. after reperfusion, 6 h following the occlusion; additional

Although the mechanisms underlying reperfusion injury doses were given daily. This regime of administration wasare not well understood, studies conducted in animal chosen deliberately to permit assessment of the drug’ssystems of transient focal ischaemia have contributed to effects on the damaging events that occur during reperfu-our knowledge. During reperfusion, free radicals — in- sion. A secondary aim of the study was to test thecluding nitric oxide (NO) and superoxide — and pro- hypothesis that aminoguanidine acts on the nitric oxideinflammatory cytokines are released from the microvascu- system in this stroke model. Three methods were used tolar endothelium and from inflammatory cells infiltrating monitor the activity of NO-generating enzymes in treatedthe region. These pro-inflammatory substances are thought and control animals: (i) measurement of NOS activity into be responsible for the recruitment of penumbral tissue brain homogenates, (ii) measurement (using a chemilumin-into the lesion [9,25]. Experimental therapies using free escence-based detection system) of NO metabolites inradical scavengers [7,18] and anti-inflammatory agents plasma, and (iii) Western blot analysis of iNOS and nNOS[8,30] have proven effective in reducing lesion volumes as proteins in brain homogenates.well as in retarding the progression of the tissue damage inreperfusion injury.

Aminoguanidine (AG) is a molecule with several phar-2. Materials and methods

macological actions which, because of its ability to retardthe chronic tissue damage associated with formation of

2.1. Chemicalsadvanced glycosylation end products (AGEs), has beenused for clinical trials in diabetes [34]. More recently,

All chemicals were purchased from Sigma (UK) unlessneuroprotective properties of AG have been discovered in

otherwise indicated.several experimental animal models of stroke and corticalinjury [5,53,56]. Importantly, AG was effective in reducingthe ischaemic lesion volume even when administered as 2.2. Animal preparationlate as 24 h after the insult [56]. On the basis of thisfinding, it is of interest to investigate whether AG can All experimental procedures were carried out in accord-attenuate cell damage subsequent to ischaemia and reperfu- ance with The Animals (Scientific Procedures) Act, Unitedsion injury in thromboembolic stroke. Kingdom, 1986.

Although the mechanism of AG neuroprotective action Overnight-fasted male Sprague–Dawley rats (Charlesremains uncertain, beneficial effects of the drug in rodent Rivers Ltd., UK) (body weight 318661 g, mean6S.D.)systems of focal ischaemia have been ascribed primarily to were initially anaesthetised with 3% halothane in oxygen/its action as a relatively selective inhibitor of inducible nitrous oxide (30:70%). Anaesthesia was maintained withnitric oxide synthase (iNOS) [17]. NO is an important 2% halothane during all surgical procedures. The leftmessenger in the brain, but it can be toxic in certain middle cerebral artery (MCA) was occluded using anconditions of pathology (for review, see [40]). For exam- intraluminal thread approach by adaptation of the methodple, excessive activation of the constitutive NOS isoform originally described by Koizumi et al. [24]. Briefly, the left— neuronal NOS (nNOS) — during unregulated elevation carotid vasculature was exposed via a 3-cm-long incisionof intracellular calcium during ischaemia is linked with in the neck. A nylon thread (3 /0 monofilament), its tipNO-mediated cell damage. Inducible NOS (iNOS), as coated in silicone rubber, was introduced via externalsuggested by its name, is not normally present in the brain; carotid arteriotomy and advanced along the internal carotidit is expressed on demand during a variety of inflammatory artery. The thread was inserted to a depth of 10 mm fromconditions including those which can occur in ischaemia the entrance of the carotid canal. Identifying the entrance[16,39]. The large quantities of NO produced by iNOS are of the carotid canal as a cranial marker assured thetoxic to cells and NO produced by this route may be accurate positioning of the tip at 1 mm beyond the originpartially responsible for the spread of tissue damage seen of the left MCA [44]. Following 2 h of MCA occlusion,in reperfusion injury. animals were briefly reanaesthetised with halothane, as

A primary aim of this study was to evaluate the above, and the thread was withdrawn from the internalneuroprotective effects of AG therapy in a rat model of carotid artery to permit normal anterograde blood flow totransient focal ischaemia using serial magnetic resonance the left MCA.imaging. The model used in our study was based on a Rectal temperature was monitored throughout theregime first published by Koizumi [24]. Here, 2 h of surgery, and maintained at 37618C with a heated electricmiddle cerebral artery occlusion were followed by reperfu- pad under the rat. Animals were given no food for the first

D. Cash et al. / Brain Research 905 (2001) 91 –103 93

6 h following the surgery but had access to drinking water permitted the application of a contour-tracing algorithm toat all times. delineate the boundaries of lesions. The total volume of

Venous blood samples (ca. 0.5 ml from the jugular vein) each brain was measured in the same fashion to assess thewere taken from all animals, under halothane anaesthesia, degree of swelling. For each selected area, in each slice,immediately before the surgery (0 h) and then at 24, 48 pixel counts were generated. These values were thenand 72 h after MCAO. Twenty-four hour and 72 h samples multiplied by the slice thickness and summed to yield total

3were taken immediately after MRI scanning while the volume in mm . To limit subjective bias, the operator wasanimals were still under anaesthesia. Rats were briefly blinded to the treatment strategy.re-anaesthetised (as before) for sampling blood at 48 h T -weighted images acquired at 6, 24 and 72 h were2

after MCAO. analysed for measurements of lesion and total brainThe animals were randomly assigned to groups, then at volume. Diffusion-weighted images acquired at 6 and 24 h

6, 24 and 48 h after the occlusion given one of the were analysed for lesion volume only. Because DWI isfollowing treatments: (i) aminoguanidine (AG) hemisul- considered more accurate than T WI in revealing the early2

phate at a dose of 100 mg/kg i.p. in 1 ml saline (AG- ischaemic lesion [37], a comparison of T WI- and DWI-2

treated group) or (ii) 1 ml 0.9% saline (control group). At derived lesion volumes at 6 and 24 h post-MCAO was48 or 72 h post-occlusion, the animals were killed by performed (Pearson correlation analysis). DWI-deriveddecapitation and the brains were removed immediately, measurements of lesion volume at 72 h after MCAO weresnap frozen and stored in liquid nitrogen for later analysis. not performed as it is well known that a hyperintensity

characterising an ischaemic lesion in DWI is often reduced2.3. MRI methods by 3 days following the ischaemic insult [2,20].

