Henipavirus Encephalitis: Recent Developments and Advances.

The genus Henipavirus within the family Paramyxoviridae includes the Hendra virus (HeV) and Nipah virus (NiV) which were discovered in the 1990s in Australia and Malaysia, respectively, after emerging to cause severe and often fatal outbreaks in humans and animals. While HeV is confined to Australia, more recent NiV outbreaks have been reported in Bangladesh, India and the Philippines. The clinical manifestations of both henipaviruses in humans appear similar, with a predominance of an acute encephalitic syndrome. Likewise, the pathological features are similar and characterized by disseminated, multi-organ vasculopathy comprising endothelial infection/ulceration, vasculitis, vasculitis-induced thrombosis/occlusion, parenchymal ischemia/microinfarction, and parenchymal cell infection in the central nervous system (CNS), lung, kidney and other major organs. This unique dual pathogenetic mechanism of vasculitis-induced microinfarction and neuronal infection causes severe tissue damage in the CNS. Both viruses can also cause relapsing encephalitis months and years after the acute infection. Many animal models studied to date have largely confirmed the pathology of henipavirus infection, and provided the means to test new therapeutic agents and vaccines. As the bat is the natural host of henipaviruses and has worldwide distribution, spillover events into human populations are expected to occur in the future.

Introduction

The Henipavirus genus in the subfamily Paramxyxovirinae (family Paramyxoviridae) that also includes the well‐known measles virus (genus: Morbillivirus) was originally created for two pathogenic viruses, Hendra and Nipah 123. Since then, nonpathogenic Cedar virus has also been included into Henipavirus 71. More recently, many newly discovered African henipaviruses and a pathogenic henipavirus from the Philippines have been reported, all of which have not been isolated or fully characterized 14, 27. Hendra virus (HeV) and Nipah virus (NiV) are enveloped, pleomorphic virions with diameters averaging about 500 nm, but may range from about 40 to 1900 nm in length 38, 55. Each virion contains a tangle of nucleocapsids with a herringbone structure typical of paramyxoviruses. As negative‐sense, single‐stranded RNA viruses, they have unusually large genomes of more than 18 000 nucleotides, comprising six genes encoding for the nucleocapsid, phosphoprotein, matrix protein, fusion glycoprotein (F), attachment glycoprotein (G) and the large polymerase (Figure 1).The viral G protein attaches to the host cell ephrin B2 and/or B3 receptor, and activates the F protein to initiate viral envelop and host membrane fusion and viral entry 4, 82, 91.

Figure 1

Diagrammatic representation of the henipavirus virion with an outer, plasma membrane‐derived envelop through which the viral G and F glycoproteins protrude. The matrix protein is found under the envelop. The central core consists of the nucleocapsid, viral RNA, phosphoprotein and polymerase.

Various aspects of HeV and NiV infections, including pathology and pathogenesis, have been reviewed previously, either individually or as henipaviruses 69, 116, 129, 130, 133, 135, 137. The purpose of this review is to revisit what is currently known and to describe recent developments and advances of emerging henipaviruses as deadly zoonotic infections with human outbreaks continue to be reported in Bangladesh and the Philippines 14, 98.

HeV was first isolated in 1994, five years before NiV, after an outbreak in horses and two humans in Australia, hence its original name, Equine Morbillivirus 81, 109. To date, numerous outbreaks in horses and a total of seven human infections with four fatalities have been reported 47, 69, 85, 92. So far, no HeV cases have been reported outside Australia.

NiV was first isolated in 1999 after outbreaks in pigs and humans in Malaysia and Singapore from 1998 to 1999 12, 19. Although not precisely known, the number of infected pigs must have been at least in the hundreds. It was estimated that there were more than 350 human cases, both asymptomatic and symptomatic 135, including at least 265 cases of acute NiV encephalitis with 105 fatalities 88, 89. From 2001 onwards, several recurrent outbreaks of NiV in Bangladesh and an outbreak in the adjacent Indian state of West Bengal have probably involved more than 250 people with a fatality rate of about 70% 49, 57, 68. The latest outbreak of nine cases in Bangladesh was reported in early 2015 98, and a henipavirus outbreak involving 17 people, more likely due to NiV, occurred in the Philippines in 2014 14.

