Transient definitive host presence is sufficient to sustain avian schistosome populations

To control swimmer’s itch in northern Michigan inland lakes, one species of bird, the common merganser (Mergus merganser), has been relocated from several lakes since 2015. Relocation efforts are driven by a desire to reduce the prevalence of the swimmer’s itch-causing parasite Trichobilharzia stagnicolae. The intention of this state-sponsored control effort was to interrupt the life cycle of T. stagnicolae and reduce parasite egg contribution into the environment from summer resident mergansers such that infections of the intermediate snail host Stagnicola emarginata declined. Reduced snail infection prevalence was expected to greatly reduce abundance of the swimmer’s itch-causing cercarial stage of the parasite in water. With no official program in place to assess the success of this relocation effort, we sought to study the effectiveness and impact of the removal of a single definitive host from a location with high definitive host and parasite diversity. This was assessed through a comprehensive, lake-wide monitoring study measuring longitudinal changes in the abundance of three species of avian schistosome cercariae in four inland Michigan lakes. Environmental measurements were also taken at these lakes to understand how they can affect swimmer’s itch incidence. Results from this study demonstrate that the diversity of avian schistosomes at the study lakes would likely make targeting of a single species of swimmer’s itch-causing parasite meaningless from a swimmer’s itch control perspective. Our data also suggest that removal of the common merganser is not an effective control strategy for the T. stagnicolae parasite, likely due to parasite contributions of migratory birds in the fall and spring. This suggests that only minimal contact time between the definitive host and the lake ecosystem is required to contribute sufficient parasite numbers to maintain a thriving population of parasite species with a high host-specificity.


149
150 Evaluating the study lakes based on the total avian schistosome population revealed clear 151 distinctions between Glen Lake and Walloon Lake. Most notable, was that in both 2019 and 152 2020, avian schistosomes were detected in abundance at Walloon Lake three weeks earlier than 153 at Glen Lake. This observation was obvious at most of the Walloon Lake sites. Additionally, the 154 density of the avian schistosome population was found to be substantially higher in May and 155 June at Walloon Lake when compared to Glen Lake. Such distinctions were not as clear when 156 comparing North and South Lake Leelanau to each other, or to Glen or Walloon Lake (Figure 1; 157 Figure S1). 158 159 At Glen Lake, 70 sites were positive throughout 2019 for avian schistosomes compared to 90 160 positive sites in 2020. Only 10% (7) and 14% (13) of the positive sites at Glen Lake were 161 collected before the end of June in 2019 and 2020 respectively (Figure 1). This is contrasted by 162 Walloon Lake, at which 33% (96) of sites and 90 (36%) of sites had detectable avian 163 schistosomes prior to the end of June (Figure 1). A comparison of the mean estimates of 164 cercariae number per 25 litres of water from all positive sites sampled at Glen (77.16 +/-SD 165 465.77) and Walloon (2975.07 +/-SD 9347.1) Lakes prior to the end of June indicated that there 166 was a significant difference (p=0.02) despite large standard deviations. The same was true for the 167 comparison in 2020, where mean estimates of avian cercariae number at Glen Lake (39.7 +/-SD 168 176.48) and Walloon Lake (13518.97 +/-SD 47491.66) yielded a statistically significant 169 difference (p=0.04) as well (Figure 1). 170 171 The number of avian schistosome-positive sites prior to July 1 at North and South Lake Leelanau 172 were similar in 2019 with 7 (23%) and 5 (17%) of sites. In 2020, North Lake Leelanau remained 173 consistent with 5 (17%) positive sites before July 1, however, South Lake Leelanau increased the 174 number of positive sites to 14 (47%).
175 T. stagnicolae persists even at lakes with 3-year long control programs. 176 177 Underpinning the higher-level distinctions between the four study lakes is the fact that the avian 178 schistosome populations are comprised of very different species of avian schistosome. T. 179 stagnicolae consistently dominated the avian schistosome population at Glen Lake and North 180 Lake Leelanau in 2019 and 2020 despite sustained control programs for three years prior to 181 2019, which targeted the definitive merganser host (Figure 2, panels A and C). At Glen Lake T. 182 stagnicolae was first detected in the water on June 11, 2019, and its presence was detected 183 consistently until October 01, 2019. The frequency of detection was highly variable between 184 sites. At some sites (notably E10, H10, H9, I3, O11, and P3) T. stagnicolae was detected 185 consistently; however, at other sites, the parasite was rarely detected (i.e.: G8, M13) ( Figure 2). 186 187 Lake-wide, T. stagnicolae signal rose sharply starting in July 2019 and 2020 at Glen Lake. 188 However, at Walloon Lake and North Lake Leelanau, which both sustained consistent 189 populations of T. stagnicolae through 2019 and 2020, the T. stagnicolae signal is detected earlier 190 (late May-early June at Walloon Lake and early June at North Lake Leelanau). T. stagnicolae 191 DNA abundance maintained consistently high concentrations continuing through August 2019 192 and September 2019 at Glen Lake and North Lake Leelanau until sharply dropping (Figure 2; 282 infection prevalence was 0% in both years. No snails were found to host T. stagnicolae at South 283 Lake Leelanau in 2019 or 2020 (Table 1).
