Transactions of the Royal Society of Tropical Medicine and Hygiene (2006) xxx, xxx—xxx

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ierhea l th .com/ journa ls / t rs t

A preliminary neuropathological study of Japaneseencephalitis in humans and a mouse model

Allison C. Germana,b, Khin Saw Aye Myintc, Nguyen Thi Hoang Maid,Ian Pomeroye, Nguyen Hoan Phud, John Tzartose, Peter Wintera,b,Jennifer Colletta,b, Jeremy Farrar f, Alan Barrettg, Anja Kiparh,Margaret M. Esiri e, Tom Solomona,b,∗

a Division of Medical Microbiology, University of Liverpool, Liverpool, UKb Division of Neurological Sciences, University of Liverpool, Liverpool, UKc Armed Forces Research Institute of Medical Sciences, Bangkok, Thailandd Centre for Tropical Diseases, Cho Quan Hospital, Ho Chi Minh City, Viet Name Department of Neuropathology, Radcliffe Infirmary, Oxford, UKf University of Oxford—Wellcome Trust Clinical Research Unit, Ho Chi Minh City, Viet Namg Department of Pathology, Center for Biodefense and Emerging Infectious Diseases, and Institute for Human Infections andImmunity, University of Texas Medical Branch, Galveston, TX, USAh Department of Veterinary Pathology, University of Liverpool, Liverpool, UK

Received 28 November 2005; received in revised form 20 February 2006; accepted 20 February 2006

KEYWORDSArbovirus;Japaneseencephalitis;Zoonosis;Blood brain barrier

Summary Japanese encephalitis virus is a mosquito-borne flavivirus that causes approxi-mately 10 000 deaths annually in Asia. After a brief viraemia, the virus enters the centralnervous system, but the means of crossing the blood—brain barrier is uncertain. We used rou-tine histological staining, immunohistology and electron microscopy to examine brain materialfrom four fatal human cases, and made comparisons with material from a mouse model. Inhuman material there was oedema, perivascular inflammation, haemorrhage, microglial nod-ules and acellular necrotic foci, as has been described previously. In addition, there was newevidence suggestive of viral replication in the vascular endothelium, with endothelial cell dam-age; this included occasional viral antigen staining, uneven binding of the vascular endothelialcells to Ulex europaeus agglutinin I and ultrastructural changes. Viral antigen was also foundin neurons. There was an active astrocytic response, as shown by glial fibrillary acidic pro-tein staining, and activation of microglial cells was demonstrated by an increase in major

∗ Corresponding author. Viral Brain Infections Group, Divisions of Medical Microbiology and Neurological Sciences, 8th Floor DuncanBuilding, Liverpool L69 3GA, UK. Tel.: +44 151 706 4381; fax: +44 151 706 5805.

E-mail address: tsolomon@liv.ac.uk (T. Solomon).

0035-9203/$ — see front matter © 2006 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.trstmh.2006.02.008

TRSTMH-463; No. of Pages 11

2 A.C. German et al.

histocompatibility complex class II expression. Similar inflammatory infiltrates and a microglialreaction were observed in mouse brain tissue. In addition, �-amyloid precursor protein stainingindicated impaired axonal transport. Whether these findings are caused by viral replication inthe vascular endothelium or the immune response merits further investigation.© 2006 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rightsreserved.

1. Introduction

Japanese encephalitis (JE) is an acute encephalitis caused byinfection with Japanese encephalitis virus (JEV), a memberof the genus Flavivirus, family Flaviviridae, closely relatedto West Nile virus (Solomon, 2004). JEV is found in Asia and,like West Nile virus, is spreading, with recent outbreaksin Nepal and Australia (Mackenzie et al., 2004; Solomon,2004; Weaver and Barrett, 2004). Numerically, JE is a moreimportant disease than West Nile Virus, with an estimated30 000—50 000 encephalitis cases and 10 000—15 000 deathsannually, mostly among children in Asia (Tsai, 2000). JEVis zoonotic and is transmitted naturally among birds, espe-cially egrets, herons and other water birds, by mosquitoes,principally Culex species (Scherer and Buescher, 1959). Pigsare also important amplifying hosts (Scherer and Buescher,1959). Humans become infected with the virus following thebite of an infected mosquito; most human infections areasymptomatic or cause a non-specific febrile illness, whichis thought to correspond to a self-limiting viraemia. How-ever, in a proportion of those infected the virus crosses theblood—brain barrier to cause central nervous system (CNS)disease. Researchers are uncertain about the mechanismby which the virus crosses the blood—brain barrier; mostdata suggest the vascular endothelium is a more likely routethan the olfactory mucosa, but whether the virus is pas-sively transported or actively replicates in the endotheliumi

cwMbtbht1hiWwm

2

2

2Staa

Nam (Solomon et al., 2002, 2003). JEV infection was con-firmed on admission by measurement of IgM antibodies inthe serum and cerebrospinal fluid (CSF) using a rapid IgMdot enzyme immunoassay (Solomon et al., 1998). Serumand CSF samples were subsequently analysed for IgM andIgG anti-JEV antibodies using a double sandwich captureELISA (Innis et al., 1989), and viral culture attempted, asdescribed previously (Solomon et al., 2002). Where permis-sion was granted, a post-mortem needle biopsy of the brainthrough the infra-occipital route, and/or an autopsy, wasperformed as soon after death as possible. Tissue sampleswere fixed in 10% non-buffered formalin for a minimum of 4weeks for histopathological examination and 2% glutaralde-hyde for electron microscopic examination.

2.1.2. MouseDuring previous studies (Cao et al., 1995; Nitayaphan etal., 1990), 3—4-week-old female outbred (NIH Swiss strain)white mice (Harlan, USA), weighing 20—25 g, were inoc-ulated intracerebrally with 20 �l of either an attenuatedderivative of the JEV Nakayama strain or JEV strain SA-14.The original Nakayama strain was attenuated by six pas-sages through HeLa cells, characterized by alterations inthe E protein. This strain (Nakayama-O/HeLa p6) failed toinduce overt encephalitis in mice when inoculated intrac-erebrally (Cao et al., 1995), whereas the SA-14 strainhad a lethal neurovirulent phenotype in mice when inoc-umil1tac

2

TsahpfiaacwauSsa

s unclear (Solomon and Vaughn, 2002).The few pathological studies of JE in humans describe

haracteristic ‘punched-out’ necrotic foci, often associatedith blood vessels (Desai et al., 1995; Johnson et al., 1985;iyake, 1964), but their nature is unknown, and there haveeen few immunohistological and/or ultrastructural inves-igations. The mouse is often cited as a useful model for JE,ecause of the similar clinical features to humans, but thereas been little work comparing the pathological changes inhe mouse model with those in human disease (Hase et al.,990a; Miyake, 1964). We therefore conducted an immuno-istological and ultrastructural study of JE in humans, look-ng in particular at the effect on the vascular endothelium.e also looked for similar changes in limited material thatas available from a previous study of JEV in the mouseodel (Cao et al., 1995).

