40 Zoonotic paramyxoviruses

Hendra virus (HeV), Nipah virus (NiV), and Menangle virus (MenV) are recently emergent paramyxoviruses that are responsible for zoonotic infections and represent potential threats to agriculture and humans. In particular, HeV and NiV cause fatal disease in animals and man, and outbreaks of NiV continue to occur almost annually in Southeast Asia. Molecular biologic studies have made substantial contributions to the characterization of these new paramyxoviruses by providing an accurate picture of their relative taxonomic positions, and molecular techniques were used to provide rapid diagnostic capabilities. In the outbreaks of NiV in Malaysia, Bangladesh, and India, molecular biological data quickly identified the etiologic agent present, and RT-PCR and serologic assays were used to rapidly confirm NiV infections in humans and animals. There has only been one report of human illness due to MenV and one study has detected an antibody response to a related rubulavirus, Tioman virus (TiV), in humans. It is interesting that all of these viruses share a common reservoir in large fruit bats. Because of their clear potential to cause severe disease in humans and animals, NiV and HeV have been designated as Class C Select Agents and have been the focus of intense study since their emergence.

Analysis of the sequences of the entire genomes of both Hendra virus (HeV) and Nipah virus (NiV) provided convincing evidence that these viruses are members of a novel genus, Henipavirus, within the subfamily Paramyxovirinae (Mayo and van Regenmortel 2000). HeV and NiV share 68–92% amino acid identity in their protein coding regions and 40–67% nucleotide homology in the non-translated regions of their genomes (Harcourt et al. 2000, 2001). Compared to the other four genera within the Paramyxovirinae, the henipaviruses are more closely related to the respiroviruses and morbilliviruses than to the rubulaviruses and avulaviruses (Fig. 40.1). Of course, the recent genetic characterization of a number of novel paramyxoviruses shows that our understanding of the amount of diversity within this group of viruses is still incomplete (Bowden et al. 2001; Chua et al. 2000a, 2001b; Franke et al. 2001; Lamb 1996; Murray et al. 1995b; Philbey et al. 1998). However, it is clear that the henipaviruses have the potential to cause severe disease in humans and animals and further studies to characterize the pathogenesis, epidemiology, and virology of these viruses are clearly warranted (Eaton et al. 2006; Halpin and Mungall 2007; Lo and Rota 2008).

The single stranded, negative sense RNA genomes of NiV and HeV have the same gene order, 3’-nucleoprotein (N)-phosphoprotein (P)-matrix protein (M)-fusion protein (F)-attachment protein (G)-RNA dependant RNA polymerse (L)- 5’, as the respiroviruses and the morbilliviruses (Fig. 40.2). HeV and NiV retain a number of genetic features found in viruses throughout the subfamily (Harcourt et al. 2001, 2005), although the henipaviruses also have several unique genetic and biochemical features.

The genomes of HeV and NiV are 18, 234 and 18, 246 nucleotides in length, which, until the characterization of Beilong and J viruses (Jack et al. 2005; Li et al. 2006), were the largest genomes among the paramyxoviruses. In contrast, the average genome size for the other members of the Paramyxovirinae is approximately 15,500 nucleotides. The increased size of the HeV and NiV genomes is mostly due to the large sizes of the open-reading frame (ORF) for the P gene and the large 3’ untranslated regions present in several genes (Harcourt et al. 2001; Wang et al. 2000). The functional significance of the large 3’ untranslated regions has not been explored. The ‘rule of six’ (interaction of the nucleocapsid (N) protein with six nucleotides on the RNA genome, i.e. N phase context) states that the total length of the genomic RNA of viruses within the subfamily Paramyxovirinae must be evenly divisible by six in order to replicate (Calain and Roux 1993) and the sizes of the genomes of both HeV and NiV are evenly divisible by six indicating that the henipaviruses adhere to the rule of six. Additional evidence is provided by the observation that for HeV and NiV the differences in genome sizes as well as the sizes of the individual genes are evenly divisible by six (Harcourt et al. 2001, 2005), and by results of studies with a minigenome replication assay (Halpin et al. 2004).

The complete nucleotide sequence of HeV was completed in 2000 (Wang 1998; Wang et al. 2000) and since that time, only a few other HeV isolates have been sequenced and found to be virtually identical (Halpin et al. 2000; Hooper et al. 2000). Analysis of the complete genomic sequences of the NiV strains associated with the outbreaks in Malaysia in 1999 (NiV-M) and Bangladesh in 2004 (NiV-B) showed that the genome of NiV-B is six nucleotides longer than NiV-M, the prototype strain of NiV, and two nucleotides shorter than HeV (Harcourt et al. 2005). The additional six nucleotides are inserted in the 5’ non-translated region of the fusion protein (F) gene. The gene order and sizes of all the ORFs except V are conserved between NiV-B and NiV-M. The overall nucleotide homology between the genomes of NiV-B and NiV-M is 91.8% and the predicted amino acid homologies between the expressed proteins are all greater than 92% (Harcourt et al. 2005). The case fatality rates associated with NiV-M was lower than the case fatality rate observed in the outbreaks in Bangladesh (Chua et al. 2000a; Hossain et al. 2008). One of the major unanswered questions is whether there are differences in pathogenecity between NiV-M and NiV-B. Of course, the limited amount of sequence variation among strains could be related to their geographic distribution. The N gene sequence of a bat isolate of NiV from Cambodia (Reynes et al. 2005) is more closely related to the sequence of NiV-M than to the sequence of NiV-B, while sequences obtained from samples from an outbreak in India were more closely related to NiV-B (Chadha et al. 2006).

