Molecular mechanisms of neuroinflammation and injury during acute viral encephalitis

Abstract

Viral infections in the central nervous system are a major cause of encephalitis. West Nile virus (WNV) and Herpes simplex virus (HSV) are the most common causes of viral encephalitis in the United States. We review the role of neuroinflammation in the pathogenesis of WNV and HSV infections in the central nervous system (CNS). We discuss the role of the innate and cell-mediated immune responses in peripheral control of viral infection, viral invasion of the CNS, and in inflammatory-mediated neuronal injury. By understanding the role of specific inflammatory responses to viral infections in the CNS, targeted therapeutic approaches can be developed to maximize control of acute viral infection while minimizing neuronal injury in the CNS.

1. Introduction

Viral encephalitis is an important cause of morbidity and mortality in the United States and throughout the world. In the U.S. viral encephalitis results in a mean of 7.3 hospitalizations per 100,000 population, with a disproportionate burden on individuals greater than 65 or less than 1 year of age (George, Schneider, & Venkatesan, 2014a, 2014b). Numerous viruses are associated with encephalitis, but increased vaccination rates and new vaccines have reduced the rates of encephalitis due to previously common pathogens including measles (Morbillivirus), mumps (Rubulavirus), and poliovirus (Picornavirus) over the last 60 years. Other pathogens that can cause encephalitis such as Western Equine Encephalitis Virus and St. Louis encephalitis virus have decreased in incidence over the last 20 years, for reasons that remain poorly understood (Go, Balasuriya, & Lee, 2014).

In studies of encephalitis etiologies, 50–62% of encephalitis cases are due to unknown causes (Glaser, Gilliam, Schnurr, et al., 2003; Granerod, Ambrose, Davies, et al., 2010; George et al., 2014a, 2014b). In these same studies, viruses account for approximately half the cases of encephalitis with an identified etiology (Glaser et al., 2003; Granerod et al., 2010; George et al., 2014a, 2014b). A variety of viruses belonging to a diverse group of viral families cause encephalitis including DNA viruses such as Herpes simplex viruses (HSV) or human cytomegalovirus (CMV) and Varicella-zoster virus (VZV); single stranded RNA viruses such as West Nile virus (WNV), St. Louis encephalitis virus, and Zika virus; and retroviruses such as human immunodeficiency virus 1 (HIV-1). Overall, HSV is the most common cause of sporadic viral encephalitis and WNV the most common cause of epidemic encephalitis in the US (George et al., 2014a, 2014b) (Table 1). While other viruses, including recently emerged Zika virus, cause acute infections of the central nervous system, these viruses are much less common when compared to WNV and HSV.

Table 1.

Viral encephalitis case numbers.

Virus	US cases (dates)	Fatalities	Mortality rate
WNV	~44,000 total cases and 20,265 Neuroinvasive cases (2000–2015) Cdc.gov/westnile	1783	4%
HSV	1.2–7.3 outof 100,000 persons (annual incidence) (George et al., 2014a, 2014b; Jouan, Grammatico-Guillon, Espitalier, et al., 2015)	N/A	5.5–11.9%
St. Louis encephalitis	92 reported cases (2004–2013) cdc.gov/sle	2	2%
Eastern equine encephalitis virus	85 reported cases (2004–2013) cdc.gov/easternequineencephalitis	34	40%
Zika virus	5040 reported cases in Continental US and 48 cases of fetal infection in 1047 Zika virus-infected pregnant women (2015–2016)	5 (Pregnancy loss)	10%a

^aCalculated as known number of fetal loss cases/total reported fetal infections.

Given the importance of HSV and WNV as the predominant identified causes of viral encephalitis in the US, this review will focus on the mechanisms of inflammation and neuronal injury during acute WNV encephalitis (WNE) and acute Herpes simplex virus encephalitis (HSE). In general, the clinical outcome of viral encephalitis is the result of a complex interplay between the direct effects of viral infection on the central nervous system (CNS) and the associated innate and adaptive host immune responses to infection. Understanding the interactions between viral infection, inflammation, and injury in the CNS informs our clinical understanding of outcomes and therapeutic interventions in patients with viral encephalitis.

2. West Nile virus encephalitis pathogenesis

West Nile virus (WNV) is a member of the genus flavivirus and is related to other clinically important neuroinvasive flaviviruses including Japanese Encephalitis virus, St. Louis encephalitis virus, and Zika virus. WNV is the leading cause of epidemic encephalitis in North America and has caused recurrent localized outbreaks of viral encephalitis since its introduction in 1999 (CDC.gov/westnile). Following its introduction into the United States, WNV rapidly spread across North America and caused over 41,000 confirmed cases of disease, and nearly 19,000 cases of neuroinvasive infection (encephalitis, meningitis, myelitis). WNV infections now occur as annual epidemics in characteristic areas in the United States (Fig. 1). Mortality among patients with WNV neuroinvasive disease is approximately 9%, with the majority of mortality occurring in patients with encephalitis.

Fig. 1.

Average annual incidence of West Nile virus neuroinvasive disease reported to the CDC by county from 1999 to 2015.

(Source: CDC ArboNet, Arboviral Diseases Branch.)