Natural Hosts and Henipavirus Transmission to Humans

The natural host of henipaviruses has been confirmed to be bats. HeV has been isolated from Pteropus alecto and a few other Pteropus species (flying foxes) found in Australia 111. NiV was isolated from P. hypomelanus and P. vampyrus (Malaysia), and P. lylei (Cambodia), and NiV RNA recovered from P. giganteus (India), P. vampyrus (Indonesia) and P. lylei (Thailand) 29, 100, 110, 120. Viral RNAs of African henipaviruses have been detected in various species of non‐pteropid bats, including Eidolon helvum 27. Zoonotic henipavirus transmission to humans can occur in several ways: (i) direct bat‐to‐human transmission; (ii) bat‐to‐intermediate host‐to‐human transmission; and (iii) human‐to‐human transmission. The best example of bat‐to‐human transmission was from Bangladesh where humans became infected when they drank raw date palm sap contaminated by NiV‐infected bats that had fed at overnight collection pots 67.

For both NiV and HeV, involvement of intermediate hosts is common. Most likely urine and oronasal secretions from infected animals were responsible for passing the infection to humans through direct or possibly airborne exposure 75, 92, 109. In the NiV outbreak in Malaysia and Singapore, the pig was the main intermediate host, and pig farmers, abattoir workers, pork sellers and army personnel involved with culling of pigs had the highest prevalence of infection 2, 13, 88, 95, 106. Bat‐to‐pig transmission could have resulted from pigs consuming half‐eaten, contaminated fruits dropped by bats 20. Other domestic animals such as dogs, cows and goats may have the potential to act as intermediate hosts for NiV 18, 76, 133. Human HeV outbreaks were closely associated with infection in horses and indeed occurred mainly among veterinarians, stable workers and other personnel in close contact with horses. Though rare, the HeV‐infected dog may also have the potential to be an intermediate host 97. Bat‐to‐horse transmission remains unclear but is likely to be due to environmental contamination from bat urine or other bat‐derived materials 31, 127. Interestingly, in the henipavirus outbreak in the Philippines, the majority of patients were involved with slaughtering infected horses and/or consuming under‐cooked horse meat 14.

Human‐to‐human transmission was firmly established in the NiV outbreaks in Bangladesh and India where a relatively high incidence was reported either in family members or other people in contact with patients, and in health care workers 44, 49, 54. Although human‐to‐human transmission may well have occurred in the Malaysia/Singapore outbreaks, the only well‐documented case was in a nurse who seroconverted but remained asymptomatic despite brain magnetic resonance imaging (MRI) that showed a few typical lesions of acute encephalitis 115, 117. In other health staff involved with patient care, serum neutralization tests were found to be negative, possibly because of good clinical prophylactic measures 78. In the Philippines' henipavirus outbreak, the clinical and epidemiological evidence also suggested human‐to‐human transmission 14. To date, bat‐to‐human or human‐to‐human HeV transmission has not been reported 69. It is perhaps not surprising that human‐to‐human henipavirus transmission can occur as virus has been detected in various excretions from patients 21, 92.

Clinical Manifestations and Sequelae

The incubation period of henipavirus infections ranges from a few days to about 3 weeks. Milder clinical features include fever, nonspecific influenza‐like illness, lethargy, myalgia and headache 14, 16, 37, 92, 109. In general, the more severe clinical features manifest as either an acute encephalitic syndrome or less frequently a pulmonary syndrome. It is also very likely that there is a considerable number of infections that remained asymptomatic but the exact prevalence is unknown.

As there have been only seven known cases of human HeV infection, much less is known about its clinical manifestations compared with NiV infection. Nonetheless, the acute encephalitic syndrome is characterized by drowsiness, confusion, ataxia, ptosis and seizures. One patient had a pulmonary syndrome described as a pneumonitis with chest radiograph findings of diffuse alveolar shadowing 69, 92, 109. However, at autopsy, this patient's brain also showed features of acute encephalitis that had not been clinically apparent 134. Two clinical encephalitis cases also had similar abnormal chest radiographs.

Severe NiV infection appears to be more frequently manifested as an acute encephalitic syndrome. In the largest cohort of 90 patients with acute NiV encephalitis admitted to a single hospital, apart from the symptoms seen in milder infections, 53% of patients had reduced consciousness 37. Clinical signs such as areflexia/hyporeflexia with hypotonia, pinpoint pupils with variable reactivity, tachycardia, hypertension and abnormal doll's eye reflex were more frequently seen in patients with reduced level of consciousness. Segmental myoclonus characterized by focal, rhythmic jerking of the diaphragm and muscles in the limbs, neck and face was present in 30% of patients, and may be unique to acute NiV encephalitis. Other clinical features (eg, meningism, generalized tonic–clonic convulsions, nystagmus and cerebellar signs) were also observed. Similar findings were also found in patient cohorts from Bangladesh and Singapore 53, 61. In Bangladesh, altered mental status and/or unconsciousness was reported in >70%, severe weakness in 67% and areflexia/hyporeflexia in 65% of cases, consistent with a high prevalence of acute encephalitis. In the Philippines' henipavirus outbreak, 11 of 17 cases (65%) were reported to have acute encephalitis and 1 case had meningitis but no encephalitis 14.