284 Common merganser relocation does not significantly impact T. stagnicolae cercariae 285 abundance. 286 287 The T. stagnicolae DNA abundance data for each lake in 2019 and 2020 (i.e.: total abundance 288 data, all sites and timepoints) was first graphed using box plots, and this visualization showed 289 slight differences between the lakes in both years, but the data had large variation, especially at 290 South and North Lake Leelanau and Walloon Lake. A single factor ANOVA indicated that there 291 were no significant differences between T. stagnicolae abundance at each of the four lakes in 292 both 2019 (p = 0.2) and 2020 (p = 0.9). Furthermore, when T. stagnicolae DNA abundance was 293 compared between each of the lakes in both years, there were no significant differences found. 294 295 The abundance of T. stagnicolae was compared between 2019 and 2020 at Glen Lake and North 296 Lake Leelanau, because in 2020 both lakes discontinued their merganser relocation programs to 297 control swimmer's itch. Whole lake (i.e.: total abundance across all sites and timepoints) was 298 compared in an unpaired two-samples Wilcox test. There was no significant difference between 299 T. stagnicolae abundance from 2019 to 2020 at North Lake Leelanau (p = 0.053). However, 300 there was a statistically significant decrease in T. stagnicolae abundance at Glen Lake between 301 2019 and 2020 (p = 0.04). Once again, after hatch-year snails collected in 2020 would reflect 302 trapping and relocation efforts in 2019 since the parasites overwinter in the snail hosts.
303 The effect of environmental variables on avian schistosomes 304 The effect of environmental variables on total cercariae concentrations was assessed using a 305 Hurdle Model. Hurdle models contain two parts, one component (the hurdle part) contains 306 factors that predict zeros (i.e.: no cercariae present in a sample), while the second part of the 307 model predicts non-zero counts. Two predictors were included in the final hurdle-part of the 308 model, temperature and wind speed (Table 2). Lower temperatures predict fewer cercariae in 309 water samples, while high wind speeds also predict fewer cercariae in the water. Every 1 degree 310 increase in temperature increases the odds of having greater than 0 cercariae by 1.60 (p < 0.05), 311 while every 1 km/hr increase in wind speed decreases the odds of having greater than 0 cercariae 312 by 0.88 (p < 0.05). The effects of wind direction, lake, and wind speed, were assessed in the non-313 hurdle part of the model. Having no wind (i.e., no reading on the handheld anemometer) 314 increased odds of cercariae concentrations by 1.28 times (p<0.05). Every 1 km/hr increase in 315 wind speed predicts a -0.026 (p < 0.01) fold change in cercariae concentrations.
