This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Rev. Med. Virol. 2012; 22: 69–87.Published online 16 November 2011 in Wiley Online Library
Received: 11 May 2011; Revised: 5 August 2011; Accepted: 9 August 2011
INTRODUCTIONBecause of rapid changes in climate and demogra-phy, vector-transmitted arboviral diseases pose anincreasing threat to global health and welfare [1,2].Among the most severe arboviral infections knownto affect human race are those caused by membersof the Flavivirus genus of the Flaviviridae. The genuscomprises over 70 different members and includesmajor human pathogens such as Yellow Fever virus(YFV), Dengue virus (DENV), Japanese encephalitisvirus (JEV), Tick-borne encephalitis virus (TBEV),and West Nile virus (WNV) [3–5].
or: Dr. Jolanda M. Smit, Department of Med-olecular Virology Section, University Medicalntonius Deusinglaan 1, 9713 AV Groningen,
Flaviviruses are single-stranded, positive-senseRNAviruses, whose genome encodes three structuraland seven NS proteins [6]. All flaviviruses circulate intransmission cycles consisting of vertebrate hosts andinsect vectors, in which humans mostly act as dead-end hosts [7]. Natural cases of human infectionalmost invariably follow the bite of an infected tickor mosquito, although incidental cases related toother transmission mechanisms, including the use ofinfected blood products and organ transplants or,in case of TBEV, oral transfer through consumptionof unpasteurized milk (products) have, infrequently,been reported as well [4,5]. A spectrum of distinctclinical syndromes is known to complicate flavivi-rus infections in humans, ranging from relativelymild fever and arthalgia to severe hemorrhagic andencephalitic manifestations [5]. In contrast to the sys-temic syndromes, the development of encephaliticpathology relies upon the ability of the virus to gainentry to the CNS, a process known as viral neuroin-vasiveness, and to infect neural cells, a phenomenonknown as neurovirulence [8]. Interestingly, both abil-ities seem to be widely dispersed among variousmembers of the Flavivirus genus (Figure 1, red). Neu-roinvasive infections frequently occur upon infectionwith emerging viruses such as JEV, WNV, and TBEV,which have a global distribution range (Figure 2) andaffect hundreds of thousands of individuals world-wide, annually [5,9]. Recently, however, they havealso increasingly been reported in the setting of
70 G. J. Sips et al.
viruses that otherwise mostly cause hemorrhagic dis-ease such asDENV [10–12]. This phenomenon of neu-roinvasive dengue was, until recent years, largelyunrecognized and interestingly suggests the existenceof a continuum concerning the pathogenesis of
Figure 1. Maximum likelihood tree demonstrating the evolutionary reated with human CNS disease are highlighted (red). The tree was constion and the hypervariable loop excluded. Sequences were assumed totransitions and transversions and the extent of among-site variation in sAdapted and reproduced with permission from Reference 5. ALF, AlUSU, Usutu; KOU, Koutango; KUN, Kunjin; WN, West Nile; YAO, YaKOK, Kokobera; STR, Stratford; BAG, Bagaza; IT, Israel Turkey meninbusu; ILH, Ilheus; ROC, Rocio; SLE, Saint Louis encephalitis; DEN, dUganda; JUG, Jugra; POT, Potiskum; SAB, Saboya; BOU, Bouboui; EHSokoluk; YOK, Yokose; GGY, Gadgets Gully; KFD, Kyasanur ForestFETBE, Far Eastern TBE; Vs, Vasilchenko; OHF, Omsk hemorrhagic feMEA, Meaban; SRE, Saumarez Reef; TYU, Tyuleniy; APOI, Apoi; BC,bat; DB, Dakar bat; RB, Rio Bravo; MML, Montana myotis leucoenceJutiapa; SP, San Perlita; TBE, Tick-borne encephalitis; WTBE, Western
flaviviral CNS infections and potential emergence ofmore neurovirulent dengue strains.
This review aims to synthesize current knowledgeon flavivirus-induced CNS disease, including theemerging concept of neurological dengue. Apart
lationships between the different flaviviruses. Flaviviruses associ-tructed using partial NS5 sequence data with the third codon posi-evolve according to the JKY85 substitution model with the rate ofubstitution rate (i.e., Gamma distribution) estimated from the data.fuy; MVE, Murray Valley encephalitis; JE, Japanese encephalitis;ounde; CPC, Cacipacore; ARO, Aroa; IGU, Iguape; NJL, Naranjal;goencephalomyelitis virus; TMU, Tembusu; THCAr, strain of Tem-engue; SPO, Spondweni; ZIK, Zika forest; KED, Kedougou; UGS,, Edge Hill; YF, Yellow Fever; SEP, Sepik; EB, Entebbe bat; SOK,
Figure 2. Approximate geographic distribution of major emerging flaviviruses associated with human CNS disease. Based on data fromReferences 9, 14, 51–54, 89, 90 and 131
71Neuroinvasive flavivirus infections
from discussing their epidemiology, symptomatol-ogy, and neuropathology, we will specifically focuson the neuropathogenesis of highly prevalent emerg-ing viruses associated with human CNS disease. Aprofound understanding of the pathogenesis ofthese syndromes is of clear importance with respectto the development of effective therapeutic strate-gies. Furthermore, the identification of specific fac-tors involved in flaviviral neuroinvasiveness, suchas distinct viral proteins or host factors, might helpto explain the (re)emergence of specific viruses, in-cluding previously unrecognized or potentiallynovel ones, as neuroinvasive pathogens. We willstart with a discussion of JEV which, historically,has been the best studied neurovirulent flavivirusand is the cause of annual large-sized epidemics inmany Eastern Asian countries.