Generally, brain MRI scans were useful for diagnosis of acute henipavirus encephalitis showing multiple, disseminated, small discrete hyperintense lesions mainly in the cerebral cortex, subcortical and deep white matter 65, 69, 92, 107. In addition to this, acute HeV encephalitis lesions may be larger and more confluent, similar to some acute NiV encephalitis cases in Bangladesh 99.

Overall, mortality is high in henipavirus infections ranging from about 40% to 70% 49, 53, 69, 88. In a large cohort of patients who survived, the majority had no or few sequelae 37. Some patients were reported to have neurological deficits, neuropsychiatric sequelae and gait/movement disorders 83, 108. Fatal intracerebral hemorrhage is a very rare complication 37. The most interesting complication is probably relapsing encephalitis which may occur weeks to years after symptomatic henipavirus infection (mild, non‐encephalitis illness or acute encephalitis) and even after asymptomatic NiV infection/seroconversion 85, 108, 114. So far, there has been only 1 case of relapsing HeV encephalitis and >20 cases of relapsing NiV encephalitis (possibly involving <10% of survivors), one of which occurred 11 years after an asymptomatic infection 1. Clinical and radiological findings suggest that relapsing NiV encephalitis is distinct from acute NiV encephalitis 107, 114. The brain MRI in relapsing henipavirus encephalitis shows more extensive and confluent hyperintense cortical lesions. Virus could not be cultured from cerebrospinal fluid, naso‐tracheal secretions, urine and brain tissue in a series of NiV relapsing encephalitis 114.

Laboratory Diagnosis

Detection of anti‐henipavirus antibodies remains a key to establishing the diagnosis both in animals and humans because the diagnosis may not be suspected initially, and disease manifestations can be rather nonspecific. Enzyme‐linked immunosorbent assays (ELISA) that were developed during the initial HeV and NiV outbreaks in Australia and Malaysia, respectively, proved to be very useful as a screening test and were widely applied to diagnose these infections in animals and humans 122. Subsequently in Bangladesh, the first NiV outbreaks were only suspected after these tests were found to be positive in patients 54. In human NiV infection, IgM seroconversion from day 4 to day 12 was about 65%–100%, persisting for at least 3 months, while IgG seroconversion by day 25 was 100% and persisted for several years 17, 101. These tests have also been used widely for investigative field studies in bats and other animals 56, 58, 120. Specific IgM or IgG has also been reported in HeV‐infected patients 47, 85, 109.

Serum neutralization tests, although accepted as the reference standard to confirm henipavirus infection, may not be done in most laboratories because this test involves handling live viruses which requires biosafety level 4 facilities. An innovative pseudotyped vesicular stomatitis virus bearing the NiV F and G proteins has been used to obviate this requirement, and was shown to be more sensitive than conventional serum neutralization tests, and applicable for both HeV and NiV infections 14, 59, 113. Specific and reliable PCR‐based methods and sequencing 30, 42, 122 to identify henipaviruses are now more widely available for outbreak investigations and diagnosis, and have begun to replace serological testing 51, 111, 140. Where infected animal or human tissues are available, immunohistochemistry using specific anti‐henipavirus antibodies is useful to confirm the diagnosis 34, 130, 133.

Pathology of Human Henipavirus Infection

The pathology of acute HeV and NiV infections appears to be similar although our current knowledge of acute henipavirus infection is based on only 3 autopsies of HeV infection, and >30 autopsies of NiV infection 69, 134, 136. Acute henipavirus infection is characterized by disseminated small vessel vasculopathy comprising endothelial infection/ulceration, vasculitis (Figure 2A,G) and vasculitis‐induced thrombosis/occlusion (Figure 2A), often resulting in parenchymal ischemia and microinfarction in the central nervous system (CNS) (Figure 2D), lung, kidney and other major organs 134, 136. In about 25% of NiV infections, endothelial multinucleated syncytia have been detected (Figure 2B). This feature is probably unique to henipavirus infection as it has not been described in other infections, including measles. Viral antigens/RNA and nucleocapsids were detectable in vascular endothelium, multinucleated giant cells and smooth muscle.

