Disclaimer: Early release articles are not considered as final versions. Any changes will be reflected in the online version in the month the article is officially released.
Author affiliation: The University of Tokyo, Bunkyo-ku, Tokyo, Japan (M. Tominaga, A. Hyodo, M. Onodera, H. Yang, Y. Kitai, Y. Akahori, H. Katoh, M. Takeda); Keio University Human Biology-Microbiome-Quantum Research Center, Tokyo (S. Kojima); Okayama University Medical School, Okayama, Japan (M. Onodera); Japan Institute of Health Security, Musashimurayama, Japan (F. Seki)
During 2006–2008, canine distemper virus (CDV) disease outbreaks occurred among macaques (Macaca spp.) in China and Japan (1–3). Despite efficient transmission and high illness and death rates, overt neurologic disease was uncommon, although infection of the central nervous system and histopathologic lesions occurred in some animals (1–3). Wild-type CDV from dogs can use both canine and macaque signaling lymphocytic activation molecule (SLAM) to enter cells; however, human SLAM is used inefficiently (1,4).
SLAM V domains in macaques and humans differ by only 2 amino acids. At position 28, macaque SLAM contains histidine, whereas in other primates, including humans, position 28 contains arginine; that amino acid residue determines whether typical wild-type CDV can use macaque or human SLAM (4,5).
Hemagglutinin (HA) protein substitutions R519S, D540G, and P541S in CDV enhanced SLAM-dependent CDV binding to Vero cells that overexpressed human SLAM (6,7). However, whether those substitutions enable efficient SLAM usage in primary human immune cells remains unknown. Host switching by morbilliviruses, including CDV, requires more than SLAM adaptation. Morbilliviruses must evade or suppress innate immunity and efficiently use nectin-4 for epithelial spread and shedding (8). Moreover, morbillivirus infections generally induce long-lasting immunity, and infection or vaccination can provide some cross-protection against related morbilliviruses, creating additional barriers to CDV establishment in humans.
During 2018–2020, multiple lethal CDV disease outbreaks occurred in free-ranging marmosets (Callithrix spp.) in Brazil (9–12). Unlike measles virus infections in humans, CDV infections in carnivores are associated with a high risk for acute encephalitis. In CDV-infected marmosets, marked virus localization was observed not only in systemic lymphoid and epithelial organs but also within the central nervous system (9–12), which is believed to account for the high pathogenicity and mortality rates observed in that species. Phylogenetic analysis indicated independent introduction of >2 distinct virus strains (11). In one 2018 outbreak, 5 free-living black-tufted marmosets (C. penicillata) died in urban Central Brazil near neighborhoods with unvaccinated domestic dogs, suggesting possible transmission from those dogs (11).
Using mammal cell expression systems, we analyzed SLAM receptor usage by marmoset- and dog-derived CDV strains in different animal species. Furthermore, we evaluated whether adaptive mutations in those CDV strains affected human SLAM usage. We did not require ethics approval for this study because no experiments involving humans or animals were performed.
Figure 1
Figure 1. Expression of SLAM in Vero cells in study of natural occurrence of primate SLAM-adapted canine distemper virus, Brazil, 2018. We assessed the expression of marmoset, canine, human, and macaque SLAM…
We reconstructed HA genes by using the genomic sequences from the CDV NHP1 strain isolated from a marmoset (GenBank accession no. PP847352.1) and the Ac96I strain isolated from a dog (accession no. AB753775). We introduced mutations in the HA genes that produced amino acid substitutions at positions 519 and 530 and evaluated functional characteristics of those HA proteins in mammal cell expression systems. Vero cell lines constitutively expressing canine, macaque, and human SLAM have been described previously (4), whereas we established the Vero cell line expressing marmoset SLAM in this study by using the same method (Figure 1). We transfected those Vero cell lines with plasmids expressing both the HA (with and without mutations) and fusion (F) proteins to induce receptor-mediated cell–cell fusion. For direct comparison of receptor use capacities of the different HA protein variants, we used the F protein derived from the Ac96I strain throughout the assays. That method enabled assessment of HA protein–mediated receptor use capacities of canine, marmoset, macaque, and human SLAM, as described previously (1,4).
