The geographic distribution, and the biotic and abiotic predictors of select zoonotic pathogen detections in Canadian polar bears

Abstract

Increasing Arctic temperatures are facilitating the northward expansion of more southerly hosts, vectors, and pathogens, exposing naïve populations to pathogens not typical at northern latitudes. To understand such rapidly changing host-pathogen dynamics, we need sensitive and robust surveillance tools. Here, we use a novel multiplexed magnetic-capture and droplet digital PCR (ddPCR) tool to assess a sentinel Arctic species, the polar bear (Ursus maritimus; n = 68), for the presence of five zoonotic pathogens (Erysipelothrix rhusiopathiae, Francisella tularensis, Mycobacterium tuberculosis complex, Toxoplasma gondii and Trichinella spp.), and observe associations between pathogen presence and biotic and abiotic predictors. We made two novel detections: the first detection of a Mycobacterium tuberculosis complex member in Arctic wildlife and the first of E. rhusiopathiae in a polar bear. We found a prevalence of 37% for E. rhusiopathiae, 16% for F. tularensis, 29% for Mycobacterium tuberculosis complex, 18% for T. gondii, and 75% for Trichinella spp. We also identify associations with bear age (Trichinella spp.), harvest season (F. tularensis and MTBC), and human settlements (E. rhusiopathiae, F. tularensis, MTBC, and Trichinella spp.). We demonstrate that monitoring a sentinel species, the polar bear, could be a powerful tool in disease surveillance and highlight the need to better characterize pathogen distributions and diversity in the Arctic.

Figure 1

The harvest location of (A) all harvest tissue sets included in the analyses (n = 68), and the positive detections for (B) E. rhusiopathiae (n = 25), (C) F. tularensis (n = 11), (D) Mycobacterium tuberculosis complex (MTBC; n = 20), (E) T. gondii (n = 12), and (F) Trichinella spp. (n = 51). All maps are overlayed with the boundaries of the polar management units that fall within Canada. These maps were generated using free and open source QGIS software (v3.30.3; https://qgis.org).

Figure 2

Prevalence estimates of the five focal pathogens (E. rhusiopathiae, F. tularensis, MTBC, T. gondii, and Trichinella spp.) based on individual status, per ice ecoregion (Divergent, Convergent, Archipelago, and Seasonal ice ecoregions), as described by Amstrup et al., (2007). The average standard error of the mean difference of prevalence is 52.57, 31.26, 36.17, and 27.04 for the Divergent, Convergent, Archipelago, and Seasonal ice ecoregions, respectively. This map was generated using free and open source QGIS software (v3.30.3; https://qgis.org).

Figure 3

The conditional inference trees (partykit; E. rhusiopathiae and F. tularensis) and the generalized linear mixed-model decision trees (glmertree; MTBC and Trichinella spp.) resulting from the recursive partitioning analyses, the grey bars representing a negative detection and the black bars representing a positive detection for the respective pathogens.