This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Research Report

Blood–arachnoid barrier disruption in experimental ratmeningitis detected using gadolinium-enhancementratio imaging

Hiroyuki Ichikawaa, Kouichi Itohb,c,⁎aThird Institute of New Drug Discovery, Otsuka Pharmaceutical Co., Ltd., Tokushima 771-0192, JapanbLaboratory of Molecular and Cellular Neurosciences, Kagawa School of Pharmaceutical Sciences,Tokushima Bunri University, 1314-1, Kagawa 769-2193, JapancLaboratory for Brain Science, Kagawa School of Pharmaceutical Sciences, Tokushima Bunri University, 1314-1, Kagawa 769-2193, Japan

A R T I C L E I N F O A B S T R A C T

Article history:Accepted 15 March 2011Available online 22 March 2011

Disruption of the central nervous system (CNS) barriers is one of themajor pathophysiologicalconsequences of bacterial meningitis. The increase in the permeability of the CNS barrierscaused by the disruption is thought to contribute to the development of adverse neurologicaloutcomes. We have established a method by which the CNS barrier permeability can bedemonstrated by the gadolinium-enhancement ratio (GER) calculated from the T1 weightedimage (T1WI) which is based on gadolinium-enhancedmagnetic resonance imaging (GdEMRI).The present study examined the disruption of CNS barriers such as blood–brain barrier (BBB),blood–cerebrospinal fluid barrier (BCSFB) and blood–arachnoid barrier (BAB) in rats withmeningitis induced by lipopolysaccharide (LPS)- or interleukin (IL)-1β. Four hours afterintracisternal injection of LPS or IL-1β, severe disruption of the BAB, but not the BBB or BCSFB,was observed. This suggests that the BAB, rather than the BBB or BCSFB, plays a key role in theinflux of blood-borne cells and substances during meningitis. The BAB is therefore morevulnerable to disruption than the BBB or BCSFB duringmeningitis in rats. In addition, GdEMRIwith GER imaging analysis appears to be useful in spatio-temporal studies on the function ofthe CNS barriers under various physiological and pathological conditions.

© 2011 Elsevier B.V. All rights reserved.

Keywords:Blood–arachnoid barrierMRIGd-DTPAMeningitisLPSIL-1β

1. Introduction

Meningitis is the inflammation of the protective membranesknown as the meninges in the central nervous system (CNS),which may be caused by infection with bacteria, and occurs in

infants, children and adults (Sáez-Llorens and McCracken,2003). Bacterial meningitis is now a top 10 infectious cause ofdeath worldwide (Grimwood et al., 2000). The mortality rate ofbacterial meningitis has remained 5–40% and causes neurolog-ical sequelae such as cranial nerve damage with cerebral

B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

⁎ Corresponding author at: Laboratory of Molecular and Cellular Neurosciences, and Laboratory for Brain Science, Kagawa School ofPharmaceutical Sciences, Tokushima Bunri University, 1314-1, Kagawa 769-2193, Japan. Fax: +81 87 894 0181.

E-mail address: itoh@kph.bunri-u.ac.jp (K. Itoh).Abbreviations: MRI, magnetic resonance imaging; BBB, blood–brain barrier; BCSFB, blood–cerebrospinal fluid barrier; BAB, blood–

arachnoid barrier; Gd, gadolinium; Gd-DTPA, gadolinium-diethylenetriamine pentaacetic acid; GdER, gadolinium-enhancement ratio;SAS, subarachnoid space; GdEMRI, gadolinium-enhanced MRI; TJ, tight junction; CSF, cerebral–spinal fluid; HRP, horseradish peroxidase;WBC, white blood cell; CNS, central nervous system; SI, signal intensity

0006-8993/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.brainres.2011.03.035

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

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

Author's personal copy

infarction, learning disabilities and/or mental retardation in upto 30% of the survivors (Mustafa et al., 1990; Quagliarello andScheld, 1992; Tunkel and Scheld, 1993). The inflammation thatoccurs in the subarachnoid space (SAS) during bacterial menin-gitis is not a direct result of bacterial infection but can beattributed to the response of the immune system to the entranceofbacteria into theCNS (Tunkel et al., 1990).Whencomponentsofthe bacterial cell membrane such as lipopolysaccharide (LPS) areidentified by the immune cells of the brain suchas astrocytes andmicroglia, they respondby releasing proinflammatorymediators,suchascytokinesandprostaglandins.Thesesubsequently lead toan increase in the permeability of the blood–brain barrier (BBB),which then triggers trans-endothelial migration of blood-bornecells such as neutrophils, mononuclear leukocytes, and macro-phages, and the leakage of plasma proteins that further lead toinflammation of the meninges, vasogenic cerebral edema, andcytotoxic edema (Tunkel and Scheld, 1993; Pfister et al., 1994;Van Furth et al., 1996). The severe pathophysiologic alterationsin meningitis are thought to initiate or contribute to adverseneurological outcomes, but how the bacteria cross the BBB andenter the CNS is not clearly understood.