Figure 2
Figure 2. Cell fusion assay in study of natural occurrence of primate SLAM-adapted canine distemper virus, Brazil, 2018. Vero cells constitutively expressing marmoset (GenBank accession no. XM_002760176), canine (accession no. AF325357), human…
Figure 3
Figure 3. Phylogenetic analysis of marmoset CDV HA protein in study of natural occurrence of primate SLAM-adapted CDV, Brazil, 2018. A) Phylogenetic tree of CDV HA proteins. Alignment was performed by using…
Functional assays revealed that the HA protein from the marmoset-derived CDV NHP1 strain not only recognized canine SLAM but also had adapted to bind to marmoset SLAM, unlike the conventional dog-derived Ac96I strain (Figure 2, panels A, B). We observed a similar overall SLAM use pattern when the homologous F protein from strain NHP1 was used. Furthermore, detailed molecular phylogenetic analysis demonstrated that the marmoset-derived NHP1 strain belongs to the Europe 1/South America 1 lineage, which is distinct from the lineages containing CDV strains that previously caused outbreaks in macaques (Figure 3, panel A). Eleven amino acid positions showed marked divergence among CDVs from marmosets and other species (Figure 3, panel B). The marmoset isolates harbored unique amino acid mutations at or adjacent to the specific residues on the HA protein that we previously identified as critical for human SLAM adaptation (6) (Figure 3, panel C).
We focused on 2 specific amino acid substitutions at positions 519 (R519I) and 530 (G530S) and analyzed whether those unique mutations found in the marmoset-derived CDV strain were responsible for their adaptation to marmoset SLAM. We demonstrated that R519I, but not G530S, marginally enhanced the use of marmoset SLAM; however, co-existence of both mutations further increased marmoset SLAM usage (Figure 2, panel A). Conversely, introduction of either the I519R or S530G mutation in the HA protein from the NHP1 strain caused a lost or greatly reduced ability to use marmoset SLAM (Figure 2, panel B).
Figure 4
Figure 4. Quantitative analysis of SLAM-dependent cell–cell fusion in study of natural occurrence of primate SLAM-adapted canine distemper virus, Brazil, 2018. We quantified cell–cell fusion mediated by hemagglutinin (HA) proteins derived from…
To quantitatively assess differences in SLAM usage, we performed dual-split reporter protein–based cell–cell fusion assays, as described previously (4). Fusion activity for each HA protein was normalized to that obtained with canine SLAM, which was set to 100%. The quantitative results were consistent with microscopic observations. The R519I mutation of the HA protein enhanced marmoset SLAM usage in the Ac96I strain background; a further increase in usage occurred when the R519I mutation was combined with G530S. The I519R or S530G reverse substitutions in the NHP1 HA background reduced marmoset SLAM usage (Figure 4).
We investigated whether the R519I or G530S adaptive mutations affected human SLAM usage. The microscopic fusion assay showed those mutations affected human SLAM usage similar to that observed with marmoset SLAM (Figure 2, panel A). The quantitative fusion assay independently confirmed that pattern. The R519I mutation in the HA protein enhanced human SLAM usage in the Ac96I background, and combination with G530S further increased usage. The corresponding I519R or S530G reverse substitutions in the HA protein from the NHP1 strain reduced human SLAM usage (Figure 4). Our findings indicate that CDV adaptation to marmoset SLAM can occur under natural conditions and that this adaptation can concurrently enhance the ability to use human SLAM.
The NHP1 strain can be characterized as a primate SLAM-adapted CDV, because the HA protein efficiently uses SLAM from macaque, marmoset, and human cells. The adaptations did not reduce the ability of the HA protein to use canine SLAM (Figure 2, panel B), which indicates no apparent trade-off in SLAM usage was associated with adaptation to primate SLAMs.
Recently, CDV has rapidly expanded its host range to include diverse noncanine species (13), including primates. Although CDV can cause severe neurologic disease in susceptible animal hosts, the clinical consequences of potential CDV infection in humans remain unknown. Animal studies have suggested that existing immunity to measles virus might provide some degree of cross-protection against CDV (14,15). Therefore, our findings should not be interpreted as evidence that CDV currently poses an imminent risk for neurologic disease in humans. Continued surveillance of circulating CDV strains will be critical for monitoring virus evolution and cross-species transmission. Maintaining high measles vaccination coverage might also be critical, especially in regions where SLAM-adapted CDV strains circulate, because of the partial cross-protection against CDV demonstrated by measles immunity in animal studies (14,15).
Ms. Tominaga is a researcher at the Graduate School of Medicine and Faculty of Medicine at The University of Tokyo, Japan. Her research interests focus on acute viral infections in humans and animals.