JAPANESE ENCEPHALITIS VIRUS
EpidemiologyJapanese encephalitis virus constitutes the mostsignificant cause of mosquito-borne encephalitisworldwide and is endemic throughout large partsof Central and Southeastern Asia, where it causesa massive total of about 30,000–50,000 reportedcases of infection, annually [9,13–15]. In endemicareas, JEV is estimated to have an asymptomatic/
symptomatic disease ratio of about 25–1000/1and about 20%–30% of all symptomatic infectionsare fatal [9,13–15]. Symptomatic infections mostlypresent as febrile syndromes that commonlyprogress into the multifocal CNS disorders thatcharacterize the disease [9,13,14]. It has been esti-mated that approximately 70% of symptomaticinfections clinically manifest as encephalitis,whereas an additional 10% present as meningitis[9,13,14]. JEV mostly affects children and nonim-mune adults and treatment remains largely sup-portive [16]. Effective vaccination schemes havebeen developed but their implementation in high-risk areas has, thus far, proven difficult for financialand logistical reasons [9,17,18].
Central nervous system diseaseNeurological symptoms typically develop after anincubation period of 5–15 days and commonly in-clude alterations of consciousness, seizures, and,specifically, the development of Parkinsonianmovement disorders and dystonias, which havebeen reported to occur in up to 60% of symptom-atic patients [13]. Another 5%–20% of patients will,additionally, present with poliomyelitis-like pyramidalmotor pathology, characterized by multifocal paralysisor paresis [13]. After an acute episode of illness, about50% of survivors retain permanent (neuro)psychiatric
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
72 G. J. Sips et al.
sequelae, which manifest as persisting cognitive ormovement disorders [13]. Although these sequelaecould result from acute disease processes, subacuteand chronic forms of JEV infection have been reportedas well. In this respect, one report has, interestingly,described relapses associated with the recovery ofinfectious virus from peripheral blood leukocytes,whereas another study described the prolonged per-sistence of IgM and viral antigens in the CSF of about5% of patients studied [19,20].
NeuropathologyAutopsy studies have identified gray matter areasof the thalamus, midbrain/substantia nigra, hippo-campus, cerebral cortices, and anterior horns of themedulla oblongata and cervical spinal cord as pri-mary JEV-affected brain regions [21–25]. Generalneuropathological alterations consist of edema,hemorrhage, vascular congestion, and widespreadperivascular inflammatory infiltrates. Characteristicof the disease are distinct foci of acellular necrotic
Figure 3. Immunohistochemical stains offlavivirus-positive brain tissue sof hippocampal pyramidal neurons in a case of JE (anti-JEVantibody, x50histochemical stain demonstratingWNV-positivity of pontine neurons inx50). Courtesy ofDr. J.Guarner (Department of Pathology and LaboratoryC Immunohistochemical stain demonstrating DENV-positivity of corticalducedwith kind permission fromReference 109. D Immunohistochemicaof TBE (anti-TBEVantibody, H/E counterstain, x40). Courtesy of Dr. E. GeVienna, Austria)
‘plaques’ confined to gray matter areas. Astrogliosisand the formation of microglial nodules, often inclose proximity to affected regions, have beendescribed as well. Furthermore, examinations of thepreviously mentioned subacute and chronic casesof infection have demonstrated diffuse calciumdeposits as well as binucleated nerve cells [21–25].Neuronal cells, particularly pyramidal (motor)neurons, clearly constitute the main cellular targetpopulation of JEV in vivo. On microscopic evalua-tion, many neurons within affected areas displayclear degenerative changes and contain viral anti-gen [21–25] (Figure 3). Infection of vascular endothe-lial cells as well as occasional ependymal cells andastrocytes have, albeit infrequently, been reportedas well [21,22].