Total brain volumes were compared between the twoFor all MRI procedures, animals were anaesthetised and groups at each time-point and within a group over time

maintained under halothane (1% in O , v/v). Throughout using a two-way analysis of variance (ANOVA) and2

the experiment, body temperature was maintained at ANOVA with repeated measures, respectively. The effect37628C using a thermostatically controlled heating blan- of AG on the changes in lesion volumes over time wasket. Saline was administered intraperitoneally (0.5–1 ml) analysed by ANOVA with repeated measures. To eliminateafter each scanning session to alleviate dehydration that the confounding effect of the pre-treatment variability in

1may have been caused by anaesthesia. H MRI was lesion sizes, baseline data (6 h post-occlusion) was in-1performed at 4.7 Tesla (200 MHz for H) using a SIS-200 cluded as a covariant in the statistical analysis.

imaging spectrometer (Spectroscopy Imaging Systems, For calculation of the T relaxation time, the data from2

USA) and a home-made, 75-mm-diameter, 8-legged bird- PDWI and T WI were fitted to mono-exponentials. This2

cage radiofrequency coil for signal transmission and was performed on a pixel by pixel basis using two pointreception. fitting routines. T values were measured in a 232-pixel2

A total of 12 animals (six AG-treated and six controls) regions of the three anatomically distinct brain areas:were imaged at 6, 24 and 72 h post-occlusion. The MRI subcortical — caudate–putamen and cortical — upperprotocol consisted of three interleaved imaging regimens: parietal cortex and lower parietal cortex in both hemi-proton density-weighted imaging (PDWI), T -weighted spheres. Those values were compared between the treated2

imaging (T WI) and diffusion-weighted imaging (DWI). and the control groups using a two-tailed, unpaired t-test.2

Repetition time (TR) was 3 s for each of the three In all of the above comparisons, P-values of ,0.05 wereregimens and echo times (TE) were 70 ms for the T WI considered significant. Data are expressed as means6S.D.,2

and DWI and 28 ms for the PDWI. For DWI, diffusion- where n5number of animals.23 2sensitising gradients (b-value 590310 s /mm ), were

applied along the inferior to superior axis of the brain. Two 2.5. Preparation of brain extracts and plasmaaverages were acquired per phase encode step. The imageswere collected at an in-plane resolution of 0.3130.31 mm To prepare the plasma, blood was collected into afrom 18 contiguous, 1-mm-thick slices, with a total syringe containing 0.1 mM EDTA and immediately cen-acquisition time of 43 min. trifuged for 5 min at ca. 10003g. Haemocytes, forming the

pellet, were discarded and the plasma was stored at 2208C2.4. MR image analysis for further analysis.

Brain extracts were prepared by homogenising the tissueVolumetric brain and lesion measurements were per- in an ice-cold buffer using an Ultra-Turrax homogeniser.

formed following a direct Fourier transform (FT) of the Homogenisation buffer (HB) contained 100 mM Tris–HClraw time domain data, for the T WI and DWI. Images (pH 7.4), 0.5 mM EDTA, 0.5 mM EGTA, 5 mM dithio-2

were displayed in the coronal plane, as collected. Lesions threitol and one dissolved Mini Protease Inhibitor Cocktailwere detected as hyperintense signals in the MR images. tablet per 10 ml (Boehringer Mannheim, UK). Homoge-Selection of appropriate seed points by the operator nates were centrifuged at 13,0003g, at 48C, for 40 min.

94 D. Cash et al. / Brain Research 905 (2001) 91 –103

The supernatant was collected and immediately used in spectroscopy of the 5 ml flow-through after having addedNOS assays or stored at 2208C for subsequent Western ca. 5 ml of liquid scintillation cocktail (Beckman, UK).blot analysis. Total protein concentrations in brain extracts In control reactions brain extract was replaced by bufferwere determined using the modified Lowry protein assay (HB). The ‘blank’ dpm values in these reactions were(Pierce Chemicals) with bovine serum albumin (BSA) as a subtracted from the total dpm values in the enzymestandard [32]. reactions. Based on the percentage of total radiolabelled

arginine converted to citrulline, a figure representing the2.6. Purification of NO synthase conversion of total available (‘cold’ plus tracer) substrate,

expressed as pmol of citrulline produced per min per mg ofNOS was partially purified from rat brain extracts using protein, was obtained. All measurements were performed

29,59-adenosine dinucleotide phosphate (ADP)–sepharose in duplicate and the data are expressed as means6S.D.,affinity chromatography [50]. Brain extracts were continu- where n5number of animals. Differences in NOS activityously mixed with ADP–sepharose (approximately 40 ml were assessed by a paired two-tailed t-test, for comparisonsettled volume per 1 ml of extract) for 30 min at 48C, then between ischaemic and contralateral hemispheres, and bycentrifuged for 5 s in a microcentrifuge. The supernatant, an unpaired, two-tailed t-test for comparison between thecontaining unretained proteins, was discarded. The pelleted two groups of animals. In both analyses, P values ofADP–sepharose was successively washed (three times in 1 ,0.05 were considered significant.ml for each wash) by resuspension and centrifugation asabove, first with HB, then with 0.2 M NaCl in HB and 2.8. Measurement of nitric oxideagain with HB. Specifically-bound proteins were elutedfrom the ADP–sepharose by mixing with approximately 8 Nitric oxide in plasma was determined as nitrite plusvolumes of 15 mM NADPH in HB for 15 min at 48C. The nitrate (NOx), by the chemiluminescence method using amixture was centrifuged for 5 s and the supernatant nitric oxide analyser (NOA, Sievers, model 280) [1,12].containing eluted NOS and other ADP-binding proteins Nitrates were first converted to nitrites by incubatingwas collected and stored at 2208C for subsequent exami- plasma samples with nitrite reductase (1 unit /ml, fromnation by sodium dodecyl sulphate–polyacrylamide gel Aspergillus species, Boehringer Mannheim, UK), NADPHelectrophoresis (SDS–PAGE). (40 mM), FAD (1 mM) and Tris–HCl (20 mM) for 60 min

at 378. Sample (100 ml) was then injected into a purge2.7. Measurement of NOS activity vessel of the NOA, containing 1.5 ml of potassium iodide,

6 ml of glacial acetic acid and 200 ml of antifoaming agentThe catalytic activity of NOS was measured by the (Sievers) in an atmosphere of nitrogen. Nitrite concen-

3 3conversion of L-[ H]-arginine to L-[ H]-citrulline by a trations were quantified by comparison to a standard curvemodification of the method described by Bredt and produced with known amounts of authentic sodium nitrite.Schmidt [4]. Citrulline synthesis was measured in reaction The sensitivity of this assay is within picomolar range withmixtures obtained by mixing, in 1:1 ratio, brain extract 1 pmol limit for detection of NO (NOA manual, Sievers).(approximately 8 mg/ml protein) and assay buffer (AB) All measurements were performed in duplicate and thecontaining NOS cofactors and substrate. The final con- data were expressed as means6S.D., where n5number ofcentrations of other reaction components were: 1 mM animals. Concentration of NOx in plasma at 0, 24, 48 andNADPH, 1 mM CaCl , 50 mM arginine hydrochloride, 10 72 h after MCAO were statistically analysed for differ-2

mM flavin mononucleotide (FMN), 10 mM flavin adenine ences between AG-treated and control groups and withindinucleotide (FAD), 50 mM tetrahydrobiopterin (H B), 50 each group over time, using ANOVA with repeated4

mM Tris–HCl (pH7.5) and 1,000,000 dpm L-[2,3,4,5- measures. P-values of ,0.05 were considered significant.3H]arginine monohydrochloride (55 Ci /mmol, from Amer-sham Int. plc). In some reactions an arginine analogue, 2.9. SDS–PAGE and Western blotting of NOSN-nitro-L-arginine methyl ester (L-NAME) (1 mM final