WNV is naturally maintained in an enzootic cycle in avian hosts, primarily passerine species of birds including crows, jays, and sparrows, with transmission by Culex and other mosquito species. Infection of humans usually occurs following the bite of mosquito but transmission can also occur by less frequent means such as blood transfusion or organ donation. Humans are typically “dead-end hosts” because infection does not produce a viremia of sufficient magnitude to infect subsequent biting mosquitoes. After inoculation via an infected mosquito, WNV replicates in dendritic cells and macrophages in the dermis near the bite site. Infected inflammatory cells spread virus to regional lymph nodes and into the lymphatic system. Following localized replication, patients develop a primary viremia with subsequent seeding of reticuloendothelial organs such as the spleen where continued replication results in a secondary viremia that facilitates spread of virus to the CNS (Petersen, Brault, & Nasci, 2013). Following entry to the CNS, WNV directly infects neurons, with a predilection for cells in subcortical regions including the basal ganglia, thalamus, midbrain, and cerebellum. WNV can also infect the motor neurons of the anterior horn of the spinal cord. Injury in these subcortical regions results in the distinctive features of WNV encephalitis that can include altered mental status, Parkinsonian features (trem- or, bradykinesia, rigidity), ataxia, myoclonus, and acute flaccid paralysis (Sejvar, Haddad, Tierney, et al., 2003; Debiasi & Tyler, 2006).

The inflammatory process following WNV infection can be divided into responses in the periphery that limit primary and secondary viremia, inflammatory processes associated with invasion of the CNS, and inflammatory responses in the CNS responsible for clearing virus. Due to extensive research in mouse models of disease, the pathophysiology of the immune response following WNV infection is increasingly well characterized (Reviewed by Suthar, Diamond, & Gale, 2013b).

3. Peripheral innate responses

Interferon responses are the primary, early responses required to control WNV infection of a susceptible host and prevent viral invasion of the CNS. Mice lacking interferon receptors (IFNAR−/−) exhibit increased susceptibility to WNV infection, as do mice lacking key components in downstream IFN signaling pathways including interferon regulatory factor-3 (IRF3) (Daffis, Samuel, Keller, et al., 2007). Recent work in mouse models of WNV encephalitis have shown that viral RNA acts as a pathogen-associated molecular pattern (PAMP) that is recognized by cytosolic innate immune pathogen recognition receptors or PRRs (reviewed in Suthar, Aguirre, & Gernandez-Sesma, 2013a). One such PRR is called the retinoic-acid inducible gene-I (RIG-I)-like receptor or RLR. RLRs recognized non-self RNA signatures and interact with mitochondrial antiviral (MAVs) signaling protein to induce type I interferon (IFN), proinflammatory cytokines, and express IFN-stimulated genes (ISGs) such as 2′–5′ oligoadenylate synthetase 1 (OAS1). In many cells types, RLRs like RIG-I or myeloma differentiation factor 5 (MDA5) are critical for detecting and responding to flavivirus infections including West Nile virus (Fredericksen et al., 2008). While the fundamentals of pathogen detection and IFN responses are present in the periphery and in the CNS, it is the peripheral IFN-dependent restriction of WNV that is critical to limiting viral invasion of the CNS.

Once type 1 interferon is induced, it can restrict additional viral replication through induction of interferon stimulated genes (ISGs) which directly restrict viral infection. For example, T1IFN-induced induction of 2′–5′-oligoadenylate synthase 1 (OAS1) and RNaseL results in direct degradation or inhibition of viral ssRNA (Zhou et al., 1997). RNaseL cleaves viral RNA and products of cleavage can serve as RLR ligands to amplify the innate immune response. The specific mechanism of OAS restriction is less well characterized but may function to inhibit synthesis of 2′–5′ oligoadenylates resulting in restriction of viral replication. The ISG family consists of a large and ever-increasing number of genes that inhibit various stages of the viral life cycle including entry, replication, assembly and release. In general, these genes are an active area of research as they are responsible for direct inhibition of viral infections in a tissue-specific and cell-specific manner.

Flaviviruses like WNV have developed approaches to inhibit or escape IFN restriction resulting in CNS invasion. An important and well-conserved flavivirus escape mechanism involves production of small non-coding RNAs made from the 3′ untranslated region (UTR) of flaviviruses (Roby, Pijlman, Wilusz, & Khromykh, 2014). Host RNA degradation machinery recognize and degrade invading viral RNA. The host cell 5′ to 3′ exonuclease called XRN1 degrades flavivirus RNA but halts at a secondary structure in the 3′ UTR (Fig. 2). The remaining RNA is processed into small flavivirus RNAs (sfRNAs) which can inhibit interferon responses. The 3′ UTR RNA secondary structure was recently solved in Zika virus revealing that flaviviruses have developed a pseudoknot structure in the 3’UTR that halts XRN1 and creates sfRNAs (Akiyama, Laurence, Massey, et al., 2016). Thus, the interplay between flavivirus pathogenesis, IFN restriction, and virus-induced IFN escape mechanisms plays a central role in the risk of CNS invasion.

Fig. 2.

Mechanism of small flavivirus (sf) RNA production following host XRN1 degradation of flavivirus RNA. A) XRN1 mediated flavivirus RNA degradation halts in the 3′ untranslated region (UTR) at a conserved viral RNA stem loop structure. The remainder of the 3’UTR is processed to become sfRNA. B) Predicted RNA secondary structure of the flavivirus 3’UTR stem loops responsible for halting XRN1-mediated degradation.