NeuropathogenesisWith respect to exact CNS entry mechanisms of JEV,both intraneural transport through the olfactorynerve, following intranasal inoculation, as well as
amples. A Immunohistochemical stain demonstrating JEV-positivity). Reproduced with kind permission from Reference 22. B Immuno-a case ofWNE (anti-flavivirus polyclonal antibody, H/E counterstain,Medicine, EmoryUniversity School ofMedicine, Atlanta, GA,USA).neurons in a case of dengue infection (anti-DENVantibody). Repro-l stain demonstrating TBEV-positivity of cerebellar neurons in a caselpi and Prof. H. Budka (Institute of Neurology,Medical University of
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
73Neuroinvasive flavivirus infections
hematogenous transport have been described[26–29]. The first method of inoculation is unlikelyto occur in vivo, but a more indirect mechanism ofentry through the olfactory nerve following initialsystemic replication and subsequent hematogenousspread to the olfactory mucosa has, interestingly,been described in studies with the closely relatedflaviviruses Saint Louis encephalitis virus (SLEV)and Murray Valley encephalitis virus (MVEV)[30,31]. In animal models, a cytokine-mediated in-creased permeability of the blood-brain barrier(BBB) has been demonstrated, which likely pre-cedes and facilitates viral transport across the BBB[27]. Several mechanisms of transport across theBBB have been described, including direct infectionof vascular endothelial cells, transcellular transport,and within infected monocytes, the so-called“Trojan-horse” mechanism [26–29]. Within the CNS,JEV has, notably, been shown to display a particulartropism for developing neurons and neuroprogenitorcellswhichmight help to explain the viral predilectionfor specific brain regions, such as the hippocampus, aswell as the severity of JEV infections and their out-come in children [32,33]. Although this distinct tro-pism might point towards the existence of specificneuronal JEV-receptors, few studies have examinedthe presence of such receptors on cells of neural origin.So far, only one study, using mouse neuroblastomacells, has been published, suggesting that heat shockprotein 70 (Hsp70) mediates viral entry into neuronsand further studies defining the nature of potentialneuronal JEV-receptors are warranted [34]. Down-stream of cellular entry, both virus-induced apoptosisaswell as necrosis, mediated by an uncontrolled over-activation of microglia and release of reactive oxygenspecies, TNFa, and nitric oxide (NO), leading towards“bystander” damage to neuronal cells, have beendemonstrated in vitro [35–37]. Furthermore, microgliaand JEV-infected leukocytes have been identified aspossible viral reservoirs and could play a role in thepathogenesis of subacute and chronic infections, aswell as the neurological sequelae, that have beenreported following JEV-infection [38,39]. The patho-genesis of these chronic forms of JEV infection, how-ever, has, thus far, not been widely investigated andis certainly in need of further study.A large number of studies have, historically, been
performed to address specific viral determinants ofneuroinvasiveness and neurovirulence. These stud-ies have shown that a large proportion, and likelythe vast majority, of epitopes that govern JEV
neuroinvasiveness and neurovirulence are locatedwithin relatively limited sections of the viral enve-lope (E) protein. Particularly, these include areaswithin the lateral surface of domain III as well asbase of domain II of E, which are believed to playcrucial roles in cellular receptor binding and fusionwith target cells, respectively [8,40–45]. Apart fromthese studies clearly indicating an important rolefor the E protein, a relatively limited number of re-cent studies have demonstrated the involvement ofother viral proteins in governing viral neuroinva-siveness and neurovirulence. Several studies have,in this respect, indicated the effect of mutations inthe viral capsid (C) and premembrane (prM) pro-teins in limiting viral neurovirulence [46,47]. Also,it has recently been shown that production of theNS1’ protein, which occurs as a result of ribosomalframe-shifting in members of the JE-serocomplex,but not in most other flaviviruses, increases viralneuroinvasiveness [48].
WEST NILE VIRUS
EpidemiologyWest Nile virus, a mosquito-borne member of theJE-serocomplex, which has historically been endem-ic throughout large parts of Africa, Asia, Australia,and Europe, caused a massive outbreak of humandisease in the New York area in 1999 and, sincethen, has rapidly spread throughout the NorthAmerican continent [49–53]. Serological studieshave indicated the circulation of the virus in anumber of Latin American countries as well, butreports on human infection have thus far remainedsparse [52,54].
Following the 1999 epidemic, WNV has becomethe leading cause of arboviral encephalitis in theUSA and, here alone, a total of about 30,000 casesof infection have been reported during the lastdecade, of which approximately 1,200 (4%) havebeen fatal [55]. It is estimated that about 80% ofinfections are asymptomatic, whereas symptomaticinfections mostly give rise to the development of aself-limited febrile syndrome known as West Nilefever (WNF) [49,50]. About 1/150 patients developCNS complications, which are usually groupedtogether under the term West Nile neuroinvasivedisease (WNND) [49,50]. Unlike some other flavi-viruses, WNV, notably, mostly appears to affectthe elderly and immunocompromised, even whenintroduced into largely naïve populations as occurred
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
74 G. J. Sips et al.
in 1999 [49,50]. Currently, treatment remains largelysupportive, although extensive scientific efforts arebeing made to develop therapies for future clinicaluse [56].
Central nervous system diseaseWest Nile neuroinvasive disease can be subclassifiedinto three main clinical syndromes of meningitis, en-cephalitis, and acute flaccid paralysis/poliomyelitis[49,50]. Of these syndromes, the clinical picture ofacute flaccid paralysis (AFP) is the most distinc-tive as well as best characterized entity. In contrastto what holds true for most other arboviral en-cephalitides, neuromuscular weakness constitutesa prominent finding in WNND, occurring in upto 50% of patients, sometimes in the absence ofother disease symptoms [49]. AFP typically pre-sents as monoplegia, asymmetric upper or lowerextremity weakness, or generalized asymmetrictetraplegia or quadriplegia. Additionally, in about70% of patients with AFP, there is involvement ofone or more of the cranial nerve(s) and a largenumber of patients require intubation or ventila-tion because of respiratory failure [49,50]. Al-though most patients with WNV meningitiswithout focal neurological deficits tend to recoverfully, the prognosis is much worse in cases of en-cephalitis or AFP, which are characterized by a10%–20% mortality rate [49]. Up to 70%–75% ofsurvivors of WNND, furthermore, retain perma-nent neurological sequelae [49]. Recently, subacuteand relapsing forms of AFP, as well as long-termpersistence of WNV associated with viral shed-ding in urine, have been reported in subsets ofpatients, which might provide clues with respectto possible mechanisms of flavivirus persistenceas well as the frequent occurrence of postinfec-tious sequelae [57,58].