Magnetic resonance imaging (MRI) scans in patients withclinical meningitis are primarily performed at a later stage to

assess and diagnose complications of meningitis (Sáez-Llorensand McCracken, 2003; Jan et al., 2003; Kastrup et al., 2005).However, its application has been limited in studies of experi-mental rat meningitis, and the few studies that have beenperformed were at low resolution (Narayana et al., 1987;Wiesmann et al., 2002). In experimental meningitis, the evalua-tion of the CNS barrier permeability has been performed byseveral methods using Evan's blue dye, horseradish peroxidase(HRP), radioactive tracers andbydeterminationof thewhitebloodcell (WBC) concentrations in cerebral–spinal fluid (CSF) (Quagliar-ello et al., 1991; Jaworowicz et al., 1998). Experimental bacterialmeningitis results in diffuse and unpredictable involvement ofthe brain vasculature and parenchyma that is often complicatedby systemic infection. MRI is able to provide quantifiable in vivodata on BBB function, and brain water distribution (Dijkhuizenand Nicolay, 2003). Conventional MRI provides a sensitivemeasure of tissue structure and water content and, through theuse of an intravenous gadolinium (Gd)MRI contrast agent suchasgadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA), canalsomeasure BBB permeability (Hawkins et al., 1990; Morrissey etal., 1996).In our previous report, we used Gd-enhanced MRI(GdEMRI) in order to sequentially quantify the permeability of theCNS barriers in experimental rat meningitis induce by LPS andinterleukin (IL)-1β (Ichikawa et al., 2010).

In the present study, we have determined the changes inCNS barriers by quantitative analysis of Gd-DTPA leakage intothe brain following IC injection with lipopolysaccharide (LPS) orhuman recombinant IL-1β (hrIL-1β) using GdEMRI and Gd-enhancement ratio imaging (GERI).

2. Results

2.1. Influence of body temperature and the infiltration ofwhite blood cells (WBCs) in experimental meningitis rats

The most typical symptom of bacterial meningitis is fever(Roos and Van de Beek, 2010). The body temperatures of ratsinjected with LPS (38.7±0.1 °C) and hrIL-1β (2.5×104 U; 38.20.1 °C and 5×104 U; 38.6±0.1 °C) were significantly higher thanthose of control (ACSF) rats (Table 1). As the hyperthermiaappears to be a critical determinant in experimental

Table 1 – Body temperaturesof rats injected intracisternallywith LPS and hrIL-1β.

Treatment group Body temperature (°C)

ACSF 37.2±0.1LPS 200 μg 38.7±0.1a)

IL-1β 2.5×104 U 38.2±0.1a)

IL-1β 5×104 U 38.6±0.0a)

IL-1β 5×104 U+Prednisolone 37.6±0.1b),c)

ACSF (5 μl), LPS (200 μg) or hrIL-1β (2.5×104, 5×104 U) was injectedinto the intracisternal cavity of the rats. Prednisolone wasadministered orally to the rats 30 min before the injection ofhrIL-1β (5×104 U). Four hours after the LPS or hrIL-1β treatment,the rectal temperature of each rat was monitored by athermometer. The results represent the means±SE; a)P<0.05,ACSF versus LPS or hrIL-1β, b)P<0.05, hrIL-1β (5×104 U) versushrIL-1β (5×104 U)+Prednisolone (3 mg/kg); and c)not significant,ACSF versus hrIL-1β (5×104 U)+Prednisolone (3 mg/kg).

Fig. 1 – Experimental schedules. ACSF (5 μl), (LPS (200 μg) or hrIL-1β (2.5×104, 5×104 U) was administered into the IC cavity ofrats. Prednisolone was administered orally to the rats 30 min before the injection of hrIL-1β (5×104 U). Four hours after eachtreatment, the rectal temperature of each rat was monitored. Then, T1WI were measured for each rat. Gd-DTPA (0.2 mmol/kg)was administered intravenously. After 5 min, the T1WI were monitored. The GERI were calculated from the T1WI.