(Source: Chapman et al., eLife 2014; 3:e01892.)

WNV has developed several other countermeasures to inhibit innate immune responses so that the virus can replicate. Since host cells depend on detection of flavivirus double-stranded RNA intermediates that are required for viral genome replication, flaviviruses form replication complexes in endoplasmic-reticulum (ER) derived membrane vesicles called replication complexes (Suthar et al., 2013a). These ER membrane-bound replication complexes have multiple functions including the exclusion of RLRs to prevent detection of viral RNA. WNV also directly antagonizes type I IFN responses using several mechanisms such as inhibition of STAT1 and STAT2 signaling, WNV nonstructural proteins that mediate degradation of the IFN receptor, and virus-induced small RNA production as mentioned above (Suthar et al., 2013a). Additional research is needed to understand the critical innate immune responses that are inhibited by flavivirus infections to develop novel therapeutic approaches and vaccine adjuvants.

IFN responses are likely to be important in the control of WNV infection in humans as well as in mice. Polymorphisms in genes associated with IFN signaling, including IRF3 and OAS1, may increase the likelihood or severity of symptomatic WNV infection and neuroinvasive disease (Bigham, Buckingham, Husain, et al., 2011). OAS1 is a member of the type-1 interferon inducible family and is involved in viral RNA degradation. OAS1 polymorphisms (SNP rs10774671) may increase susceptibility to WNV infection and increase the likelihood of more severe clinical disease (Lim, Lisco, McDermott, et al., 2009; Bigham et al., 2011). Overall, the peripheral innate immune response is a major point of restriction for WNV invasion into the CNS and is a primary determinant of outcome in many patients. Studies are ongoing to determine novel approaches that augment the innate immune response to inhibit or prevent WNV invasion of the CNS.

4. WNV CNS entry and inflammation

The mechanism of WNV entry into the CNS is not clearly defined, and virus may utilize more than one pathway to enter the CNS (Fig. 3). The blood-brain-barrier (BBB) is a major impediment to the entry of pathogens into the CNS. Endothelial cells linked by tight junctions, a basement membrane, and astrocyte foot processes together form the BBB, which acts as a barrier between the intravascular compartment and the CNS tissue. Viruses can utilize several mechanisms to bypass this barrier. WNV can infect cerebral microvascular endothelial cells resulting in inflammation and breakdown of tight junctions that help form the BBB (Verma, Kumar, Gurjav, et al., 2010). WNV infection in the periphery may also result in enhanced permeability of the BBB following the production of pro-inflammatory cytokines during acute systemic infection (King, Getts, Getts, et al., 2007; Petersen et al., 2013). Studies in mice have identified specific cytokines that may contribute to alterations in BBB permeability that facilitate WNV neuroinvasion. For example, Toll-like receptor – 3 (TLR3) mediated production of TNFalpha may play an important role in WNV neuroinvasion (Wang, Town, Alexopoulou, et al., 2004; Daffis, Samuel, Suthar, et al., 2008). Mice lacking TLR3 (TLR3−/−) are more resistant to WNV CNS infection compared to wild-type control mice. A decrease in WNV infection of the CNS appears to result from decreased peripheral production of TNF-alpha and related pro-inflammatory cytokines, which increase BBB permeability. Mice lacking TLR3 show heightened levels of infection in the periphery, due to the absence of TNF-alpha and other pro-inflammatory cytokine responses, but have less severe neuroinvasive disease. Interestingly, attenuated CNS disease in TLR3−/− mice is not due to a direct effect within the CNS since TLR3−/− mice are as susceptible as wild-type controls to WNV inoculation directly into the brain.

Fig. 3.

Possible Mechanisms of WNV entry into the Central Nervous system. A) WNV and other related viruses may gain entry to the CNS following infection of olfactory neurons that have a single synapse that separates these neurons from the olfactory bulb in the CNS. B) WNV and other related viruses may also gain access to the CNS through the blood stream by direct infection of endothelial cells, break down of tight-junctions by MMP and TNFalpha, and/or by infecting infiltrating immune cells. MMP(matrix metalloproteinase), TNF (tumor necrosis factor).

Adapted from Kristensson, 2011 and Cho & Diamond, 2012.

Although virus-induced cytokine activation may lead to BBB disruption and increase WNV infection in the CNS, the acute cytokine response can also decrease BBB permeability and prevent CNS invasion. WNV-induced IFN-λ production in the brain may improve or enhance BBB function and inhibit WNV neuroinvasion (Shrestha et al., 2006). Adjunctive IFN-λ treatment in WNV-inoculated mice enhanced survival compared to non-treated controls, even when IFN-λ was administered 48 h after infection. Overall, these studies suggest that targeting pathways that support neuroinflammation but decrease BBB permeability may be a desirable approach to decrease the severity of acute WNV infections.

Once in the CNS, WNV infection can further alter the BBB through WNV-induced activation of matrix metalloproteinase 9 (MMP-9). Mice lacking MMP9 exhibited delayed WNV growth in the CNS and have less BBB disruption compared to MMP9+ control mice despite equivalent viral titers in the periphery (Wang, Dai, Bai, et al., 2008). Since MMP9 is likely produced by resident astrocytes, these data imply that viral infection of astrocytes or paracrine activation of astrocytes may result in increased MMP9 production and increased BBB permeability.