NeuropathologyHistologically, WNND is characterized by apattern of microglial nodules, perivascular in-flammatory infiltrates and reactive astrogliosiscombined with neuronal loss, necrotic foci, andneuronophagia [59–61]. Topographically, there isa clear predilection for gray matter areas of thebrainstem, particularly the medulla, and spinalcord [59–61]. Brain areas that might, additionally,be affected include the cerebellum, temporallobes, basal ganglia, and thalamus [59–61]. This
predilection in terms of affected brain areas hasbeen confirmed by immunohistochemical stud-ies [59–61] (Figure 3). The virus primarily infectsneurons, mostly pyramidal motor neurons ofthe anterior horns and cerebellar Purkinje cells,although there have been occasional reports ofinfection of astroglial and monocytic cells aswell [59–62].
NeuropathogenesisThe recent North American epidemics have greatlyfuelled WNV-related research and, as a conse-quence, led to a vast increase in our knowledge ofits neuropathogenesis as well as, potentially, thatof other neuroinvasive flaviviruses.
Importantly, several adverse effects of innate andadaptive systemic antiviral immune responseshave, during recent years, been described, whichlead towards increased permeability of the BBBand, hence, likely facilitate viral entry into theCNS [63]. Specifically, increases in brain endothelialcapillary permeability have been reported, inducedby the TLR3-mediated release of TNFa, as well asby macrophage migration inhibitory factor (MIF),intercellular adhesion molecule-1 (ICAM-1), andmatrix metallopeptidase 9 (MMP9). Dysregulationsof TLR3-responses have specifically been demon-strated to compromise BBB-integrity in the elderly[64–68]. Following this breakdown, the virus hasbeen suggested to cross the BBB via severalmechanisms, including transcellular transport,paracellular transport, direct infection of endothe-lial cells, or Trojan-horse mechanisms of entry[69]. Another pathway, which was shown to di-rectly induce AFP in animal models of infection,includes retrograde axonal transport throughperipheral motor nerves [70].
Upon entry of the CNS, WNV displays a particu-lar tropism for (anterior horn motor) neurons. Notmany studies have, so far, been undertaken to iden-tify possible neuronal WNV-receptors, althoughone study, interestingly, described the presence of aplasma membrane glycoprotein of Mr 105,000 thatfacilitated viral entry in murine neuroblastoma cells[71]. Once infected, neural cells have been demon-strated to undergo various mechanisms of apopto-sis [72–75]. Bystander damage, resulting fromimmunopathological effects of the CD8+ T-cell re-sponse as well as the recruitment of inflammatorymonocytes mediated by chemokine (C-C motif)ligand-2 (CCL-2), has been described, as well
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
75Neuroinvasive flavivirus infections
[76,77]. A number of recent studies have investigatedthe pathogenesis of persistent WNV infections.These studies have demonstrated persistence ofWNV in the CNS and peripheral organs, particularlythe urinary tract, and correlated neurological se-quelae to persistent viral disease activity in theCNS [78–80].As for JEV, the E protein appears to be particu-
larly important in governing the neuroinvasivenessand neurovirulence of WNV [40,81–83]. Additionalroles for the C as well as several NS proteins havebeen described in governing viral neurovirulence,but further studies addressing their roles in flavi-viral neuropathogenesis would be required [62,84].Interestingly, a number of predisposing host fac-
tors have now been identified. Of these, genetic riskfactors include mutations in the C—C chemokinereceptor type 5 (CCR5) and 2–5 oligoadenylatesynthethase (OAS) genes, which play importantroles in antiviral immune responses [85,86]. Fur-thermore, a number of acquired, age-specific, T-celldefects in both CD4 as well as CD8 subsets havebeen described, which, at least in animal models,greatly increase host susceptibility to severe andneuroinvasive WNV infection [87].
DENGUE VIRUS
EpidemiologyThe mosquito-borne DENV is by far the mostimportant arbovirus known to affect mankind andconstitutes a significant public health problem,particularly in the developing world. Accordingto the WHO, DENV is now endemic in over 100different countries, where some 2.5 billion peopleare at risk of getting infected. In these regions, ap-proximately 50–100million cases of DENV infec-tion occur annually, about 250,000 to 500,000 ofwhich are cases of severe dengue hemorrhagicfever (DHF) [88,89]. Outside of the more traditionalregions, a re-emergence of autochtonous DENVtransmission within Europe has, after a long periodof absence, recently, been described [90]. Approx-imately 50% of infections are asymptomatic,whereas symptomatic infections can present witha variety of clinical syndromes, ranging from aspe-cific or mild-febrile disease to the aforementionedDHF or dengue shock syndromes (DSS) [88,89,91].Treatment of DENV remains supportive and thereis a strong and urgent need for effective therapeuticand vaccination strategies [88,89,91,92].
Central nervous system diseaseApart from hemorrhagic disease symptoms, den-gue can also present with a number of less typicalsymptoms and, in this respect, there has, interest-ingly, been a recent re-appraisal of its neurologicalcomplications [10,11,93]. Although a large propor-tion of neurological complications probably resultfrom the consequences of systemic infection, andhave been termed “dengue encephalopathy”, ithas, during recent years, become apparent thatDENV indeed also causes true neuroinvasive dis-ease in subsets of infected individuals [10,11,93].According to various large epidemiological studies,neurological manifestations make up part of theclinical picture of approximately 1%–5% of all casesof symptomatic DENV infection and, in endemicareas, DENV might represent a significant andpotentially underreported cause of viral enceph-alitis [12,94–101]. Notably, these figures are roughlycomparable to those of WNV in the Westernhemisphere and the large numbers of individualsannually affected by DENV turn the concept ofDENV-induced CNS disease into a potentiallyworldwide phenomenon that is of considerableclinical significance.