. Materials and methods

.1. Patients and samples

.1.1. Humanspecimens were collected from four fatal JE cases admit-ed during prospective clinical studies of JE between 1995nd 1997, at the paediatric and adult intensive care unitst the Centre for Tropical Diseases, Ho Chi Minh City, Viet

lated by the intracerebral route (Hase et al., 1993). Threeice (one infected with Nakayama-O/HeLa p6 and two

nfected with SA-14) were euthanized by cervical spine dis-ocation 5 d post-infection, and the brains were fixed in0% non-buffered formalin for histopathological examina-ion. Two mice inoculated intracerebrally with an equiv-lent volume of phosphate-buffered saline were used asontrols.

.2. Histology and immunohistology

issues were routinely embedded in paraffin wax and 5 �mections cut and stained with haematoxylin and eosin (HE)nd Toluidine Blue for histological examination. For immuno-istological examination, 5 �m sections were mounted onoly-lysin coated slides. Briefly, sections were deparaf-nized and then rehydrated through a series of gradedlcohols. Optimal conditions for antigen retrieval, blockingnd antibody dilutions were independently determined byomparative titration experiments. Endogenous peroxidaseas blocked through incubation with hydrogen peroxide andntigen retrieval pre-treatment performed as necessary,sing formic acid and/or heated citrate buffer (pH 6.0).ections were incubated with normal serum to avoid non-pecific binding of antibodies and then incubated for 15—18 ht 4 ◦C with the primary antibodies. For human tissues,

Japanese encephalitis in humans and a mouse model 3

antibodies directed against JEV antigen (1:2000; polyclonalanti-JEV murine ascitic fluid, UTMB, Texas); the majorhistocompatibility complex (MHC) (1:25; mouse anti-humanMHC II, HLA-DR, clone TAL.1B, DAKO Cytomation, UK); themyeloid/histiocyte antigen of monocytes/macrophages andneutrophils (1:200; mouse anti-myeloid/histiocyte antigen,clone MAC387, Serotec, UK); glial fibrillary acidic protein(GFAP) of astrocytes (1:1500; polyclonal anti-GFAP, DakoCytomation, UK) and �-amyloid precursor protein (�-APP),which accumulates in damaged axons (1:1500; monoclonalmouse anti-�-APP, clone LN27, Zymed Laboratories Inc, UK).Mouse tissues were assessed with antibodies for GFAP, �-APP,JEV antigen, and CD45 for lymphocytes (1:100; monoclonalmouse anti-CD45, clone CBL 464, Cymbus Biotechnology,UK). Endothelial cell glycoprotein expression was assessedusing biotinylated Ulex europaeus agglutinin 1 (UEA-1) forhuman tissues (1:100 UEA-I, Vector, UK) and biotinylatedGriffonia simplicifolia Lectin 1 (GSL-1) for mouse tissues(1:100 GSL-I, Vector, UK).

The peroxidase-anti-peroxidase (PAP) method wasapplied for MHCII and JEV antigen, following previouslypublished protocols (Kipar et al., 2001), with modificationsfor the detection of JEV antigen. Initial detection of JEVantigen for diagnostic purposes was performed as previouslyreported (Myint et al., 1999). The avidin biotin complexmethod, using biotinylated secondary antibodies detectedwith avidin and biotin solutions (ABC Kit, Vector, UK), wasused for the remaining markers, with the MOM Kit (Vector,

2.3. Controls

Control sections were incubated in parallel for each stainingreaction. A formalin-fixed and paraffin-embedded Vero cellpellet infected with JEV strain JKT2363 (Chen et al., 1992)was used as a positive control for JEV antigen. Human tonsilwas used as a positive control tissue for lymphocyte andmacrophage markers. Marmoset experimental autoimmuneencephalomyelitis (EAE) brains were used as positive controltissue for �-APP staining. Normal human brain tissue wasused as a control for UEA-I and GFAP staining. Consecutivesections incubated with TBS or PBS-Tween instead of theprimary antibody served as negative controls. Uninfectedmouse brains were used as control tissues for the mousestudies.

2.4. Electron microscopy

Tissues were fixed in 2% glutaraldehyde in 0.1 mol/l phos-phate buffer (pH 7.4) and processed as described previously(Pongponratn et al., 2003). Thin sections were stained withuranyl acetate and lead citrate prior to examination with aJeol 1200 EX II transmission electron microscope.

3. Results

3.1. Human patients

3TcbacT

nts w

An

Se

Ig

1

—

19

42

lasgo

UK) used for mouse on mouse assays. Visualization wasachieved by incubation with 0.05% 3,3′-diaminobenzidinetetrahydrochloride (DAB, Sigma-Aldrich Company Ltd,Poole, UK) in 0.1 mol/l imidazole/HCl buffer (ph 7.1), anda haematoxylin counterstain was applied. Finally, sectionswere dehydrated through ascending ethanol and xylenebaths and mounted in DPX Mountant Medium (R A Lamb Ltd,Eastbourne, UK).

Table 1 Clinical and demographic details of the study patie

Patientidentity, age(years), sex

Illness lengthbefore admission(days)

Presentingclinicalfeatures

CNS 48, 19, M 4 Fever, confusion,coma, GCS 4

CNS 307, 18, M 3 Fever, confusion,generalizedtonic-clonic seizures,GCS 6

CNS 712, 9, F 3 Fever, headache,hemiparesis, rigidityspasm, GCS 11

PIF 44, 10, F 3 Fever, confusion,focal thengeneralized seizures,status epilepticus,coma, GCS 4

JEV: Japanese encephalitis virus; CSF: cerebrospinal fluid; GCS: GNote: Patient CNS 48 also had JEV isolated.

.1.1. Demographic and clinical featureswo adults and two children were studied. All had a typi-al history for JE, with a brief febrile prodrome followedy neurological disease (Table 1). All deteriorated soonfter admission and were ventilated because of worseningoma (three patients) or status epilepticus (one patient).he time interval between admission and death ranged

ith Japanese encephalitis

ti-JEV ELISA (units) Time todeath(hours)

Time tosampling

JEV viralantigen

rum CSF

M IgG IgM IgG

4 0 10 48 30 min (Bx) +

— — — 156 5 min (Bx),5 h(autopsy)

+

0 52 0 96 15 min (Bx) −

13 142 1 68 30 min (Bx) +

w Coma Score (out of 15); Bx: post-mortem needle biopsy.

