Human actions commonly alter wildlife populations. A classic example of an alteration is hunting, which often has density and demographic effects on a population (Milner-Gulland et al. 2003; Whitmanet al. 2004; Packer et al. 2009; Treves 2009). However, the consequences of these actions on pathogen transmission and evolution are largely unknown, and the few available studies report contradictory findings. Theory predicts that for density-dependent pathogens hunting-induced changes to density should decrease transmission rates yet make little difference to transmission dynamics for frequency-dependent pathogens. However, in practice models suggest that reducing host density can decrease (Lloyd-Smith et al. 2005; Potapov et al. 2012) or even increase pathogen transmission and prevalence (Choisy & Rohani 2006; Beeton & McCallum 2011). The complex interplay between host density, demography and behavior also makes predicting the impacts of hunting on pathogen dynamics complex. Limited empirical work shows that population reduction can increase pathogen prevalence via a ‘perturbation effect’ (Woodroffe et al. 2004, 2006; Carter et al. 2007; Carr et al. 2019). For example, culling induced changes or ‘perturbations’ to badger (Meles meles ) territorial behavior was considered a driver of increased bovine tuberculosis transmission between badgers and cattle (e.g., Woodroffeet al. 2006). However, there is also evidence that population reduction has little impact on canine rabies (Morters et al.2013) or Tasmanian devil facial tumor disease (Lachish et al.2010) dynamics. Recent advances in high-resolution pathogen sequencing and analytic approaches can now elucidate patterns of pathogen transmission and evolution (Smith et al. 2015; Didelot et al. 2017; Grubaugh et al. 2019) that were previously out of reach. Here we address the effects of hunting on pathogen dynamics by capitalizing on pathogen sequences collected from a detailed study on the demographic effects of hunting (Logan & Runge 2020b) as well as from sequences obtained over the same time period in a region where little hunting occurred. Our approach enables us to provide insights into the cascading consequences of hunting on pathogen-host dynamics.
RNA viruses are ideal agents for examining the effect of hunting on pathogen transmission and evolution. Genomic variation rapidly accrues in RNA viruses, enabling estimation of fine-scale epidemiological processes (such as transmission between hosts) and the basic reproduction number R0 (the average number of secondary cases per infection, Biek et al. 2015; Didelot et al.2017). Altered transmission dynamics and the arrival of new lineages can imprint distinctive evolutionary signatures on RNA viruses as they adapt quickly to changes in host populations they encounter (Woolhouseet al. 2014; Pybus et al. 2015). For example, if the cessation of hunting led to a higher frequency of transmission events, we expect that the transmission bottleneck would lead to high purifying selection since within-host mutations are lost with transmission (e.g., Pybus & Rambaut 2009). Conversely, if new mutations entering the host population allow the pathogen to escape immune detection, we may expect an increase in diversifying selection. Altered transmission dynamics and new lineages will also shape the phylogenetic diversity of the pathogen (Fountain-Jones et al. 2018). For example, if novel pathogen lineages are frequently arriving into a host population with limited transmission, we would expect to see a pattern of phylogenetic dispersion (i.e., higher phylogenetic diversity than expected by chance (Webb 2000)). In contrast, phylogenetic clustering (i.e., lower phylogenetic diversity than expected by chance (Webb 2000)), may be a marker of increased transmission events within a population.
Here we leverage viral data collected from closely monitored puma (Puma concolor ) in two areas in Colorado during the same time period: a ‘treatment region’ in which hunting pressure changed over time and a ‘stable management region’ acting as a control (hereafter ‘stable region’). We sequenced viral genes sampled from captured puma for an endemic RNA retrovirus, puma feline immunodeficiency virus (FIVpco), which is a host-specific pathogen considered relatively benign and not associated with overt disease outcomes (Bieket al. 2003). Even though FIVpco is endemic in puma populations, novel infections can spread in susceptible and previously infected individuals (Malmberg et al. 2019). Evidence suggests FIVpco is transmitted via aggressive interactions in most instances, although vertical transmission is also possible (Biek et al. 2003; Fountain-Jones et al. 2017). We analyzed these viral data in both regions using a transmission network approach (Didelot et al. 2017; Fountain-Jones et al. 2018) that incorporates a stochastic epidemiological model with pathogen genomic data to trace transmission between individual puma. The treatment region consisted of puma in a ~12000km2 area in western Colorado in which hunting prior to our study was common practice (see Logan & Runge 2020a). Hunting was excluded for a five-year period (2004-2009, “no-hunting period”) and reinstated for a further five years afterwards (2009-2014, “hunting period”). During the no-hunting period in the treatment region, the population of independent pumas (i.e., adults and sub-adults) increased from an estimated 23 (2005) to 57 (2009) individuals with much of this growth occurring 2007-2010 (Logan & Runge 2020a, i.e., after a two year lag 2004-6 hereafter “Lag 1”). Adult and sub-adult male survival was significantly higher in the no-hunting period (Logan & Runge 2020a). When hunting resumed, the overall population declined after a lag of two years (hereafter ‘Lag 2’). However, the decline in abundance and demographic effects on males were severe and rapid with males > 6 years old apparently eliminated from the population after two hunting seasons (Logan & Runge 2020b). In contrast, over the same period, the stable region in the Front Range of Colorado experienced minimal hunting pressures and no change in management practice. Nearly all the individuals sampled in both regions were adults and both sexes were evenly represented. While density was higher in the stable region (see Table S1), individual survival probabilities in the stable region were unaltered across years (Moss et al. 2016). By comparing the treatment and stable regions, we were able to test how demographic changes caused by hunting cessation and reinstatement perturb viral transmission networks and epidemiological parameters (e.g., R0), and also alter pathogen diversity and evolution. In doing so we begin to untangle the complex interplay between wildlife management and pathogen transmission, which is crucial for pathogen-orientated conservation and disease management strategies.