In general, neurological dengue can presentwith a wide variety of CNS manifestations, whichcommonly include non-specific alterations of con-sciousness, seizures, headache, and meningealsigns but, in analogy to JEV and WNV, may alsoinclude paralytic or Parkinsonian symptoms[10,11,93]. Generally, neurological dengue is associ-ated with a poor outcome. Risk factors for the dis-ease include infection with serotypes 2 or 3 andthe age of the patient, younger children carrying ahigher risk of developing neurological disease thanolder ones [12,93].
NeuropathologyLittle is known about the exact pathology of neu-rological dengue. Relatively few studies haveaddressed this phenomenon, although their num-ber has vastly increased during recent years. Anumber of early autopsy studies have reportedgeneral neuropathological alterations, such asedema, vascular congestion, and perivascular lym-phocytic infiltration, in the CNS of patients withdengue [102,103]. These studies, furthermore,reported distinct neuronal abnormalities, manyneurons being acidophilic or displaying a clear
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
76 G. J. Sips et al.
shrinkage of cytoplasm. A number of recent studieshave demonstrated high positivity rates of CSFsamples for DENV RNA and DENV-specific IgMor IgG in patients with neurological dengue, indi-cating that direct neuroinvasion might occur in aconsiderable fraction of these patients [99,104,105].Another study on patients with neurological dengueinterestingly reported infiltration of both gray andwhite matter areas with DENV-positive macro-phages that were often found in close proximity toneurons demonstrating clear cytopathic alterations[106]. Additionally, various studies have demon-strated the presence of DENV antigens in neurons,astrocytes, microglia, endothelial and perivascularcells, or recovered viral RNA by RT-PCR from braintissue samples [94,107–109] (Figure 3).
NeuropathogenesisThe concept of neuroinvasive dengue has arisenrelatively recently and, as a consequence, its neuro-pathogenesis largely remains elusive. Much of ourpresent knowledge on this topic, interestingly,comes from animal models that were originallyaimed at studying hemorrhagic disease. In manyof these models, DENV was shown to induce neu-rological instead of hemorrhagic syndromes and,therefore, they have been very successful in iden-tifying a number of possible neuropathogenicmechanisms and underlying virus-host interactions[110]. With respect to entry into the CNS, a cytokine-mediated breakdown of the BBB and Trojan-horsemechanism of entry have been suggested [106,111].Furthermore, a distinct viral tropism for neurons ofthe anterior horns, hippocampus, cerebral cortex,and olfactory bulb has been demonstrated in vivo,and DENV-triggered apoptosis has been shown tooccur in human and murine neurons both in vivoand in vitro [112–117]. Interestingly, as was, in aslightly different way, suggested for JEV, Hsp70,together with Hsp90, has been shown to form acandidate receptor complex governing DENV entryin human monocytes as well as neuroblastoma cells[118]. Additionally, a possible DENV receptor ofMr 65,000 has been identified on human andmurineneuroblastoma cells as well, although the role ofboth proteins as potential neuronal dengue recep-tors in vivo requires further elucidation [119]. Asfor WNV and JEV, mutations within severaldomains of the DENV E protein have been shownto mediate DENV neuroinvasiveness and neuro-virulence in animal models of infection [120–125].
OTHER MOSQUITO-BORNE CAUSES OFCENTRAL NERVOUS SYSTEM DISEASEApart from the viruses discussed so far, two othermembers of the Japanese encephalitis serocomplex,SLEV and MVEV, as well as a member of the Ntayaserocomplex, Rocio virus (ROCV), have been asso-ciated with the development of human CNS dis-ease [5]. Although these viruses have causedconsiderable epidemics in the past, they have notdone so during recent decades, the reasons forwhich are not understood. Occasional cases of hu-man infection, particularly for SLEV, continue tobe reported and proof of the continuous circulationof all of these viruses in various vertebrate hosts inwildlife exists [126–129]. Factors and mechanismsexplaining why these viruses apparently have not,in recent years, re-emerged as major human patho-gens, whereas closely related flaviviruses have, areimportant topics of further study and will likelyprovide more general insight into flavivirus ecol-ogy and virus-host interactions.
TICK-BORNE ENCEPHALITIS VIRUS
EpidemiologyTick-borne encephalitis virus is the most commoncause of arboviral encephalitis in Europe and, interms of annual morbidity, second only to JEVamong the neurovirulent flaviviruses [130–132].Historically, TBEV has been endemic in many partsof Central Europe, the former Soviet Union andAsia, but, more recently, has emerged in an increas-ing number of Western European countries as well[130–133]. Phylogenetically, European (TBEV-Eu),Siberian (TBEV-Sib), and Far-Eastern (TBEV-FE)TBEV subtypes are recognized, which togetherhave accounted for annual averages of about 9000reported cases of infection during the past two dec-ades [130–133]. It is estimated that about 70%–95%of all cases of TBEV infection occur asymptomati-cally [132]. Neurological disease manifests as men-ingitis in 50% and (meningo)encephalitis in theother 50% of symptomatic cases [132]. Mortalityrates have been reported to range from 0.1%–4%,upon infection with TBEV-Eu, to up to 20%–40%following infection with TBEV-FE [130–132]. LikeWNV, TBEV mostly affects the elderly [131]. Anumber of antiviral vaccines are available andlarge-scale vaccination programs have, so far withvarying rates of success, been implemented in anumber of countries where TBEV is endemic [133].