Figure 2

Pathology of henipavirus infection: vasculopathy in Nipah virus encephalitis showing vasculitis, thrombosis (A) and endothelial multinucleated syncytia with viral inclusion (B, arrow). There are numerous Nipah viral inclusions/antigens within neurons (C), especially around necrotic plaques (D). Necrotic plaques may also have evidence of adjacent vascular thrombo‐occlusion (D, arrow). Vasculitis in Hendra encephalitis may manifest as mild endotheliitis (G). Hendra viral inclusions (H), antigens (F,I) and RNA (E) can be demonstrated in neurons. Viral antigens may also localize to the ependyma (F, arrow). Hematoxylin and eosin stain (A,B,D,G, H). Immunohistochemistry with new fuchsin (C) and DAB (F,I) chromogens, hematoxylin counterstain. In situ hybridization with NBT/BCIP substrate, and hematoxylin counterstain. Original magnification: ×40 (A–I). Permissions were obtained from reference 133 for A and B; reference 136 for C and D; reference 134 for E–I.

Vasculopathy was most severe in the CNS and frequently associated with discrete necrotic or more subtle vacuolar, plaque‐like lesions (Figure 2D,F). As neuronal viral inclusions/antigens/RNA were often demonstrable (Figure 2C,E,F,H,I), the lesions were probably the result of microinfarction and/or neuronal infection. Similar white matter necrotic lesions were probably pure microinfarcts as glial cells were far less susceptible to infection. Focal neuronophagia, microglial nodule formation, clusters of foamy macrophages, perivascular cuffing and meningitis can also be found but extravascular paramyxoviral Warthin–Finkeldey multinucleated giant cells were absent. The dual pathogenic mechanisms of vasculopathy and parenchymal cell infection in multiple organs may be unique to acute henipavirus infection 136.

The pathological features of relapsing henipavirus encephalitis were based on an autopsy case of HeV 134 and autopsies of relapsing NiV encephalitis 114, 136. Relapsing HeV and NiV encephalitides are similar where pathology was confined to the CNS. Confluent and extensive parenchymal necrosis, edema and inflammation were seen mainly in neuronal areas where there was prominent perivascular cuffing, severe neuronal loss, reactive gliosis and neovascularization. Focal viral inclusions/antigens/RNA and nucleocapsids were demonstrated mainly in surviving neurons. Severe meningitis was found in many areas. Notably, vasculitis, endothelial syncytia and thrombosis seen in acute henipavirus encephalitis were absent and blood vessels were all negative for antigens/RNA. The presence of viral inclusions, nucleocapsids, antigens and RNA confirmed relapsing henipavirus encephalitis as a recurrent infection rather than postinfectious encephalitis 114. The absence of vasculopathy and extra‐CNS organ involvement suggests reactivation of latent viral foci within the CNS rather than virus reentry from an extra‐CNS site.

The risk factors for relapsing henipavirus encephalitis are unknown. Clinically and pathologically, relapsing henipavirus encephalitis appears to share some similarities with subacute sclerosing panencephalitis (SSPE). Moreover, like relapsing henipavirus encephalitis, SSPE is not invariably fatal and recurrences have been reported 17, 26, 114. Similar measles virus mutations associated with SSPE 11 have so far not been reported in relapsing NiV encephalitis 131. Measles virus is known to cause immune suppression 39, and being paramyxoviruses, henipaviruses could possibly cause immune suppression leading to relapsing encephalitis.

Pathology of Acute Henipavirus Infection in Animals

Animal studies have contributed to our understanding of henipavirus pathology and pathogenesis 23, 34, 75, 126, 133. Many mammalian species can be infected either under natural or experimental conditions. The range of susceptible includes the definitive host (bat), known intermediate hosts (horse, pig, dog) and other experimental animal models (mouse, hamster, cat, guinea pig, ferret, chicken embryo, African green monkey and squirrel monkey) 7, 33, 43, 52, 63, 70, 73, 74, 104, 119, 127, 128, 132. Overall, the findings in animals have confirmed the findings in human autopsies that the acute infection is a disseminated, multi‐organ disease characterized by vasculopathy and parenchymal cell infection. This extensive dual pathogenic mechanism of tissue injury is responsible for the severity and high mortality of henipavirus infection.

