|Home | About | Journals | Submit | Contact Us | Français|
Nipah virus (NiV) and Hendra virus (HeV) are zoonotic paramyxoviruses capable of causing severe disease in humans and animals. These viruses require biosafety level 4 (BSL-4) containment. Like other paramyxoviruses, the plaque reduction neutralization test (PRNT) can be used to detect antibodies to the surface glycoproteins, fusion (F) and attachment (G), and PRNT titers give an indication of protective immunity. Unfortunately, for NiV and HeV, the PRNT must be performed in BSL-4 containment and takes 5–7 days to complete. Thus, we have developed a neutralization assay using VSV pseudotype particles expressing the F and G proteins of NiV (pVSV-NiV-F/G) as target antigens. This rapid assay, which can be performed at BSL-2, was evaluated using serum samples from outbreak investigations and more than 300 serum samples from an experimental NiV vaccination study in swine. The results of the neutralization assays with pVSV-NiV-F/G as antigen showed a good correlation with those of standard PRNT. Therefore, this new method has the potential to be a rapid and cost-effective diagnostic method, especially in locations that lack high containment facilities, and will provide a valuable tool for basic research and vaccine development.
Nipah virus (NiV) and Hendra virus (HeV) are recently emergent, related paramyxoviruses that can cause severe diseases in humans and animals. These viruses are the only two members of a newly named genus, Henipavirus, within the family Paramyxoviridae (Bellini et al., 2005; Wang et al., 2001). HeV emerged in Australia in 1994 and caused severe respiratory disease in humans and horses (Murray et al., 1995). There have been sporadic cases of HeV through 2006 (Anonymous, 2008; Field et al., 2007; Westbury, 2000a; Westbury, 2000b), but the virus has not been detected outside of Australia. The first outbreak of NiV occurred in 1998–1999 in Malaysia and Singapore, and was manifested as a respiratory disease in swine and fatal, febrile encephalitis in humans (Chua et al., 2000). In 2001, NiV outbreaks occurred in Siliguri, India (Chadha et al., 2006; Harit et al., 2006). From 2001–2004, four outbreaks of NiV took place in Bangladesh involving 93 patients with a 73% case fatality rate (Hossain et al., 2008). In the Malaysian outbreak, pigs served as the amplifying host and transmitted the virus to humans (Chew et al., 2000; Chua et al., 2000; Lam and Chua, 2002; Paton et al., 1999). Unlike Malaysia (Mounts et al., 2001), person-to-person transmission and food-borne infection of NiV was reported in Bangladesh (Gurley et al., 2007a; Gurley et al., 2007b; Luby et al., 2006). Nucleotide sequencing studies showed that NiV strains in India were more closely related to those in Bangladesh than Malaysia (AbuBakar et al., 2004; Harcourt et al., 2005). Fruit bats were found to be the main reservoir of both HeV and NiV (Chua et al., 2002; Halpin et al., 2000). Five species of bats in peninsular Malaysia had antibodies to NiV (Yob et al., 2001). More recently, antibodies to Nipah-like viruses were detected among bats in Bangladesh (Hsu et al., 2004), Cambodia (Olson et al., 2002; Reynes et al., 2005), Thailand (Wacharapluesadee et al., 2005), Madagascar (Lehle et al., 2007), India (Epstein et al., 2008), China (Li et al., 2008) and Ghana (Hayman et al., 2008). The widespread distribution of fruit bats suggests that spillover of NiV from this reservoir into the human population could occur in many countries within Southeast Asia and sub-Saharan Africa with serious consequences for both human health and agriculture (Epstein et al., 2006; Hayman et al., 2008).
Enzyme-linked immunoassay (ELISA) and RT-PCR were used as a first-line diagnostic test in suspected NiV outbreaks (Daniels et al., 2001). However, since there is an approximately 2 percent false positive rate in the ELISA, confirmatory testing was performed by using a serum neutralizing antibody test (Daniels et al., 2001). The neutralization test requires live NiV and must be performed at biosafety level 4 (BSL-4) creating a significant challenge for countries lacking high containment facilities. Therefore, there is a need to develop alternative diagnostic methods for detecting antibodies to NiV that can be performed at BSL-2 conditions.