The genomes of paramyxoviruses contain a number of conserved cis-acting signals that regulate gene expression and replication (Lamb 2001). The cis-acting signals on the genomes of HeV and NiV including the gene start sites, gene stop sites, RNA editing sites, genomic termini and intergenic sequences are nearly identical between these viruses, and very closely related to the corresponding sequences within the genomes of respiroviruses and morbilliviruses (Harcourt et al. 2000, 2001; Wang 1998; Wang et al. 2000). The development of minigenome replication assays and reverse genetic systems for NiV (Freiberg et al. 2008; Halpin et al. 2004; Yoneda et al. 2006) will permit more detailed studies on the genetics and pathogenesis of the henipaviruses. NiV has been rescued from plasmid DNA and wild-type NiV and the rescued NiV showed similar percentages of mortality is a hamster infection model (Yoneda et al. 2006).

The P is an essential component of the replication complex for all paramyxoviruses including HeV and NiV. The P protein of NiV contains binding domains for the N at both its amino and carboxyl termini (Chan et al. 2004). The coding strategy for the P gene of the henipaviruses is similar to that found in the respiroviruses and morbilliviruses. In each case, a faithful transcript of the P gene codes for the P protein, while the transcript encoding the V protein is produced by RNA editing. RNA editing refers to the insertion of non-templated guanosine (G) nucleotides into the mRNA of the P gene to permit access to additional ORFs (Thomas et al. 1988). The V proteins of the respiroviruses, morbilliviruses and henipaviruses share the same amino terminus as their respective P proteins which, at the editing site, are joined to a unique, carboxyl-terminal cysteine rich ORF that is unique to V. The P genes of the henipaviruses also code for a C protein, which is produced by ribosomal choice from an overlapping reading frame located near the 5’ terminus of the P gene mRNA. As in the case of the morbilliviruses, the translational start site for the C protein of HeV and NiV is located downstream of the start codon for the P/V protein (Harcourt et al. 2000; Wang 1998). The P genes of HeV and NiV also have the capacity to code for a protein that is analogous to the W protein described for Sendai virus; W is expressed from an mRNA with a 2 G insertion at the editing site (Vidal et al. 1990).

The henipaviruses produced edited transcripts at twice the frequency of other paramyxoviruses such as measles virus (MeV), Sendai virus (SeV), and Newcastle disease virus (NDV) (Bankamp et al. 2008; Hausmann et al. 1999; Kato et al. 1997; Mebatsion et al. 2003; Vanchiere et al. 1995) with the exception of human parainfluenza virus-3 (hPIV-3), which edits approximately half of its P gene mRNA transcripts (Galinski et al. 1992), and bovine parainfluenza virus (Pelet et al. 1991). In NiV, approximately two-thirds of all P gene transcripts were edited and 50% of all transcripts encoded for P, 25% for V, and 25% for W. The relative number of mRNA transcripts encoding P protein compared to those encoding the V and W proteins decreased over the course of the first 30 hours of infection (Kulkarni et al. 2009). The P, V, W, and C were detected in both infected cells and in sucrose gradient purified virions. The P protein, while localized in the cytoplasm, concentrated near the plasma membrane. The V protein was evenly distributed throughout the cytoplasm and the W protein exclusively localized to the nucleus. The C protein was distributed in the perinuclear areas in a punctuate pattern (Lo et al. 2009; Shaw et al. 2005). The V and C proteins of paramyxoviruses have been shown to affect viral replication (Horikami et al., 1996; Witko et al. 2006). The NiV C, V, and W proteins inhibited NiV minigenome transcription and replication in a dose-dependent manner (Sleeman et al. 2008).

The C and V proteins of paramyxoviruses can inhibit the induction of type I IFNs and also block IFN signalling (Fontana et al. 2008; Lamb and Parks 2007). Several studies utilizing eukaryotic plasmid expression systems have provided insight into possible mechanisms by which NiV C, V, and W proteins block the host antiviral response. The NiV V protein inhibits IFN signal transduction by sequestering STAT1 and STAT2 in high molecular weight complexes in the cytoplasm and by inhibiting STAT1 phosphorylation (Rodriguez et al. 2002). The domain of V responsible for binding STAT1 resides between amino acids 100 and 160. Despite finding a STAT2 binding site on V (amino acids 230–271), STAT2 binding to V required the presence of STAT1 (Rodriguez et al. 2004).

Transient expression of NiV V, W, or C was able to rescue the replication of an IFN-sensitive NDV containing a GFP reporter gene in IFN-treated chicken embryo fibroblasts. The V and W proteins rescued GFP expression in a robust manner, while the level of GFP rescued by the C protein was less pronounced. Both the V and W proteins inhibited the expression of a luciferase reporter gene under the control of an IFN stimulated response element (ISRE) promoter, which demonstrated the ability to block IFN signalling (Park et al. 2003). The STAT1 binding domain was mapped to amino acids 50-150 of the N-terminus shared by the P, V, and W proteins (Rodriguez et al. 2004; Shaw et al. 2004). Interestingly, in spite of the shared N-terminus, P, V, and W have differential abilities to inhibit induction of ISRE promoters, with W having the most inhibitive capacity, followed by V, and then P (Shaw et al. 2004).

