To examine whether mammals are naturally infected with Oz virus, we collected serum samples from 24 hunters and 240 wild animals (40 Japanese macaques [Macaca fuscata], 124 wild boars [Sus scrofa], and 76 sika deer [Cervus nippon]) captured in Yamaguchi prefecture, Japan, during 2013–2019. Because Yamaguchi prefecture is close to Ehime prefecture and the environment in Yamaguchi is very similar to that in Ehime, we used stocked samples in Yamaguchi prefecture for the first surveillance of Oz virus infection. To test for the presence of Oz virus antibodies in the serum samples, we performed a PRNT₈₀ (80% plaque-reduction neutralizing test) using Oz virus (Table 1). Among humans, 8.3% of the serum samples had Oz virus neutralization (VN) antibodies; VN titers were 1:40 and 1:80. In wild animals, serum from 47.5% of macaques, 60.5% of wild boars, and 73.7% of sika deer in Yamaguchi had Oz virus VN antibodies (Table 1).

We applied ELISA protocol used for serosurveillance of many infectious diseases (11–13) to detect Oz virus antibodies in the serum samples from wild animals. We extracted proteins from Oz virus or mock-infected Vero cells and used the extracts as ELISA antigens. To prepare the primary antibody, we diluted serum samples 1:100 in phosphate-buffered saline containing 0.05% Tween 20 and 0.4% Block Ace. We used peroxidase conjugated recombinant protein A/G (Thermo Fisher Scientific, https://www.thermofisher.com) as the secondary antibody and KPL ABTS peroxidase substrate (SeraCare Life Sciences, https://www.seracare.com) as the detection reagent. We measured absorbance using a spectrophotometer (Bio-Rad Laboratories, https://www.bio-rad.com) with a 405 nm filter and subtracted the value of the corresponding control mock-infected cells from all values.

Figure 1

Figure 1. Dot plot comparison between VN test and ELISA against Oz virus in serum samples from wild animals in Yamguchi prefecture, Japan. A) Macaques (n = 40); B) wild boar (n...

To determine ELISA cutoff values, we tested serum samples from the 40 macaques, 124 wild boars, and 76 sika deer captured in Yamaguchi prefecture. We compared the optical density values of the ELISA to the results of the VN test by 2-graph receiver-operating characteristic (ROC) curve analysis as described elsewhere (14). In macaques, the correlation coefficient between the ELISA and VN test was 0.9163, and an ELISA cutoff value of 0.2245 produced 100% sensitivity and specificity. In wild boars, the correlation coefficient was 0.8807, with 88.0% sensitivity and 89.8% specificity at an ELISA cutoff value of 0.1965. In sika deer, the correlation coefficient was 0.7569, sensitivity 78.6%, and specificity 80.0% at an ELISA cutoff value of 0.3165 (Figure 1).

Figure 2

Figure 2. Collection sites of serum samples from macaques, wild boars, and sika deer for study of Oz virus seroprevalence in Japan. Gray shading indicates prefectures in which samples were collected; black...

Next, we surveyed Oz virus infection among macaques, wild boars, and sika deer in many prefectures in Japan using the established ELISA (Table 2; Figure 2). Among 197 macaques captured during 2007–2019, seropositivity rates were 47.5% in Yamaguchi, 33.3% in Wakayama, and 6.3% in Mie prefectures. Among 879 wild boars captured during 2007–2014, seropositivity rates were 10.3% in Oita, 55.8% in Yamaguchi, 34.8% in Wakayama, 10.5% in Gifu, and 0% in both Toyama and Tochigi prefectures. Among 450 sika deer, seropositivity rates were 37.8% in Yamaguchi, 11.1% in Wakayama, 8.3% in Gifu, and 30% in Chiba prefectures.

First, we applied PRNT₈₀ to detect Oz virus antibodies in humans and wild animals in Yamaguchi prefecture. The results showed that 60.5% of wild boars, a major host of A. testudinarium, and 73.7% of sika deer in Yamaguchi prefecture during 2013 and 2015 had Oz virus VN antibodies, indicating that the virus was infecting wild animals in the western part of Japan. Next, we examined wild macaques for Oz virus infection; 48% were infected, and the antibody titers were high. In addition, 2 persons who hunted wild boars and sika deer in Yamaguchi prefecture had Oz virus antibodies. These results indicate that humans and macaques are also exposed to Oz virus.

We compared results from an ELISA, established for this study, using an Oz virus–infected cell extract for the surveillance of Oz virus infection among many mammalians, with those from the VN test to determine correlation between the 2 tests. The correlation coefficient was 0.9163 for macaques, 0.8807 for wild boars, and 0.7569 for sika deer, suggesting that the ELISA is effective for serosurveillance of Oz virus infection in samples from many animal species. However, because its sensitivity and specificity differed among animal species and values were lower for sika deer, in particular, cutoff values should be determined for each animal species. In addition, VN testing should be performed to confirm the presence of Oz virus antibodies.

Our nationwide surveillance of Oz virus infection in Japan indicated that many wild animals were positive for Oz virus antibodies. However, wild boars in Toyama and Tochigi prefectures did not have Oz virus antibodies, suggesting that the virus might not be distributed in the northern and eastern parts of Japan. A. testudinarium is the major tick species that infests humans in the southern and western parts of Japan (15), and because we found Oz virus mainly in those areas, it appears that the distribution of Oz virus–infected animals correlates with the habitat of the tick. In addition, because 2 hunters in Yamaguchi prefecture tested positive for Oz virus antibodies, further investigation is needed to determine whether Oz virus might be a zoonotic pathogen, especially because intracerebral inoculation of the virus in suckling mice causes lethal disease (10).