Serologic Evidence of Human Exposure to Bat-Borne Zoonotic Paramyxoviruses, Cambodia

Abstract

Fruit bats in the genus Pteropus are the natural reservoirs for zoonotic paramyxoviruses, notably henipaviruses and pararubulaviruses, which are found across Southeast Asia and Oceania. The genetic and antigenic diversity of viruses in both genera, and region specificity, are ill-defined, limiting health security measures aimed at minimizing spillover. For example, Nipah virus has been isolated from bats in the Battambang province of western Cambodia, and surveys suggest bat foraging behaviors occur in close proximity to human settlements. However, there have been no historical cases of Nipah virus in Cambodia. Here, we use a multiplex microsphere immunoassay to identify cryptic human exposure to selected henipaviruses and pararubulaviruses in Cambodia. Convalescent human sera from persons presenting with acute respiratory illness were screened to detect the presence or absence of antibodies reactive with attachment glycoprotein antigens from Nipah virus, Hendra virus, Cedar virus, and Ghana virus, and a hemagglutinin-neuraminidase antigen from Menangle virus. In this sero-survey, we detected antibodies that were specifically reactive with Cedar virus and Menangle virus, including one serum sample that neutralized a recombinant Cedar virus. Additionally, we detected a pattern of cross-reactivity with Hendra virus, Cedar virus, and Ghana virus, suggesting previous infection by an antigenically-related henipavirus. We did not detect high antibody reactivity with the NiV glycoprotein. Future studies should expand serological surveillance for these transboundary pathogens, including genetic surveillance to aid in henipavirus discovery, and focused biosurveillance where interfaces with livestock and humans occur.

Keywords: Cambodia; Cedar virus; Menangle virus; henipaviruses; multiplex serology; surveillance.

Figure A1

K-means and k-medoids clustering approaches identified distinct groups of sero-reactivity. (A) Cluster count optimization by elbow visualization method. (B) Validated optimized cluster count by average silhouette width. (C) Visualization of data along primary axes of principal components. (D) Same data along principal components with cluster boundaries visualized.

Figure A2

Validation of sero-reactive k-medoid clustering.

Figure A3

Detection of neutralizing antibodies in Cedar virus-binding antibody positive samples.

Figure A4

Probability density distributions observed from univariate mixture modeling for threshold analysis of anti-Ghana virus G antibodies. The curves identify the subpopulation with (green) higher MFI values as compared to the subpopulation with (red) lower MFI values. The distinction between these groups is presented as the threshold for seroreactivity.

Figure 1

Identification of sero-groups reactive to zoonotic paramyxovirus antigens. Human serum samples (n = 1469) were analyzed by k-means clustering. A two-component model was selected and sub-groups of cluster 1 (n = 17)- and cluster 2 (n = 1452)-containing samples are shown. Lines connect individual serum samples. MFI, median fluorescence intensity; G, henipavirus attachment glycoprotein; HN, pararubulavirus hemagglutinin-neuraminidase antigen; NiV, Nipah virus; HeV, Hendra virus; CedV, Cedar virus; GhV, Ghana virus; and MenV, Menangle virus.

Figure 2

Sero-reactivity profiles of human sera against selected zoonotic paramyxoviruses. Radar charts demonstrating human (n = 1469) sero-reactivity with five paramyxovirus G and HN antigens and a mock control protein. Radial axes represent each of the antigens used as the serological target for detection of antibodies. Scales are a continuous linear measurement of median fluorescence intensity that represent antibody levels. Connecting lines represent the individual clusters based on k-medoids clustering after principal component analysis to six components where cumulative explained variance is >70%.