There is a nuclear localization signal (NLS) in the unique C-terminus of the NiV W protein, which requires basic residues at amino acid positions 439, 440, and 442 (Shaw et al. 2005). The NLS interacts with the nuclear importins, karyopherin-α3 and karyopherin-α4. While NiV V and W could inhibit an IRF-3 responsive promoter activation both by SeV infection and cytoplasmic dsRNA, only W could inhibit IRF-3 activation via dsRNA stimulation of TLR-3. The addition of a non-viral nuclear localization signal to the V protein resulted in the same inhibitory effect observed with W. This study also demonstrated that plasmid expression of the W protein in cells reduced levels of phosphorylated IRF-3, indicating a distinct mechanism of IFN evasion that was unique to W. There was a demonstrable difference in the ability of the V and W proteins to inhibit the IKK-like kinases, IKKε and TBK-1. Both V and W inhibited IKKε–mediated activation of ISG transcription, while only W was able to inhibit TBK-1-mediated activation of ISG transcription. The authors concluded that because W localizes to the nucleus, it could inhibit both pathways leading to IRF-3 activation, the TLR-3 pathway and the virus/dsRNA pathway (Shaw et al. 2005).

A single amino acid change at residue 125 abolished the ability of NiV V to bind STAT1 and to block IFN signalling. NiV was able to block IFN signalling in cell lines from numerous mammalian species, indicating that this ability does not constrain the virus from crossing species (Hagmaier et al. 2006). More recently, it has been shown that plasmid expressed NiV V; along with many other paramyxovirus V proteins bind the helicase Mda-5 via its cysteine-rich C-terminal domain, and blocks activation of the IFN-α promoter by preventing the oligomerization of Mda-5 upon binding to dsRNA (Childs et al. 2007, 2009). A novel interaction between NiV V and a host cell kinase Polo-like kinase 1 (PLK1) has recently been demonstrated (Ludlow et al. 2008). The binding site for PLK1 overlaps with the STAT1 binding region. By constructing point mutants in the shared binding region, STAT1 binding was abrogated independently of PLK1 binding to NiV V. Modifying the binding sites to STAT1 and PKL1 on the NiV P protein did not affect NiV minigenome replication, which indicates the possibility of attenuating NiV by altering the ability of viral proteins to interact with host proteins.

The genes coding for the RNA dependent RNA polymerase (L protein) of HeV and NiV have a linear domain structure that is conserved in all of the Mononegavirales (Poch et al. 1990). In domain III, all of the negative stranded RNA viruses have a predicted catalytic site with the amino acid sequence GDNQ. The sequence, QDNE, is found only in HeV, NiV and Tupaia paramyxovirus (Harcourt et al. 2005; Tidona et al. 1999). However, substitution of the E for Q did not affect the function of the L protein of NiV in a minigenome replication assay (Magoffin et al. 2007).

The two membrane glycoproteins of HeV and NiV, F and G, serve the same functions as the membrane glycoproteins of the morbilliviruses and respiroviruses. Both G and F are required for cell fusion and heterotypic mixtures of the G and F proteins of HeV and NiV are also fusion competent (Bossart et al. 2002; Tamin et al. 2002).

The F proteins of the Paramyxovirinae are type I membrane glycoproteins that facilitate the viral entry process by mediating fusion of the virion membrane with the plasma membrane of the host cell. F proteins are synthesized as inactive precursors, F₀, that are converted to biologically active subunits, F₁ and F₂, following proteolytic cleavage by a host cell protease (Lamb 2001). The fusion peptide, located at the amino terminus of the F₁ protein, is highly conserved within the Paramyxovirinae (Langedijk et al. 1997) and the fusion peptides of HeV and NiV are related to the fusion proteins of other paramyxoviruses with the exception that HeV and NiV have leucine at the first position while almost all of the other viruses have phenylalanine (Harcourt et al. 2000). However, substitution of phenylalanine for leucine in the F₁ of NiV does not affect its ability to form syncytia (Moll et al. 2004a).

Among the paramyxoviruses, the carboxyl terminus of F₂ protein subunits contains either single basic, or multiple basic, amino acids that comprise the cleavage site between F₁ and F₂. F proteins with multiple basic amino acids are cleaved by furin-like protease during exocytosis from the host cell. F proteins of viruses with a single basic amino acid are cleaved at the cell surface by trypsin like proteases and these viruses usually require the addition of exogenous trypsin to replicate in cell culture. While HeV and NiV have a single basic residue at the cleavage site, both produce productive infections in a variety of cell lines in the absence of exogenous trypsin. In addition, the cleavage site of the F proteins of NiV and HeV do not contain a furin-like protease consensus sequence (R-X-R/K-R) found in most morbilliviruses, rubulaviruses, and pneumoviruses (Lamb and Jardetzky 2007; Langedijk et al. 1997), and the basic amino acids are not required for cleavage (Moll et al. 2004b). Cleavage of the F protein of NiV and HeV occurs by a novel mechanism involving clatherin mediated endocytosis via a tyrosine dependant signal on the cytoplasmic tail (Michalski et al. 2000; Vogt et al. 2005). The F proteins of both HeV and NiV require the endosomal protease, cathepsin L, for proteolytic processing (Pager and Dutch 2005). N glycans of the F protein of NiV are required for proper proteolytic processing and these glycans may modulate access to neutralization epitopes (Aguilar et al. 2006; Moll et al. 2004b).