4 A.C. German et al.

Figure 1

Figure 3

Figure 2

Figure 4

Japanese encephalitis in humans and a mouse model 5

from 2 to 7 days. The diagnosis of JE was based on ele-vated CSF IgM to JEV for two patients (in one of whomJEV antigen was also detected by immunohistology), pos-itive virus isolation and antigen detection for one patientand antigen detection alone for one patient. A post-mortemneedle biopsy was performed for all four patients 5 to15 minutes after death, and for one (CNS 307) autopsyof the brain was also possible. This revealed a grosslyswollen brain, weighing 1500 g, with injected meninges andmicrohaemorrhages visible over the frontal and parietallobes.

3.1.2. HistopathologyThe most striking observations on routine histopathol-ogy were the widespread inflammatory infiltrates andfocal areas of necrosis, giving a ‘punched-out’ appearance(Figures 1 and 2). In addition there was vascular congestionand haemorrhage, and in some patients microglial noduleformation.3.1.2.1. Inflammatory changes. The inflammatory infil-trates were widespread throughout most sections of brainsfrom all patients, particularly in CNS 307 and PIF44. In CNS307, mild focal perivascular accumulations of erythrocytes,perivascular, lymphocyte-dominated cuffing (Figure 1A) andmild, diffuse leptomeningeal mononuclear cell infiltrationswere observed in the frontal cortex. Perivascular cuffs inPIF 44 were often loose and irregular and consisted of lym-phocytes and macrophages (Figure 2A); such accumulationswere only occasionally observed for CNS 712. The anti-myeloid/histiocyte antigen marker detected macrophagesthroughout the brain parenchyma, with focal and perivas-cular accumulations in the frontal, parietal, occipital andtemporal cortices of CNS 307 (Figure 7A) and in the needle

Figure 1 Characteristic histopathology in Japanese encephalitisinfiltrate and necrosis in pontine tissue (original magnification (OMbrain/thalamus (OM ×200). (C) Cerebellum, showing well-preservacellular necrotic lesion is visible in the molecular layer, associatestriking necrotic foci with little peripheral inflammation (OM ×100vessel with mild lymphocytic infiltrate (OM ×400).

Figure 2 Histopathological alterations in patients with Japanesefrontal cortex; (E,F) CNS 307, pons. (A) Moderate perivascular monon(OM) ×400). (B) Small glial nodule (OM ×400). (C) Necrotic foci wiwith gitter cells (arrows) (OM ×400). (E) Perivascular lymphocytic(F) Immunohistochemical stain for Japanese encephalitis virus antiendothelium (arrow) (OM ×100).

Figure 3 Toluidine Blue staining of pontine tissue for myelin in Jcentrally, surrounded by oedema (white regions) and secondarily d×200). (B) Damaged myelin, myelin splitting (arrows) and fragmeRegion showing better preservation of myelin for comparison with (Bby arrows (OM ×200).

Figure 4 Expression of major histocompatibility complex (MHC) II aencephalitis. (A—C) Patient CNS 307 frontal cortex; (D) patient PIFmorphology of macrophages (gitter cells, arrow) and in cells with thnecrotic foci (original magnification (OM) ×400). (C) JEV antigen wasfinely granular cytoplasmic reaction (arrows) (OM ×400). Positive nseen only in scattered cell processes suggestive of neural processes

biopsies from CNS 712 and PIF 44, but not CNS 48. Mor-phologically, in HE sections, there was evidence for diffusemicroglial activation. Microglial nodules were observed inPIF 44 alongside activated microglial cells (Figure 2B), butnot in CNS 307.3.1.2.2. Acellular necrotic foci. Patient CNS 307 showedwidely distributed foci of necrosis, giving a characteristic‘punched-out’ appearance (Figures 1C, D and 2C). Thesefoci were often seen in association with vessels (Figure 1C)and contained axons and microglial cells (in various stagesof necrosis), with occasional gitter cells and neuronophagia(Figures 1B, E and 2D). There were no acellular necrotic fociin the hippocampus, although there were diffuse foci in thesubiculum (Figure 1E). Neurons in the CA4 layer of the hip-pocampus were pyknotic. In other areas of hypoxic damage,axonal spheroids were observed as condensed eosinophilicregions alongside degenerating neurons (Figure 1B and E).Staining for neuronal structure with Toluidine Blue revealedoedematous, dilated neurons with thinned myelin sheaths(Figure 3A—C), particularly in areas with inflammatory infil-trates and/or oedema. In some areas, focal demyelinationwith gitter cells was observed. Oedematous changes wereparticularly marked in CNS 307 (Figure 3A).3.1.2.3. Vascular changes. Patient CNS 307 showedregions of intense hyperaemia, vascular damage (Figure 1F)and mild perivascular haemorrhages. Occasionally, fibrinplugs were observed blocking blood vessels (Figure 2E).Haemorrhagic regions and congested vessels were especiallyp

33p

(patient CNS 307, HE). (A) Damaged vein with perivascular) ×100). (B) Ischaemic, shrunken, damaged neurons in mid-

ed Purkinje cells, compared with the granular cells. A focald with an end vessel (OM ×40). (D) Internal capsule showing). (E) Necrotic neurons in subiculum (OM ×200). (F) Damaged

encephalitis (HE). (A,B) PIF 44, cortical tissue; (C,D) CNS 307uclear cuffing and haemorrhage (arrow) (original magnificationthin the white matter (arrows) (OM ×200). (D) Necrotic focusinfiltrate around a vein containing a fibrin plug (OM ×100).

gen (red). Note the presence of viral antigen in the vascular

apanese encephalitis, patient CNS 307. (A) Congested vesselamaged myelin sheaths (arrows) (original magnification (OM)ntation of axons around swollen capillary (*) (OM ×200). (C)) (OM ×200). (D) Focus of acellular necrosis; border indicated

nd Japanese encephalitis virus (JEV) antigen in human Japanese44. (A,B) MHC II antigen expression is seen in cells with the

e morphology of activated microglial cells (arrowhead) withinobserved in several neurons, which exhibited a variably intense

rominent in PIF 44 (Figure 2A).

.1.3. Immunohistology

.1.3.1. Expression of JEV antigen. JEV antigen wasarticularly prevalent in the medulla, midbrain, left

eurons appear morphologically unaltered. (D) JEV antigen wasin patient PIF 44 (OM ×400).

