Dynamics of viral index infections in novel hosts

Nardus Mollentze; Daniel G. Streicker; Pablo R. Murcia; Katie Hampson; Roman Biek

Emerging disease epidemics often result from a pathogen establishing transmission in a novel host 27 species. However, most cross-species transmissions fail to establish in the newly infected species 28 for reasons that remain poorly understood. Examining cross-species inoculations involving rabies, 29 a widespread viral zoonosis, we show that mismatches in virulence, which are predictable from 30 host and viral factors, make sustained transmission in the novel host less likely. In particular, 31 disease progression was accelerated and virus excretion decreased when the reservoir and novel 32 host were physiologically or genetically more dissimilar. These mechanistic insights help to 33 explain and predict host shift events and highlight meta-analyses of existing experimental 34 inoculation data as a powerful and generalisable approach for understanding the dynamics of index 35 infections in novel species. Features inherent to the virus lineage involved (virus effects), irrespective of the infected 91 host, likely due to adaptation of individual lineages to reservoir host species. For example, 92 incubation periods were shortened by both increased dose and by inoculation routes which were 125 closer to the brain, although the effect size estimate for inoculation distance included zero ( Figure 126 2A). More importantly, the duration of incubation periods was also influenced by features of the 127 virus as well as its interaction with the inoculated host. Specifically, significant differences in 128 incubation period duration were associated with reservoir type (bat vs. carnivore) and with body 129 temperature differences between source and inoculated hosts (Figure 2A ). Both effects depended 130 on viral dose. At low doses, viruses from bat reservoirs were associated with shorter incubation 131 periods compared to viruses from carnivores, though this effect diminished at higher doses ( Figure 132 2B). A similar effect was seen when the inoculated species was a known rabies reservoir ( Figure 133 2C) although the 95% highest posterior density interval (HPD) of this effect size included zero 134 ( Figure 2A ). The difference in typical body temperature between the virus reservoir and the 135 inoculated species had a more marked effect on incubation period durations ( Figure 2A ). The onset 136 of symptoms was delayed when the virus was inoculated into species with a warmer body 137 temperature than its reservoir (negative values in Figure 2D ), although this delay reduced with 138 higher doses. The opposite was also true -in hosts with lower body temperatures than the virus 139 reservoir, the incubation period tended to be shorter ( Figure 2D , Supplementary figure S 1). Models 140 fitting effects for inoculated and reservoir species body temperature separately allowed us to 141 explore this temperature effect further. Inoculated species with higher typical body temperatures 142 tended to have longer incubation periods (Supplementary figure S 2). However, viruses from 143 reservoirs with higher body temperatures were associated with shorter incubation periods, 144 suggesting that these viruses had adapted to counteract any losses in efficiency caused by the body 145 temperature of their reservoir host. Importantly, there was no correlation between phylogenetic 146 distance and body temperature difference among species (Supplementary figure S 3), indicating 147 that the observed temperature effects were not explainable by the level of t