We examined broad-scale descriptors of coastal watersheds, local marine habitat characteristics and individual animal diet preferences, and found that each is correlated with sea otter S. neurona seroprevalence risk. The highest S. neurona seroprevalence risk was found in animals living in habitats with low kelp cover and adjacent to developed lands, areas of denser human settlement, cropping land and wetlands. Within California, the relationship of S. neurona prevalence with human population density and land use remained statistically significant even after accounting for variation in marine habitat/substrate, sex, and diet.
In agreement with earlier research on the California sea otter population8, male animals and older age classes demonstrated markedly higher seroprevalence, as did animals with at least 10% clams in their diet. Also re-affirmed in this larger dataset was the association between soft-sediment marine habitat and high S. neurona prevalence8. All high seroprevalence areas in the present study were characterised by soft sediment habitat, and many were located near the mouths of large rivers or estuaries. Conversely animals using kelp-dominated habitat with a preponderance of rocky substrate at Big Sur, Santa Barbara Channel exhibited average or below-average seroprevalence (Table S2). Outside the period from 2002–2005 (including the 2004 outbreak – Table S2), the rocky Monterey Peninsula area also exhibited low seroprevalence. Consumption of high-risk prey (clams) that predominate in soft sediment habitat most likely explains this relationship, though terrestrial factors could also be synergistic. Furthermore, within California, low S. neurona seroprevalence was observed among sea otters living in areas with high kelp cover. This association is likely not due to a protective effect of kelp, but rather as an indicator of substrate/habitat type – high kelp cover does not occur in predominantly soft substrate areas. These findings suggest a role for clams or other prey living in soft sediment habitats in concentrating S. neurona sporocysts, thereby increasing the likelihood of ingestion by a sea otter.
Invertebrate bioconcentration of S. neurona sporocysts by clams or other bivalves has biological precedent. An analogous but distinct process is believed to occur with T. gondii oocysts, mediated by aggregation in marine polymers, adhesion to kelp blades and ingestion by turban snails (Tegula spp.)28,29,30. Our findings suggest that a similar process might occur with this related apicomplexan parasite, but in soft sediment habitats, rather than in kelp forests, and with a resident bivalve rather than a gastropod. This difference may be attributable to differences in the surface properties, aggregation and movement of the S. neurona sporocysts compared to T. gondii oocysts in estuarine and marine environments, or may simply reflect the differing features of locations where large numbers of the respective species of protozoa are discharged into the ocean. Further study of the surface properties of sporocysts in fresh, brackish and saline water and more detailed spatial modeling, including both terrestrial hosts and sea otters, will be needed to answer this question. Further insights into exposure mechanisms may come from studies of sea otter foraging behaviour in different habitats.
The observed association of high S. neurona seroprevalence in animals living adjacent to watersheds with high wetland area was consistent even after accounting for diet and marine habitat type. Wetlands are complex ecosystems with many features that could both positively and negatively influence the risk of infectious S. neurona sporocysts reaching sea otters. Firstly, wetlands are potential habitat for the definitive host of S. neurona, the Virginia opossum. The observed association of increased S. neurona seroprevalence among sea otters living near watersheds with high wetland area may reflect high definitive host density in these watersheds. Available data on opossum densities and habitat associations in North America are not adequate to determine with confidence if this is the case. Most studies on habitat associations are from the eastern USA (Table S6), and generally suggest that opossums favor human-modified environments (including developed areas and cropland) at certain times of year due to availability of anthropogenic food sources, which increase winter survival31. A proclivity for wetland areas is not generally noted, though some sources have indicated a preference for low-lying moist environments32 and streams33. West of the Rocky Mountains, in coastal California in particular, a markedly different climate regime means that opossums in this region are not subject to the same physiological limitations on survival imposed by cold winters in the north-eastern USA or the upper Midwest31, so habitat associations may differ significantly from those observed in the native range. Opossums in California have been reported to favor habitats with greater human land use intensity, small habitat fragments and proximity to urban edge areas (Table S6). Reanalysis of data from Rejmanek et al.34 showed greater success trapping opossums near habitat with open water, near developed areas and poorer success near cropping land and open grassy areas (see Supplementary Analyses and Table S7).
In addition to being potential habitat for definitive hosts, previous experimental and modeling studies have indicated that wetlands may act as filters for some terrestrial-origin pathogens. Oocysts of T. gondii and other similar particles appear to be trapped and retained in sediment of vegetated wetlands where they are ultimately degraded and deactivated rather than reaching the ocean35,36. The practical importance of this process depends on many interacting factors, not least of which is the structure of the wetland and its hydrological context. Highly canalised wetlands and those with no vegetation, little water throughput or very rapid transit time are unlikely to play a significant role in filtering infectious particles. Wetland classification for the current study was based on widely available, but imperfect data, so some misclassification (such as former wetlands that have been drained or converted to agricultural uses) may have contributed to the observed association with wetlands.