21concentration), was included to block NOS activity. Ca - Aliquots of ADP–sepharose eluates were analysed byindependent NOS activity was measured by excluding electrophoresis on 7.5% sodium dodecyl sulphate–poly-calcium and adding EGTA (3 mM final concentration) to acrylamide gels (SDS–PAGE) according to the discontinu-the reactions. The samples were incubated for 5–30 min at ous method of Laemmli [26] using a Bio-Rad Mini-Protein378C and reactions were terminated by adding 9 volumes II electrophoresis cell (Bio-Rad Laboratories Ltd., UK).of ice-cold stop buffer (50 mM HEPES containing 5 mM Gels were equilibrated in transfer buffer (25 mM Tris–EDTA, pH 5.5). Samples were analysed immediately by HCl, 192 mM glycine, pH 8.4) for 15–20 min at 48C prior

1applying them to Dowex AG 50W-X8 columns (Na to electrophoretic transfer of proteins to PVDF mem-form; BioRad Laboratories Ltd., UK; 2 ml bed volume) branes. After transfer, membranes were incubated for 45which were eluted into scintillation vials with ca. 5 ml of min at room temperature (RT) with ‘blotto’ buffer (5%

3water. H-citrulline was quantified by liquid scintillation wt /v non-fat dry milk, 125 mM sodium chloride, 0.05%

D. Cash et al. / Brain Research 905 (2001) 91 –103 95

Tween 20, 20 mM Tris–HCl, pH 7.4). Membranes werethen incubated for 1 h at RT with either anti-humanncNOS polyclonal antibody (Santa Cruz BiotechnologyInc., Germany), diluted 1:100 in blotto, or anti-mouseiNOS polyclonal antibody (Transduction Laboratories,USA), diluted 1:10,000 in blotto. Following this, mem-branes were incubated with appropriate secondary anti-bodies conjugated to alkaline phosphatase (Santa CruzBiotechnology Inc., Germany) for 1 h at RT. Between eachstage, the membranes were washed in three changes ofTBS/Tween (20 mM Tris–HCl, pH 7.4, 125 mM sodiumchloride, 0.05% Tween 20). Immunoreactive proteins weredetected by incubating membranes in a BCIP/NBT solu-tion until bands appeared. Positive control samples con-taining iNOS (mouse macrophage lysate) or nNOS (ratpituitary lysate) (Transduction Laboratories, USA) were Fig. 1. Ischaemic lesion volumes, derived from the T -weighted MRI2

run on the SDS–PAGE along with the ADP–sepharose images, of AG- and saline-treated rats. Lesion volume at 6 h representsthe pre-treatment baseline; AG treatment was initiated at 6 h aftereluates.MCAO. Data are expressed as mean6S.D. of n56 in each group. Nosignificant difference in lesion volumes between the groups was revealedby one-way ANOVA. AG, aminoguanidine; MCAO, middle cerebral

3. Results artery occlusion.

3.1. Lesion volumesAnalysis of the lesions’ temporal evolution showed that

Two hours of middle cerebral artery occlusion (MCAO) lesion growth over the 72-h period was attenuated by theresulted in the formation of ipsilateral ischaemic lesions in AG treatment. In Fig. 2, mean lesion volumes at 24 and 72all 12 animals. The lesions were revealed by MR imaging h are expressed as a change from respective lesionand were confirmed by post mortem visual observations of volumes at the 6-h time-point. For both saline- and AG-the brains. We have shown previously that DWI and treated rats, mean lesion volumes increased during thisT WI-derived lesion volume measurements at 24 h corre- interval. However, the rate of increase for the saline-2

late well with those observed with 2,3,5-triphenyltetra- treated group was significantly greater than for the AG-3zolium chloride (TTC) histological staining [48]. To treated group (lesions increased by 1.760.6 mm /h in

3ensure that an accurate assessment of lesion volume was AG-treated rats and by 3.060.2 mm /h in saline-treatedderived from T W MRI, a correlation analysis between the rats, P,0.05). Lesions continued increasing between 242

two imaging methods at 6 and 24 h post-MCAO wasperformed. A significant correlation between DWI- and T2WI-derived lesion measurements was observed, with Pear-son correlation coefficients of 0.47 (P,0.05) and 0.96(P,0.001), for 6 and 24 h after MCAO, respectively.

Mean brain volumes were determined by the analysis of3T WI (control group 1638.6638.8 mm ; AG-treated group2

31633.1636.8 mm ). Because the total brain volumes didnot differ between the groups or between time-points(P50.943), it was not necessary to normalise lesionvolumes to total brain volumes. Comparison of lesionvolumes in the two groups revealed that lesions of treatedanimals appeared smaller than in controls at 72 h, but thedifference failed to reach statistical significance (AG group

3 3150.7616.34 mm ; control group 173.50614.72 mm ;P50.07; Fig. 1). Animals had been randomly assigned to Fig. 2. Development of ischaemic lesion volumes, with time, aftereither the control or treated group and, by chance, mean MCAO. Results (mean6S.D.; n56 in each group), derived from the

T -weighted MRI images, are expressed as difference from the pre-lesion volumes were slightly larger in the AG group at 6 h. 2

treatment, baseline lesion volume (i.e. lesion volume at 6 h post-occlu-This discrepancy, considered together with the small sizession). A significant difference in the lesion evolution between 6 and 72 h

of the groups (n56 for each treated and control group), is after MCAO was observed between the two groups (*P,0.05, repeatedlikely to contribute to underestimation of the effects of the measures ANOVA). MCAO, middle cerebral artery occlusion; AG,drug on the lesion volume. aminoguanidine.

96 D. Cash et al. / Brain Research 905 (2001) 91 –103

3and 72 h post-MACO (at a rate of 10.560.03 mm /h),while no further increase after 24 h was observed in the

3AG-treated group (lesions decreased by 20.160.1 mm /h). The difference between these changes was also foundto be statistically significant. In individual rats, lesionvolumes decreased after 24 h in five out of six AG-treatedrats, while they increased in five out of six saline-treatedrats. A series of slices, showing typical lesion topologies at6, 24 and 72 h post-MCAO are illustrated in Fig. 3A and Bfor AG- and saline-treated rats, respectively. Differencesbetween corresponding images are noticeable at 24 h butbecome clearly visible at 72 h.