WNV CNS entry may also involve transneural spread or even a combination of transneural and blood-borne pathways. WNV can spread to the CNS using transneuronal routes. Following peripheral inoculation of WNV, virus is capable of trans-synaptic transport and can infect motor neurons resulting in paralysis (Samuel, Wang, Siddharthan, et al., 2007b; Samuel, Wang, Siddharthan, Morrey, & Diamond, 2007c). Blocking transneuronal transport decreased neuroinvasive infection by WNV and decreased rates of paralysis.

Taken together, these data imply a delicate interaction between peripheral infection and inflammatory control of viral infection and subsequent changes in BBB permeability. As details of these interactions and additional information regarding mechanisms of CNS entry are discovered, targeted therapeutics to decrease neuroinvasion can be developed and tested.

5. Innate immune responses to WNV infection in the CNS

Cytokines and chemokines are produced during WNV infection of the CNS and expression impacts inflammation in the brain and progression of disease (Yang, Ramanathan, Muthumani, et al., 2002; van Marle, Antony, Ostermann, et al., 2007; Lim & Murphy, 2011; Cho & Diamond, 2012). Similar to the peripheral immune response reviewed above, the innate immune responses in the brain play a critical role in the inhibition of WNV infection in the CNS. For example, IFN-dependent signaling influences WNV infection through differential expression of interferon-stimulated genes (ISGs) in different regions of the brain. IFi2712a, an important ISG in the CNS, is expressed in a heterogeneous pattern in specific CNS regions, and mice lacking IFi2712a exhibited enhanced viral burden in neurons of the cerebellum and brain stem but not other regions (Lucas, Richner, & Diamond, 2016). These data suggest that regional variations in innate immune responses in the CNS could be a factor in determining viral tropism.

In addition to IFN-dependent restriction, other constitutively expressed neuronal proteins such as alpha-synuclein (αsyn) can inhibit viral infection of the CNS independent of IFN responses and T-cell responses (Beatman, Massey, Shives, et al., 2015). αSyn expression in neurons inhibits WNV growth, injury, and caspase-induced apoptosis in the CNS. Knockout mice for the neuronal protein αsyn (snca(−/−)) also exhibited higher mortality due to WNV CNS disease at a much earlier time point than matched controls and exhibited significantly increased WNV titer in the brain compared to heterozygote or wild-type control mice (Beatman et al., 2015). Based on these data, neurons express both IFN-dependent and IFN-independent innate mechanisms of virus-restriction in the brain.

6. West Nile virus-induced neuronal injury

WNV infects neurons directly and causes neuronal cell death through apoptosis. WNE pathology is likely a result of both viral-induced apoptosis and secondary effects from neuroinflammation (Samuel, Morrey, & Diamond, 2007a). Multiple viral proteins have been implicated in initiating the pro-apoptotic signaling cascade. Caspase 3 is a major effector caspase responsible for the initiation of apoptosis via cleavage of key regulatory proteins such as poly (ADP-ribose) polymerase (PARP) (Soldani, Lazzè, Bottone, et al., 2001). The WNV NS3 protein can directly trigger caspase-8 and caspase-3 activity (Ramanathan, Chambers, Pankhong, et al., 2006) and these apoptotic pathways are activated in the brains of WNV infected mice (Clarke, Leser, Quick, et al., 2013). The WNV capsid protein also activates mitochondrial apoptotic pathways leading to activation of caspase-9 (Yang et al., 2002). Inhibition of caspase activity prevents WNV-induced cell death and enhances survival in caspase-3 knockout mice (Samuel et al., 2007a). In human cases of WNV encephalitis apoptosis is a primary mechanism of neuronal injury and death (Kleinschmidt, Michaelis, Ogbomo, et al., 2007).

Analysis of a human case of WNE using next generation sequencing evaluated minority variation in WNV sequences in brain tissue. Interestingly, analysis of WNV-infected brain tissues revealed that neuronal apoptotic death was increased in regions with higher expression levels of innate immune transcripts and increased diversity of WNV escape variants (Grubaugh, Massey, Shives, et al., 2016). These data imply that neuroinflammation contributes to injury patterns in the CNS and that WNV develops escape mutations to maintain replication in the presence of innate immune responses.

While WNV infects primarily neurons, studies have also evaluated the roles of other cell types in the CNS in the pathogenesis of WNV infection. For example, neurovirulent WNV can replicate in astrocytes but strains that are neuroinvasive but non-neurovirulent lack the ability to productively infect astrocytes (Hussmann, Samuel, Kim, et al., 2013). Other studies show that granular cell neurons (GCNs) of the cerebellum are spared during WNE due their relative responsiveness to interferon stimulation (Omalu, Shakir, Wang, et al., 2003; Cho, Proll, Szretter, et al., 2013).