6 A.C. German et al.

Figure 5

Figure 7

Figure 6

Figure 8

Japanese encephalitis in humans and a mouse model 7

cerebellum, temporal and frontal cortex and pons of CNS307. A variably intense, finely granular cytoplasmic reactionwas observed in several neuronal cell bodies, with moreintense staining of the perinuclear cytoplasm (Figure 4C).Although intensity of JEV antigen staining was weak,specificity of the antibody was assured through staining ofcontrols. PIF 44 and CNS 48 demonstrated viral antigen onlyin cell processes that were probably axons (Figure 4D), andCNS 712 was negative for JEV antigen. The distribution ofviral antigen was not related to the inflammatory infiltratein any of the cases. Viral antigen was not detected withinthe acellular necrotic foci. It was occasionally observed inthe vascular endothelium for CNS 307 during preliminarystudies (Figure 2F), but was not consistently observed onrepeated investigations. Antigen-positive neurons did notshow degenerative changes. For CNS 307, �-APP stainingwas performed but was negative for all neurons, despitegood staining of the positive control tissue.3.1.3.2. Astrocytic response. The astrocytic response wasassessed by staining astrocytes for GFAP. PIF 44, CNS 712and CNS 48 all displayed numerous areas of increased cel-lularity, with some very large swollen reactive astrocytes(Figure 5A), compared with control sections (Figure 5D). Thefrontal and parietal cortex of CNS 307 showed very unusualastrocyte staining, with a punctuate pattern instead of thetypical morphology described above (Figure 5C). Many ofthe astrocytes in this section appeared to have lost their

Figure 5 Cell-specific immunohistology in patients with Japanese(GFAP) staining, showing astrocyte cell bodies and multiple thin ptemporal cortex: intermediate GFAP beading pattern (OM ×400). (blood vessels may indicate astrocytic foot processes, with intermittprocesses or disrupted staining pattern (OM ×400). (D) Healthy humandifficult to identify and just the dense network of astrocyte processeagglutinin I staining of endothelial cells (OM ×400). Staining intensitof the same vessel, suggesting loss of endothelial integrity. (F) Normendothelial cells (OM ×400). Typical blood vessel, showing intense u

Figure 6 Histopathology of murine CNS, infected with Japanesestrain SA-14; cortex. (D—F) Mouse infected with JEV strain Nakayaminfiltrates and microgliosis (original magnification (OM) ×200); (B) sevcytes) with perivascular haemorrhage (arrow) (OM ×400); (C) necrotineuron (arrowhead) (OM ×400); (D) perivenous mononuclear infiltraand microgliosis. The vessel is infiltrated by subendothelial macroph(arrow) (×100); (F) microglial nodule with ‘vacuolation’ (arrow) of p

Figure 7 Inflammation (human and mouse) and astrocytic proliferanervous system infection. (A) Patient CNS 307 temporal cortex: MA(original magnification (OM) ×100). (B) Mouse infected with Nakayregion of inflammation and neuronal damage (OM ×400). (C—F) Glmild astrocytic response around region of sham inoculation (OM ×pronounced astrocytosis (OM ×40). (E) Control mouse, showing normNakayama-O/HeLa p6. Astrocyte cell bodies are more prominent andin the control mouse (OM ×400).

Figure 8 Axonal damage associated with areas of inflammation inp6-infected mouse (original magnification (OM) ×100, HE). (B) Samaxons in the inflamed region (OM ×100). (C,D) Non-inflamed regiostaining for �-APP (HE and �-APP) (OM ×100).

processes, and the majority of blood vessels displayed aperipheral ring of positive-staining material (Figure 5C). Inthe temporal and occipital cortices of CNS 307 there wasbeading of astrocytic processes (clasmatodendrosis [Medanaet al., 2002]), possibly representing an intermediate stageof astrocyte degeneration (Figure 5B).3.1.3.3. MHC II expression. For the most part, few MHC II-positive microglial cells were identified, scattered through-out the brain parenchyma. However, in CNS 307 theacellular necrotic foci contained MHC II-expressing, acti-vated microglial cells and some MHC II-positive gitter cells(Figure 4A and B).3.1.3.4. Vascular endothelium integrity. Because of thesigns of vascular damage on HE staining, we investigatedthe integrity of the vascular endothelium using UEA-I, whichbinds to glycoproteins and glycolipids containing �-linkedfucose residues and has been established as an excellentmarker for healthy human endothelial cells (Medana etal., 2002). CNS 307 and CNS 712 showed variable stain-ing intensity in different vessels and also between dif-ferent areas of the same vessel (Figure 5E), rather thanthe consistent staining seen in the control brain sections(Figure 5F).

3.1.4. Electron microscopyExamination of pontine tissue (a major site of damage inJE) from CNS 307 did not reveal viral particles. However,

encephalitis. (A) Patient PIF 44: glial fibrillary acidic proteinrocesses (original magnification (OM) ×400). (B) Patient 307C) Patient 307 frontal cortex: abnormal GFAP staining aroundent staining throughout section representing either sectionedbrain, showing normal GFAP staining. Individual astrocytes are

s can be defined (OM ×400). (E) Patient PIF 44: Ulex europaeusy is variable between vessels and also between different areasal human brain tissue: Ulex europaeus agglutinin I staining ofniform staining of the endothelium.

encephalitis virus (JEV) (HE). (A—C) Mouse infected with JEVa-O/HeLa p6; cortex. (A) Focal cell-rich area with perivascularere perivenous mononuclear infiltration (macrophages, lympho-c area, showing degenerate ganglion cell (arrow) and apoptoticte (lymphocytes, macrophages, some degenerate cells; arrow)ages (OM ×400); (E) necrotic region with microglial activationarenchyma (×200).

tion (mouse) in response to Japanese encephalitis virus centralC387 staining of macrophages at a site of focal inflammationama-O/HeLa p6 (CD45+ stain). Perivascular inflammation in aial fibrillary acidic protein stain. (C) Control mouse, showing40). (D) Mouse infected with Nakayama-O/HeLa p6, showingal astrocyte morphology (OM ×400). (F) Mouse infected with

the processes are thicker and more extensive than those seen

mouse brain. (A) Region of inflammation in Nakayama-O/HeLae section as (A), showing accumulation of �-APP in damaged

n in Nakayama-O/HeLa p6-infected mouse, showing a lack of

8 A.C. German et al.

there was expansion of the rough endoplasmic reticulum(RER) within the neurons, with numerous RER-derivedvesicles. Similar expansion and disruption of the RERwas seen in the vascular endothelial cells (data notshown).

3.2. Mouse model

The pathological changes seen in mouse brains infected witheither strain of JEV (SA-14 [Figure 6A—C] or Nakayama-O/HeLa p6 [Figure 6D—F]) were similar to those observedin humans, with perivascular cuffs, cellular infiltrates andmild vascular damage. Despite a variation in the phenotypeof the inoculated viruses, all mice showed histopathologi-cal evidence of JEV infection and disturbance in neuronalfunction. There was a greater degree of perivascular haem-orrhage and inflammation seen in the SA-14 mice, althoughastrocytosis was more severe in the Nakayama-O/HeLa p6mice. There were small areas of necrosis with degenerateneurons and/or glial cells, but not the ‘punched-out’ acel-lular necrotic foci seen in human tissue. Additionally, �-APPstaining was positive in some neurons in all sections frominfected mice; control mouse tissues were negative.