Measurements of the ‘spin to spin’ relaxation time (T )2

are shown in Fig. 4. There were no differences in T2

between the treated and the control group regions ofinterest at any time-point, indicating that AG treatment didnot change the relaxation characteristics of the brain tissueand that the observed differences in T WI-derived lesion2

volumes result exclusively from the changes in the lesionsize and cannot be attributed to variation in the T contrast.2

3.2. NOS activity

NOS catalytic activity was determined by measuringcitrulline synthesis in rat brain homogenates in vitro. Totalsoluble NOS activity was measured in reactions containingall NOS cofactors and calcium. To confirm that NOS wasresponsible for citrulline synthesis, we included in controlreactions the L-arginine analogue, L-NAME, a known NOSinhibitor. In all such assays, citrulline synthesis was almostcompletely inhibited (598.460.1%, mean6S.D.; see Fig.6).

Fig. 5 shows the specific activities of total NOS inhomogenates from ischaemic and contralateral hemi-spheres of both AG- and saline-treated rats. Although theNOS activity in saline-treated rats was slightly higher thanthat of the AG-treated group (AG group — ischaemic120.4626.4, contralateral 113.4621.3 and saline group —

Fig. 3. A series of coronal, T -weighted MR images of an AG-treated (A)2ischaemic 146.4616.71, contralateral 154.3626.58 pmoland a saline-treated (B) rat brain at 6, 24 and 72 h (top to bottom in each

of citrulline /min /mg protein), this difference was not panel) after MCAO. Slices shown are approximately 0.48, 20.4, 22.3statistically significant. Within-group analysis revealed no and 24.3 mm from bregma, left to right. MR parameters were: TR 3s, TEsignificant difference in the total NOS activity between the 70 ms, 0.313 0.31 mm in-plane resolution, 1 mm slice thickness. The

region of hyperintensity in the left (top) hemisphere of each imagetwo hemispheres.21 delineates the area of ischaemic lesion. Notice that the total area ofCa -independent NOS activity, attributable to the

hyperintensity appears to increase between 24 and 72 h in the controlinducible NOS isoform, was measured in the control brains animal (B), while it remains similarly sized in the AG-treated rat (A).21 21using Ca -free assay buffers which contained the Ca AG, aminoguanidine; MCAO, middle cerebral artery occlusion.

21chelator, EGTA. Under these conditions all Ca -depen-dent NOS activity is inhibited and any activity remainingcould have been produced only by iNOS [45]. The data plasma of AG- and saline — treated rats are given in Table(Fig. 6) show that there was no appreciable iNOS activity 1. When the relative changes in NOx levels over time werein either ischaemic or contralateral hemispheres of saline- compared, no statistically significant difference betweentreated rats. saline- and AG-treated rats was found. Within group

analysis revealed that levels of NOx in plasma were not3.3. Nitric oxide in plasma significantly elevated over baseline at any time-point

following MCAO, suggesting the lack of enhanced releaseThe concentrations of nitrites and nitrates (NOx) in the of NO between 24 and 72 h after MCAO.

D. Cash et al. / Brain Research 905 (2001) 91 –103 97

Fig. 4. Comparison of T relaxation times at 6, 24 and 72 h after MCAO in AG- and saline-treated rats. T values were measured in a 232 pixel region of2 2

two cortical brain areas (upper and lower parietal cortex, top and middle panels) and one subcortical area (caudate–putamen, bottom panel). T maps were2

calculated from PDWI (TE 28 ms, TR 3s) and T WI (TE 70 ms, TR 3s) data using a two point fitting routine. Data are expressed as mean6S.D. of n56 in2

each group. There were no differences in T between the treated and the control group at any time-point (t-test), indicating that AG treatment did not2

change the relaxation characteristics of the tissue. AG, aminoguanidine; MCAO, middle cerebral artery occlusion; PDWI, proton density-weighted imaging;T WI, T -weighted imaging.2 2

3.4. NOS immunoblotting dominant band of ca. 160 kD was revealed in the rat brainextracts (Fig. 7, top panel). The M of this protein isr

To determine whether iNOS was present in the brain consistent with that expected for nNOS [3] and identical inextracts of the control (saline-treated) rats, Western blot this electrophoresis system to that found in the positiveanalysis was performed. In each experiment, identical control (rat pituitary extract). Other bands of lower M arer

samples of the ADP–sepharose eluates were run on typical [49] and likely to represent enzymatically truncatedcorresponding lanes of the two gels and each blot was forms of nNOS — artefacts of tissue extraction.probed with a different antiserum. Samples analysed When probed with anti-iNOS antiserum, none of theincluded normal brain (n52), ischaemic and contralateral eluates contained immunoreactive proteins (Fig. 7, bottomhemispheres of rats killed at 48 (n52) and 72 (n52) h panel). Only the positive control was immunoreactive,after MCAO and positive controls containing iNOS and revealing a band of ca. 138 kD, consistent with publishednNOS proteins. A typical result is shown in Fig. 7. values for iNOS [31,52]. Importantly, anti-nNOS did not

When probed with the anti-nNOS antiserum, a pre- cross-react with the iNOS control preparation, nor did

98 D. Cash et al. / Brain Research 905 (2001) 91 –103

Table 1Effect of aminoguanidine on NO metabolites in the rat plasma

Time following Plasma NOx concentration (mM)MCAO (h)

Saline Aminoguanidine

0 11.6164.5 10.1164.224 9.8462.3 7.9262.748 13.3161.6 11.2264.972 14.9864.0 10.6564.3

Results are expressed as mean6S.D. for n56 in each group. Plasma NOxdid not differ significantly between the groups at any time-point(ANOVA). MCAO, middle cerebral artery occlusion; NO, nitric oxide;NOx, nitrates1nitrites.

Fig. 5. Total, soluble NOS activity in brain extracts of AG- (n54) and 4. Discussionsaline- (n56) treated rats (mean6S.D.) at 72 h after MCAO. The catalytic

3activity of NOS was measured by monitoring the conversion of L-[ H]- Studies of cerebral perfusion and blood flow have shown3arginine to L-[ H]-citrulline. All measurements were performed in dupli-that a hypoperfused, but temporarily viable penumbracate. Data are expressed as mean6S.D., n5number of animals. Theresurrounds the irreversibly damaged core of ischaemic brainwere no significant differences in NOS activity between the two groups or

between the hemispheres within each group (t-test). AG, aminoguanidine; tissue produced subsequent to thromboembolic strokeMCAO, middle cerebral artery occlusion; NOS, nitric oxide synthase. [2,20]. Over time this penumbral region may be recruited

into the ischaemic core, inflicting additional permanenttissue damage [13,33,36]. Paradoxically, post-ischaemicdamage may be exacerbated by reperfusion, which can

anti-iNOS reveal nNOS (i.e. lane 7 of panel A is blank and occur either spontaneously or as a result of thrombolyticlane 8 of panel B is blank). That iNOS was not ‘lost’ therapy [14]. The aim of this study was to evaluate theduring ADP–sepharose chromatography was confirmed; neuroprotective effects of aminoguanidine (AG) therapy onimmunoblotting of crude extracts gave qualitatively similar post-ischaemic brain damage in a rodent model of transientresults, but had more non-specific, lower M bands, i.e. focal ischaemia. In an attempt to mimic a reasonabler

more artefacts of proteolysis (data not shown). These clinical scenario, we started the AG treatment 6 h after theresults indicate that iNOS was not expressed in these initial ischaemic event. Our results show that AG at-brains at either 48 or 72 h following the ischaemic insult. tenuated the progression of tissue damage when compared