7. Cell mediated immune responses to WNV infection in the CNS

The cell-mediated immune response to WNV infection in the CNS is an important component of viral control, innate signaling amplification, and a source of inflammation-induced injury in the CNS. The role of infiltrating inflammatory cells is evident in mice lacking CCR5, which modulates leukocyte trafficking into the CNS. These mice exhibit significantly higher viral titers in the brain and enhanced mortality that parallels their lack of CNS inflammatory cell infiltration. (Shirato, Kimura, Mizutani, et al., 2004; Glass, Lim, Cholera, et al., 2005; Durrant, Daniels, Pasieka, et al., 2015). A similar circumstance may occur in humans who have naturally occurring deletions in the CCR5 gene. Individuals with the CCR5 Δ32 mutation appear to have enhanced susceptibility to neuroinvasive WNV disease (Glass, McDermott, Lim, et al., 2006; Diamond & Klein, 2006; Lim, Louie, Glaser, et al., 2008; Lim, McDermott, Lisco, et al., 2010). Indeed several gene polymorphisms in humans are linked to increased CNS viral infections (Table 2). Unfortunately, much of this data has been sporadic and difficult to replicate due to the inherent heterogeneity of patient populations and resulting biases when enrolling for studies of uncommon diseases such as viral encephalitis.

Table 2.

Human polymorphisms associated with viral encephalitis.

Causative agent	Human polymorphism
WNV	CCR5 Δ 32(Glass et al., 2006; Lim et al., 2008; Lim et al., 2010) OAS1b SNP rs10774671(Lim et al., 2009; Bigham et al., 2011)
HSV-1	TLR3-pathwaya Unc93B1(Casrouge, Zhang, Eidenschenk, et al., 2006) TLR3(Zhang, Jouanguy, Ugolini, et al., 2007; Guo, Audry, Ciancanelli, et al., 2011; Lim, Seppänen, Hautala, et al., 2014) TRIF(Sancho-Shimizu, Pérez de Diego, Lorenzo, et al., 2011) TBK1(Herman, Ciancanelli, Ou, et al., 2012) TRAF(Pérez de Diego, Sancho-Shimizu, Lorenzo, et al., 2010) IRF3(Andersen, Mørk, Reinert, et al., 2015) IKK complex regulation pathway IKBKG/NEMO(Audry, Ciancanelli, Yang, et al., 2011) Type 1 IFN response STAT1(Dupuis, Jouanguy, Al-Hajjar, et al., 2003)

^aBoth WNV and HSV exhibit TLR3-dependent inhibition of viral spread in the CNS but not in peripheral tissues.

In humans and in mouse models, cytokine IL-1β is elevated in plasma during the course of WNV infection (Ramos, Lanteri, Blahnik, et al., 2012). IL-1β expression was shown to control WNV infection in the CNS in mouse models. Mice that lack expression of IL-1β, the IL-1 receptor, or components of the NLRP3 inflammasome pathway are significantly more susceptible to WNV infection and have increased amounts of virus in the CNS but not peripheral tissues (Kumar, Roe, Orillo, et al., 2013). Studies with IL-1 receptor (IL1R) knockout mice showed that loss of IL1R had no impact on BBB integrity; however, pro-inflammatory chemokine expression (CCL2, CCL3, CCL4, CCL5, CCL7, CXCL9, CXCL10, CCR1, CCR2, and CXCR3) was increased compared to wild-type control mice (Durrant, Daniels, & Klein, 2014). Infiltrating CD11b + CD45hi macrophages secrete IL-1β in the CNS in response to WNV infection to promote CXCR4-mediated T cell adhesion to microvasculature endothelial cells and support inflammation in the brain. IL-1R1 is also linked to CD11c + cell-mediated T cell reactivation in the CNS and loss of IL-1R1 expression in mice results in a decreased CD4 TH1 response and decreased control of WNV infection in the CNS while early injury patterns were decreased due to decreased TH1-induced neuroinflammatory pathology (Durrant, Robinette, & Klein, 2013). These data demonstrate that a delicate balance occurs in the CNS between cell-mediated immune responses which are vital to support WNV clearance from the CNS and limit injury from virus and the immune responses in the CNS.

In addition to T-cell responses, pro-inflammatory monocytes (Ly6+) are recruited to the CNS during WNE which can contribute to inflammation-induced pathology (Getts, Terry, Getts, et al., 2008; Getts, Terry, Getts, et al., 2012; Terry, Getts, Deffrasnes, et al., 2012). Levels of pro-inflammatory cytokine production correlate with the severity and outcome of WNV infection in mice. Mice inoculated with virulent WNV (NY99) produce more pro-inflammatory cytokines, more type-1 interferon, and develop more severe disease when compared to mice inoculated with an attenuated strain of WNV called Eg101 (Kumar, Roe, O’Connell, & Nerurkar, 2015). Eg101 elicits a more robust innate and adaptive immune response than NY99, leading to better control of viral replication and improved survival. Following WNV Eg101 neuroinvasion in a mouse model, mice exhibited enhanced WNV-specific CD8+ T-cells responses. This correlated with increased IFN-γ and TNF-α in the brain and reduced monocyte/macrophage accumulation in Eg101 infected mice compared to NY99. Differences in immune recruitment and the production of specific cytokines led to enhanced control of virus in the CNS and improved survival in Eg101-infected mice compared to NY99-infected mice (Kumar et al., 2015). This work demonstrates the importance of the adaptive immune response and the recruitment of immune cells to control WNE. Studies such as these suggest the possibility of therapeutic manipulation of the inflammatory response and immune cells recruitment and activation in the CNS in order to improve patient outcomes.