The bat as the definitive and reservoir host for henipaviruses demonstrated little evidence of pathology except for focal mild vasculitis 126. Perhaps the hamster and the African green monkey may be the best known animal models for both NiV and HeV because the pathological findings appear to most accurately recapitulate changes in humans, in particular CNS vasculopathy and neuronal infection. Although vasculopathy was invariably present, neuronal infection and encephalitis were not prominent in the ferret, pig, cat, horse and guinea pig (except in high HeV doses) models. NiV infection in mice was subclinical 28, but older mice were susceptible to HeV for reasons possibly related to interferon I response after viral entry 25 or perhaps to other host responses such as CXCL10, a chemokine induced by NiV 72. An interesting but rare finding was neuronal syncytia formation in the hamster and chicken embryo CNS 118, 133. In addition to vasculopathy and lung parenchyma infection seen in most animal models, the pig also showed prominent infection and syncytia formation in the bronchial epithelium 63, 133. The dog has not been studied as model for henipavirus infection but naturally infected animals have some similarity to human pathology 52, 133.

Apart from the obvious hematogenous spread into the CNS via damaged blood vessel/blood–brain barrier, recent animal studies in pig, mouse and hamster models have demonstrated neuroinvasion via infected nasal mucosa, and then into the olfactory bulb to enter the CNS 28, 80, 124. So far, there has been no direct evidence for this in human infection 136. Despite NiV strains from Bangladesh and Malaysia sharing about 91.8% genomic sequence homology 48, the infectious disease pathology may still differ. To address this question, two separate studies using the hamster model were performed. In one study, the Malaysian strain caused a more accelerated infection 24, but in another study there were no apparent significant differences 3. A study in ferrets also showed no pathological differences except for higher viral titers in oral secretions following infection with the Malaysian strain 22. Further studies by manipulating NiV genome using a recently described infectious clone may be helpful to further understand the differences, if any, between Bangladesh and Malaysian strains 142. The only study so far that directly compared HeV and NiV infections in an animal model (hamster) did not appear to demonstrate significant differences in the two viruses either, except for the finding that NiV initially replicated in the upper respiratory tract epithelium, whereas early HeV replication was in lung interstitium 105.

Therapeutic and Vaccine Development

To date, there are no approved vaccines or therapeutic agents to treat human henipavirus infection. Various approaches being pursued as part of ongoing research have been reviewed 8. During the NiV outbreak in Malaysia, ribavirin (a broad spectrum antiviral drug) was given empirically, and subsequently thought to be useful for patients 15. However, its effect on human HeV infections, if any, was uncertain 69. Nonetheless, as it was found to have some effect on henipaviruses experimentally 36, 139, ribavirin was further tested for its anti‐henipavirus properties either alone or in combination with a potentially useful antimalarial drug called chloroquine 94. Unfortunately, in the hamster, ferret and African green monkey models, ribavirin or ribavirin/chloroquine combinations were found to be ineffective against infection and disease 32, 86, 104. Other drugs have been suggested as having possible anti‐henipavirus effects 69.

Passive immunization strategies using cross reacting, henipavirus‐specific antibodies raised against F and G viral proteins have been investigated in hamster, ferret and African green monkey models and found to be effective 5, 7, 35, 40, 41, 43. In fact, on humanitarian grounds, two patients with HeV infection have been given these antibodies with no adverse effects but the antiviral effectiveness has not been verified 9. Many new vaccine candidates for active immunization using different live recombinant vaccine, subunit vaccine and virus‐like particle platforms that mainly express henipavirus F and/or G proteins have been successfully developed and tested in various animal models including hamsters, pigs, cats, ferrets and African green monkeys models 6, 40, 60, 66, 77, 79, 87, 93, 96, 121, 125, 141. One of the recombinant subunit vaccines that contained the HeV G glycoprotein has now been licensed for use by the equine industry in Australia 103.

Conclusion

As far as newly emerging viruses, including henipaviruses, Ebola virus, SARS coronavirus, are concerned, one of the most important natural hosts is the bat 10, 112, 138. Because of the worldwide distribution and their flying range that can cover huge areas of human habitat in Southeast Asia, China, Japan, Oceania, Indian subcontinent, Australia and Africa, pteropid bats, in particular, are highly effective in virus dissemination 20, 29, 45, 46, 50, 54, 62, 64, 84, 102, 120. Hence, future henipavirus outbreaks in new areas/countries are probably inevitable as shown by the recent outbreak in the Philippines 14. A group of Africans who were involved with butchering bats for bush meat was recently found to be seropositive, thus providing evidence of previously unknown henipavirus spillover events in a continent that has yet to witness a henipavirus outbreak 90. The concept of One Health that integrates activities and contributions of international scientists, ecologists, veterinarians, health professionals, politicians and other stakeholders may be needed to minimize future outbreaks of bat‐borne zoonoses 112.