143B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

Author's personal copy

meningitis, rats intracisternally (IC) injected with LPS andhrIL-1β can be used as an experimental model for themeningitis. The hyperthermia induced by LPS was equivalentto that by hrIL-1β. Furthermore, the use of prednisolone toclinically treat bacterial meningitis significantly attenuatedthe hyperthermia. In addition, we examined the infiltration ofWBCs into the SAS in these model animals. The infiltration ofWBCs into the CSF collected at 4 h after hrIL-1β (2.5×104 and5 × 104 U) injections were 3.15 ± 0.50 × 103/ml and 4.05±0.74×103/ml, respectively, although they were detected atless than 1×103 cells/ml in the control (ACSF) rats (Fig. 3).Prednisolone also significantly inhibited the infiltration of

WBCs into the CSF of the hrIL-1β 5×104 U-injected rats (2.64±0.78×103/ml) (Fig. 3). These findings indicated that the ratsinjected with LPS and hrIL-1β develop conditions that mimicthe processes in the SAS after bacterial infection causingbacterial meningitis.

2.2. Influence on GERI following injection of LPS or hrIL-1β

In order to determine the effects of injection of LPS or hrIL-1β onthe CNS barriers, semi-quantitative analyses were performedusing GERI based onGdEMRI. The effects on the BBB, BCSFB andBAB were evaluated by the signal intensity (SI) measured at thecerebral cortex (Cx), lateral ventricle (Lv), and peri-Lv as brainparenchyma, andSAS, respectively. TheSI at theperi-Lvand theCxwerenot increasedby the injectionof LPSorhrIL-1β (Fig. 4; K–O,Q, S, U,W).On the other hand, theSI in the Lv (Fig. 4; K–O,Q, S,U,W) and SAS (Fig. 4; K–O, P, R, T, V)were significantly increasedat 4 h after the injection of LPS or hrIL-1β. No Gd-DTPA leakageinto the brain parenchymawas seen regardless of the high SI inthe Lv in rats with experimental meningitis. These resultsindicated that failure of the BBB and BCSFB were not caused bythe experimental meningitis, but that the BAB permeabilityincreased in the SAS.

2.3. Quantitation of CNS barrier breakdown by LPS andhrIL-1β

The dynamic changes in the CNS barriers were quantified by theGd-DTPA concentrations. In the Lv space, the Gd-DTPA concen-trations, 90.1±10.4 μMin theLPS-injected rats, and128.3±8.6 and146.5±4.5 μM, in the hrIL-1β-injected rats (2.5×104 and 5×104 U,respectively) were 4–7 fold higher than that (20.8±7.5 μM) ofcontrol (ACSF) animals (Fig. 5). The concentrations of Gd-DTPA inthebrainparenchymaofperi-LvandCxwerenot increased in theexperimental meningitis models, similar to the results of thesemi-quantitative analyses described above. Using the Gd-DTPAtracer, the BBB and BCSFB were found to be impermeable at 4 hafter LPS and hrIL-1β injections.

Fig. 2 – Reconstruction of the GERI by GdEMRI. The GERI were reconstructed from the MRI without Gd-DTPA and GdEMRIaccording to the formula: GERI=(T1WIw/ Gd-DTPA−T1WIw/o Gd-DTPA)/(T1WIw/o Gd-DTPA)×100.

Fig. 3 – The effects of prednisolone on the WBC infiltrationinto the CSF following IC injection of hrIL-1β. ACSF (5 μl) orhrIL-1β (2.5×104, 5×104 U) was administered into the ICcavity of rats. Prednisolone was administered orally to therats 30 min before the injection of hrIL-1β (5×104 U).□=ACSF, =hrIL-1β (2.5×104 U), =hrIL-1β (5×104 U),■=hrIL-1β (5×104 U)+Prednisolone (3 mg/kg). The resultsrepresent the means±SE. N.D=not detected. *P<0.05,hrIL-1β (5×104 U) versus hrIL-1β (5×104 U)+Prednisolone(3 mg/kg).

144 B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

Author's personal copy

In the SAS, the concentration of Gd-DTPA in the LPS-injected rats was 90.1±10.4 μM, and the concentrations in the2.5×104 and 5×104 U hrIL-1β-injected rats were 96.2±4.5 and122.4±5.9 μM, respectively (Fig. 5). The extent of the BABdisruption after the LPS (200 μg) injection was almost equiv-alent to that of the hrIL-1β (2.5×104 U)-injected rats. Therefore,the data presented here indicate that BAB breakdown, but notbreakdown of the BBB or BCSFB, occurred at 4 h after LPS andhrIL-1β injections.

2.4. Effect of prednisolone on the BAB disruptionby hrIL-1β

To evaluate the effects of prednisolone, a steroidal anti-inflammatory drug used to treat the bacterial meningitis, onthe BAB disruption induced by hrIL-1β (5×104 U), the rats were

administered oral prednisolone 30 min before the hrIL-1βinjection. Prednisolone significantly inhibited the Gd-DTPAleakage into the SAS of rats with experimental meningitisinduced by hrIL-1β (vehicle: 122.4±5.9 μM, versus hrIL-1β plusprednisolone; 70.2±7.2 μM) (Fig. 5).