Currently, there are no approved vaccines or therapeutic options for the treatment of acute WNV infection, and clinical treatment of WNE is currently limited to supportive care. Future therapies that reduce apoptotic-cell death in the CNS, improve BBB integrity, and target host pathways responsible for pathogenic inflammation may be potential targets for novel antiviral therapy and would likely synergize with drugs inhibiting viral replication or act as adjuvants to vaccination programs against flaviviruses like WNV.

8. Herpes simplex virus encephalitis (HSE)

Herpes simplex virus encephalitis (HSE) is the most common cause of sporadic encephalitis in the US and accounts for 5–10% of encephalitis cases worldwide (Tunkel, Glaser, Bloch, et al., 2008). HSE can occur in patients of all ages as a result of both primary infection and reactivation of latent infection. HSE appears most commonly in patients aged less than 20 years (33%) and over 50 years of age (50%) (Whitley & Gnann, 2002). HSV-1 encephalitis is more common in adults, and HSV-2 encephalitis is more common in neonates due to mother-to-infant transmission of genital HSV-2 during delivery (Tunkel et al., 2008; Whitley, Soong, Linneman, et al., 1982b; Whitley, Lakeman, Nahmias, & Roizman, 1982a)). In cases of HSV-1 encephalitis, 30% of cases are due to primary infection (first exposure to virus) while the remaining 70% of cases are due to the reactivation of latent virus (Nahmias, Whitley, Visintine, et al., 1982).

9. Inflammation and HSV CNS entry

The exact mechanism of HSV CNS invasion is not known. Studies in mice have demonstrated that olfactory neurons and vomeronasal chemosensory neurons are both potential routes into the CNS for HSV (Mori, Goshima, Ito, et al., 2005; Mori, Goshima, Watanabe, et al., 2006). Necropsies in human patients have demonstrated the presence of HSV in the olfactory nerves and olfactory bulbs of fatal HSE cases, further supporting the model of nasal entry and olfactory nerve spread to the CNS (Twomey, Barker, Robinson, & Howell, 1979; Dinn, 1980; Ojeda, Archer, Robertson, & Bucens, 1983).

In mice, mucocutaneous and intranasal inoculation of HSV-1 results in trigeminal ganglion infection, where HSV then establishes a latent infection (Kastrukoff, Hamada, Schumacher, et al., 1982; Mori et al., 2005). In humans, HSV is latent in trigeminal ganglia and reactivation from this site may result in spread via tentorial nerves to the middle cranial fossa (Fig. 4). As with WNE, defects in TLR3 expression are associated with enhanced neuroinvasion of HSV suggesting that TLR3-mediated innate immune signaling pathways function to restrict HSV spread in the CNS and maintain BBB integrity (Li, Ye, Wang, et al., 2012; Reinert, Harder, Holm, et al., 2012). Human astrocytes express TLR3 at both the cell surface and at intracellular sites. Stimulation of TLR3 by HSV results in microglial activation, which indicates that astrocytes may play a role in restricting HSV entry into the CNS (Jack, Arbour, Manusow, et al., 2005). Consistent with these experimental studies, multiple defects in human TLR3 signaling pathways are associated with susceptibility to HSVE in humans (Piret & Boivin, 2015). TLR3 deficiency in humans appears to specifically modulate CNS susceptibility to HSV-1 and have no effect on non-neuroinvasive viruses (Perales-Linares & Navas-Martin, 2013). Overall, the mechanism of HSV neuroinvasion seems more consistent with a transneuronal pathway from the olfactory bulb and/or the trigeminal ganglion.

Fig. 4.

Possible mechanism of Herpes simplex virus type 1 entry into the CNS with subsequent spread to orbital-frontal and temporal lobes. A) HSV may gain entry via the olfactory nerves resulting in infection of the olfactory bulb and subsequent spread to adjacent temporal lobes (dashed arrow). HSV may also spread from the trigeminal ganglion to small sensory fibers that supply the basilar dura of the anterior and middle fossa (red arrows). B) Diagram of olfactory nerve synaptic pathways to the CNS. Olfactory rods of the receptor cells are directly exposed at the mucosal surface which synapses directly with mitral cells in the olfactory bulb in the CNS.

