Encephalitis caused by flaviviruses

Abstract

The genus Flavivirus, family Flaviviridae, contains some of the most important arboviral pathogens of man. The genus includes several aetiological agents of encephalitis, the most significant being Japanese encephalitis virus, West Nile virus and tick-borne encephalitis virus. In each case, the majority of exposed individuals will not develop disease, but a minority will develop a severe illness with a significant chance of permanent neurological damage or death. The factors that determine this are numerous, involving complex interactions between virus and host and are still being actively uncovered. In many cases it appears that the immune response, while crucial to containing the virus and limiting spread to the brain, is also responsible for causing neurological damage. Innate responses can limit viral replication but may also be responsible for generating pathological levels of inflammation. Neutralizing antibody responses are protective but take time to develop. The role of T cells is less clear, and may be either protective or pathogenic. This review summarizes recent developments in the understanding of the pathogenesis of encephalitis caused by flaviviruses.

Although there are about 70 members of the genus Flavivirus (family Flaviviridae), the most important causes of encephalitis are Japanese encephalitis virus (JEV), West Nile virus (WNV) and Tick-borne encephalitis virus (TBEV). The genus is named after yellow fever virus (in Latin, yellow is flavus), and also includes dengue viruses, which cause fever and rash, as well as St Louis encephalitis virus (SLEV) and Murray Valley encephalitis virus (MVEV).

This review will briefly consider the epidemiology and clinical features of flavivirus encephalitis before examining recent advances in the immunology and pathogenesis; although flaviviruses cause encephalitis on five continents, the review will focus on JEV, WNV and TBEV, drawing on recent data from clinical studies and mouse models.

Epidemiology and clinical features

Since they are transmitted by arthropods (insects or ticks) these viruses are known as arboviruses (arthropod-borne viruses), and except for dengue, they are all zoonotic (i.e. animal viruses that spill over into humans). JEV, WNV, SLEV and MVEV are transmitted principally among birds by Culex mosquitoes that breed in muddy water. Although they are genetically closely related viruses, they are found in geographically different parts of the globe; however, they are tending to spread and can cause unexpected outbreaks. JEV occurs in the Asia-Pacific region, MVEV in Australia and nearby islands and SLEV is confined to the Americas¹; WNV was found in Africa and the Middle East, but in recent years has caused outbreaks in Southern Europe, and reached America, where it rapidly spread across the continent.² TBEV has quite different ecology and epidemiology. It is transmitted between small mammals in the forests of central Europe and Russia by hard (ixodid) ticks.

Many infections with these viruses are asymptomatic, or cause a non-specific febrile illness. Neurological presentations include aseptic meningitis, febrile convulsions in children, encephalitis or myelitis—which may cause a polio-like flaccid paralysis. Since it is more abundant, JEV tends to infect children, and so most disease is in this group; WNV, SLEV and MVEV are more likely to cause disease in adults.

Pathogenesis

The pathogenesis of flavivirus encephalitis remains incompletely understood. Aspects of viral replication, central nervous system (CNS) invasion and the nature of the immune response are still being uncovered. Flaviviruses are small single-stranded, positive-sense enveloped RNA viruses. The degree of sequence diversity seen in flaviviruses is less than that observed in other RNA viruses that establish chronic infection, such as HIV and hepatitis C virus. Although viral variation does play an important role in flavivirus pathogenesis, the host immune response is probably a more critical determinant of clinical outcome, at least in humans. An overview of the pathogenesis of flavivirus encephalitis is presented in Figure 1.

Figure 1.

Pathogenesis of flavivirus encephalitis. Key mechanisms in immune control or immunopathology of flavivirus infection are shown in capitals. TLR, toll-like receptor; IL-6, interleukin 6; TNF, tumour necrosis factor; CXCL, C-X-C motif-containing chemokine ligand.

Viral entry and replication

Following inoculation from the bite of an infected mosquito, virus is thought to undergo replication in the local tissues mainly in Langerhans cells. For WNV, replication in keratinocytes, which gives a much larger target cell reservoir, has also been shown.³ Following replication, virus disseminates to the local lymph nodes where further replication in monocyte lineage cells takes place, causing a viraemia that is followed later by spreading to the brain. In mice, high rates of JEV replication within dendritic cells are linked to higher overall mortality⁴ and stopping early viral replication in monocyte-lineage cells using small-interfering RNAs prevent development of WNV and SLEV encephalitis.⁵ Interestingly, insertion of brain-specific microRNA complementary sequences into flavivirus genomes abolished virulence suggesting similar technology can in principle work in the brain as well.⁶ Autophagy (internal phagocytosis for recycling cell components), which is known to be important in dengue virus and HCV replication, is also involved in early JEV replication. Enhancement of autophagy increased JEV replication and inhibition reduced JEV replication.⁷

How the neurotropic flaviviruses gain access to the CNS remains incompletely understood, but interactions at the blood–brain barrier (BBB) are probably critical. Proposed mechanisms include active replication within endothelial cells, passive transfer across the BBB or within leucocytes that migrate across the barrier.