3. Discussion

In the current investigation, we found that LPS and hrIL-1βincreased the permeability of the BAB, but not that of the BBBor BCSFB at 4 h after injection in rats.

The mechanisms of pathogenesis of meningitis caused bybacteria are largely unknown. After bacteremia, pathogenshave been shown to penetrate into the CSF through the BBBand/or BCSFB to enter the SAS by several mechanisms

Fig. 4 – GERI of the brain taken after LPS and hrIL-1β injection following Gd-DTPA bolus injection. ACSF (5 μl; A, F, K, P, Q), LPS(200 μg; B, G, L, R, S) or hrIL-1β (2.5×104; C, H, M, T, U, 5×104 U; D, I, N, V, W) was administered into the IC cavity of rats.Prednisolone (3 mg/kg) was administered orally to the rats 30 min before the injection of hrIL-1β (5×104 U) (E, J, O, X, Y). Fourhours after each treatment, Gd-DTPA (0.2 mmol/kg) was administered intravenously. After 5 min, the GERIwere calculated fromthe T1WI. The ROI was put in the various brain regions on the GERI (1: Lv, 2: peri-Lv, 3: Cx, 4: SAS). The location of each SAS inGERI (K, L, M, N, O) is shown in magnified GERI (P, R, T, V, X), and the location of each Lv, peri-Lv and Cx in GERI (K, L, M, N, O) isshown in magnified GERI (Q, S, U, W, Y).

145B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

Author's personal copy

(Gottschalk and Segura, 2000; Kim, 2003). The robust inflam-mation elicited by bacterial products such as LPS persists afterthe bacteria are destroyed by the host responses and antibiotictherapy (Kaplan et al., 2004). The LPS produced by meningealpathogens leads to the production of different inflammatorymediators including IL-1β, IL-6, tumor necrosis factor-α andplatelet activating factor, as well as other immune responses,including production of nitric oxide, matrix metalloprotei-nase-2, and prostaglandins (Ramilo et al., 1990; Pignol et al.,1990; Leib and Tauber, 1999). Cytokines such as IL-1β producedby LPS-stimulated cells are thought to be implicated in thefailure of CNS barriers in the meningitis induced by LPS(Jaworowicz et al., 1998; Garabedian et al., 2000).

In the present study, the number of WBCs that penetratedinto the CSF increased 4 h after IC injection of hrIL-1β, andprednisolone significantly suppressed the hrIL-1β-inducedWBCs penetration into the SAS. The ensuing influx of WBCsand altered CNS barrier permeability results in the release ofproteolytic products and toxic reactive oxygen species. CNSbarrier failures allow toxic substances to enter the ventricularsystem and subsequently reach the brain parenchyma. In thepresent experiments, the model rats used represented theseprocesses of CNS barrier failure after inoculation of bacteriainto the CSF of the SAS.

CNS barriers consist of at least the BBB, BCSFB, and BAB,which differ from each other in morphology, constituent cells,locations, and functions (Saunders et al., 2008). As is wellknown, the BBB and BCSFB are composed of endothelial,choroidal plexus, and glial cells that restrict the traffic of most

molecules in and out of the CNS. However, little is knownabout the BAB, which has been the least studied and appearsto be the structurally most complex of the CNS barriers(Saunders et al., 2008).

Although the pathogenesis of bacterialmeningitis is involvedin the dysfunctions of the BBB and BCSFB, it was not knownwhether the BAB was affected during bacterial meningitis(Tunkel et al., 1990). Sometimes, the BAB was confused with orincluded in the BBB and/or BCSFB in previous reports. In thepresent study, we focused on the BAB in the SAS, because theprimary target site of bacterial meningitis is the SAS. The BAB iscomposed of an arachnoid barrier cell layer with numerous tightjunctions (TJs), and is considered to represent an effectivemorphological and physiological meningeal barrier betweenthe CSF in the subarachnoid space and the blood circulation inthedura. The arachnoid barrier layer is always characterizedby adistinct continuousbasal laminaon its inner surface towards theinnermost collagenous portion of the arachnoid (Vandenabeeleet al., 1996). Given the complexity of these structures and thesmall spaces involved, it has been difficult to observe the SAS inliving animals. Our previous report showed the possibility ofdetermining BAB dysfunction in the SAS using GdEMRI/GERI(Ichikawa et al., 2010).