10. HSV-induced injury in the CNS

Once HSV enters the CNS, it infects and injures neurons resulting in clinical disease. HSV infection of neurons leads to apoptosis, which is a major contributor to HSV pathogenesis in the CNS. Notably, both HSV-1 and HSV-2 cause neuroinvasive infections but only HSV-1 induces apoptosis in hippocampal cultures as measured by TUNEL assay, caspase-3 activation, and c-Jun N-terminal kinase (JNK) activation. On the other hand, HSV-2 activates pro-survival MEK/ERK pathway by the viral protein ICP10, which prevents apoptotic cell death in cell culture (Perkins, Pereira, Gober, et al., 2002; Perkins, Gyure, Pereira, & Aurelian, 2003). Treatment of hippocampal cultures with the JNK inhibitor SP600125 inhibits HSV-1-induced apoptosis (Perkins et al., 2003), indicating that the inhibition of apoptotic pathways mediated by JNK may serve as a future therapeutic target for the treatment of HSV-1 encephalitis to limit necrotic cell death and CNS injury. Additional studies using HSV-2 mutant virus lacking a virus-encoded protein kinase called US3 protein kinase (US3) demonstrate that US3 functions as an anti-apoptotic factor in neurons by inhibiting JNK signaling (Mori, Goshima, Koshizuka, et al., 2003). In an US3-disrupted HSV mutant JNK signaling was enhanced resulting in increased levels of apoptosis. Subsequent work with US3-disrupted HSV mutants inoculated by an intranasal route in a mouse model demonstrated enhanced apoptosis in olfactory neurons. However, the US3-disrupted HSV mutant was unable to penetrate the CNS and cause encephalitis indicating that HSV-specific modulation of apoptosis is critical for successful CNS invasion (Mori et al., 2005; Mori et al., 2006). When inoculated directly into the brain, both WT- and US3-disrupted HSV cause similar levels of mortality, suggesting that US3 functionality plays a role in facilitating infection from the PNS to the CNS (Mori et al., 2006).

Since HSV must undergo latency in neurons it must also support survival of latently infected cells. HSV supports cell survival by inhibiting apoptosis and interacting with essential host cell pathways that regulate cell survival such as macroautophagy, referred to as autophagy hereafter. HSV directly targets the host autophagy pathway to support viral replication in neurons. Autophagy is a cellular process that degrades proteins and organelles in the cell as well as invading pathogens and can function as a protective response against certain infections in the CNS (Orvedahl & Levine, 2008). HSV avoids autophagy-mediated degradation by binding to the autophagy regulator Beclin-1, resulting in autophagy inhibition and decreased host cell clearance of the pathogen. The protein ICP34.5 in HSV was originally identified as a viral protein non-essential for growth in culture but necessary for neurovirulence (Chou, Kern, Whitley, & Roizman, 1990; Bolovan, Sawtell, & Thompson, 1994). Mutant HSV-1 lacking the Beclin-1 binding domain in ICP34.5 did not inhibit autophagy in neurons and exhibited impaired neurovirulence (Orvedahl, Alexander, Tallóczy, et al., 2007). Neurovirulence was restored when the mutant virus was introduced in protein kinase R knockout (pkr−/−) mice demonstrating that PKR-regulation of autophagy induction during HSV-1 infection in neurons is a key host defense strategy against HSV infection in the CNS.

11. Immune response to HSV in the CNS

Both the innate and cell-mediated immune responses are important to control HSV infection once in the CNS. Defects in the innate sensing pathway downstream of TLR3, such as interferon regulatory transcription factor 3 (IRF-3) (Andersen et al., 2015), Unc93B1(Casrouge et al., 2006), TIR-domain-containing adapter-inducing interferon-beta (TRIF) (Sancho-Shimizu et al., 2011), TANK-binding kinase 1 (TBK1) (Herman et al., 2012), and TNF receptor-associated factor (TRAF) (Pérez de Diego et al., 2010), result in a defective immune response to HSV in the CNS and reduced control of viral replication (see Table 2). These polymorphisms are found in familial and recurrent cases of HSE, demonstrating the clinical relevance of these molecular pathways in the development of HSE as well as the importance of the TLR3/IFN-I/III signaling axis in the control of HSV in the brain. However, it is not currently known how these host factors control HSE in sporadic cases and if deficiencies in TLR3/IFN pathways are the basis for sporadic HSV reactivation in the CNS.

The innate immune system and TLR3 sensing of HSV in astrocytes also leads to the production of IFN-λ resulting in IFN-λ mediated anti-HSV activity (Li et al., 2012). This effect was further linked to the IFN regulatory factors IRF3 and IRF7. Both IRF3 and IRF7 regulate TLR3-mediated IFN-λ1 expression, while IRF7 knockdown impaired only the TLR3-mediated IFN-λ2/3 expression. These results indicate that both IRF3 and IRF7 are involved in TLR3-mediated expression of IFN-λ, but signaling through IRF3 or IRF7 result in different isotypes of IFN-λ production. The production of cytokines and chemokines in the CNS also play a major role in the pathology and control of HSE (Lokensgard, Hu, Sheng, et al., 2001; Wickham, Lu, Ash, & Carr, 2005; Rosato & Leib, 2015). Similar to WNE, the innate immune response is vital to limit early disease in cases of HSV encephalitis.

The cell-mediated response to HSV infection in the CNS is critical for acute control and long-term elimination of HSV from the brain. Recruitment of NK cells, activated T-cells, and dendritic cells are all important for the control of HSV-1 in the CNS. A major mediator of this immune cell recruitment is the chemoattractant CXCL10. In mice that lack expression of CXCL10, HSV-1 specific CD8+ T cells are not recruited to HSV-infected brain tissue due to dysregulated CXCR3 signaling (Wuest & Carr, 2008). Further study with mice deficient in CXCR3 (the CXCL10 receptor) expression demonstrated that CXCR3−/− mice had higher HSV titer in brain tissue, increased expression of CXCL10, CXCL9, CCL5 and IFN-γ, and a two-fold increase in CD8+ cells in the brain stem compared to control mice (Wickham et al., 2005). Despite the elevated viral loads in the brains of CXCR3−/− mice compared to controls, knockout mice demonstrated better survival than control mice which was related to enhanced expression of the proinflammatory cytokine IL-6. Additional studies have demonstrated the importance of IL-6 expression for the control and survival of viral encephalitis in mice, but further studies must be conducted to determine if modulation of IL-6 expression during HSE could provide a therapeutic benefit for human cases of HSE (Pavelko, Howe, Drescher, et al., 2003; Bissonnette, Klegeris, McGeer, & McGeer, 2004; Dvorak, Martinez-Torres, Sellner, et al., 2004).

