Up until the present, all the reported cases or outbreaks of infection with henipaviruses were within the geographical distribution of Pteropus
]. These bats appear to settle into subpopulations with limited interactions among colonies [1
]. NiV and HeV apparently remain separate within their hosts and respective regions with little overlap [1
]. On the other hand, other families of bats were often found coexisting in the same colony with Pteropus.
Antibodies reactive with Nipah virus were found in Eidolon dupreanum
in Madagascar [35
]. Furthermore, other novel henipavirus-like sequences and cross-reacting antibodies have recently been identified in Eidolon helvum
from Ghana, West Africa [36
], which potentially implicates much wider endemic regions of henipaviruses than previously known.
Here we report two monoclonal antibodies that recognized native conformations of N and P/V/W proteins of henipaviruses. In previous studies, monoclonal antibodies were produced either through phage display library screening, or using chemically inactivated virus/recombinant viral protein immunizations [22
]. The hybridomas in our study originated from mice immunized with γ-irradiated virus, and the secreted antibodies recognized native viral proteins. Linear epitope mapping on NiV N sequence indicates the epitope of 1A11 C1 is located within the last C-terminal 23 amino acids (a. a.) of the N protein (509-532). Alignment data of the C terminal area of N protein between NiV and HeV further indicated that the last 14 a.a. (520-532) were identical between the two viruses [41
], and would likely represent the actual epitope. Interestingly, this epitope is located right after the N-P interaction site (a.a. 468-496) on the N protein [41
]. Epitope mapping of other monoclonal antibodies to recombinant N protein or phage library screening of infected swine sera did not identify the C terminus of N protein as a site of antibody recognition [40
]. In addition, linear epitope mappings against the NiV N protein sequence using ascites fluids, sera from immunized animals or infected humans/pigs were performed and compared. The C terminal peptide of N protein was found to be a strong epitope recognized by all of polyclonal antibodies tested except for NiV infected human convalescent sera.
Our antigen capture ELISA, using plates coated with anti-N 1A11 C1 antibodies was capable of detecting HeV and NiV at a lower limit of detection of 4000 pfu/mL which is comparable to detection sensitivities reported in other antigen detection ELISAs [43
]. The 1A11 C1 or 2B10 p4 antigen detection assay also demonstrated its specificity by low background signal cutoff values for uninfected Vero cell lysate and Lassa virus. In addition, when using high dilutions of NiV/HeV infected Vero cell lysates (≥ 1: 300), non-specific signals were kept below background levels when polyclonal antibodies against MHF or Crimean-Congo hemorrhagic fever (CCHF) virus (data not shown) were coated on the plate.
Monoclonal antibody 1A11 C1 was also shown to capture NiV from one frozen pig lung specimen from the initial Malaysian NiV outbreak. However, corresponding RT-PCR and virus isolation results obtained at the time of the outbreak suggest the assay failed to identify another infected pig. A previous study had shown virus titers of 5 × 107
, 2.5 × 103
, and 6.3 × 105
pfu/g of swine CSF, lung, and spleen tissues, respectively [45
]. In our antigen detection assay, 10% tissue suspension (wt/vol) was used as the origin of serial dilutions. Virus loads in some tissue may be too low to be detected at this dilution range. Furthermore, viral degradation resulting from long term storage may further compromise the results of antigen capture ELISA. Unfortunately, we were unable to confirm this possibility by repeating RT-PCR and virus isolation since tissue samples were irradiated.
A previous study has described using monoclonal antibodies against N and M proteins to differentiate NiV from HeV by Western blot [38
]. In our study, 2B10 p4 antibody specifically captured native HeV P/V/W proteins and could only detect NiV proteins at high virus concentration or by Western blot. These results suggest that the binding affinity of 2B10 p4 could be influenced by how its epitope was presented by the P proteins of NiV and HeV. In fact, we were unable to identify linear epitope of 2B10 p4 from the NiV Malaysia P sequence despite knowing it should be located within the shared N-terminal sequence of P/V/W protein (a.a. 1-407). In contrast, polyclonal HMAF raised against NiV was able to recognize 6 individual plate-coated peptides in this region by direct ELISA (data not shown).
NiV and HeV soluble recombinant G proteins coated on beads had previously been developed and utilized in a Bio-Plex protein array to differentiate between infections with these viruses [37
]. As the G proteins of these viruses share 83.3% homology [19
] and the assay would be assessing polyclonal antibodies present in clinical specimens, it is unclear to what extent the ability to differentiate between the viruses would be maintained on analysis of diverse specimens from humans, bats or pigs. One of the advantages of the antigen capture assay described here is that it is based on the P protein of NiV and HeV which is highly diverse and share only 67.6% homology [19
], which may facilitate the ability to robustly differentiate among infections with henipaviruses.
Real time RT-PCR has been shown to detect Nipah virus with high sensitivity and specificity [28
]. However, nested RT-PCR using broad-range primers and sequencing may be required to identify newly emerging henipaviruses [36
]. Although further validation for our antigen capture assay will be needed once HeV and NiV infected tissue specimens are available, this relatively inexpensive and robust diagnostic tool could be useful in broad spectrum surveys for detection of henipaviruses.