The attachment proteins of the Paramyxoviridae are type II membrane glycoproteins and are responsible for binding to receptors on host cells (Lamb and Jardetzky 2007; Lamb 1996, 2001). Unlike many other paramyxoviruses, neither of the henipaviruses has been shown to have erythrocyte binding or neuraminidase activities. The G proteins of the henipaviruses are most closely related the haemagglutinin neuraminidase (HN) proteins of the respiroviruses (Yu et al. 1998). The conservation of most of the structurally important amino acids suggests that the G proteins of HeV and NiV would have structures that are very similar to the structure proposed for the attachment proteins of other paramyxoviruses (Langedijk et al. 1997). EphrinB2, the membrane-bound ligand for the EphB class of receptor tyrosine kinases, specifically binds to G proteins of henipaviruses and is a functional receptor for HeV and NiV (Bonaparte et al. 2005) While EphrinB3 has also been shown to be a functional receptor for both viruses, the binding of NiV to EphrinB3 is much more efficient than the binding of HeV (Negrete et al. 2005, 2006) and the G protein of NiV has distinct binding regions for EphrinB2 and EphrinB3 (Negrete et al. 2006). EphrinB3 but not EphrinB2 is expressed in the brain stem, so the difference in the abilities of HeV and NiV to bind to these cellular receptors is consistent with the neuroinvasiveness of NiV (Negrete et al. 2005, 2006). NiV infection does not appear to down regulate cell surface expression of EphrinB2 or EphrinB3 (Sawatsky et al.2007).

Analysis of the crystal structures of the ephrinB2 and ephrinB3 interactions with NiV G (Bowden et al. 2008b; Xu et al. 2008) confirmed the accuracy of the functional mapping studies (Negrete et al. 2006), and also provided a comprehensive analysis of critical residues composing the hydrophobic binding cleft which interacts with high affinity to ephrinB2 and ephrinB3. The unbound form of NiV G contains highly processed complex-type glycans with negligible amounts of oligomannose-type glycans. Interestingly, the N-acetylglucosamine (GlcNAc) β12Man terminal structures on NiV G were noted as a potential ligand for LSECtin, a C-type lectin expressed on sinusoidal endothelial cells of lymph nodes and the liver (Bowden et al. 2008a). A receptor binding activation site in the stalk region of NiV G triggers fusion by the F protein (Aguilar et al. 2009). This finding is consistent with the attachment and fusion proteins of paramyxoviruses (Corey and Iorio 2007, 2009; Lamb and Parks, 2007; Lee et al. 2008; Melanson and Iorio 2004, 2006).

As for the other paramyxoviruses, the NiV surface glycoproteins are the primary targets for neutralizing antibodies (Crameri et al. 2002; Guillaume et al. 2006; Tamin et al. 2002). Recombinant vaccinia viruses expressing NiV F and G proteins elicit neutralizing antibodies against NiV and protect Syrian hamsters and pigs against lethal NiV challenge (Guillaume et al. 2006; Tamin et al. 2002; Weingartl et al. 2006) and cats are protected from a lethal challenge by a soluble NiV G (McEachern et al. 2008). Antibodies to F or G also provided passive protection in the hamster challenge model (Guillaume et al. 2004a).

The sequences of the complete genomes of these antigenically related viruses clearly showed that both MenV and TiV are members of the rubulaviruses (Bowden and Boyle 2005; Bowden et al. 2001; Chua et al. 2002b). Like other members of the genus, these viruses lack the SH gene that is found in mumps virus and simian virus 5. The coding strategy for the P genes is also the same as that found in the other rubulaviruses. The unedited transcript codes for the V protein while a 2 G insertion at the RNA editing site produces the transcript coding for the P protein. MenV lacks detectable neuraminidase and erythrocyte binding activity (Paton et al. 1999) and the HN proteins of MeV and TiV lack the hexapeptide, NRKSCS, that is proposed to be essential for neuraminidase activity (Bowden et al. 2001). Compared to the other rubulaviruses, TiV and MeV also have some unique genetic features in their RNA start sites and intergenic regions (Chua et al. 2001b; Chua et al. 2002b).

All of the outbreaks of HeV have occurred in Australia and all have involved horses with or without spillover into humans. The first incident occurred in September 1994 at a stable in Hendra, a suburb of Brisbane. An outbreak of acute respiratory disease resulted in 14 horse deaths (Murray et al. 1995a, 1995b; Selvey et al. 1995). Approximately 1 week after exposure to the index horse case, two humans who had contact with the horses developed influenza-like illness and one died after a seven day illness. The first isolate of HeV was isolated from this fatal case. The second incident in 1995 involved the death of a farmer who lived near Brisbane (O’Sullivan et al. 1997) and who had assisted a veterinary surgeon during treatment of horses. The patient developed meningitis shortly after he assisted in the autopsies of two horses that died of acute respiratory distress and rapid-onset neurological symptoms, respectively. Both horses were retrospectively diagnosed with HeV (Hooper et al. 1996). The patient recovered completely and remained symptom-free for 13 months before his fatal illness, which is believed to have resulted from persistent infection with HeV. A third fatal case of HeV infection occurred in 2008 in a veterinarian who had been treating sick horses at a veterinary clinic outside of Brisbane (PROMED-MAIL 2008). In 2004, a veterinarian working near Cairns became infected with HeV after treating a terminally ill horse. The illness was mild and the patient recovered and remains well (Hanna et al. 2006).