3.2.1. Histopathology and immunohistologyInflammatory infiltrates were clearly visible in the mousebac((moacaonwt1putam(

aNippsaraiise

4. Discussion

With the increasing spread of JEV across the Asia-Pacificregion, and related viruses such as West Nile virus else-where across the globe, there is an urgent need to betterunderstand the pathogenesis of flavivirus encephalitis, andto develop treatments. Critical to the pathogenesis is viraltransmission across the blood—brain barrier. This barrier iscomposed of specialized endothelial cells joined by tightjunctions, on a basal lamina, surrounded by pericytes andastrocytic foot processes. In some animal models of infectionwith the flavivirus St Louis encephalitis virus, CNS inva-sion occurs via the olfactory route, where the barrier isimpaired (Monath et al., 1983). Entry of virus via cranialnerves I and V was also important in the mouse model of thealphavirus Venezuelan equine encephalitis virus, even whenvirus was inoculated into the footpad (Charles et al., 1995).However, in most studies of flavivirus infection, viraemiathen spread across the blood—brain barrier is thought tobe the mechanism of entry (Desai et al., 1995; Hase etal., 1990a; Johnson and Mims, 1968; Miyake, 1964; Mukherjiand Biswas, 1976), and indeed, the ‘punched-out’ necroticlesions that characterize the pathological changes in humanJE support a haematogenous route of spread. While mostagree there is haematological spread, what is not clear iswhether this spread is passive or involves viral replicationin the vascular endothelium, and whether this might con-tribute to the pathogenesis. A ‘Trojan horse’ theory for JEVpwtl(jpotstopam

p1((Spstoiootsteba

rains (Figure 6) and were associated with neuronal dam-ge. Sections were characterized by marked perivascularuffing (Figure 6A and B), mild perivascular haemorrhageFigure 6B) and a mild diffuse microglial infiltrationFigure 6A and D—F). For the Nakayama-O/HeLa p6 infectedouse, veins were often infiltrated with one or more layers

f subendothelial macrophages (Figure 6D). Small necroticreas with microglial activation and perivascular mononu-lear infiltration were seen in the cortex (Figure 6D—F),nd mild focal leptomeningeal lymphocyte infiltration wasbserved. Microglial nodules were present throughout theeuropil (Figure 6F). Most infiltrating inflammatory cellsere identified as lymphocytes and macrophages, based on

heir expression of CD45+ (Figure 7B). Mice infected with SA-4 showed similar changes, but the macrophage-dominatederivascular inflammation was more pronounced, partic-larly in the white matter (Figure 6A and B). In addition,here was a moderate increase of microglial cells, oftendjacent to affected vessels (Figure 6A and B), with severalicroglial nodules. Necrotic areas were also identified

Figure 6C).Interestingly, no sections were positive for JEV antigen,

n observation that has been previously reported for theakayama O/HeLa p6 virus in mice (Cao et al., 1995). Stain-

ng for GFAP confirmed widespread astrocytic activation,articularly in the mouse infected with Nakayama-O/HeLa6 (Figure 7D and F). The sham-infected control mousehowed only occasional areas with activated astrocytes and

few ischaemic neurons (Figure 7C and E), most likelyelated to the inoculation. Staining with �-APP revealedxonal damage in the infected mouse brains in areas ofnflammation (Figure 8). Unfortunately, non-specific stain-ng of microglial inflammatory cells occurred with GSL-1taining, so assessment of the integrity of the blood vesselndothelium could not be interpreted.

assage across the blood—brain barrier has been proposed,hereby JEV-infected monocytes or macrophages infiltrate

he CNS, and neurons subsequently become infected fol-owing release of virions from the monocytes/macrophagesYang et al., 2004). Our study does not support a ‘Tro-an horse’ theory, as viral antigen was not detected in theerivascular cell infiltrates. However, it does show signsf endothelial damage, which could reflect viral replica-ion, immune-mediated damage or non-specific changeseen in other severe diseases. Our study was limited byhe availability of human material. For cultural reasons,btaining autopsy material is difficult in Asia, althoughost-mortem needle biopsy can be more acceptable, ands we found in this study, provides useful additionalaterial.The characteristic foci of acellular necrosis in the brain

arenchyma in JE were seen as long ago as 1933 (Miyake,964) and have since been described using a range of termsrarefaction necrosis, malacic foci, ‘punched-out’ lesions)Desai et al., 1995; Johnson et al., 1985; Miyake, 1964;hankar et al., 1983). In our series we saw necrotic foci inatient CNS 307, who died on day 10 of illness, but did notee them in the other patients. This may reflect the limita-ions of the needle biopsy sampling in these other patients,r the fact that they died earlier in the illness; previouslyt has been reported that foci develop from day 4 of illnessnwards (Shankar et al., 1983). In a similar way, the lackf necrotic foci in the mouse tissue may reflect the facthat they are euthanised relatively soon after developingymptoms of encephalitis; alternatively this may be becausehe virus was inoculated intracerebrally rather than periph-rally. Interestingly, similar necrotic lesions have recentlyeen described in West Nile virus encephalitis (Guarner etl., 2004). The pathogenesis of acellular foci in flavivirus

Japanese encephalitis in humans and a mouse model 9

encephalitis remains controversial, and has been ascribedto a generalized toxic inflammation, direct toxic actions ofthe virus, or vascular spasms secondary to circulatory distur-bances (Miyake, 1964; Shankar et al., 1983). Several lines ofevidence indicate that vascular pathology is critical in thedevelopment of these necrotic foci. Firstly, they occur closeto or around vessels, particularly end vessels (as seen inFigure 1C). In addition, the necrosis of the vascular endothe-lium (Figure 1F), capillary congestion, ring haemorrhagesand occasional thrombus formation (Figure 1E), that we andothers have seen in JE (Li et al., 1988; Miyake, 1964; Shankaret al., 1983; Zimmerman, 1946) support the idea that thevascular route may be important for viral disseminationand/or replication. Perivascular haemorrhage was also seenin the mouse infected with the more virulent JEV strain (SA-14). These changes are similar to vascular alterations seenin other viral encephalitides, including encephalitis causedby alphaviruses such as eastern equine encephalitis virus andthat caused by the morbillivirus Nipah virus, in which similarnecrotic foci are seen (Garen et al., 1999; Paessler et al.,2004; Wong et al., 2002). In Nipah virus encephalitis, viralreplication in endothelial cells and vascular smooth muscle,with an associated vasculitis, is critical to the pathogenesis(Wong et al., 2002). In JE, impairment of the blood—brainbarrier is also implicated by the widespread perivascularoedema. In our study, this was particularly well demon-strated by the toluidine blue staining, which also showeddamaged myelin and areas of demyelination.