In the present study, the Gd-DTPA concentrations in the Lvwere significantly increased by LPS or hrIL-1β, but there wasminimal leakage into thebrainparenchyma (Fig. 4). These resultssuggest that Gd-DTPAmay not be able to pass into the CSF of theLv, and the high SI on GERI may be caused by Gd-DTPA incapillary vessels of the choroid plexus. This is because if the Gd-DTPA passes through the BCSFB from the capillary vessels of thechoroid plexus, it can diffuse easily into the cerebral parenchy-ma. Therefore, these findings demonstrate that the BBB andBCSFB are not disrupted at 4 h after LPS or hrIL-1β IC injections(Figs. 4 and 5). Other investigations have reported that thedisruption of BBB and BCSFB in these model rats appeared at 6–40 h later (Blamireet al., 2000; Leib et al., 2000).On theotherhand,the Gd-DTPA concentrations in the SAS were significantlyincreased by LPS or hrIL-1β (Figs. 4 and 5). The SAS is surroundedby the arachnoid and the pia mater, having arteries running onthe brain surface as well as veins connected to the venous sinusof the dura mater, and the BAB exists in the subarachnoid andthearachnoidbarrier cell layers (Peters et al., 1991). These vesselswith TJs in the SAS, but without TJs in the dura mater, aresurrounded by the leptomeninges (Friede and Schachenmayr,1978). Although systemically injected Gd-DTPA diffuses into theepidural and intradural spaces, Gd-DTPA is not normallydetected in the SAS. Thus, these findings suggest that thedisruption of the BAB is firstly involved in the inflammation ofthemeninges by inflammatorymediators such as LPS or IL-1β intheSAS, andsecondary leads to severedamage toBBBandBCSFBand cytotoxic cerebral edema.

While the BAB breakdownwas concluded to lead to leakageof Gd-DTPA from the intradural space into the SAS, hyperper-meability of the vessels in the SAS could not be ruled out. Inthis study, it was not clear which route was involved in theleakage of Gd-DTPA into SAS. It is therefore possible that theblood vessels with TJs and the pia mater of SAS function asCNS barriers. Taken together, we propose that there exists ablood–pia mater barrier (BPMB) or a blood–leptomeningesbarrier (BLMB) including the BAB and BPMB as comprehensive

Fig. 5 –TheGd-DTPAconcentration in the variousbrain regionsas determined by the GERI. ACSF (5 μl), LPS (200 μg), or hrIL-1β(2.5×104, 5×104 U) was administered into the IC cavity of rats.Prednisolonewas administered orally to the rats 30minbeforethe injection of hrIL-1β (5×104 U). Four hours after eachtreatment, Gd-DTPA (0.2 mmol/kg) was administeredintravenously. After 5 min, the Gd-DTPA concentration of eachROI (1: Lv, 2: peri-Lv, 3: Cx, 4: SAS) was calculated from theGdEMRI (Fig. 4).□=ACSF, =LPS (200μg), =hrIL-1β(2.5×104 U), =hrIL-1β (5×104 U), ■=hrIL-1β (5×104 U)+Prednisolone (3mg/kg). The differences between the treatmentand control groups were assessed by a one-way ANOVA followedby the two-tailed Dunnett's test. The results represent the means±SE. a)P<0.05, ACSF versus LPS or hrIL-1β b)hrIL-1β (5×104 U)versushrIL-1β (5×104 U)+Prednisolone (3mg/kg). c)Not significant,versus control.

146 B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

Author's personal copy

terms. However, to resolve the morphological existence andfunction of the BPMB or BLMB in the SAS, further ultrastruc-tual and functional studies will be required.

In conclusion, we have demonstrated that disruption of theBAB was present at 4 h after IC injection of LPS or hrIL-1β, butthat the BBB and BCSFB were not disturbed in the rat GdEMRIwith GERI analyses. These findings suggest that the disruptionof the BAB is caused by inflammation of the meninges afteracute infection.

4. Experimental procedures

4.1. Materials

Human recombinant IL-1β (hrIL-1β, 2×107 units/mg protein)was synthesized at the Otsuka Pharmaceutical Co., Ltd.(Tokushima, Japan). Lipopolysaccharide (LPS; E. coli serotype055:B5) was purchased from Sigma-Aldrich Co (St. Louis, MO,USA). Prednisolone was purchased from Wako Pure ChemicalIndustries, Ltd. (Osaka, Japan). Gadolinium-diethylenetriaminepentaacetic acid (Gd-DTPA; Magnevist®) was purchased fromBayer Health Care Pharmaceuticals Inc. (Berlin, German).