Despite the potential for HeV to infect a wide variety of animals under experimental conditions (Williamson et al. 1998, 2000), horses appear to be the primary source of human HeV infection (O’Sullivan et al. 1997; Selvey et al. 1995). After the first two incidents, no evidence of infection was found in 22 other persons who reported feeding or nursing sick horses or participating in their autopsies and more than 110 other persons associated with, or living near, the affected stables (McCormack et al. 1999). These data indicate that transmission of infection from horses to humans is not very efficient and requires very close contact. Laboratory experiments suggest that the urine and saliva from infected horses are important in disease transmission (Williamson et al. 1998), whereas respiratory spread is less likely. Human-to-human transmission of HeV has not been documented, either among domestic contacts or among health care workers (McCormack et al. 1999; Selvey et al. 1995).

The pathology of HeV infections and factors in disease production are incompletely characterized because of the small number of cases. Following initial infection with the virus, possibly through the oral or respiratory route or through direct inoculation of cutaneous abrasions with infectious secretions, viremia develops resulting in spread to various organs, including the central nervous system. The incubation period of acute disease is approximately 5–7 days. Autopsy of the horse trainer who died in the first HeV incident showed a severe interstitial pneumonia with both lungs congested, hemorrhagic, and edematous. Histologic examination showed focal necrotizing alveolitis, with giant cells, syncytium formation, and viral inclusions. HeV was isolated from post-mortem lung, liver, kidney, and spleen specimens (Murray et al. 1995b; Selvey et al. 1995). The farmer from Mackay who died more than one year after initial infection with HeV (O’Sullivan et al. 1997) showed leptomeningitis with prominent lymphocyte and plasma cell infiltration. There were discrete foci of necrosis in the neocortex, basal ganglia, brainstem, and cerebellum and multinucleate endothelial cells were observed in the brain, liver, spleen, and lungs.

HeV infections begin with abrupt onset of an influenza-like illness, which includes myalgia, headache, lethargy, and vertigo (Hanna et al. 2006; Murray et al. 1995a, 1995b). The first patient developed nausea and vomiting on the fourth day of illness, and deteriorated rapidly in the next 2 days, requiring admission to an intensive care unit and mechanical ventilation. He died on the seventh day of illness. Laboratory results showed thrombocytopenia, increased levels of creatine phosphokinase, lactic dehydrogenase, aspartate aminotranferase, alanine aminotransferase, and glutamyltransferase, and features of dehydration and acidosis (Murray et al. 1995b). Chest radiographs showed diffuse alveolar shadowing. No laboratory abnormalities were detected in the patient who survived.

The affected farmer in the second HeV incident primarily had neurologic manifestations (O’Sullivan et al. 1997). He initially presented with features of meningitis, including headache, drowsiness, vomiting, and neck stiffness. Thirteen months following complete recovery, the patient presented again with a 2-week history of irritable mood and low back pain, 3 episodes of focal seizures of the right arm, and an episode of generalized tonic-clonic seizures. In the following week, he continued to have a low-grade fever, focal and generalized seizures. By day 7, he developed dense right hemiplegia, signs of brainstem involvement, and depressed consciousness, requiring intubation. The patient remained unconscious and febrile until he died 25 days after admission. Cerebrospinal fluid examination showed an elevated protein level, normal glucose level, and mononuclear pleocytosis. MRI imaging of the brain showed multifocal cortical lesions sparing the subcortical white matter that became more pronounced and widespread prior to death.

The first known human infections with NiV occurred during an outbreak of severe encephalitis in Southeast Asia in 1998–1999. Two hundred sixty-five patients (40% fatal) and 11 patients (1 fatal) with laboratory-confirmed NiV disease were reported in Peninsular Malaysia and Singapore, respectively (Anon. 1999a, 1999b; Chua et al. 2000a; Eaton et al. 2006). This outbreak began in October 1998 in northern Malaysia and then spread southward in conjunction with the movement of pigs, resulting in at least three other clusters of human disease in Malaysia. In Singapore, abattoir workers who slaughtered pigs imported from outbreak-affected areas in Malaysia were exclusively affected (Chew et al. 2000; Paton et al. 1999). Adult males who were primarily involved in pig farming activities accounted for more than three-fourths of cases in Malaysia. Infections were also documented in abattoir workers (Chew et al. 2000; Paton et al. 1999), in veterinary personnel, and in military personnel involved in pig culling activities.

Since the initial outbreak, human cases of NiV encephalitis have occurred in several small outbreaks in India, 2001 (Chadha et al. 2006), and Bangladesh, 2001 to 2007 (Anon. 2003; Eaton et al. 2006; Hossain et al. 2008; Hsu et al. 2004; World Health Organization (WHO) 2001, 2004a, 2004b, 2004c). These smaller outbreaks had a marked increase of case fatality rate (CFR) which ranged from 67% to 92% and a number of epidemiologic features associated with these outbreaks differed from those of the initial outbreak. In the outbreak in Meherpur, Bangladesh in 2001, both close contacts with infected patients as well as with sick cows were associated with NiV infection, although samples from cows were not available for testing (Hsu et al. 2004). Eight households were affected, and of the 13 cases reported, 9 were related either by blood or by marriage to the index patient. Person-to-person contact was also a primary risk factor during the 2003 outbreak in the Nagoan district of Bangladesh, but in this instance there were no blood relationships between affected households (Hossain et al. 2008). In the outbreak in Faridpur, Bangladesh in 2004, person-to-person transmission was implicated, and the possibility of nosocomial transmission was demonstrated by the detection of NiV RNA on hospital surfaces (Gurley et al. 2007a, 2007b). Retrospective analysis of an outbreak in Siliguri, India in 2001 confirmed that nosocomial transmission resulted in the amplification of the outbreak (Chadha et al. 2006). Ingestion of NiV-contaminated date palm sap was reported as the primary risk factor during an outbreak in the Tangail district of Bangladesh in 2005, and this practice was potentially linked to more recent outbreaks of NiV in the Manikganj and Rajbari districts of Bangladesh (Luby et al. 2006; PROMED-MAIL, 2008). The increase in CFR in the Bangladesh outbreaks may be due to the inherent strain-specific differences between Malaysian and Bangladeshi strains of NiV, or the comparatively lower level of supportive care available in Bangladesh compared to Malaysia and Singapore (Harcourt et al. 2005).