1991). This may indicate damage to, or functional impair-ment of, the blood—brain barrier due to direct or indirecteffects of virus replication.

Previously, in studies using HE staining, an astrocyticreaction was not thought to be a prominent feature in JE(Desai et al., 1995). However, in our study, identifying astro-cytes on the basis of their GFAP expression revealed numer-ous areas with large reactive astrocytes, particularly in areasof neuronal damage; a similar astrocytic reaction was alsoseen in mouse tissue. Although an astrocytic reaction isa non-specific change that occurs whenever degenerationtakes place in the CNS, the patchy distribution we saw ischaracteristic of encephalitis (Booss and Esiri, 2003). Irre-versible injury of astrocytes due to acidosis and ischaemia,characterized by blunting and loss of the distal processesand the appearance of cell fragments or ‘filling bodies’has been termed clasmatodendrosis (Hulse et al., 2001).This phenomenon, observed in CNS 307 (Figure 5B and C),has been reported in patients with ischaemic brain lesionsand Alzheimer’s disease (Tomimoto et al., 1997). Beadingof astrocytic processes has also been reported in cerebralmalaria (Medana et al., 2002).

The preliminary data from the mouse model of JEshow perivascular haemorrhage, astrocytic activation andneuronal degeneration similar to those observed in humantissues. Increased haemorrhage and inflammatory infiltrateswere associated with a more neurovirulent phenotype inmice. Others have found intraperitoneal inoculation of SA-1tiIco(tWiiasldahalbdtaiui(c

rrJtt

As others have done previously (Desai et al., 1995), wefound most JEV antigen staining was in neurons of the brain-stem and cortex. Previously, when JEV antigen has been seenin vascular endothelial cells and the perivascular zone, thiswas attributed to phagocytosis rather than viral replication(Desai et al., 1995; Johnson et al., 1985). In addition, insome in-vitro and mouse studies of flavivirus encephalitis,phagocytosis of virus and passive or active transport acrossthe endothelium appeared more important than viral repli-cation in the endothelium (Dropulic and Masters, 1990; Haseet al., 1990a; Huang and Wong, 1963; Liou and Hsu, 1998).However, in our study of humans, there were ultrastruc-tural changes (expansion of the RER and vesicle formationin endothelial cells) that were consistent with virus replica-tion, but could alternatively be non-specific changes. Thesechanges were similar to those seen in neurons, and to thosereported for flavivirus replication both in vitro and in themouse model (Boulton and Webb, 1971; Hase, 1993; Hase etal., 1990b; Steele et al., 2000; Wang et al., 1997; Westawayet al., 1997). For the alphavirus Semliki Forest virus, suchchanges are a prerequisite for viral RNA synthesis and mat-uration (Mehta et al., 1990; Pathak and Webb, 1978). Wedid not observe any viral particles. This may be because ofsampling error, a temporal effect or a low replication rate.JEV particles have never been reported previously in post-mortem studies of human CNS tissue, and visualization ofrelated flavivirus particles in humans by electron microscopyis rare. However, there have been occasional observationsof viral particles in the endoplasmic reticulum of neuronsinfected with West Nile virus (Hayes et al., 2005). Thedecrease in binding of UEA-I in the vascular endothelium,indicating a loss in integrity of the glycoproteins, was similarto that reported in the brain endothelium in AIDS (Buttner etal., 1996) and in cytomegalovirus vasculitis (Shintaku et al.,

4 virus generated a more marked inflammatory responsehan intracerebral (Hase et al., 1990a), although intranasalnoculation failed to induce encephalitis (Cao et al., 1995).deally, a peripheral inoculation route in mice would morelosely reflect viral invasion observed in humans. However,ur strain of SA-14 did not show neuroinvasive propertiesCao et al., 1995). Route of inoculation therefore influenceshe observed phenotype and warrants further investigation.e observed evidence of axonal dysfunction in areas of

nflammation, as determined by �-APP staining. �-APPs transported by fast axonal transport and accumulatest sites of axonal injury, thus enabling the detection ofubtle alterations in neuron function within 2 hours fol-owing injury, particularly in the absence of morphologicalegenerative change in HE-stained sections (Medana etl., 2002). Why similar �-APP staining was not seen in theuman material is not certain, but may be due to a samplingrtifact, the age of the tissues examined, or the relativeong duration of illness in CNS 307. Alternatively, it maye that axonal transport was not occurring at all in theseamaged tissues. Staining for JEV antigen was negative inhe mouse tissues. This may be due to prolonged fixation,s we were also unable to demonstrate CD3 and CD45Rn the sections, or may be related to the strain of virussed (Cao et al., 1995). However, in fatal human disease,t is not uncommon for antigen not to be demonstratedDesai et al., 1995), which may be because virus has beenleared.

In summary, our preliminary studies on limited mate-ial from humans and mice provides evidence that viraleplication in endothelial cells may be important in humanE. Although viral infection of the brain parenchyma ishought to be the main pathogenic process, our data suggesthat virus may also directly or indirectly cause damage

10 A.C. German et al.

to the vascular endothelial cells, leading to the devel-opment of necrotic foci that are so characteristic of thisdisease.

5. Ethics

The free and informed consent of the legal guardians of thesubjects was obtained for inclusion of the patients into thestudy group and for post-mortem sampling. Local (Viet Nam)and UK (Liverpool School of Tropical Medicine) ethical reviewboards approved the investigation.

Conflicts of interest statementThe authors have no conflicts of interest concerning the workreported in this paper.

Acknowledgements

We thank the Director and staff of the Centre for TropicalDisease, Ho Chi Minh City for their support; the familiesof the Vietnamese patients agreeing to participate inthis study; the Pathology Department for help with theautopsies; Gareth Turner for help setting up the study; thetechnical staff in the Department of Veterinary Pathology,Liverpool for assistance with immunohistological staining;David Ferguson at the Department of Neuropathology,ODtaNT(C

R

B

B

B

C

C

C

D

D

Garen, P.D., Tsai, T.F., Powers, J.M., 1999. Human eastern equineencephalitis: immunohistochemistry and ultrastructure. Mod.Pathol. 12, 646—652.

Guarner, J., Shieh, W.J., Hunter, S., Paddock, C.D., Morken, T.,Campbell, G.L., Marfin, A.A., Zaki, S.R., 2004. Clinicopathologicstudy and laboratory diagnosis of 23 cases with West Nile virusencephalomyelitis. Hum. Pathol. 35, 983—990.