316 Discussion 317 318 The avian schistosome life cycle was first described in Michigan in 1928 [20]. Since this 319 discovery, the state has implemented various strategies to control swimmer's itch over the past 320 80 years. In the 1980s, studies were conducted to gauge the efficacy of treating mallards (Anas 321 platyrhynchos) and common mergansers (M. merganser) with Praziquantel, an anthelmintic drug 322 that is also used to treat human schistosomiasis. These studies suggested Praziquantel was 323 effective at reducing the infection prevalence of schistosomes in birds and indicated that yearly 324 treatment of birds at specific sites may reduce swimmer's itch in localized areas [8,9]. These 325 studies led to the waterfowl-based control efforts that are outlined in this study, that focus on 326 lake-wide relocation of a top-tier predatory bird from numerous inland Michigan lakes. While 327 the principle of parasite life cycle interruption by host relocation seems sound in theory, here we 328 present evidence that suggests that lake-wide relocation of resident common mergansers had 329 little to no measurable impact on the population of T. stagnicolae on any of the inland lakes that 330 participated in our comprehensive study. 331 A resident definitive host population is not required to maintain a large avian schistosome 332 population. 333 334 We found no statistically significant differences between T. stagnicolae abundance between 335 lakes with resident common merganser control programs (Glen and North Lake Leelanau), the 336 negative control lake (Walloon Lake) which had resident common mergansers present on the 337 lake in 2018-2020, or the positive control lake (South Lake Leelanau) which does not support 338 resident mergansers. There was not a significant difference between Glen Lake and Walloon 339 Lake in 2019, nor in 2020. If merganser relocation were effective, we would have expected the 340 cercariae abundance at Glen Lake to be similar to that of South Lake Leelanau, and to differ 341 from Walloon Lake. 342 343 In 2020, Glen and North Lake Leelanau discontinued relocating summer resident common 344 merganser broods. We compared T. stagnicolae cercariae abundance in 2019 and 2020 and 345 found that Glen Lake had a moderately statistically significant reduction in T. stagnicolae 346 abundance when summer resident common mergansers remained onto the lake. However, since 347 schistosomes over-winter in snails and we only collected after hatch-year snails, cercariae data 348 reflect previous year infection logistics.  (Table 2), suggest this 353 effect is likely due to water temperatures at this time of the year becoming conducive to cercariae 354 emergence from snails. This observation corroborates findings from previous laboratory studies 355 that have found cercariae emergence to be temperature dependent in numerous digenean 356 trematode systems [21-24]. T. stagnicolae abundance remains high until late September at lakes 357 with and without control programs (Figure 3). 358 359 This study shows that T. stagnicolae maintains its life cycle in the absence of a continuous 360 definitive host presence for most of the transmission season. Our group collected additional field 361 data in summer 2020 on Higgins Lake (Roscommon Co., MI) that would support this study. Not 362 only were all merganser broods removed since 2015, but they also systematically eliminated 363 adult and breeding mergansers through two years of spring merganser duck hunts (2015 and 364 2016), filling nesting holes, and by euthanizing breeding hens for ancillary research projects. 365 This resulted in a stable summer resident merganser population of 9 broods in 2015 to 0 broods 366 in 2020. Even with these definitive host mitigation efforts, numerous cases of swimmer's itch 367 were reported in 2020 and average levels of T. stagnicolae were found in the water in July [25]. 368 369 If we assume that the common merganser is the only suitable definitive host for T. stagnicolae at 370 inland Northern Michigan lakes, as is suggested by controlled challenge studies [7], then 371 migratory common mergansers in the spring and fall must have been contributing enough 372 parasites into the ecosystems of Glen Lake and South and North Lake Leelanau to sustain T. 373 stagnicolae populations in the intermediate host S. emarginata until fall, when the migratory 374 birds return. Since all broods were trapped <4 weeks from hatching, ducklings could not have 375 contributed to the parasite load on lakes implementing merganser control. In addition, non-376 nesting adult mergansers usually vacate these inland lakes by early-mid summer, thereby 377 reducing the likelihood of them contributing to parasite transmission. Thus, birds that are not 378 resident contributors to the biodiversity of a specific lake ecosystem nonetheless can contribute 379 and sustain the parasite biodiversity of an ecosystem in the case of this avian schistosome 380 species. Underpinning these observations are our observations of the intermediate snail host 381 community. North Lake Leelanau, Glen Lake and Walloon Lake have higher densities of S. 382 emarginata than South Lake Leelanau. It appears that lakes with relocation programs (i.e., Glen 383 Lake and North Lake Leelanau) that have higher snail densities will still have high cercariae 384 concentrations, due to the exponential amplification of the parasite within the intermediate host.
385 Similarly, this exponential amplification explains why South Lake Leelanau, which has a low S. 386 emarginata density and does not have resident birds, still has T. stagnicolae cercariae in the 387 water. We also noted decreases in S. emarginata density at Glen Lake and North Lake Leelanau 388 between 2019 and 2020, whereas the population of this snail increased at South Lake Leelanau 389 and Walloon Lake. However, there was not a corresponding increase in snail infection 390 prevalence. Changes in the snail populations undoubtably underpin changes in cercariae 391 abundance from year to year, and future research from our group will explore how changes in S. 392 emarginata snail density at these lakes over the course of three years are related to changes in T. 393 stagnicolae parasite populations.