Sequence analysis of different strains of NiV has also provided some information about the transmission patterns of the virus. Molecular biological data suggest that there were at least two introductions of NiV into pigs prior to the outbreak on 1999 (AbuBakar et al. 2004). Only one of these variants was associated with the explosive spread within pig farms and subsequent transmission to humans suggesting that a single spillover from the reservoir triggered the outbreak. In contrast, the sequence heterogeneity observed between samples obtained from the outbreak in Bangladesh in 2004 suggesting multiple spillovers between the reservoir and humans (Harcourt et al. 2005).

During the outbreaks in Malaysia and Singapore, close contact with pigs was the primary source of human NiV infection (Mohd Nor et al. 2000; Parashar et al. 2000). In pigs, extensive infection of the upper and lower airways is seen with evidence of tracheitis, and bronchial and interstitial pneumonia and a harsh, non-productive cough was a prominent clinical feature (Chua et al. 2000a; Hooper and Williamson 2000). Vasculitis of small vessels in the kidney was also seen (Chua et al. 2000a; Hooper and Williamson 2000), and viral antigen was detected by IHC studies as focal staining in renal tubular epithelium. Therefore, exposure to respiratory secretions and possibly the urine of infectious pigs likely resulted in transmission of virus among pigs and to humans. In experimental studies, transmission among pigs occurred through both oral and in-contact exposure (Hooper and Williamson 2000). Serologic studies demonstrated evidence of infection among other species of animals in Malaysia, including dogs and cats (Chua et al. 2000a; Hooper and Williamson 2000). It is unclear whether humans are at risk from exposure to infected animals other than pigs, but this possibility cannot be excluded because some patients reported no direct contact with pigs (Goh et al. 2000; Parashar et al. 2000). In Bangladesh, pigs did not act as intermediate hosts of the virus, and transmission directly from bats or from fruits or commodities such as date palm syrup that were contaminated by bats may have caused the primary infections in the small outbreaks occurring there (Anon. 2004; Luby et al. 2006; Montgomery et al. 2008).

Although NiV is excreted in respiratory secretions and urine of patients (Chua et al.2001a), a survey of health care workers demonstrated no evidence of human-to-human transmission in Malaysia (Mounts et al. 2001). Transmission from patients to relatives or caregivers in contact with patients during the course of disease was strongly implicated in at least some of the cases in Bangladesh and perhaps in India (Anon. 2004; Chadha et al. 2006; Gurley et al. 2007a, 2007b; Montgomery et al. 2008). The risk of person-to-person transmission may be increased in settings in which standard infection control measures are not the usual practice.

The incubation period among Bangladeshi patients with NiV infection who had well-defined exposure to another case, was 9 days (range, 6–11 days) (Hossain et al. 2008), though longer incubation periods were noted in Malaysia (Parashar et al. 2000). A multiorgan vasculitis associated with infection of endothelial cells was the major pathologic feature of NiV infection (Goh et al. 2000). Occasionally, multinucleate giant cells characteristic of paramyxovirus infections were observed in the affected vascular endothelium. Infection was most pronounced in the central nervous system, where a diffuse vasculitis characterized by segmental endothelial cell damage, mural necrosis, karyorrhexis, and infiltration with polymorphonuclear leukocytes and mononuclear cells was noted. The lesions are primarily seen in the cerebral cortex and brain stem with extension to parenchymal tissue, where extensive areas of rarefaction necrosis were seen. Eosinophilic, mainly intracytoplasmic, viral inclusions with a ‘melted-tallow’ appearance were seen in the affected neurons and parenchymal cells. Evidence of endothelial infection and vasculitis were also seen in other organs, including the lung, heart, spleen, and kidney (Wong et al. 2002). NiV has been isolated from cerebrospinal fluid, tracheal secretions, throat swab, nasal swab, and urine specimens of patients (Goh et al. 2000) (Parashar et al. 2000).

Limited data are available on the immune response to NiV infection, correlates of immune protection and disease resolution. A serum IgM response occurs shortly after onset of illness, with 50% of patients being antibody-positive on the first day of illness and 100% being antibody-positive by the third day (Ramasundrum 2000) with persistent IgM detectable up to 3 months after symptom onset. An IgG antibody response is seen in 10–29% of patients in the first 10 days of illness, and in 100% of patients after days 17–18 of illness.