During acute HSE, intrathecal synthesis of anti-HSV antibodies increases (Schultze, Weder, Cassinotti, et al., 2004), yet the role of anti-HSV antibodies in the prevention of HSE neuronal injury is not fully understood (Ramakrishna, Openshaw, & Cantin, 2013). Recent studies have shown that higher titers of anti-HSV antibodies in the CNS are indicative of a positive prognosis (Bell, Davies, & Thompson, 2003). Patients who recovered fully or with minor sequelae from HSE demonstrated robust, monoclonal antibody production to HSV, while patients with severe sequelae post-infection or who succumbed to infection showed a weak polyclonal response (Bell et al., 2003). Studies in mice have demonstrated that anti-HSV antibodies are protective against fatal encephalitis (Morrison & Knipe, 1997). Mice were inoculated with a replication deficient isolate of HSV-1 and then sera harvested for passive transfer into HSV-1 naïve mice. Mice that received sera from HSV-1 infected mice showed a high resistance to the development of encephalitis and enhanced survival (75–100%) compared to mice treated with control sera, which had a survival rate of 0% (Morrison & Knipe, 1997). Based on the human and mouse-model observations, antibodies to HSV-1 may contribute to host responses to acute HSE and improve clinical outcomes.

Despite the importance of the immune response in the control of acute HSV infection, infection of the central nervous system is still rare even in cases of significant immune suppression such as advanced HIV infection or allogeneic stem cell bone marrow transplantion. In one series of 281 allogeneic hematopoietic stem cell transplants, the authors described 5 cases of HSV type 1 encephalitis (Wu, Huang, Jiang, et al., 2013). While acyclovir prophylaxis may partially explain the low rate of HSV encephalitis despite profound immune suppression, this likely doesn’t fully explain the persistently low rates of disease despite profound immune suppression in all cases. Since alpha-herpes viruses have co-evolved with humans, it is likely that the nervous system has developed innate restriction mechanisms, not yet well described, that prevent viral invasion of the central nervous system. Ongoing research in this field will likely reveal novel innate mechanisms of restriction that prevent virus infections of the central nervous system.

12. Treatment of HSV encephalitis and inflammation

Since the inflammatory response to HSV may contribute to pathology in the CNS, several studies have evaluated the role of modulating the immune response using corticosteroid therapy co-administered with acyclovir during acute HSE. In murine models of HSE, treatment with acyclovir in combination with corticosteroids does not increase the severity of HSE compared to acyclovir treatment alone (Thompson, Blessing, & Wesselingh, 2000) and may be beneficial during acute HSE (Kamei, Sekizawa, Shiota, et al., 2005). Several published studies of human cases of HSE also suggest a possible benefit of adjuvant corticosteroid treatment for pediatric (Musallam, Matoth, Wolf, et al., 2007) and adult HSE (Lizarraga, Alexandre, Ramos-Estebanez, & Merenda, 2013). A meta-analysis of published studies has determined that there may be a clinical benefit to treating HSE patients with corticosteroids during acute HSE but randomized, blinded clinical trials are required to further define the role of adjuvant corticosteroid therapy for acute HSE (Ramos-Estebanez, Lizarraga, & Merenda, 2014).

To improve treatment outcomes, development of novel treatment approaches and adjunctive therapies for HSE is of high importance (Piret & Boivin, 2014). Since the inflammatory response contributes to HSE pathology, targeted manipulation of inflammation is a rational approach for future therapeutics. Additional approaches to limit CNS injury during HSE may involve targeting chemokine expression to enhance immune cell recruitment to the CNS and control viral replication, or therapeutic targeting of IL-6 production to alter levels of inflammation in the CNS. Clinical trials for HSE are difficult to complete due to the sporadic nature and relative low absolute number of HSE cases.

13. Conclusions

WNV and HSV are important causes of CNS infections throughout the world and North America. The data reviewed here provide the current understanding of the mechanisms of direct neuronal injury and neuroinflammation involved in these important infections in the CNS. With a more thorough understanding of the pathogenesis underlying these viral infections, we can elucidate better pathogen-specific therapies and develop neuroprotective therapeutic approaches. While WNV and HSV are distinct viruses, they share common mechanisms of inflammation and cellular injury that may provide common neuroprotective therapeutic approaches for patients with possible viral encephalitis. A broad-spectrum, neuroprotective approach to therapy for these patients is attractive since many patients present with an undifferentiated syndrome of altered mental status, fever, and neuroinflammation. Early institution of neuroprotective strategies targeted at common neuroinflammatory pathways while waiting diagnostic testing may improve clinical outcomes and decrease morbidity and mortality that is due viral encephalitis.