394 Avian schistosome C is an important contributor to avian schistosome populations in 395 northern Michigan Lakes, while T. physellae is a minor contributor. 396 397 In 2018 our group discovered a novel avian schistosome-like species emerging from a P. 398 trivolvis (=Helisoma trivolvis) snail. As reported in McPhail et al. [19], this proved to be a novel 399 genus of avian schistosome which utilizes definitive host Branta canadensis (and possibly 400 others) and caused dermatitis under laboratory conditions. Avian Schistosome C made up a 401 significant component of the avian schistosome community and was the most abundant avian 402 schistosome at Walloon Lake and South Lake Leelanau (Figure 2; Figure 3). This parasite is 403 detectable earlier and persists later into the season than T. stagnicolae. Interestingly, lakes with 404 control programs aimed at controlling T. stagnicolae have lower abundance of Avian 405 Schistosome C. These lakes also have lower densities of P. trivolvis, the intermediate host of 406 Avian Schistosome C. As might be expected, South Lake Leelanau and Walloon Lake have 407 higher densities of P. trivolvis and consequently higher relative abundances of Avian 408 Schistosome C. 409 410 Trichobilharzia physellae was sporadically detected throughout the season at all four lakes. The 411 intermediate host of T. physellae in Northern Michigan lakes is Physellae parkeri, and the 412 definitive hosts are mallard ducks and common mergansers. Lakes in our study that trapped and 413 relocated summer resident common mergansers have low abundance and more sporadically 414 appearing T. physellae populations than the control lakes. However, the control lakes also 415 support low abundances of this parasite. Interestingly, Glen Lake has a high density of P. 416 parkeri, despite having a low reactive abundance of T. physellae. Because we do not know the 417 prevalence of this parasite in bird populations, it is possible this parasite has a lower infection 418 prevalence in birds and consequently could produce lower percent abundance in water samples. indicates that 423 onshore winds represent an increased risk for higher cercariae concentrations, and that onshore 424 winds increase the daily incidence of swimmer's itch. While we found that onshore winds do 425 increase the odds of cercariae concentrations by 1.21 times, this was not statistically significant. 426 This can likely be explained by how the water samples were collected for this study. By 427 sampling along the length of a dock, rather than solely along the shore, it is likely that the effect 428 of onshore winds on cercariae concentration was diluted.
429 430 This study also shows that temperature and wind speed act as "hurdles" predicting an absence of 431 cercariae at beaches (Table 2). Numerous laboratory studies have demonstrated that cercariae 432 emergence is temperature dependent [22,26,27], and as such it is unsurprising that temperatures 433 less than 15C are predictive of 0 cercariae detected in the water. However, temperature was not 434 a significant predictor of non-zero values in our model. While temperature has been extensively 435 demonstrated to drive cercariae survival and production [27], this effect varies between different 436 species of trematode and their snail hosts. Therefore, our pan-avian qPCR approach to measuring 437 cercariae in the water would mask this effect. 438 439 The results of temperature and their effect on cercariae concentrations is corroborated in 440 Sckrabulis et al [14]. These authors did not find temperature to be predictive of increases in 441 swimmer's itch incidence. Windspeed was a significant predictor in our model, with increased 442 wind speed resulting in a decrease in cercariae concentrations at local beaches. In addition, 443 strong winds likely congregate the floating cercariae very near shore (<1m) which was inside our 444 sampling zone. This study also highlights the variable nature of cercariae concentrations at 445 individual locations (Figure 1; Figure 2), and highlights that the risk of encountering cercariae in 446 water samples differs spatially and temporally at locations across a lake, with some sites having 447 relatively low concentrations of cercariae, while others are nearly constantly inundated with 448 cercariae. In conclusion, the swimmer's itch problem in Michigan is complex, and has resulted in 484 researchers and riparians alike searching for a solution for the better part of the last century. 485 Even with the long history of research in this area, there continue to be discoveries. These 486 include the presence of a novel avian schistosome species found in the state that has been 487 revealed as another major contributor to the schistosome community at these lakes. Furthermore, 488 with this study, we demonstrate that removing the resident definitive host of the dominant avian 489 schistosome, T. stagnicolae, from the lakes did not have the hypothesized effect of interrupting 490 the life cycle and reducing the numbers of this parasite found in the water. As such, swimmer's 491 itch research is ongoing, and future research will continue to focus on these avian schistosomes, 492 and the contributions made to the schistosome community by non-resident migrants. In addition, 493 we predict a practical paradigm shift away from whole-lake control to individual prevention will 494 occur, and we plan to assess the myriad prevention strategies already employed by riparians.