The onset of NiV disease is abrupt, usually with the development of fever. Often, patients deteriorate rapidly, requiring hospitalization 3–4 days after onset of symptoms. Severe encephalitis is the most prominent clinical manifestation. Fever, headache, dizziness, vomiting, and reduced level of consciousness are the most common features; acute respiratory failure was noted in some of the outbreaks in Bangladesh (Goh et al. 2000; Hossain et al. 2008). Several other features of neurologic involvement, particularly signs of brain-stem dysfunction, were noted in patients during the course of illness. NiV disease was fatal in up to one-third of hospitalized patients in Malaysia (Chua et al. 2000b; Goh et al. 2000). Residual neurological deficits occurred in 10%–15% of patients (Goh et al. 2000; Paton et al. 1999; Sarji et al. 2000). Recurrence of neurologic dysfunction was seen in some patients, including neurologic relapse with seizures and/or cognitive impairment or focal signs such as isolated cranial nerve dysfunction (Sarji et al. 2000). In another study, delayed progression to neurologic illness following NiV infection was not observed, but persistent fatigue and functional impairment were frequent (Sejvar et al. 2007). Neurologic dysfunction may persist for years after acute infection.

Antibodies to HeV antibodies were detected in several fruit bat species primarily of the Pteropus genus in Queensland and the virus isolated from a fruit bat was indistinguishable from that isolated from horses and humans (Halpin et al. 1999, 2000). Transmission of HeV to horses may occur through the ingestion of pasture recently contaminated by the urine or infected fetal tissue of fruit bats (Field et al. 2000; Halpin et al. 2000). Despite the ability of NiV to infect many mammalian species (dogs, cats, ferrets, pigs, horses), the absence of neutralizing antibody in non-infected hosts indicated that these were ‘dead-end hosts’ (Mohd et al. 2000). Since HeV had been detected in fruit bats, they were the logical reservoir for NiV (Halpin et al. 2000). Neutralizing antibodies to NiV were found primarily in Pteropus hypomenalus and Pteropus vampyrus during initial surveillance studies, but virus was not isolated (Yob et al. 2001). The first isolation NiV from bats was from Pteropus hypomenalus on Tioman Island, Malaysia (Chua et al. 2002a). Since then, antibodies to henipaviruses have been detected in other Pteropus species (Pteropus lylei, Pteropus giganteus,) as well as in non-Pteropus species (Hipposideros larvatus, Scotophiilus kuhlii) at much lower frequencies, in Cambodia, China, Thailand, India, Indonesia, Bangladesh, to Madagascar (Epstein et al. 2008; Hsu et al. 2004; Li et al. 2008; Reynes et al. 2005; Sendow et al. 2006; Wacharapluesadee et al. 2005). Antibodies to henipaviruses have been detected in Eidolon helvum in west Africa (Hayman et al. 2008).

In 1997 a piggery in New South Wales, Australia noticed a decline in the farrowing rate of sows, which was associated with an increase in the proportion of malformed, mummified, and stillborn piglets, with occasional abortions (Philbey et al. 1998, 2007). The affected piglets had craniofacial and spinal abnormalities and degeneration of the brain and spinal cord. A new paramyxovirus, MenV, was isolated from the brain, heart, and lung specimens of several affected piglets. No disease was seen in postnatal pigs of any age, but a high proportion of serum specimens (>95%) collected from these animals contained high titres of antibodies that neutralized the virus. Evidence of infection with MenV was also detected in porcine sera from two other associated piggeries that received weaned pigs from the affected piggery, but not in sera from several other piggeries throughout Australia (Kirkland et al. 2001).

A serologic survey of persons who came into contact with the affected piglets (Chant et al. 1998) detected a high titre of neutralizing antibodies in two workers, one at the affected piggery and one at an associated piggery. Both workers had an influenza-like illness at the same time as the outbreak in pigs and no alternative cause for their infection was identified despite serologic testing. Thus, the illness was attributed to MenV virus infection.

Close contact with infected piglets appears to be the primary mode of transmission of MenV to humans (Chant et al. 1998). A large breeding colony of fruit bats roosted within 200 m of the affected piggery and sera from several bats had antibodies that neutralized MenV (Philbey et al. 1998). Serum samples collected from bats before the outbreak also had detectable antibodies to MenV, while all other serum samples collected from a variety of wild and domestic animals in the vicinity of the affected piggery tested negative for antibodies to the virus. Fruit bats are likely the primary reservoirs of MeV.

Both of the affected workers had a similar illnesses which were characterized by abrupt onset of fever, malaise, chills, drenching sweats, and severe headache (Chant et al. 1998). On the fourth day of illness, both developed a spotty, red, nonpruritic rash. Bilateral hypochondrial tenderness was present in one patient, and an abdominal ultrasound conducted 2 months after the illness showed splenomegaly and liver size at the upper limit of normal. Both patients recovered after approximately 10 days of illness.

Very little is known about Tioman virus (TiV) which was isolated from fruit bats on Tioman Island, Malaysia (Chua et al. 2001b). Several pigs experimentally inoculated with TiV developed pyrexia, produced neutralizing antibodies and shed virus (Yaiw et al. 2008). Serologic testing of 169 inhabitants of Tioman Island showed that a small percentage (3.0%) had serologic evidence of infection by TiV or a related virus (Yaiw et al. 2007).

The development or characterization of animal models to study henipavirus infections is critical for understanding their pathogenesis and development of new therapeutics or vaccines. Both cats and golden hamsters have been used as small animal models and both develop fatal disease after challenge with NiV. In cats, virus is mostly present in the respiratory epithelium, while hamsters develop neurologic disease (Guillaume et al. 2004a, 2006; Mungall et al. 2006a). NiV in pigs causes a febrile respiratory illness with or without neurological signs (Middleton et al. 2002; Mohd Nor et al. 2000; Weingartl et al. 2005). Infection of fruit bats with NiV did produce clinical signs; some of the bats seroconverted with intermittent excretion of low levels of virus (Middleton et al. 2007). Golden hamsters are highly susceptible to HeV infection (Guillaume et al. 2009).