3 3Fig. 6. NOS activity in brain extracts of control (saline-treated) rats, as revealed by conversion L-[ H]-arginine to L-[ H]-citrulline. Grey bars representNOS activity in the ischaemic hemisphere and white bars represent activity in the contralateral (non-lesioned) hemisphere. Assays were performed: in the

21 21presence of all NOS cofactors (n56), measuring total soluble NOS activity (neuronal1inducible); in the absence of Ca and in the presence of a Ca21chelator, EGTA (n53), measuring inducible NOS activity only and in the presence of NOS inhibitor, L-NAME (n54). Notice that in absence of Ca

(EGTA), NOS activity was completely abolished and the values did not differ from those obtained in assays containing L-NAME. All measurements wereperformed in duplicate. Data are expressed as mean6S.D., n, number of animals. MCAO, middle cerebral artery occlusion; NOS, nitric oxide synthase;L-NAME, N-nitro-L-arginine methyl ester.

D. Cash et al. / Brain Research 905 (2001) 91 –103 99

Fig. 7. Western blots of control (saline-treated) rat brain extracts. Extracts were enriched for soluble NOS by ADP–sepharose affinity chromatography. Theblot in the upper panel has been probed with anti-nNOS antiserum; in the lower panel with anti-iNOS antiserum. Lanes show: 1, normal brain; 2, ischaemichemisphere at 48 h post-MCAO; 3, contralateral hemisphere at 48 h post-MCAO; 4, ischaemic hemisphere at 73 h post-MCAO; 5, contralateral hemisphereat 73 h post-MCAO; 6, iNOS-positive control; 7, nNOS-positive control; 8, buffer only. The anti-nNOS antiserum revealed a predominant band of ca. 160kD in all brain extracts (upper panel, lanes 1–5), which is of identical M to the positive control (lane 7). The anti-iNOS antiserum did not detect anyr

immunoreactive proteins in any of the samples (lower panel); only the positive control was immunoreactive, revealing a band of ca. 138 kD (lane 6).Anti-nNOS antiserum did not cross-react with the iNOS control preparation, nor did anti-iNOS reveal nNOS (i.e. lane 6 of the upper blot is blank and lane7 of the lower blot is blank). The results confirm the absence of iNOS at either 48 or 72 h following the MCAO. MCAO, middle cerebral artery occlusion;NOS, nitric oxide synthase (n, neuronal; i, inducible).

100 D. Cash et al. / Brain Research 905 (2001) 91 –103

with saline controls (Fig. 2) and suggest that AG may have results, together with the results of measurements of NOxprotected portions of the ischaemic penumbra from ir- and NOS activity, confirm the absence of iNOS in thereversible damage. These results are in agreement with brains of the rats in our study, between 24 and 72 hother studies of cortical injury [53] and ischaemia [17,56] following MCAO.in which neuroprotection by AG has been reported. The absence of any detectable iNOS activity in our

Neuroprotection by AG has been ascribed primarily to system directly contrasts with the findings of other re-its inhibitory effect on the inducible isoform of nitric oxide search groups working with transient MCAO in ratssynthase (iNOS) [17,28]. Previous studies have implicated [16,56] and mice [10,41]. It is important to note that iNOSnitric oxide (NO) as a contributor to the neuronal damage activity observed in the lesioned cortex in these studies isearly in stroke, i.e. during acute ischaemic events [6,46]. attributed primarily to iNOS expressed in vascular endo-During this phase, persistent activation of the constitutive, thelium of the blood–brain barrier and infiltratingnNOS isoform resulting from excessive glutamate sig- leukocytes. The entry of activated leukocytes is facilitatednalling produces cytotoxic doses of NO. iNOS is not by endothelial damage, apparently caused by inflammatoryimplicated in NO-mediated cytotoxicity during this early, events during or subsequent to ischaemia. Leukocyteacute phase of ischaemia, as this isoform is ordinarily infiltration is, however, not necessarily a consequence ofexpressed in cells only subsequent to induction by immune transient brain ischaemia [43]; such may be the case in oursystem effectors [39]. Aforementioned studies by Iadecola model as well. In transient ischaemia, it is possible thatet al. and Zhang et al. [16,56] using a rodent model have differences in surgical protocols to produce MCAO, espe-indicated that inflammation occurring during reperfusion cially with respect to the invasiveness of the technique,may trigger induction of iNOS within the ischaemic may be of critical importance in terms of the developmentterritory, suggesting that NO produced by this isoform may and progression of local inflammatory reactions. In ourcontribute to post-ischaemic tissue damage. AG treatment model, surgical intervention is both minimal and gentle soof the rats, in addition to reducing lesion volumes (ob- that rats normally recover within 3 to 4 days. (In thisserved by histology), also inhibited iNOS activity (as particular study, we suffered no losses of animals duringshown in biochemical assays of brain homogenates), any of the procedures employed.) The method for occlud-leading these workers to conclude that iNOS is a target of ing the MCA employed by Zhang et al. and Iadecola et al.AG’s neuroprotective action. [16,17,56] (developed by Zea Longa [55]) involves in-

Prompted by these findings, we have investigated traluminal insertion of a thread, the tip of which has beenwhether AG exerts a neuroprotective effect via inhibition rounded by flame. We have used the method developed byof iNOS in our model of ischaemia. Because iNOS activity Koizumi et al. [24] for MCAO, in which the thread tip isis expected to increase levels of NO metabolites in plasma coated with silicone rubber. Laing et al. [27] point out that[11,53], we measured the concentration of nitrates and the Zea Longa thread is more difficult to pass along annitrites in the plasma of AG-treated and control rats. The internal carotid canal and that this method results in aresults of our study show that plasma NOx did not higher incidence of vessel perforation. It is conceivablesignificantly change during 3 days following the MCAO that the flame-rounded thread is more likely to damage thecountering the prediction that NO production would mark- endothelium, triggering local ‘downstream’ inflammatoryedly increase under the conditions of our study. To address reactions which could trigger iNOS induction. We proposethis point more directly, we used a biochemical assay to that the silicone-tipped thread employed in our studymeasure soluble NOS activity in the brains of AG- and produces less endothelial damage, and consequently lesssaline-treated rats. Our results show that total soluble NOS inflammation. This could account for the lack of iNOSactivity was similar in the brains of both AG- and saline- during the first 72 h after MCAO.treated rats (Fig. 5). A small (but not significant) elevation It should further be noted that while iNOS has beenin the contralateral hemisphere of both groups may reflect detected in various CNS pathologies, the temporal profiledecreases in nNOS activity in the ischaemic hemisphere, of its expression varies according to the nature and thedue to neuronal damage. To measure the levels of iNOS in severity of the injury. For example, iNOS activity isthe homogenates, we performed parallel assays in which detectable within 8 h following inflammation induced by