Since the only known definitive host of S. neurona in North America is the Virginia opossum, it is logical to expect that infection will not occur in locations far from any opossum habitat. Though little data on precise distributions and population densities is available in the literature, this species likely inhabits all areas of California included in this study, with the exception of San Nicolas Island, as well as Washington’s Olympic Peninsula. Opossums are not known to live in any part of Alaska, nor in British Columbia, except in the lower Fraser Valley and on Hornby Island32. Opossums have not been reported to reside on the west coast of Vancouver Island, near sea otter habitat, but the population on Hornby Island is the closest, being adjacent to the east coast of Vancouver Island. Two opossum sightings were reported in Victoria, BC, at the south-eastern tip of Vancouver Island in 1992, but none since that time32. As expected, low S. neurona seroprevalence (<10%) was recorded at all Alaskan study sites, and otters there did not have titers higher than 1:320. In contrast, the seroprevalence at the northern-most of the two Vancouver Island study sites – Nuchatlitz Inlet (BC), was 13% (4/30), and all positive animals had titers of at least 1:2560. This finding, combined with a recent report of 2 dead stranded sea otters from Vancouver Island with positive S. neurona PCR results11, indicates that infection with this pathogen is occurring in sea otters in the Vancouver Island area. The possibility of cross-reaction of the IFAT with other little-known Sarcocystis species occurring in the marine environment exists, but reported sequence data from the two PCR-positive stranded sea otters confirmed the presence of S. neurona in this region11. Although we cannot exclude the possibility that sea otters were infected in WA and moved to BC, such movements are rare, and infected animals were detected only in the northernmost, more remote site in BC. Therefore, the most likely explanation for our findings are that (i) opossums have colonised remote parts of Vancouver Island but remain undetected, (ii) long-distance transport of viable sporocysts is occurring (either directly or within a transport host), or (iii) a hitherto unknown (possibly marine) definitive host of S. neurona exists.
Long-distance transport of S. neurona sporocysts has not been described, but the co-occurrence of high apparent seroprevalence in Washington (both in the present study and previously reported marine mammal cases) and a predominantly westward estuarine flow out the strait of Juan de Fuca, then northwest along the coast of Vancouver Island are consistent with this possibility37. Two further features of this parasite necessary to render this transport process possible also exist – S. neurona sporocysts are highly resilient in the environment38 and a possible concentration mechanism has been identified in the form of bivalve filtration. Though individual-level diet data were not available for Vancouver Island sea otters, many of the animals at this location were observed to forage on the outer coast (dominated by rocky substrate) during the summer, but shift to more sheltered bays and inlets during winter. These areas are dominated by soft sediment habitat where clams are a common prey item39.
The seasonality of S. neurona infection risk is a factor that was not explored in this study, principally because it is not possible to determine how long ago a seropositive animal encountered the pathogen that elicited the antibody response. In California, a distinct pattern of S. neurona mortalities occurring in late spring and early summer has been identified8, and a strong temporal correlation has been identified between runoff events and S. neurona associated-mortalities of sea otters living in Estero Bay6. Given the seasonal nature of precipitation and river flows in California, Shapiro et al.6 attribute this seasonality in part to a ‘first flush’ effect. Sporocyst shedding by opossums also appears to peak in the spring34. Furthermore, seasonal variation in habitat use by opossums may exist, and habitat use in winter and spring is likely most relevant to sea otter infection risk. Future research should also focus on validating the indirect fluorescent antibody test (IFAT) in sea otters, establishing the relationship between S. neurona infection status and pathology and quantifying its population-level impact.
Studies at the intersection of ecology and epidemiology are enabling new uses of ecological data to clarify patterns of disease transmission, and the effects of ecosystem processes on infectious disease distribution. The sea otter has been the source of unique scientific insights into how pathogen pollutants enter the marine environment and move through marine food webs due to certain unique characteristics. Chief among these are its nearshore habitat, small home range, high trophic level and metabolic rate, and notably its tractability for observational studies – especially the fact that its unique feeding behavior allows direct quantification of individual diets40,41. In this study we have elucidated not only specific risk factors for exposure to one parasite, but also contributed to a more nuanced understanding of the physical processes underlying marine pathogen pollution. Our results highlight the importance of ocean physical processes in shaping spatial patterns of parasite infection risk, as well as the importance of long-term large-scale sampling in understanding these patterns. We have also shown that S. neurona occurs in areas outside its previously known range, raising important questions about mechanisms of parasite transmission to sea otters. Future work should seek to determine whether observed relationships with human population density, wetlands and farming are related to opossum population density or watershed modification; whether clams have a unique capability to concentrate S. neurona sporocysts; and, perhaps above all, whether a viable physical and biological transmission pathway exists (at least in British Columbia) on a scale not previously imagined.
Source: Ecology - nature.com