Like other paramyxoviruses, neutralizing antibodies to NiV are targeted to the fusion (F) and the attachment (G) glycoproteins (Bossart et al., 2002; Tamin et al., 2002). This manuscript reports on the development of a novel assay to detect neutralizing antibodies to NiV. The target antigen is a vesicular stomatitis virus (VSV) pseudotype particle (pVSV) displaying NiV F and G (pVSV-NiV-F/G) and expressing a luciferase reporter gene. The inhibition of entry of the pseudotype particle by antibodies to the surface glycoproteins of NiV can be detected by measuring the level of luciferase activity following infection of Vero cells. Because no infectious NiV is used, this assay can be performed at BSL-2 and completed within 48 hours and the results compared favorably to those of a standard plaque reduction neutralization test (PRNT).
Vero cells were maintained in Dulbecco’s Modified Essential Medium (DMEM) containing 10% fetal calf serum (FCS), L-glutamine, penicillin, and streptomycin (GIBCO BRL). The production of pVSV-NiV-F/G was described previously (Aguilar et al., 2007; Negrete et al., 2005; Negrete et al., 2006). The experimental NiV vaccine based on recombinant canarypox (CP) expressing either the F (CP-F), or G (CP-G) proteins of NiV was described previously (Weingartl et al., 2006).
The details of experimental design, vaccination and sample collections schedule were described previously (Weingartl et al., 2006). Briefly, pigs were vaccinated intramuscularly with 108 plaque forming units (pfu) (high dose) or 107 pfu (low dose) of CP-F, CP-G, or a mixture of both (four animals per group) under BSL-3 conditions. The pigs were boosted 21 days post vaccination (dpv) and serum samples were collected from each animal before vaccination (day 0) and at 7, 14, 21, 28, 35 and 48 dpv, and the samples were heat-treated at 56°C for 45 minutes prior to the neutralization test.
pVSV-NiV-F/G or pVSV were diluted in PBS and added to duplicate wells of a flat-bottom 96 well tissue culture plate and incubated at 37°C for one hour. The plate was washed 5 times with PBS-Tween buffer (0.1M PBS with 0.05% Tween 20), and primary antibodies from humans or mice diluted 1:200 in blocking buffer (5% skim milk, 0.1M PBS, 0.05% Tween-20) were added to each well, and incubated at 37°C for one hour. The plates were washed 5 times with PBS-Tween buffer, and the corresponding secondary antibodies were added (goat anti-human-HRP, goat anti-mouse-HRP) and incubated at 37°C for one hour. The plates were washed 5 times with PBS-Tween buffer, 75 μl per well of TMB substrate was added and incubated at room temperature for 15 minutes. The reaction was stopped by adding 75 μl per well of 2M phosphoric acid. The absorbance was then read in an EIA reader at 450 nm (TECAN, Inc.).
The dilution of pseudotype virus used was calculated to produce 1500–2000 relative light units (RLUs) in control wells. Four fold serial dilutions of the serum samples were prepared in a 96 well plate beginning with an initial dilution of 1:20 in Minimal Essential Medium (MEM). An equal volume (55 μl) of pVSV-NiV-F/G was added to each well and the plates were incubated at 37°C for 45 minutes. One hundred microliters of the pseudotype particle-serum mixtures were transferred onto 96 well plates containing a monolayer of Vero cells. Plates were incubated at 37°C and 5% CO2 for 2 hours for adsorption after which the innocula were replaced with 150 μl of DMEM with 5% fetal bovine serum (FBS) and 1% penicillin-streptomycin (PS), and the plates were incubated at 37°C and 5% CO2 for 17 hours. After incubation, the medium was removed by aspiration, and the plates were washed twice with 200 μl phosphate-buffered saline (0.01M PBS) per well. Cell lysis and detection of Renilla luciferase were performed according to the protocol recommended by the manufacturer (Promega, Inc). Briefly, Renilla Luciferase Assay Lysis Buffer was thawed to room temperature; diluted 1:5 with ddH20, and 30 μl was added to each well. Plates were placed on an orbital shaker for 1–2 hrs with periodic mechanical manipulation, or by freeze-thawing 3X to ensure complete lysis of the cells. Following lysis, 25 μl of the cell lysates was transferred to a Fluoroskan Ascent FL Fluorometer (Thermo Corporation, Finland) which was pre-programmed to read for 10 sec with a 2 sec delay after 100 μl of Renilla Luciferase Assay Reagent was added to the well. Percent inhibition was calculated by dividing the average RLU number in triplicate wells with the test antiserum samples by the average RLU of triplicate wells infected but without antiserum (positive control). A signal decrease of 90% or greater was considered to be inhibition. The titer of the pVSV-NiV-F/G neutralization assay was the reciprocal of the highest dilution of serum showing 90% or higher inhibition. Serum samples that failed to show at least 90% neutralization at the 1:40 initial dilution were considered negative.