Innate immunity

Flaviviruses interact with several host cytosolic pathogen-recognition receptors that detect RNA viruses, such as toll-like receptor (TLR)3, TLR7 and the retinoic acid-inducible gene (RIG)-I. For WNV, TLR3 and intracellular adhesion molecule (ICAM)1 (an adhesion molecule found on the vascular endothelium) have been implicated in viral CNS entry through increasing BBB permeability,^8,9 though the detrimental role of TLR3 in mice has been questioned more recently.¹⁰ In human primary macrophages from young donors TLR3 is downregulated upon infection with WNV, but not in elderly donors (in which age group WNV infection is more commonly fatal).¹¹ The decrease in TLR3 expression resulted in turn in reduced production of pro-inflammatory cytokines in younger donors, consistent with the inflammatory BBB breakdown model of CNS entry in human flavivirus encephalitis. In monocyte-derived dendritic cells the opposite is true, however; younger donors exhibit higher TLR3 expression in response to WNV infection and greater type I interferon (IFN) production.¹² Furthermore, in the case of JEV, for many years clinical and pathological studies have suggested that co-infection with neurocysticercosis (which compromises the BBB), increasing the risk of developing encephalitis following infection with the virus.¹³

After infection is established, neurotropic flaviviruses interfere with innate immunity at a number of levels. Flaviviruses are well known to interfere with IFN antiviral pathways. More recently, WNV was shown to inhibit complement activation by at least two mechanisms, both in solution¹⁴ and on the cell surface.¹⁵ Several genes are known to influence immunity to flaviviruses in animal models. Recently, evidence has accumulated that several of these genes also predispose to disease in humans. 2′–5′ oligoadenylate synthetases (2′–5′ OAS) promote degradation of human and viral RNA. Single-nucleotide polymorphisms (SNPs) in genes encoding for these proteins are associated with susceptibility to infection with WNV in humans¹⁶ (although SNPs in OAS1, not OASL as originally reported, appear to be functionally important).¹⁷ OAS1 SNPs are also associated with susceptibility to WNV disease in horses (which have a similar OAS gene cluster and response to WNV as humans).¹⁸ More recently, OAS2 and OAS3 from the same gene family have been implicated in susceptibility to TBEV disease in humans.¹⁹ In parallel to the findings in vitro and in animal models described for WNV and TLR3 above, a functional TLR3 gene has been associated with increased susceptibility to TBEV infection,²⁰ suggesting that in humans, overall the role of TLR3 may be pathogenic.

Inflammatory changes

Once in the CNS, there are many mechanisms by which flaviviruses induce neuronal damage and cell death leading to the clinical manifestations of disease. Early histological studies showed that there is a striking inflammatory response in the brain of humans and animals that have died of flavivirus encephalitis. There is typically perivascular inflammation of lymphocytes and macrophages, with glial cell upregulation. In humans with JE, but not in animal models, there are punched out necrotic lesions, often close to blood vessels.²¹ In JEV infection, autoimmune demyelination may also occur.²²

How much brain damage is due to virus-induced cell death, how much is due to (protective?) apoptosis and how much is due to pathological inflammation are not clear. In the case of WNV there appears to be a fine balance whereby immune factors can mediate viral clearance^23–25 but also damage within the CNS.^26,27 JEV-infected mouse peripheral macrophages release mediators that induce apoptosis of uninfected neurons in vitro.²⁸ Attenuating the CNS inflammatory response to JEV in mice with minocycline results in improved survival.^29,30

Critical components of adaptive immunity

In recent years, several key experiments in animal models of flavivirus encephalitis have highlighted some of the key components of adaptive immunity that result in virus clearance. It has been known for many years from passive immunization experiments that antibody protects from flavivirus encephalitis in mice. Not surprisingly, therefore, when such experiments were repeated using B-cell-deficient mice the animals were highly susceptible to WNV and JEV infection.^31,32 The role of T cells in flavivirus encephalitis is less clear. This is in part due to variation between different viruses and in the dose, route of administration, mouse strain and age of the mice. In WNV infection CD8⁺ T-cell knockout mice have increased lethality and viral burden.²³ For JEV, adoptive transfer of cells from immune mice can be clearly shown to be protective in several models, although for optimal protection all components of the immune system should be present.³² Further evidence for a cooperative function of adaptive immune components was shown by the role of Th2-type CD4 T cells in protection from JEV through help for antibody responses.³³ However, some observations suggest an important role for CD8⁺ T cells as well. The dominant cell type accumulating in the CNS of JEV-infected mice is a CD8⁺ T cell.³² The T-cell receptor usage is highly clonal suggesting that this happens in an antigen-driven manner³⁴; recently this has been shown to be the case in WNV and TBEV as well.³⁵ Moreover, JEV is able to broadly inhibit CD8⁺ T-cell responses by interfering with antigen presentation, an effect reversible by the addition of TLR3, four or nine ligands.³⁶ There is also a modest degree of protection seen when immune CD8⁺ T cells are transferred alone into naïve mice,³² and though the protective effect of CD8⁺ T cells is more marked in the case of WNV²³, CD8⁺ T cells can also be pathogenic under some circumstances in WNV infection.²⁷ Together, these data suggest that there may be a dual role of CD8⁺ T cells in both clearing viral infection but also mediating immunopathology within the CNS. Particularly interesting in this regard is the implication that CD8⁺ T cells provide greater protection from WNV than from JEV, given that JEV is the more pathogenic virus for humans.

Conclusion

In conclusion, flaviviruses are important causes of encephalitis across the globe, and they are spreading. Recent studies on the pathogenesis have shown critical involvement of both innate and adaptive immune responses. However, for both types of response there is evidence for protective and detrimental elements. A major future research question will be to understand the precise balance between immune protection and immunopathology. This will be crucial in developing new, more targeted immunomodulatory therapies for established encephalitis.

Funding

Further evidence for a cooperative function of adaptive immune components was shown by the role of Th2 type CD4 T cells in protection from JEV through help for antibody responses. L.T. is a Wellcome Trust MB PhD postdoctoral research training fellow.

Conflict of interest: None declared.