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
77Neuroinvasive flavivirus infections
Central nervous system diseaseCommon neurological symptoms have been report-ed to include ataxia and tremors. Approximately10%–15% of symptomatic cases are complicated bythe development of poliomyelitic pathology thatmostcommonly affects the upper limbs [131]. Neurologicalsequelae develop in about 20%–50% of survivors andchronic forms of TBEV have been reported as well[130–132]. Most cases of chronic TBEV infection havebeen linked to specific TBEV (TBEV-Sib) subtypes,possibly pointing towards the importance of specificviral factors in the pathogenesis of chronic flaviviralCNS disease [134].
NeuropathologyTick-borne encephalitis virus induces widespreadinflammatory changes, characterized by diffuseinflammatory infiltrates in combination with astro-gliosis, the formation of microglial nodules, neurono-phagia, and varying degrees of neuronal loss [135]. Ingeneral, the most frequently affected areas, in declin-ing order, include the anterior horns of the spinal cord,brainstem, cerebellum, and basal ganglia [135]. Im-munohistochemical experiments have demon-strated that many large neurons within affectedareas contain viral antigen, although, interestingly,an inverse correlation between the number ofinfected neurons and magnitude of the infiltratingimmune response was observed, suggesting un-derlying immunopathogenic mechanisms [135](Figure 3).
NeuropathogenesisVery little is known about the exact route of entry ofTBEV into the CNS. This is likely mostly hematoge-nous, because a high level of peripheral viremiaappears to be a prerequisite for the developmentof neurological symptoms [136]. Individual reports,however, correlating tick bites of the upper trunk tothe development of localized shoulder girdle paral-ysis and paresis, suggest direct entry via peripheralnerves might take place as well [137]. Within theCNS, neurons are the most affected cell types anda number of human neural cell lines have been dem-onstrated to undergo apoptosis as well as necrosisupon infection with TBEV in vitro [138]. Autopsystudies, however, have been inconsistent in demon-strating the occurrence of neuronal apoptosisin vivo and, furthermore, imbalances between viral
loads and the magnitude of the infiltrating immuneresponses have, as mentioned previously, beendemonstrated, indicating potentially underlyingimmunopathogenic mechanisms [135]. Indeed, amore detailed examination of the anti-TBEV im-mune response in post-mortem tissue sectionsindicated that CD8+ granzyme B-releasing cytotoxicT-cells might significantly contribute to neuronaldamage in vivo via the induction of bystander dam-age [139]. This pathological role of CD8+ T cells, aswell as adverse effects of an elevated TNFaresponse, was later confirmed by in vivo animalexperiments, suggesting that both viral as well asimmunological factors determine the eventual out-come of TBEV-infections [140,141]. A number ofspecific viral and host factors have now been iden-tified. Important host factors, as for WNV, havebeen demonstrated to include genetic alterationsin the CCR5, OAS, and TLR3 genes, which play cru-cial roles in antiviral immune responses [142–144].Compromised T-cell responses have been sug-gested to significantly contribute to the develop-ment of chronic TBEV infections, and, furthermore,in some of these chronic cases autoantibodiesagainst axonal neurofilaments were found whichwere absent in acute cases of TBEV [145,146]. Anumber of mutations in the viral genome havebeen demonstrated to mediate viral neuroinva-siveness and/or neurovirulence. Most promi-nently, these include mutations within the lateralregion of domain III of the TBEV E protein, as isthe case for many mosquito-borne flavivirusesdiscussed so far as well [147–149]. Additionally,mutations within the 3’-noncoding region (NCR)of the TBEV genome, probably affecting viralRNA replication, as well as the viral C protein, in-terfering with virus assembly, have been shown toalter viral neuroinvasiveness as well as neuroviru-lence in animal models [150]. A recent study,furthermore, provided evidence that TBE virusesnaturally exist as quasispecies populations andthat attenuation of the viral virulence profiledepends upon selection out of this pre-existingpool of viruses rather than upon random muta-genesis. The nonstructural NS2B and NS3 pro-teins have been suggested to play an importantrole in this selection process [151]. Furthermore,another nonstructural protein, NS1, of particularstrains of TBEV-Sib has, interestingly, beendemonstrated to play a role in the developmentof chronic TBEV [134].
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
78 G. J. Sips et al.
OTHER TICK-BORNE CAUSES OF CENTRALNERVOUS SYSTEM DISEASEApart from TBEV, two other members of the TBE-serocomplex, Powassan virus (POWV) and Loupingill virus (LIV), have been associatedwith the develop-ment of human CNS disease [5]. As is true for theless common mosquito-borne viruses, these viruseshave not caused any large human outbreaks recently,despite their continuous circulation among varioustypes of vertebrate hosts in wildlife [152,153].