4.2. Induction of experimental meningitis in rats

The protocols for all animal experiments were approved by theGuidelines for Animal Care and Use at Otsuka PharmaceuticalCo., Ltd. Forty-five maleWistar rats (300–350 g) were purchasedfrom Charles River (Tsukuba, Japan) and were used in theseexperiments. Ratswere housed in a room (23.0±1.0 °C) that waslit for 12 h (07:00–19:00) daily. Rats were allowed free access totap water and pellet food (MF, Oriental Yeast Co., Ltd., Tokyo,Japan). All efforts were made to minimize the number ofanimals used and their suffering. The rats were anesthetizedwith Tribromoethanol (Avertin®; 25mg/kg; Wako Pure Chem-ical Industries, Ltd., Osaka, Japan) and placed in a stereotaxicframe (SR-5R, Narishige, Japan). A microinjection pump wasused to inject 5 μl of artificial CSF (ASCF); containing LPS (200 μg)or hrIL-1β (2.5×104 and 5×104 U) into the IC cavity over 5 min(Fig. 1) (Jaworowicz et al., 1998). After the infusion, the needlewas left in theplace for 10 minand then removed.Theboneholewas sealed with bone wax, the scalp wound was sutured, andthe animal was placed in a box with free access to food andwater. The control rats were injected with the same volume ofACSF (120mM NaCl, 3.0 mM KCl, 1.2 mM CaCl2, 1.2 mM MgCl2,0.67 mM NaH2PO4, and 0.3 mM Na2HPO4, pH 7.4). Body temper-ature was monitored and maintained (37.5±0.5 °C) throughoutthe procedure using a rectal thermocouple probe and athermostatically controlledpad (BWT-100, Bio-ResearchCenter,Nagoya, Japan).

4.3. Measurements of body temperature in experimentalmeningitis rats

Four hours after the IC injection of LPS or hrIL-1β, the rectal(core) temperature of each ratwasmonitored by a thermometer(BWT-100) for 20 s (Fig. 1). The copper-constantan thermocoupleas a rectal probe was inserted into the colon about 1 cm beyondthe anus.

4.4. In vivo MRI acquisitions

MRI measurements were performed using the Inova 300Imaging System (7 T) with the VNMRj 1.1d software package(Varian, Inc., Palo Alto, CA, USA). A volume coil and a surfacecoil (RAPID Biomedical GmbH, Germany) were used for signaltransmission and detection, respectively. During MRI mea-surements, a rat was anesthetized by inhalation of 1–1.5%halothane in a N2/O2 (70:30) mixture gas via a facemask.During the MRI measurements, the body temperature wasmeasured using a rectal thermocouple and it was keptconstant with a feedback-controlled warm-water pad (Yama-shita Tech System, Tokushima, Japan) connected to a rectalprobe (Photon Control Inc., Burnaby, BC, Canada). Four hoursafter each treatment, Gd-DTPA at a dose of 0.2 mmol/kg wasadministrated intravenously. Acquisition of Gd-enhancedMRI(GdEMRI) images was obtained before, and at 5 min after theadministration of Gd-DTPA (Fig. 2). The T1 weighed images(T1WI) were acquired by a 2-dimensional multi-slice spin echosequence using the following parameters: repetition time (TR)500 ms; echo time (TE)=10 ms; number of scans=2; slicethickness=2 mm; matrix size=256×256; the images werezero-filled to 512×512; and the field of view was 25×25 mm.

4.5. Imaging of GER

The GERI was calculated from the MRI without Gd-DTPA andGdEMRI according to the formula (Fig. 2) (Erwin et al., 2002;Ichikawa et al., 2010):

GER = T1WIw=Gd−DTPA−T1WIw=oGd−DTPA� �

= T1WIw=oGd−DTPA� �

× 100:

4.6. Measurements of signal intensities in the region ofinterest

To evaluate the disruption of the CNS barriers from GERI, theregionof interest (ROI)wasselected in theLv, theperi-Lvand theCx as brain parenchyma, as well as the SAS space (Fig. 4C;Arrows 1, 2, 3 and 4). The SI in the ROI wasmeasured at 5, 15, 30and 60min after injection with Gd-DTPA. The SI reached thehighest at 5 min, and then decreased 30% at 15min. After30 min, the SI by Gd-DTPAwas not detectable in SAS. Each SI inthe ROI was measured for the GERI using the Image Brawersoftware program (Varian, Inc., Palo Alto, CA, USA).

4.7. Measurements of the Gd-DTPA concentration

The quantitative assessment of the Gd-DTPA concentrationsin the brain regions was calculated from the SI of GERIaccording to a previously described method (Ichikawa et al.,2010). A linear correlation between the signal intensities ofGERI and Gd-DTPA concentrations (0–150 μM) was calculated.This relational formula showed that the GdER value can beused to quantitatively evaluate the Gd-DTPA concentration.

4.8. Administration of prednisolone in hrIL-1β-inducedmeningitis rats

Prednisolone was suspended in 0.5% carboxymethyl cellulose(CMC) at a concentration of 0.6 mg/ml. The solution was

147B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

Author's personal copy

administered orally to the rats 30 min before the IC injection ofartificial CSF containing hrIL-1β (Fig. 1). The CMC (0.5%)solution was administered to the control group.