21no Ca was present (a requirement for nNOS but not lipopolysaccharide and interferon gamma [23], although in21iNOS activity). In the absence of Ca , NOS activity was a model of traumatic brain injury, iNOS activity does not

undetectable in either brain hemisphere of saline-treated appear until 73 h after the insult [51]. In ischaemia, iNOSrats (Fig. 6). This result confirmed that all of the activity mRNA was first detectable at 12 h post-occlusion and

21detected in assays containing Ca arises exclusively from iNOS protein at 24 h, with activity peaking between 24the neuronal isoform. Finally, Western blot analysis of and 48 h [15,16,56]. In a murine model of transientsoluble proteins in the brain homogenates revealed the ischaemia iNOS activity is first detected at 48 h afterpresence of only nNOS (Fig. 7). iNOS was not detected in reperfusion, but does not peak until 7 days [10]. In thiseither ischaemic or contralateral brain hemispheres of the study, we did not monitor NO/NOS prior to 24 h post-saline-treated rats, either at 72 or 48 h after MCAO. These MCAO (first post-MCAO plasma sample). It is therefore

D. Cash et al. / Brain Research 905 (2001) 91 –103 101

possible that we would have missed the appearance of reported that AG-mediated inhibition of PAO and theiNOS had it been expressed only transiently during the first resultant decrease in toxic by-product, 3-aminopropanal,24 h. Extensive survey of the literature on iNOS expres- was associated with neuroprotection in the ischaemic brainsion in brain injury and ischaemia indicate that, where damage [19]. In addition, several authors suggest thatpresent, iNOS activity typically persists until at least 24 h aberrant polyamine metabolism may be a consequence offollowing the insult, and in most cases remains detectable ischaemia [22,42] so, by inhibiting key enzymes in poly-until 2 to 3 days later [15,16,23]. To conclude, based on amine metabolism, AG may produce a neuroprotectivethe evidence from three separate approaches — measure- effect. Another mechanism by which AG may effectment of plasma NOx, Western blotting of homogenate and neuroprotection is by free radical scavenging. AG has beenNOS activity assays — it is clear in our study that iNOS shown to scavenge free radicals (H O and peroxynitrite)2 2

was undetectable (if not absent) from 24 h afterwards. in vitro [54], although its in vivo efficacy as a free radicalBased on precedents in the literature, it is our opinion that scavenger has to be investigated.transient expression of iNOS restricted to the first 24 h in It should also be emphasised that the methodology usedthe rats of our study was highly unlikely. to assess the degree of neuroprotection may contribute to

Several recent investigations have suggested that differences in experimental findings between studies. Inhalothane, which was the anaesthetic used in this study and the aforementioned AG studies, ischaemic lesions weresimilar anaesthetics may interact with the NO signalling measured using histological methods. This methodologysystem. But the results of these studies are contradictory, cannot examine changes in an individual animal over timeand are unlikely to bear on the outcome of our work. For and may thus overlook subtle, but potentially clinicallyexample, Sjakste et al. [47] reported an increase in cerebral significant, changes in the volume of the damaged area. WeNO concentration upon anaesthesia with halothane, as have attempted to circumvent this limitation by using serialmeasured by electron paramagnetic resonance. In contrast, MRI to monitor lesion progression over a 72-h period. TheLoeb et al. compared isoflurane and halothane anaesthesia, application of non-invasive imaging techniques is advan-finding no changes in basal NO concentration under tageous in that it substantially reduces the number ofhalothane while isoflurane caused a small but significant animals required to determine a drug’s efficacy. Also,increase [29]. In addition to gaseous anaesthetics, several importantly, the use of MRI facilitates comparison withmechanistically dissimilar agents have also been found to similar studies of human stroke and the extrapolation ofinteract with NO system; these widespread effects of data from animal experimental systems to human clinicalanaesthetics are summarised by Johns [21]. While these situations. Quantitative MRI is now increasingly beingstudies emphasise the need for care in choice of anaes- implemented in the clinical setting and, provided that carethetic in studies of NO metabolism, they do not com- is taken with the design of research protocols, this methodpromise the findings of the present study. Here we can serve as a bridge between the laboratory and the cliniccompared the response to AG of two identically treated to evaluate the extent of ischaemic damage and thegroups of rats and any effect of halothane on basal NO response to therapy.would have contributed equally to the physiological state In summary, although underscoring our limited knowl-in both groups. edge of AG’s mechanism of action, the data from our

The results of this study contradict suggestions that high study supports the findings by other workers that AGlevels of NO generated by iNOS account for much of the exerts neuroprotective actions in stroke. Given that AG isneuronal damage in the late, post-ischaemic phase of considered relatively non-toxic and is already under in-MCAO. Nonetheless, AG was neuroprotective in our vestigation in human trials, examination of its potentialsystem. Although this does not preclude a role for AG as a therapeutic value in stroke patients seems plausible in theneuroprotective agent via its ability to inhibit iNOS, our near future.findings indicate that other targets for the neuroprotectiveactions of AG must be present. It is worth noting that in amodel of permanent occlusion, AG treatment administered Acknowledgementsat 1 or 2 h after the ischaemia leads to a significantreduction in lesion volumes at 24 h after the occlusion [5]. The authors are grateful to Drs Maurice Elphick, IanBecause iNOS induction normally requires at least 8 to 12 Jones and Richard Melarange of Queen Mary and West-h, the neuroprotection observed here by AG is likely to field College (QMW), University of London, and Nicholashave been exerted on a target other than iNOS. AG is Bradley, Audrey Dooley and Mike Jacobs of Royal Freeknown to act on a variety of other cellular metabolic Hospital School of Medicine, University of London, forpathways and some of these actions may be relevant in providing the facilities and assistance with nitric oxidebiochemical events during ischaemia. For example, AG experiments. Assistance with animal care was provided byinhibits polyamine metabolising enzymes, polyamine oxi- Sue Keefe and Zoe Coade; assistance with MRI wasdase (PAO) and diamine oxidase which evidently are provided by Drs Paul Kinchesh and Sean Smart (QMW).targets for neuroprotective effects of AG. It has been The imaging spectrometer was provided by the University

102 D. Cash et al. / Brain Research 905 (2001) 91 –103

Batliwalla, J.M. Dubinsky, C. Iadecola, H. Wang, P.K. Gregersen,of London Intercollegiate Research Service scheme and isJ.W. Eaton, K.J. Tracey, Cerebral ischemia enhances polyaminelocated at QMW. This work was supported by a Medicaloxidation: identification of enzymatically formed 3-aminopropanal

Research Council, UK, project grant (no. G9426413N). as an endogenous mediator of neuronal and glial cell death, J. Exp.PMWB was Wolfson Senior Lecturer in Stroke Medicine Med. 188 (1998) 327–340.and is the Stroke Association Professor of Stroke Medi- [20] Q. Jiang, M. Chopp, Z.G. Zhang, R.A. Knight, M. Jacobs, J.P.