The method for PRNT using live NiV was described previously (Weingartl et al., 2006), and all procedures with live virus were performed under BSL-4 conditions. Briefly, PRNT titers were determined based on a 90% plaque reduction cut-off starting with a 1:20 dilution of serum.
RIP assays were carried out as described previously (Tamin et al., 1994) with the single exception that the 35S-methionine-labeled vv/NiV-F and vv/NiV-G infected cells were resuspended in RIPA buffer (1% sodium deoxycholate, 1% Triton X-100, 0.2% SDS, 150 mM NaCl, and 50 mM Tris-HCl at pH 7.5). The human serum samples used were from the recent outbreaks of NiV in Bangladesh (Hossain et al., 2008). Normal human serum and NiV specific hyperimmune mouse ascites fluid (HMAF) were used as negative and positive controls, respectively. The ratio of antibody to 35S-methionine labeled lysate was 1:40.
Infection of Vero cells with pVSV-NiV-F/G was inhibited by monospecific antiserum to NiV-F and NiV-G, HMAF prepared against NiV, or a human serum sample from a patient infected with NiV, all of which also neutralized NiV by standard PRNT (Table 1). Negative control serum samples included normal human and normal mouse serum and monospecific antiserum to NiV-N protein and measles virus F protein (Table 1). The results confirmed that pVSV-NiV-F/G particles had incorporated both the NiV F and G proteins and the infectivity of the particles was neutralized by serum samples that neutralized NiV in PRNT.
The pseudotype particles were able to serve as NiV-specific glycoprotein antigens in a standard enzyme immunoassay (Figure 1). HMAF prepared against NiV (Figure 1A), serum from mice vaccinated with vaccinia recombinants expressing NiV F and G (Figure 1B), and a human serum sample from a patient infected with NiV (Figure 1B) specifically recognized pVSV-NiV-F/G bound to the plastic plate. The second antibody-conjugate that was used to detect the mouse antibody gave lower OD readings than the second antibody that was used to detect the human antiserum. There was no background reactivity when pVSV particles were used as antigen (Figure 1A). Serum samples recognizing the surface glycoproteins of measles virus and normal human serum did not give a signal in the enzyme immunoassay format (Figure 1B). These results indicated that pVSV-NiV-F/G particles were antigenic and can serve as antigens in neutralization assays and enzyme immunoassays.
Representative serum samples from pigs vaccinated with CP-F, CP-G or CP-F and CP-G recognized NiV-F and NiV G in a manner that was consistent with the vaccination scheme (Figure 2). However, the NiV proteins were not detected in the RIP assays until 1 week after the booster vaccination on day 21 with the exception of the day 21 serum samples from pigs vaccinated with CP-F and CP-G (Figures 2 and and4).4). Serum samples from humans infected with NiV recognized both NiV-G and NiV-F (Figure 3A). These results confirmed that the serum samples to be evaluated in the neutralization assay included antibodies specific to the surface glycoproteins of NiV.
The human serum samples containing antibodies to NiV F and G completely inhibited the luciferase production in Vero cells infected with pVSV-NiV-F/G at dilutions of up to1:640 (Figure 3B). Control serum samples did not inhibit luciferase production at the starting dilution of 1:40. These data showed that the human serum samples were able to neutralize the infectivity of the pseudotype particles.
All of the serum samples from the pigs vaccinated with CP-F, CP-G or CP-F and CP-G were tested in both the standard PRNT and the assay using pVSV-NiV-F/G as an antigen (Figure 4). Both assays consistently detected neutralizing antibodies to NiV in the serum samples obtained after the booster vaccination on day 21, and the assays based on the pVSV-NiV-F/G demonstrated high sensitivity and specificity compared to standard PRNT (Figure 5). However, the sensitivity of the pVSV-NiV-F/G assay was decreased to approximately 50% with serum samples collected 7, 14 and 21 dpv. PRNT detected low levels of antibodies more efficiently than the pseudotype assay or RIP. Overall, the specificity of the pseudotype assay was extremely high (94–100%) and none of the pre-vaccination serum samples were positive in either assay (Figures 4–5). When samples were scored as positive or negative, there was a 96% correlation between the results of the pseudotype assay and PRNT with only 16 discordant results for 363 serum samples, 15 of which occurred for samples taken 7–21 dpv (Figure 5). There was nearly 100% correlation for all of the pre-bleed samples and for samples taken 28–49 dpv. Comparing the titers for the PRNT and pseudotype assay gave a coefficient of correlation of 0.51 for all samples collected on days 28–49. Serum samples from the pigs receiving the high dose of the vaccines had an overall correlation coefficient of 0.62 and 0.68 for the samples from days 28 and 35 post vaccination, respectively.