Hase, T., 1993. Virus-neuron interactions in the mouse braininfected with Japanese encephalitis virus. Virchows Arch. B. CellPathol. Incl. Mol. Pathol. 64, 161—170.

Hase, T., Dubois, D.R., Summers, P.L., 1990a. Comparative studyof mouse brains infected with Japanese encephalitis virus byintracerebral or intraperitoneal inoculation. Int. J. Exp. Pathol.71, 857—869.

Hase, T., Summers, P.L., Dubois, D.R., 1990b. Ultrastructuralchanges of mouse-brain neurons infected with Japaneseencephalitis virus. Int. J. Exp. Pathol. 71, 493—505.

Hase, T., Dubois, D.R., Summers, P.L., Downs, M.B., Ussery,M.A., 1993. Comparison of replication rates and pathogenici-ties between the Sa-14 parent and Sa-14-14-2 vaccine strains ofJapanese encephalitis virus in mouse brain neurons. Arch. Virol.130, 131—143.

Hayes, E.B., Sejvar, J.J., Zaki, S.R., Lanciotti, R.S., Bode, A.V.,Campbell, G.L., 2005. Virology, pathology, and clinical mani-festations of West Nile virus disease. Emerg. Infect. Dis. 11,1174—1179.

Huang, C.H., Wong, C., 1963. Relation of the peripheral multiplica-tion of Japanese B encephalitis virus to the pathogenesis of theinfection in mice. Acta Virol. 7, 322—330.

Hulse, R.E., Winterfield, J., Kunkler, P.E., Kraig, R.P., 2001. Astro-

I

J

J

K

L

L

M

M

M

M

M

xford, for help with electron microscopy; Ananda Nisalak,avid Vaughn and colleagues at AFRIMS, Bangkok for assis-ance with diagnostic studies; and Dr V Ravi, Dr Anita Desaind Dr S. Shankar, National Institute of Mental Health andeuro Sciences, Banglaore, India, for helpful discussions.his work was funded by the Wellcome Trust of Great BritainGrant no. 054682). Tom Solomon is a UK Medical Researchouncil Senior Clinical Fellow.

eferences

ooss, J., Esiri, M.M., 2003. Viral Encephalitis in Humans, first ed.American Society for Microbiology Press, Washington D.C.

oulton, P.S., Webb, H.E., 1971. An electron microscopic study ofLangat virus encephalitis in mice. Brain 94, 411—418.

uttner, A., Mehraein, P., Weis, S., 1996. Vascular changes in thecerebral cortex in HIV-1 infection. II. An immunohistochemi-cal and lectinhistochemical investigation. Acta Neuropathol. 92,35—41.

ao, J.X., Ni, H.L., Wills, M.R., Campbell, G.A., Sil, B.K., Ryman,K.D., Kitchen, I., Barrett, A.D.T., 1995. Passage of Japaneseencephalitis virus in HeLa cells results in attenuation of viru-lence in mice. J. Gen. Virol. 76, 2757—2764.

harles, P.C., Walters, E., Margolis, F., Johnston, R.E., 1995. Mech-anism of neuroinvasion of Venezuelan equine encephalitis virusin the mouse. Virology 208, 662—671.

hen, W.R., Ricohesse, R., Tesh, R.B., 1992. A new genotype ofJapanese encephalitis virus from Indonesia. Am. J. Trop. Med.Hyg. 47, 61—69.

esai, A., Shankar, S.K., Ravi, V., Chandramuki, A., Gouriedevi, M.,1995. Japanese encephalitis virus antigen in the human brainand its topographic distribution. Acta Neuropathol. 89, 368—373.

ropulic, B., Masters, C.L., 1990. Entry of neurotropic arbovirusesinto the central nervous system: an in vitro study using mousebrain endothelium. J. Infect. Dis. 161, 685—691.

cytic clasmatodendrosis in hippocampal organ culture. Glia 33,169—179.

nnis, B.L., Nisalak, A., Nimmannitya, S., Kusalerdchariya, S.,Chongswasdi, V., Suntayakorn, S., Puttisri, P., Hoke, C.H., 1989.An enzyme-linked immunosorbent-assay to characterize dengueinfections where dengue and Japanese encephalitis co-circulate.Am. J. Trop. Med. Hyg. 40, 418—427.

ohnson, R.T., Mims, C.A., 1968. Pathogenesis of viral infections ofthe nervous system. N. Engl. J. Med. 278, 23—30.

ohnson, R.T., Burke, D.S., Elwell, M., Leake, C.J., Nisalak, A.,Hoke, C.H., Lorsomrudee, W., 1985. Japanese encephalitis —immunocytochemical studies of viral antigen and inflammatorycells in fatal cases. Ann. Neurol. 18, 567—573.

ipar, A., Hetzel, U., Armien, A.G., Baumgartner, W., 2001. Bilateralfocal cerebral angiomatosis associated with nervous signs in acat. Vet. Pathol. 38, 350—353.

i, Z.S., Hong, S.F., Gong, N.L., 1988. Immunohistochemical studyon Japanese B encephalitis. Chin. Med. J. 101, 768—771.

iou, M.L., Hsu, C.Y., 1998. Japanese encephalitis virus is trans-ported across the cerebral blood vessels by endocytosis in mousebrain. Cell Tissue Res. 293, 389—394.

ackenzie, J.S., Gubler, D.J., Petersen, L.R., 2004. Emerging fla-viviruses: the spread and resurgence of Japanese encephali-tis, West Nile and dengue viruses. Nat. Med. 10 (Suppl.), S98—S109.

edana, I.M., Day, N.P., Hien, T.T., Mai, N.T., Bethell, D., Phu, N.H.,Farrar, J., Esiri, M.M., White, N.J., Turner, G.D., 2002. Axonalinjury in cerebral malaria. Am. J. Pathol. 160, 655—666.

ehta, S., Pathak, S., Webb, H.E., 1990. Induction of membraneproliferation in mouse CNS by gold sodium thiomalate with ref-erence to increased virulence of the avirulent Semliki Forestvirus. Biosci. Rep. 10, 271—279.

iyake, M., 1964. The pathology of Japanese encephalitis. A review.Bull. World Health Organ. 30, 153—160.

onath, T.P., Cropp, C.B., Harrison, A.K., 1983. Mode of entryof a neurotropic arbovirus into the central nervous sys-tem. Reinvestigation of an old controversy. Lab. Invest. 48,399—410.

Japanese encephalitis in humans and a mouse model 11

Mukherji, A.K., Biswas, S.K., 1976. Histopathological studies ofbrains (and other viscera) from cases of JE virus encephali-tis during 1973 epidemic at Bankura. Indian J. Med. Res. 64,1143—1149.