Windham, D. Peck, J.R. Ewing, K.M. Welch, The temporal evolu-cine.tion of MRI tissue signatures after transient middle cerebral arteryocclusion in rat, J. Neurol. Sci. 145 (1997) 15–23.

[21] R.A. Johns, Nitric oxide, cyclic guanosine monophosphate, and theReferences anesthetic state, Anesthesiology 85 (1996) 457–459.

[22] T.D. Johnson, Polyamines and cerebral ischemia, Prog. Drug Res.50 (1998) 193–258.[1] S. Archer, Measurement of nitric oxide in biological models,

[23] Y. Kitamura, H. Takahashi, Y. Matsuoka, I. Tooyama, H. Kimura, Y.FASEB J. 7 (1993) 349–360.Nomura, T. Taniguchi, In vivo induction of inducible nitric oxide[2] A.E. Baird, S. Warach, Magnetic resonance imaging of acute stroke,synthase by microinjection with interferon-gamma and lipopolysac-J. Cereb. Blood Flow Metab. 18 (1998) 583–609.charide in rat hippocampus, Glia 18 (1996) 233–243.[3] D.S. Bredt, P.M. Hwang, C.E. Glatt, C. Lowenstein, R.R. Reed, S.H.

[24] J. Koizumi, Y. Yoshida, T. Nakazawa, G. Oonedda, ExperimentalSnyder, Cloned and expressed nitric oxide synthase structurallystudies of ischaemic brain oedema. 1. A new experimental model ofresembles cytochrome P-450 reductase, Nature 351 (1991) 714–cerebral embolism in which recirculation can be introduced in the718.ischaemic area, Jpn. J. Stroke 8 (1986) 1–8.[4] D.S. Bredt, H.H.H.W. Schmidt, The citrulline assay, in: M. Feelisch,

[25] W.J. Koroshetz, M.A. Moskowitz, Emerging treatments for stroke inJ.S. Stamler (Eds.), Methods in Nitric Oxide Research, J. Wiley,humans, Trends Pharmacol. Sci. 17 (1996) 227–233.New York, 1996, pp. 177–186.

[26] U.K. Laemmli, Cleavage of structural proteins during the assembly[5] K.M. Cockroft, M. Meistrell, G.A. Zimmerman, D. Risucci, O.of the head of bacteriophage T4, Nature 227 (1970) 680–685.Bloom, A. Cerami, K.J. Tracey, Cerebroprotective effects of amino-

guanidine in a rodent model of stroke, Stroke 27 (1996) 1393–1398. [27] R.J. Laing, J. Jakubowski, R.W. Laing, Middle cerebral artery[6] T. Dalkara, M.A. Moskowitz, The complex role of nitric oxide in occlusion without craniectomy in rats. Which method works best?

the pathophysiology of focal cerebral ischemia, Brain Pathol. 4 Stroke 24 (1993) 294–297.(1994) 49–57. [28] L. Lecanu, C. Verrecchia, I. Margaill, R.G. Boulu, M. Plotkine, iNos

[7] D.K. Das, N. Maulik, Antioxidant effectiveness in ischemia-reperfu- contribution to the Nmda-induced excitotoxic lesion in the ratsion tissue injury, Methods Enzymol. 233 (1994) 601–610. striatum, Br. J. Pharmacol. 125 (1998) 584–590.

[8] T.J. DeGraba, The role of inflammation after acute stroke: utility of [29] A.L. Loeb, N.R. Raj, D.E. Longnecker, Cerebellar nitric oxide ispursuing anti-adhesion molecule therapy, Neurology 51 (1998) increased during isoflurane anesthesia compared to halothane anes-S62–S68. thesia: a microdialysis study in rats, Anesthesiology 89 (1998)

[9] U. Dirnagl, C. Iadecola, M.A. Moskowitz, Pathobiology of is- 723–730.chaemic stroke: an integrated view, Trends Neurosci. 22 (1999) [30] S.A. Loddick, N.J. Rothwell, Neuroprotective effects of human391–397. recombinant interleukin-1 receptor antagonist in focal cerebral

[10] M. Grandati, C. Verrecchia, M.L. Revaud, M. Allix, R.G. Boulu, M. ischaemia in the rat, J. Cereb. Blood Flow Metab. 16 (1996)Plotkine, Calcium-independent NO-synthase activity and nitrites / 932–940.nitrates production in transient focal cerebral ischaemia in mice, Br. [31] C.J. Lowenstein, C.S. Glatt, D.S. Bredt, S.H. Snyder, Cloned andJ. Pharmacol. 122 (1997) 625–630. expressed macrophage nitric oxide synthase contrasts with the brain

[11] M.B. Grisham, G.G. Johnson, J.R. Lancaster, Quantitation of nitrate enzyme, Proc. Natl. Acad. Sci. USA 89 (1992) 6711–6715.and nitrite in extracellular fluids, Methods Enzymol. 268 (1996) [32] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein237–246. measurement with a folin phenol reagent, J. Biol. Chem. 193 (1951)

[12] V. Hampl, C.L. Walters, S.L. Archer, Determination of nitric oxide 265–275.by the chemiluminescence reaction with ozone, in: M. Feelisch, J.S. [33] T.L.C. Luvisotto, G.R. Sutherland, Animal models of stroke, in: H.Stamler (Eds.), Methods in Nitric Oxide Research, J. Wiley, New Bachelard (Ed.), Advances in Neurochemistry, Vol. 8, Plenum Press,York, 1996, pp. 177–186. New York, 1997, pp. 117–143.

[13] W.D. Heiss, R. Graf, M. Grond, J. Rudolf, Quantitative neuroimag- [34] Z. Makita, H. Vlassara, E. Rayfield, K. Cartwright, E. Friedman, R.ing for the evaluation of the effect of stroke treatment, Cerebrovasc. Rodby, A. Cerami, R. Bucala, Hemoglobin-age — a circulatingDis. 8 (1998) 23–29. marker of advanced glycosylation, Science 258 (1992) 651–653.