4.9. Infiltration of white blood cells into the CSF

The CSF was collected by aspiration from the cisterna magnavia puncture of the allanto-occipital membrane at 4 h after thehrIL-1β injection (Fig. 1). The CSF samples were used to countwhite blood cells (WBCs) by the trypan blue exclusion method(Tunkel and Scheld, 1993).

4.10. Statistical analysis

The results were expressed as the means±SE. The statisticalanalyses were performed using the SAS® software package (SASInstitute, Japan). Thedifferences between the treated and controlgroups were assessed by one-way ANOVA followed by the two-tailed Dunnett's test, with statistical significance set at P<0.05.

Acknowledgments

The authors thank Dr. Koichi Ishikawa (Institute for Molecularand Cellular Regulation, Gunma University) for valuable adviceon the CNS barriers, Dr. Makoto Ishikawa (Otsuka Pharmaceu-tical Co., Ltd) for support inperforming theMRIexperimentsandDr. Hironobu Ishiyama (Otsuka Pharmaceutical Co., Ltd) forvaluable suggestions on this project. This research was finan-cially supported by Otsuka Pharmaceutical Co., Ltd.

R E F E R E N C E S

Blamire, A.M., Anthony, D.C., Rejagopalan, B., Sibson, N.R., Perry,V.H., Styles, P., 2000. Interleukin 1β-induced changes inblood–brain barrier permeability, apparent diffusioncoefficient, and cerebral blood volume in the rat brain: amagnetic resonance study. J. Neurosci. 20, 8153–8159.

Dijkhuizen, R.M., Nicolay, K., 2003. Magnetic resonance imaging inexperimental models of brain disorders. J. Cereb. Blood FlowMetab. 23, 1383–1402.

Erwin, L.A., Nicolay, B.K., Goldschmeding, R., Koomans, H.A., Joles,J.A., 2002. Reduction of cerebral injury in stroke-pronespontaneously hypertensive rats by amlodipine. Eur. J.Pharmacol. 444, 75–81.

Friede, R.L., Schachenmayr, W., 1978. The origin of subduralneomembranes: II. Fine structural of neomembranes. Am. J.Pathol. 92, 69–84.

Garabedian, B.V., Lemaigre-Dubreuil, Y., Mariani, J., 2000. Centralorigin of IL-1beta produced during peripheral inflammation:role of meninges. Mol. Brain Res. 75, 259–263.

Gottschalk, M., Segura, M., 2000. The pathogenesis of themeningitis caused by Streptococcus suis: the unresolvedquestions. Vet. Microbiol. 76, 259–272.

Grimwood, K., Anderson, P., Anderson, V., Tan, L., Nolan, T., 2000.Twelve year outcomes following bacterial meningitis: furtherevidence for persisting effects. Arch. Dis. Child. 83, 111–116.

Hawkins, C.P., Munro, P.M., MacKenzie, F., Kesselring, J., Tofts, P.S.,du Boulay, E.P., Landon, D.N., McDonald, W.I., 1990. Durationand selectivity of blood–brain barrier breakdown in chronicrelapsing experimental allergic encephalomyelitis studied bygadolinium-DTPA and protein markers. Brain 113, 365–378.

Ichikawa, H., Ishikawa, M., Fukunaga, M., Ishikawa, K.,Ishiyama, H., 2010. Quantitative evaluation ofblood–cerebrospinal fluid barrier permeability in the rat withexperimental meningitis using magnetic resonance imaging.Brain Res. 1321, 125–132.

Jan, W., Zimmerman, R.A., Bilaniuk, L.T., Hunter, J.V.,Simon, E.M., Haselgrove, J., 2003. Diffusion-weighted imagingin acute bacterial meningitis in infancy. Neuroradiology 45,634–639.

Jaworowicz Jr., D.J., Korytko, P.J., Singh, L.S., Boje, K.M., 1998. Nitricoxide and prostaglandin E2 formation parallels blood–brainbarrier disruption in an experimental rat model of bacterialmeningitis. Brain Res. Bull. 46, 541–546.

Kaplan, S.L., Mason Jr., E.O., Wald, E.R., Schutze, G.E., Bradley, J.S.,Tan, T.Q., Hoffman, J.A., Givner, L.B., Yogev, R., Barson, W.J.,2004. Decrease of invasive pneumococcal infections in childrenamong 8 children's hospitals in the United States after theintroduction of the 7-valent pneumococcal conjugate vaccine.Pediatrics 113, 443–449.

Kastrup, O., Wanke, I., Maschke, M., 2005. Neuroimaging ofinfections. NeuroRx. 2, 324–332.

Kim, K.S., 2003. Pathogenesis of bacterial meningitis: frombacteraemia to neuronal injury. Nat. Rev. Neurosci. 4, 376–385.