Myint, K.S.A., Raengsakulrach, B., Young, G.D., Gettayacamin, M.,Ferguson, L.M., Innis, B.L., Hoke, C.H., Vaughn, D.W., 1999. Pro-duction of lethal infection that resembles fatal human disease byintranasal inoculation of macaques with Japanese encephalitisvirus. Am. J. Trop. Med. Hyg. 60, 338—342.

Nitayaphan, S., Grant, J.A., Chang, G.J.J., Trent, D.W., 1990.Nucleotide sequence of the virulent Sa-14 strain of Japaneseencephalitis virus and its attenuated vaccine derivative, Sa-14-14-2. Virology 177, 541—552.

Paessler, S., Aguilar, P., Anishchenko, M., Wang, H.Q., Aronson, J.,Campbell, G., Cararra, A.S., Weaver, S.C., 2004. The hamster asan animal model for eastern equine encephalitis — and its usein studies of virus entrance into the brain. J. Infect. Dis. 189,2072—2076.

Pathak, S., Webb, H.E., 1978. An electron-microscopic study of avir-ulent and virulent Semliki forest virus in the brains of differentages of mice. J. Neurol. Sci. 39, 199—211.

Pongponratn, E., Turner, G.D.H., Day, N.P.J., Phu, N.H., Simp-son, J.A., Stepniewska, K., Mai, N.T.H., Viriyavejakul, P., Looa-reesuwan, S., Hien, T.T., Ferguson, D.J.P., White, N.J., 2003. Anultrastructural study of the brain in fatal Plasmodium falciparummalaria. Am. J. Trop. Med. Hyg. 69, 345—359.

Scherer, W.F., Buescher, E.L., 1959. Ecologic studies of Japaneseencephalitis virus in Japan. Parts I-IX. Am. J. Trop. Med. Hyg. 8,644—722.

Shankar, S.K., Rao, T.V., Mruthyunjayanna, B.P., Devi, M.G., Desh-

Solomon, T., Dung, N.M., Kneen, R., Thao, L.T.T., Gainsborough,M., Nisalak, A., Day, N.P.J., Kirkham, F.J., Vaughn, D.W., Smiths,S., White, N.J., 2002. Seizures and raised intracranial pressurein Vietnamese patients with Japanese encephalitis. Brain 125,1084—1093.

Solomon, T., Dung, N.M., Wills, B., Kneen, R., Gainsborough,M., Diet, T.V., Thuy, T.T.N., Loan, H.T., Khanh, V.C., Vaughn,D.W., White, N.J., Farrar, J.J., 2003. Interferon alfa-2a inJapanese encephalitis: a randomised double-blind placebo-controlled trial. Lancet 361, 821—826.

Steele, K.E., Linn, M.J., Schoepp, R.J., Komar, N., Geisbert, T.W.,Manduca, R.M., Calle, P.P., Raphael, B.L., Clippinger, T.L.,Larsen, T., Smith, J., Lanciotti, R.S., Panella, N.A., McNamara,T.S., 2000. Pathology of fatal West Nile virus infections in nativeand exotic birds during the 1999 outbreak in New York City, NewYork. Vet. Pathol. 37, 208—224.

Tomimoto, H., Akiguchi, I., Wakita, H., Suenaga, T., Nakamura, S.,Kimura, J., 1997. Regressive changes of astroglia in white mat-ter lesions in cerebrovascular disease and Alzheimer’s diseasepatients. Acta Neuropathol. 94, 146—152.

Tsai, T.F., 2000. New initiatives for the control of Japaneseencephalitis by vaccination: minutes of a WHO/CVI meeting,Bangkok, Thailand, 13—15 October 1998. Vaccine 18 (Suppl. 2),1—25.

Wang, J.J., Liao, C.L., Chiou, Y.W., Chiou, C.T., Huang, Y.L., Chen,L.K., 1997. Ultrastructure and localization of E proteins in cul-tured neuron cells infected with Japanese encephalitis virus.Virology 238, 30—39.

Weaver, S.C., Barrett, A.D.T., 2004. Transmission cycles, host range,evolution and emergence of arboviral disease. Nat. Rev. Micro.

W

W

Y

Z

pande, D.H., 1983. Autopsy study of brains during an epidemicof Japanese encephalitis in Karnataka. Indian J. Med. Res. 78,431—440.

Shintaku, M., Inoue, N., Sasaki, M., Izuno, Y., Ueda, Y., Ikehara, S.,1991. Cytomegalovirus vasculitis accompanied by an exuberantfibroblastic reaction in the intestine of an AIDS patient. ActaPathol. Jpn 41, 900—904.

Solomon, T., 2004. Flavivirus encephalitis. N. Engl. J. Med. 351,370—378.

Solomon, T., Vaughn, D.W., 2002. Pathogenesis and clinical fea-tures of Japanese encephalitis and West Nile virus infections.in: Mackenzie, J., Barrett, A., Deubel, V. (Eds.), JapaneseEncephalitis and West Nile Viruses, vol. 267. Springer-Verlag,Berlin, pp. 171—194.

Solomon, T., Thao, L.T.T., Dung, N.M., Kneen, R., Hung, N.T.,Nisalak, A., Vaughn, D.W., Farrar, J., Hien, T.T., White, N.J.,Cardosa, M.J., 1998. Rapid diagnosis of Japanese encephalitisby using an immunoglobulin M dot enzyme immunoassay. J. Clin.Microbiol. 36, 2030—2034.

2, 789—801.estaway, E.G., MacKenzie, J.M., Kenney, M.T., Jones, M.K.,

Khromykh, A.A., 1997. Ultrastructure of Kunjin virus-infectedcells: colocalization of NS1 and NS3 with double-stranded RNA,and of NS2B with NS3, in virus-induced membrane structures. J.Virol. 71, 6650—6661.

ong, K.T., Shieh, W.J., Kumar, S., Norain, K., Abdullah, W.,Guarner, J., Goldsmith, C.S., Chua, K.B., Lam, S.K., Tan, C.T.,Goh, K.J., Chong, H.T., Jusoh, R., Rollin, P.E., Ksiazek, T.G.,Zaki, S.R., 2002. Nipah virus infection: pathology and pathogen-esis of an emerging paramyxoviral zoonosis. Am. J. Pathol. 161,2153—2167.

ang, K.D., Yeh, W.T., Chen, R.F., Chuon, H.L., Tsai, H.P., Yao, C.W.,Shaio, M.F., 2004. A model to study neurotropism and persistencyof Japanese encephalitis virus infection in human neuroblastomacells and leukocytes. J. Gen. Virol. 85, 635—642.

immerman, H.M., 1946. The pathology of Japanese B encephalitis.Am. J. Pathol. 22, 965.