CONCLUSIONS AND FUTURE PERSPECTIVESAs demonstrated in this review, the Flavivirus ge-nus of the family Flaviviridae consists of a groupof highly important human pathogens, many ofwhich possess the capacity to induce a range ofspecific CNS diseases in infected hosts. Here, wehave reviewed the epidemiology, symptomatology,pathology, and, specifically, pathogenesis of neu-roinvasive flavivirus infections, combining andcomparing current knowledge of all major emerg-ing flaviviruses associated with human CNSdisease (summarized in Tables 1 and 2). In thisrespect, an interesting and, thus far, not much-studied phenomenon is the pathogenesis of neu-rological dengue. DENV is an example of a highlyprevalent flavivirus, which, under most circum-stances, displays a relatively low tendency to cause
Table 1. Overview of the epidemiology, symptoCNS diseases caused by major emerging flavivir
Virus Endemic areas CNS disease(% symptomatic disease
clinically overt CNS infections but appears to do soin a subset of cases, when specific conditions aremet.This suggests the existence of a kind of continuumwith respect to the pathogenesis of flavivirus-induced CNS disease. A number of common themes,both in terms of the neuropathogenesis as well asneuropathology of neuroinvasive flaviviruses, can,indeed, be identified and provide interesting ave-nues for future research (Figure 4, Table 2).
Synthesis of the reviewed data reveals that allneuroinvasive flaviviruses infect a relatively limitednumber of highly specific brain regions involved inmotor control, including the thalamus, basal ganglia,brainstem, and anterior horns of the spinal cord,resulting in distinct neurological disease symptoms.It, therefore, remains of clear interest to study theseas well as other viruses that specifically target thesebrain areas with respect to the etiology and patho-genesis of, especially transient, sporadic, or idio-pathic cases of, motor disorders of unknown originin which identical brain regions are affected [154].
Viral entry into the nervous system plays a keyrole in the pathogenesis of flavivirus-inducedCNS disease. (Severe) systemic infections, resultingin a mass release of inflammatory factors and cyto-kines, might pave the way for CNS infections bycompromising BBB-integrity. In this process, hostfactors governing antiviral immune responses
matology and current treatment options ofuses associated with human CNS disease
Figure 4. Overview of different entry mechanisms (A) as well as commonly affected brain areas (B) in flavirus-induced CNS disease
80 G. J. Sips et al.
might, as has now been demonstrated, play an im-portant role. Another interesting mode of viralentry into the CNS is provided by axonal transport,where viruses hitchhike along existing neuronaltransport pathways. Viral spread within the CNSmight occur via these axonal pathways as well, asmany affected brain areas have been demonstratedto be interconnected structures involved in motorfunctioning. Indeed, flaviviruses have been demon-strated to possess a specific tropism for (motor)neu-rons, which have been demonstrated to undergovarious mechanisms of apoptosis and/or necrosisfollowing infection. Although specific neuronalreceptors have thus far not clearly been identified,the flaviviral E protein, which is directly involved incellular receptor recognition, has repeatedly beendemonstrated to be highly important in governing vi-ral neuroinvasiveness and neurovirulence (Figure 5).Increased knowledge of flaviviral neuropatho-
genesis, as reviewed here, has significantly contrib-uted to stronger evidence-based preventive andtherapeutic options and has considerably im-proved our insight into the structural and geneticalmechanisms that have enabled these viruses to(re)emerge as neuroinvasive pathogens. Mutationstudies revealing the structural regions within theviral genome that determine viral neuroinvasive-ness and neurovirulence, including the lateralsurface of domain III as well as the base of domainII of the E protein, have, in this respect, been highlyimportant for the development of safe, attenuatedvaccine strains, many of which are now beinginvestigated in (pre)clinical trials [17,92]. Further-more, the natural occurrence of mutations leading
to successfully replicating wild-type strains ofincreased neuroinvasiveness might explain whycertain flaviviruses have evolved and (re)emergedas specific neuroinvasive pathogens. Host and eco-logical factors have probably played a role in thisprocess as well, because closely related viruses thatdo have neurovirulent potential are probably, forother reasons, not circulating to a sufficient extentin the human population, while, at the same time,not all individuals infected with widely prevalentneurovirulent viruses eventually develop CNSsymptoms. The association of dengue with neu-roinvasive disease and emergence of WNV as thecause of WNND in the Western hemisphere are,in light of the (re)emergence of neuroinvasive flavi-viral strains, interesting examples, as WNV hadhistorically mostly been associated with relatively“mild” and dengue, in its severe manifestations,with “hemorrhagic” disease. A particular insightinto the neuropathogenesis of neurological dengueis, in this respect, warranted.
Increased knowledge on flaviviral neuropatho-genesis will be crucial for the development of thera-peutic approaches aimed at mitigating seriousneurological disease complications as well. A spe-cifically important prerequisite of these therapiesshould be their ability to cross the BBB and becomelocally available within the CNS. In this respect, asan important example, humanized monoclonalantibodies have now been described for WNV thatwere not only able to prevent, but also treatneurological complications once infection of theCNS had established [155]. Despite their hematoge-nous administration, these antibodies were,
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
Figure 5. Structural overviewof the envelope of amatureflavivirus virion (A) aswell as severalmolecular determinants of neuroinvasiveness andneurovirulence locatedwithin specific sections of theEprotein (B). FigureA illustrates the organization of Eproteins on the viral envelope.Groupsof three parallel homodimers are clustered within dictinct rafts which, together, form a typical herringbone pattern on the viral surface. A singleraft is highlighted and symmetry axes as well as the respective domains of the E proteins are indicated (domain I: red; domain II: yellow; domainIII: blue and fusion loop: green) Figure B provides a structural close-up of an individual homodimer in which several amino acids that have, ex-perimentally, been shown to alter neuroinvasiveness and neurovirulence are highlighted (numbered arrows). Note that many of the indicatedmutations map to the lateral surface of domain III or the base of domain II of E (circles), indicating the potential role of specific cellular receptorsor attachment factors in the pathogenesis of flavivirus-induced CNS disease. Adapted and reproduced with kind permission from Reference 45.The amino acid numbers as well as their approximate positions, as depicted in Figure 5B, are based on data from Reference 8
81Neuroinvasive flavivirus infections
furthermore, able to prevent neuronal spread andthe subsequent development of AFP in animalmodels of infection. In light of flaviviral neuro-pathogenesis, this is highly relevant as various fla-viviruses might spread towards, and within, theCNS through axons, potentially rendering purelyperipherally acting treatment methods ineffective[70,156]. Trials with these antibodies have now beenstarted and it will be interesting to analyze their ef-fectivity in the clinical setting [56]. Additionally,again fuelled by a detailed knowledge of neuro-pathogenic mechanisms, several approaches aimedat maintaining the integrity of the BBB, possibly by
directly acting on compromising factors such asMMP9, as well as inhibiting different mechanismsof neuronal apoptosis, are currently being investi-gated in in vitro as well as in vivomodels of neuroin-vasive infection [73,157–160]. It will be interestingto pursue these and similar lines of research furtherand examine whether they can potentially be ex-trapolated to other (neuroinvasive) flaviviruses orcombined in order to develop more effectivetreatments.
