Figure 1

Figure 1. Clinical signs (arrows) in frogs with Elizabethkingia miricola infection in Hunan Province, China. A) Diseased frogs had neurologic signs of torticollis. B–D) Clinical signs with different appearances showing cataracts, proptosis, or...

Video

Video. Diseased frogs sampled at Yiyang, Hunan Province, China. The diseased frogs had symptoms including neurologic signs of torticollis, disorientation, and agitation or lethargy.

Since May 2016, many black-spotted frogs in farms in Hunan Province in south-central China have experienced an emerging, contagious disease characterized mainly by severe neurologic dysfunction. The first clinical sign is intermittent swimming in circles. Thereafter, the frogs develop signs of torticollis (Figure 1, panel A), disorientation (Video), and anepithymia or meteorism (Figure 1, panel E). These signs are followed by cataracts (Figure 1, panel C); proptosis or hyperemia (Figure 1, panels B, D); agitation or lethargy; and, ultimately, death. The frogs are farmed in artificial ecologic wetlands or ponds with running water and shelter (Technical Appendix Figure 1). Most ponds in 1 farm, which share a common water supply, were infected sequentially within a short time. More than 60% of the frogs in the infected farms had signs of varying appearance, and 60%–90% of the diseased frogs died in the next few days or weeks. The disease continued until hibernation and returned the following spring.

During July–October 2016, we collected 213 abnormal frogs from 7 separate farms in Hunan Province, China (Technical Appendix Figure 2). Histopathologic examination showed severe meningitis with denatured, incrassate meninges. We observed inflammatory infiltrates, moderate multifocal gliosis, and perivascular cuffing in the cerebellum (Technical Appendix Figure 3). Results of the diagnostic tests for Batrachochytrium dendrobatidis and ranaviruses were negative (Table 1). Although we observed Myxosporidia protozoa in the gallbladder and some protists in the intestine, they were not identified as the etiologic agents, considering the proportion of infection Technical Appendix Figure 4).

We confirmed bacterial infections in 190 (89.2%) of the 213 frogs; 90% were E. miricola according to the 16S rRNA gene sequence, which shared 99.36%–99.86% similarity with E. miricola DSM14571 (Technical Appendix). We selected bacterial strain FL160902, isolated from frog no. 160, as the representative isolate and conducted experimental pathogenicity testing by various infection routes, including intramuscular injection, immersion infection, and cohabitation with infected frogs. All animal handling was done in compliance with the National Institutes of Health protocols (Technical Appendix). After 2 weeks of observations (Table 2), we found that the cumulative mortality (10%–70%) increased with dose in the injection trial and that 100% of frogs exposed to E. miricola by immersion died. In the cohabitation studies, 30% mortality was recorded, indicating cross-infection. Koch’s postulates were satisfied by identification of isolates from dead frogs as E. miricola, identical to FL160902.

Figure 2

Figure 2. Maximum-likelihood phylogenetic tree of Elizabethkingia miricola FL160902 from an infected frog in Hunan Province, China, and reference genomes. The tree was constructed by using the single-copy orthologous genes of all the...

To characterize E. miricola FL160902, we conducted WGS with the Illumina HiSeq 2500 platform (Illumina Inc., San Diego, CA, USA), producing 2 × 150-bp paired-end reads. We assembled the trimmed reads using SOAPdenovo (http://soap.genomics.org.cn/soapdenovo.html). We constructed a phylogenetic tree (Figure 2) of orthologous genes using RAxML (9) with 100 bootstrap replicates to examine the evolutionary relatedness between E. miricola FL160902 (GenBank accession no. NHPR00000000) and other Elizabethkingia genomes. The results showed that FL160902 was most closely related to CSID_3000517120, a clinical isolate of E. miricola from the United States sequenced by the Centers for Disease Control and Prevention (CDC) (10), revealing the potential of E. miricola FL160902 for pathogenicity in humans.

Before WGS was commonly used, E. meningoseptica (previously Flavobacterium meningosepticum) was found to be separated into 2 main hybridization groups, UBI and UBII, that were ≈40%–55% interrelated; UBII could be further divided into 4 subgroups (11,12). However, because the isolates from different groups are phenotypically very similar, these genomic groups remain assigned at this time to E. meningoseptica (13). In our phylogenetic tree, UBI group E. meningoseptica isolates did not group with the other Elizabethkingia spp. and were distantly related to UBII. Considering the low DNA–DNA relatedness (<70%) between the 2 groups and phylogenomic analysis based on WGS (3,11,12), we propose that UBII are not E. meningoseptica. The UBII subgroups branching separately supports the view that they are different Elizabethkingia species (3). The UBII:1 group species E. anophelis and E. endophytica formed a clade with strong support of 100%, favoring the suggestion that E. endophytica is a later subjective synonym of E. anophelis (14). Our FL160902 isolate grouped with E. miricola, which is thought to be closely related to UBII:2 (3,10). The taxonomic status of E. miricola ATCC 33958 and BM10 should be reconsidered because they clustered with UBII:3 and not with UBII:2 E. miricola species. Our results agree with Eriksen’s conclusion about the genetic diversity in Elizabethkingia; a more comprehensive taxonomic system is needed to clarify the Elizabethkingia genus (3).

In this natural outbreak of meningitis-like disease in cultured frogs in Hunan Province, China, in 2016, E. miricola was the most predominant pathogen. The neurologic signs and pathologic brain lesions suggested that E. miricola could break through the blood–brain barrier and damage the nervous system. The etiologic analyses combined with the results of experimental challenge support the conclusion that the E. miricola strain represented by isolate FL160902 is highly contagious for frogs, especially by immersion infection. We suspect that contaminated water is the primary vehicle of transmission, considering the infection assay and the epidemiology in 1 farm with different ponds. However, diverse transmission routes might be involved because there is no obvious interconnectivity among independent farms, which needs to be investigated further. Close attention should be paid to whether this disease affects the wild population of amphibians. Our results indicated the gradual expansion of its host and suggest that amphibians may serve as a reservoir for infection in humans. Black-spotted frog farming is a major aquaculture industry in south-central China; thus, animals and humans that have close contact with infected frogs should be continually monitored for emerging E. miricola infections, even though no human E. miricola infection cases were reported related to frog consumption or farming in Hunan in 2016. Our results demonstrate a contagious disease in frogs caused by E. miricola that poses a potential zoonotic threat to humans, generating a need for consideration of the role of Elizabethkingia bacteria in public health.