Several strategies have been employed to develop new diagnostic techniques to detect infections with the henipaviruses with an emphasis on assays that do not require BSL-4 containment. Expression systems were used to produce antigens for enzyme immunoassays including the NiV N protein (Chen et al., 2006; Yu et al., 2006), a truncated phosphoprotein antigen (Chen et al., 2007), or the glycoproteins of NiV (Eshaghi et al., 2005a; Eshaghi et al., 2004; Eshaghi et al., 2005b). Other assays described the use of monoclonal antibodies for immunohistochemical based diagnosis (Tanimura et al., 2004; Xiao et al., 2008), or a modified plaque assay in which the plates are prepared in BSL-4 and inactivated by gamma irradiation before staining at BSL-2 (Crameri et al., 2002). Soluble G proteins and/or ephrin B2 coupled with a bioplex protein array have been used to measure antibody responses to the glycoproteins of henipaviruses (Bossart et al., 2007).
The pVSV-NiV-F/G was initially developed to support studies to identify the cellular receptors for the henipaviruses and has recently been used as antigen for a neutralization test (Wang et al., 2006). A similar approach using pseudotype particles or reporter virus particles (RVP) was used in a neutralization assay for SARS (Temperton et al., 2005) and West Nile virus (Pierson et al., 2006). This report describes the first comparison between the neutralization assay using pVSV-NiV-F/G and PRNT. The VSV-NiV-F/G neutralization assay offers several significant advantages over PRNT. The assay can be performed entirely at BSL-2, completed in less than 48 hours, and relies on automated read-outs in a 96-well format instead of manual counting of plaques. Overall, the results obtained by the assay using pVSV-NiV-F/G were consistent and correlated well with the results from PRNT. Both methods (Figure 4) confirmed previous reports of the ability of NiV G to induce higher neutralizing titers than NiV-F (Bossart et al., 2005; Guillaume et al., 2006; Mungall et al., 2006; Tamin et al., 2002). Also, the pseudotype particles were immunogenic when used as an antigen in an indirect enzyme immunoassay format suggesting that this may be a promising technique for developing glycoprotein specific enzyme immunoassays for NiV and other paramyxoviruses.
There is considerable variation in all diagnostic methods, including PRNT, in detecting low levels of antibodies. In this study, the results of the PRNT were regarded as the gold standard. Fifteen of the 16 discordant results between PRNT and the pseudotype neutralization assay were recorded from serum samples collected 7, 14 and 21 dpv when antibody levels were very low. With only one exception, the PRNT titers did not exceed 40 for any of the samples that appeared to give false negative results in the pseudotype assay. Some of these low titers could have been due to toxic effects of the serum in the PRNT assay. All of the five apparent false positive results in the pseudotype assay were from the serum samples collected on day 21 and these had the titers of 40.
Henipaviruses are zoonotic viruses, and their natural reservoir is fruit bats of the genus Pteropus (Halpin et al., 2000; Chua et al., 2002). The geographical distribution of Pteropus fruit bats includes parts of West Africa, South Asia and Southeast Asia, the Pacific Islands, and the eastern coast of Australia. This vast area includes some of the most heavily populated areas in the world including many developing countries (Epstein et al., 2006). Unlike other paramyxoviruses, these viruses have the ability to infect a wide range of host species (Bellini et al., 2005). Therefore, there is always the potential for spillover of henipaviruses from the natural reservoir into humans or farm animals. Good laboratory tests, including both serologic and virus detection methods are needed to rapidly detect outbreaks of henipaviruses.
The standard first-line serologic test for henipaviruses has been enzyme immunoassay along with neutralization tests used for confirmation. The pseudotype neutralization assay described here could be used to replace PRNT as the confirmatory serologic test in areas lacking high containment facilities. Pseudotype particles could easily be prepared in advance and stored frozen until required. The local laboratory would need only cell culture capacity and a luminometer. This new novel, high-throughput assay could also be a valuable tool for performing serologic surveillance of various bat species and domestic animals. Finally, this assay would provide a convenient and safe way to measure protective antibodies induced by experimental vaccines against henipaviruses.
The authors wish to acknowledge financial support from NIH: R21 AI058038 (to JAR), R01 AI60694 (to BL), and T32 AI07323 (to MCW). MCW also acknowledges support from the UCLA Warsaw Fellowship.
Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.