Leib, S.L., Tauber, M.G., 1999. Pathogenesis of bacterial meningitis.Infect. Dis. Clin. North Am. 13, 527–548.

Leib, S.L., Leppert, D., Clements, J., Täuber, M.G., 2000. Matrixmetalloproteinases contribute to brain damage inexperimental pneumococcal meningitis. Infect. Immun. 68,615–620.

Morrissey, S.P., Stodal, H., Zettl, U., Simonis, C., Jung, S., Kiefer, R.,Lassmann, H., Hartung, H.P., Haase, A., Toyka, K.V., 1996. Invivo MRI and its histological correlates in acute adoptivetransfer experimental allergic encephalomyelitis:quantification of inflammation and oedema. Brain 119,239–248.

Mustafa, M.M., Ramilo, O., Sáez-Llorens, X., Olsen, K.D., Magness,R.R., MCxracken Jr., G.H., 1990. Cerebrospinal fluidprostaglandins, interleukin 1 beta, and tumor necrosis factor inbacterial meningitis: clinical and laboratory correlations inplacebo-treated and dexamethasone-treated patients. Am. J.Dis. Child. 144, 883–887.

Narayana, P.A., Wolinsky, J.S., Jensen, D.J., 1987. Magneticresonance imaging of experimental mumpsmeningoencephalitis in suckling hamsters. Magn. Reson. Med.4, 597–601.

Peters, A., Palay, S.L., Webster, H., 1991. The Fine Structure of theNervous System: Neurons and Their Supporting Cells, 3rd ed.Oxford University Press Inc, New York, NY, pp. 400–403.

Pfister, H.W., Fontana, A., Tauber, M.G., Tomasz, A., Scheld, W.M.,1994. Mechanisms of brain injury in bacterial meningitis. Clin.Infect. Dis. 19, 463–479.

Pignol, B., Henane, S., Chaumeron, S., Mencia-Huerta, J.M.,Braquet, P., 1990. Modulation of the priming effects ofplatelet-activating factor on the release of interleukin-1 fromlipopolysaccharide-stimulated rat spleenmacrophages. J. LipidMediat. 2 (Suppl), S93–S99.

Quagliarello, V.J., Scheld, W.M., 1992. Bacterial meningitis:pathogenesis, pathophysiology, and progress. N. Engl. J. Med.327, 864–872.

Quagliarello, V.J., Wispelwey, B., Long Jr., W.J., Scheld, W.M., 1991.Recombinant human interleukin-1 induces meningitis andblood–brain barrier injury in the rat. Characterization andcomparison with tumor necrosis factor. J. Clin. Invest. 87,1360–1366.

Ramilo, O., Sáez-Llorens, X., Mertsola, J., Jafari, H., Olsen, K.D.,Hansen, E.J., Yoshinaga, M., Ohkawara, S., Nariuchi, H.,McCracken Jr., G.H., 1990. Tumor necrosis factoralpha/cachectin and interleukin 1 beta initiate meningealinflammation. J. Exp. Med. 172, 497–507.

148 B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9

Author's personal copy

Roos, K.L., Van de Beek, D., 2010. Bacterial meningitis. Handb. Clin.Neurol. 96, 51–63.

Saunders, N.R., Ek, J.C., Habqood, M.D., Dzieqielewsa, K.M., 2008.Barriers in the brain: a renaissance? TrendNeurosci. 31, 279–286.

Sáez-Llorens, X., McCracken, G.H., 2003. Bacterial meningitis inchildren. Lancet 361, 2139–2148.

Tunkel, A.R., Scheld,W.M., 1993. Pathogenesis and pathophysiologyof bacterial meningitis. Clin. Microbiol. Rev. 6, 118–136.

Tunkel, A.R., Wispelwey, B., Scheld, W.M., 1990. Pathogenesis andpathophysiology of meningitis. Infect. Dis. Clin. North Am. 4,555–581.

Van Furth, A.M., Roord, J.J., Van Furth, R., 1996. Roles ofproinflammatory and anti-inflammatory cytokines inpathophysiology of bacterial meningitis and effect ofadjunctive therapy. Infect. Immun. 64, 4883–4890.

Vandenabeele, F., Creemers, J., Lambrichts, I., 1996. Ultrastructureof the human spinal arachnoid mater and dura mater. J. Anat.189, 417–430.

Wiesmann, M., Koedel, U., Bruckmann, H., Pfister, H.W., 2002.Experimental bacterial meningitis in rats: demonstration ofhydrocephalus and meningeal enhancement by magneticresonance imaging. Neurol. Res. 24, 307–310.

149B R A I N R E S E A R C H 1 3 9 0 ( 2 0 1 1 ) 1 4 2 – 1 4 9