NiV is internationally classified as a biosafety level 4 (BSL-4) agent, thus clinical specimens must be handled with caution. Propagation of viruses from clinical specimens known to be infected with NiV is not recommended without appropriate containment facilities. The Centers for Disease Control and Prevention, Atlanta, GA USA (CDC) and the Australian Animal Health Laboratory, Geelong, Australia (AAHL) have adopted the approach that primary virus isolation from specimens of outbreaks not already proven to be NiV takes place at BSL-3. However, if the results of cell culture suggest the presence of these agents, cultures should be transferred to BSL-4 to conform to biosafety guidelines (Daniels et al. 2001).

During the initial NiV outbreak, enzyme-linked immunosorbent assays (ELISAs) specific for detecting anti-HeV IgM and IgG were used to diagnose NiV infection. A NiV-specific ELISA was eventually transferred to surveillance labs in Malaysia (Daniels et al. 2001). For cases in which the ELISAs gave equivocal results, negative-stained cerebrospinal fluid (CSF) specimens were subjected to transmission electron microscopy to visually confirm the presence of NiV (Chow et al. 2000). Immunohistochemistry was crucial in detecting NiV antigen in ex vivo tissues of humans, dogs, and pigs (CDC 1999a, 1999b; Chua et al. 2000a). Complete molecular characterization of both the Malaysian and Bangladeshi strains of NiV has enabled the development of both standard and real-time RT-PCR assays to detect viral RNA from serum, urine, and CSF (AbuBakar et al. 2004; Guillaume et al. 2004b; Wacharapluesadee and Hemachudha 2007). Recently-developed ELISAs now utilize recombinant-expressed purified NiV antigens, and a number of high-affinity monoclonal antibodies have been generated for diagnostic and prophylactic purposes (Chen et al. 2006; Eshaghi et al. 2004, 2005a; Eshaghi et al. 2005b; Juozapaitis et al. 2007; Kashiwazaki et al. 2004; Tan et al. 2004; Tanimura et al. 2004a, 2004b; Yu et al. 2006; Zhu et al. 2006, 2008). While serum neutralization tests (SNTs) have long been a reference standard, the next generation of SNTs will circumvent the use of live virus, and be able to differentiate between NiV and HeV specific infection (Bossart et al. 2007; Chen et al. 2007; Kaku et al. 2009; Tamin et al. 2009).

In Malaysia, ribavirin treatment was shown to reduce mortality rates (Chong et al. 2001). The interferon inducer poly(I)-poly(C₁₂U) prevented mortality in 5 of 6 animals in a hamster model of NiV infection, while a 5-ethynyl analogue of ribavirin and several other OMP-decarboxylase inhibitors were shown to have anti-NiV activity in vitro (Georges-Courbot et al. 2006). Recent developments for NiV antivirals focus on inhibitors of fusion and receptor-binding. Peptides corresponding to the C-terminal heptad repeat of the F protein from HeV, NiV, and human parainfluenzavirus 3C have been shown to inhibit HeV and NiV infection in vitro (Bossart et al. 2005b; Porotto et al. 2006, 2007). Soluble versions of the G glycoprotein and Ephrin B2 have been shown to inhibit NiV envelope-mediated infection, and could be used as therapeutics (Bonaparte et al. 2005; Bossart et al. 2005a; Negrete et al. 2005).

No passive immunoprophylaxis, antiviral chemoprohylaxis, or vaccine is currently available for henipavirus infections. The principal means of preventing human infections are early recognition of disease and use of standards protective precautions to avoid exposure. Since interruption of transmission to horses or pigs from the natural reservoir of these viruses, presumably fruit bats, is difficult to prevent, early identification of infected animals and use of appropriate personal protective measures to prevent transmission are keys to reducing the risk to humans.

The G glycoprotein of NiV shared with 83.3% homology with the G protein of HeV virus, whereas the F proteins of the two viruses have 88.1% homology (Harcourt et al. 2000; Wang et al. 2001). Recombinant expressed, soluble versions of the G glycoprotein (sG) from NiV (Bossart et al. 2005a) were used to vaccinate cats and produced high antibody titres along with complete protection from NiV challenge (McEachern et al. 2008; Mungall et al. 2006b). Purified sG retains a number of important native structural, functional and antigenic features, making it potentially suitable as a vaccine candidate, although consistent large scale expression has been problematic. Expression of NiV F, G and M proteins leads to production VLPs (Ciancanelli and Basler 2006; Patch et al. 2007). Since VLPs have been able to generate a protective immune response against Ebola and Marburg viruses (Swenson et al. 2005) they might serve as a more efficient method of immunization against NiV. DNA vaccination with plasmids containing either NiV G or F genes stimulated considerable IgG responses in mice (Wang et al. 2006), while canarypox virus-based vaccine vectors expressing NiV G or F (or both) protected pigs against challenge and prevented viral shedding (Weingartl et al. 2006). Vaccination with these canarypox virus-based vaccines appeared to stimulate both type 1 and type 2 cytokine responses, suggesting this approach may be highly effective for the prevention of livestock infection.