Further research into flaviviral neuropathogen-esis is, in light of these therapeutic efforts as wellas the unparalleled prevalence and impact of the
Rev. Med. Virol. 2012; 22: 69–87.DOI: 10.1002/rmv
82 G. J. Sips et al.
respective viruses, of major global importance. Itshould specifically include the study of, thus far,less intensively investigated topics, including thepotential existence of specific neuronal receptors,prevalence, potential clinical significance, andpathogenesis of chronic and persistent flaviviralCNS infections, and exact characteristics of neuro-logical dengue. Taking into account their full zoo-notic spectrum as well as potential to (re)emerge,future research should not be limited to majorcauses of human encephalitis but should includethe study of other, even currently less prevalentor significant, flaviviral causes of CNS disease inhumans and other vertebrates alike.
CONFLICT OF INTERESTThe authors have no competing interest.
ACKNOWLEDGEMENTSThe authors would like to thank dr. J. Guarner(Department of Pathology and Laboratory Medi-cine, Emory University School of Medicine,Atlanta, GA, USA) as well as dr. E. Gelpi andprof. H. Budka (Institute of Neurology, MedicalUniversity of Vienna, Austria) for kindly providingus with representative immunohistochemical stainsof WNV-infected and TBEV-infected human braintissue samples, respectively (as shown in Figure 3).Furthermore, we would like to thank all otherauthors and publishers who have kindly allowedus to use previously published materials, in partor in whole, as part of this manuscript and apolo-gize to those colleagues whose work could not beincluded in this review because of lengthconstraints.
REFERENCES
1. Morens DM, Folkers GK, Fauci AS. The
challenge of emerging and re-emerging
infectious diseases. Nature 2004; 430:
242–249. DOI: 10.1038/nature02759
2. Mackenzie JS, Gubler DJ, Petersen LR.
Emerging flaviviruses: the spread and re-
surgence of Japanese encephalitis, West
Nile and dengue viruses. Nature Medicine
2004; 10: S98-109. DOI: 10.1038/nm1144
3. Weissenbock H, Hubalek Z, Bakonyi T,
et al. Zoonotic mosquito-borne flavi-
viruses: worldwide presence of agents
with proven pathogenicity and potential
candidates of future emerging diseases.
Veterinary Microbiology 2010; 140: 271–280.
DOI: 10.1016/j.vetmic.2009.08.025
4. Dobler G. Zoonotic tick-borne flaviviruses.
Veterinary Microbiology 2010; 140: 221–228.
DOI: 10.1016/j.vetmic.2009.08.024
5. Gould EA, Solomon T. Pathogenic flavi-
viruses. Lancet 2008; 371: 500–509. DOI:
10.1016/S0140-6736(08)60238-X
6. Mukhopadhyay S, Kuhn RJ, Rossmann
MG. A structural perspective of the flavi-
virus life cycle. Nature Reviews Microbiol-
ogy 2005; 3: 13–22. DOI: 10.1038/
nrmicro1067
7. Weaver SC, Barrett AD. Transmission
cycles, host range, evolution and emer-
gence of arboviral disease. Nature Reviews
Microbiology 2004; 2: 789–801. DOI:
10.1038/nrmicro1006
8. McMinn PC. The molecular basis of viru-
lence of the encephalitogenic flaviviruses.
Journal of General Virology 1997; 78(Pt 11):
2711–2722.
9. Solomon T. Flavivirus encephalitis. The
New England Journal of Medicine 2004; 351:
370–378. DOI: 10.1056/NEJMra030476
10. Varatharaj A. Encephalitis in the clinical
spectrum of dengue infection. Neurology
India 2010; 58: 585–591. DOI: 10.4103/
0028-3886.68655
11. Murthy JM. Neurological complication
of dengue infection. Neurology India 2010;
58: 581–584. DOI: 10.4103/0028-3886.
68654
12. Solomon T, Dung NM, Vaughn DW, et al.
Neurological manifestations of dengue in-
fection. Lancet 2000; 355: 1053–1059. DOI:
10.1016/S0140-6736(00)02036-5
13. Misra UK, Kalita J. Overview: Japanese enceph-
alitis. Progress in Neurobiology 2010; 91: 108–120.