[14] W.D. Heiss, A. Thiel, M. Grond, R. Graf, Which targets are relevant [35] H. Memezawa, H. Minamisawa, M.L. Smith, B.K. Siesjo, Ischemicfor therapy of acute ischemic stroke? Stroke 30 (1999) 1486–1489. penumbra in a model of reversible middle cerebral artery occlusion

[15] C. Iadecola, X.H. Xu, F.Y. Zhang, E.E. Elfakahany, M.E. Ross, in the rat, Exp. Brain Res. 89 (1992) 67–78.Marked induction of calcium-independent nitric-oxide synthase [36] H. Memezawa, H. Minamisawa, M.L. Smith, B.K. Siesjo, Ischemicactivity after focal cerebral-ischemia, J. Cereb. Blood Flow Metab. penumbra in a model of reversible middle cerebral artery occlusion15 (1995) 52–59. in the rat, Exp. Brain Res. 89 (1992) 67–78.

[16] C. Iadecola, F.Y. Zhang, R. Casey, H.B. Clark, M.E. Ross, Inducible [37] J. Mintorovitch, M.E. Moseley, L. Chileuitt, H. Shimizu, Y. Cohen,nitric oxide synthase gene expression in vascular cells after transient P.R. Weinstein, Comparison of diffusion- and T2-weighted MRI forfocal cerebral ischemia, Stroke 27 (1996) 1373–1380. the early detection of cerebral ischemia and reperfusion in rats,

[17] C. Iadecola, F.Y. Zhang, X.H. Xu, Inhibition of inducible nitric Magn. Reson. Med. 18 (1991) 39–50.oxide synthase ameliorates cerebral ischemic damage, Am. J. [38] J.P. Mohr, J.C. Gautier, D. Hier, R.W. Stein, Middle cerebral artery,Physiol. Regul. Integr. Comp. Physiol. 37 (1995) R286–R292. in: H.J.M. Barnett, B.M. Stein, J.P. Mohr, F.M. Yatsu (Eds.), Stroke.

[18] S. Imaizumi, V. Woolworth, R.A. Fishman, P.H. Chan, Liposome- Pathophysiology, Diagnosis and Management, Vol. 1, Churchillentrapped superoxide dismutase reduces cerebral infarction in Livingstone, New York, 1986, pp. 377–450.cerebral ischemia in rats, Stroke 21 (1990) 1322–1327. [39] C. Nathan, Q.W. Xie, Regulation of biosynthesis of nitric oxide, J.

[19] S. Ivanova, G.I. Botchkina, Y. Al-Abed, M. Meistrell III, F. Biol. Chem. 269 (1994) 13725–13728.

D. Cash et al. / Brain Research 905 (2001) 91 –103 103

[40] L. Packer, H.M.T. Yoshikawa (Eds.), Free Radicals in Brain [49] D.R. Springall, V. Riveros-Moreno, L. Buttery, A. Suburo, A.E.Physiology and Disorders, Academic Press, San Diego, London, Bishop, M. Merret, S. Moncada, J.M. Polak, Immunological de-1996. tection of nitric oxide synthase(s) in human tissues using hetero-

[41] S. Parmentier, G.A. Bohme, D. Lerouet, D. Damour, J.M. Stut- logous antibodies suggesting different isoforms, Histochemistry 98zmann, I. Margaill, M. Plotkine, Selective inhibition of inducible (1992) 259–266.nitric oxide synthase prevents ischaemic brain injury, Br. J. Phar- [50] D.J. Stuehr, O.W. Griffith, Purification, assay and properties ofmacol. 127 (1999) 546–552. mammalian nitric oxide synthases, in: M. Feelisch, J.S. Stamler

[42] W. Paschen, Polyamine metabolism in reversible cerebral ischemia, (Eds.), Methods in Nitric Oxide Research, J. Wiley, New York,Cerebrovasc. Brain Metab. Rev. 4 (1992) 59–88. 1996, pp. 177–186.

[43] E.E. Peters, L. Marks, G. Gartshore, W.L. Maxwell, D.I. Graham, [51] K. Wada, K. Chatzipanteli, S. Kraydieh, R. Busto, W.D. Dietrich,I.M. Macrae, No evidence for significant leukocyte involvement in Inducible nitric oxide synthase expression after traumatic braintransient focal ischaemia, J. Cereb. Blood Flow Metab. 19 (1999) injury and neuroprotection with aminoguanidine treatment in rats,S574. Neurosurgery 43 (1998) 1427–1436.

[44] S.A. Roussel, N. van Bruggen, M.D. King, J. Houseman, S.R. [52] Q.W. Xie, H.J. Cho, J. Calaycay, R.A. Mumford, K.M. Swiderek,Williams, D.G. Gadian, Monitoring the initial expansion of focal T.D. Lee, A. Ding, T. Troso, C. Nathan, Cloning and characteriza-ischaemic changes by diffusion-weighted MRI using a remote tion of inducible nitric oxide synthase from mouse macrophages,controlled method of occlusion, NMR Biomed. 7 (1994) 21–28. Science 256 (1992) 225–228.

[45] M. Salter, R.G. Knowles, S. Moncada, Widespread tissue dis- [53] K. Yamanaka, E. Kumura, K. Iwatsuki, T. Yoshimine, Y. Masana, T.21tribution, species distribution and changes in activity of Ca - Hayakawa, T. Shiga, H. Kosaka, Increase in plasma nitric oxide

21dependent and Ca -independent nitric oxide synthases, FEBS Lett. end-products following rat cortical injury, Neurosci. Lett. 194291 (1991) 145–149. (1995) 124–126.

[46] B.K. Siesjo, K. Katsura, Q. Zhao, J. Folbergrova, K. Pahlmark, P. [54] G. Yildiz, A.T. Demiryurek, I. Sahin-Erdemli, I. Kanzik, Com-Siesjo, M.L. Smith, Mechanisms of secondary brain damage in parison of antioxidant activities of aminoguanidine, methylguanidineglobal and focal ischemia: a speculative synthesis, J. Neurotrauma and guanidine by luminol-enhanced chemiluminescence, Br. J.12 (1995) 943–956. Pharmacol. 124 (1998) 905–910.

[47] N. Sjakste, L. Baumane, D. Meirena, L. Lauberte, M. Dzintare, I. [55] E. Zea Longa, P.R. Weinstein, S. Carlson, R. Cummins, ReversibleKalvins, Drastic increase in nitric oxide content in rat brain under middle cerebral artery occlusion without craniectomy in rats, Strokehalothane anesthesia revealed by EPR method, Biochem. Pharmacol. 20 (1989) 84–91.58 (1999) 1955–1959. [56] F.Y. Zhang, R.M. Casey, M.E. Ross, C. Iadecola, Aminoguanidine

[48] S.C. Smart, J.S. Beech, D.J. Lythgoe, P. Kinchesh, J.R. Everett, A.G. ameliorates and L-arginine worsens brain damage from intraluminalSwanson, J. Urenjak, S.C.R. Williams, In vivo assessment of MK- middle cerebral artery occlusion, Stroke 27 (1996) 317–323.801 efficacy in a rat model of transient focal cerebral ischaemia, in: [57] J.A. Zivin, Neuroprotective therapies in stroke, Drugs 54 (1997)J. Krieglstein (Ed.), Pharmacology of Cerebral Ischaemia, 1996. 83–88.