Interactions between two parasites of brown trout (Salmo trutta) : Consequences of preinfection

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY 4.0 https://creativecommons.org/licenses/by/4.0/ Interactions between two parasites of brown trout (Salmo trutta) : Consequences of preinfection © 2018 the Authors Published version Gopko, Mikhail; Chowdhury, Motiur; Taskinen, Jouni Gopko, M., Chowdhury, M., & Taskinen, J. (2018). Interactions between two parasites of brown trout (Salmo trutta) : Consequences of preinfection. Ecology and Evolution, 8(20), 9986-9997. https://doi.org/10.1002/ece3.4406 2018 9986 | Ecology and Evolution. 2018;8:9986–9997. www.ecolevol.org Received: 9 November 2017 | Revised: 22 June 2018 | Accepted: 29 June 2018 DOI: 10.1002/ece3.4406 ORIGINAL RESEARCH Interactions between two parasites of brown trout (Salmo trutta): Consequences of preinfection Mikhail Gopko1 | M. Motiur R. Chowdhury2 | Jouni Taskinen2 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2018 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. 1A.N. Severtsov Institute of Ecology and Evolution, Russian Academy of Sciences, Moscow, Russia 2Department of Biological and Environmental Science, University of Jyväskylä, Jyvaskyla, Finland Correspondence Gopko Mikhail, A.N. Severtsov Institute of Ecology and Evolution, Russian Academy of Sciences, Leninskij prosp., 33, 119071 Moscow, Russia. Email: [email protected]om Funding information The work was supported by the Academy of Finland mobility grant 311033/2016 to JT, CIMO Fellowship grant TM-14-9506 to JT, Ella and George foundation mobility grant (2017) to MG, RFBR grant (17-04-00247a), Russian Science Foundation grant (14-1401171) to MG, and Maj ja Tor Nessling Foundation research grant to MC. Abstract Preinfection by one parasitic species may facilitate or by contrast hamper the subsequent penetration and/or establishment of other parasites in a host. The biology of interacting species, timing of preinfection, and dosage of subsequent parasite exposure are likely important variables in this multiparasite dynamic infection process. The increased vulnerability to subsequent infection can be an important and often overlooked factor influencing parasite virulence. We investigated how the preinfection by freshwater pearl mussel Margaritifera margaritifera glochidia could influence the success of subsequent infection by the common trematode Diplostomum pseudospathaceum in brown trout Salmo trutta and vice versa whether preinfection by the trematode made fish more susceptible to glochidia infection. The first experiment was repeated twice with different (low and high) exposure doses to initiate the subsequent trematode infection, while in the second experiment we varied the timing of the preinfection with trematodes. The preinfection with glochidia made fish more vulnerable to subsequent infection with trematodes. Since the trematodes penetrate through the gills, we suggest that increased host vulnerability was most likely the result of increased respiration caused by the freshwater pearl mussel glochidia encysted on gills. In turn, brown trout preinfected with trematodes were more vulnerable to the subsequent glochidial infection, but only if they were preinfected shortly before the subsequent infection (20 hr). Fish preinfected with trematodes earlier (2 weeks before the subsequent infection) did not differ in their vulnerability to glochidia. These effects were observed at moderate intensities of infections similar to those that occur in nature. Our study demonstrates how the timing and sequence of exposure to parasitic species can influence infection success in a host–multiparasite system. It indicates that the negative influence of glochidia on host fitness is likely to be underestimated and that this should be taken into consideration when organizing freshwater pearl mussel restoration procedures. KEYWORDS community ecology, Diplostomum, experimental infection, freshwater pearl mussel, host– parasite interactions, multiple infections | 9987 MIKHAIL et AL. 1 | INTRODUCTION Individuals of freeliving organisms are usually infected by several parasite species. However, a majority of host–parasite studies are concentrated around single parasite and single host interactions (Rigaud, PerrotMinnot, & Brown, 2010). This is partly because studies concerning multiple parasite species interactions are much harder to perform and interpret. Indeed, interactions between parasitic species within host organisms can take very different forms even in the simplest case when a host harboring one parasite is subsequently infected with another parasite (see Cox, 2001; Thomas, Adamo, & Moore, 2005; Cézilly, PerrotMinnot, & Rigaud, 2014; Vaumourin, Vourc’h, Gasqui, & VayssierTaussat, 2015 for review and theoretical considerations). Thus, the presence of certain parasitic species in a host can be more or less beneficial or deleterious to other parasites that subsequently attempt to infect that same host depending on the transmission ecology of the cooccurring parasite species and hostmediated effects (Cox, 2001; Thomas et al., 2005; Vaumourin et al., 2015). A parasite which has already entered a host organism, for example, can influence (directly or indirectly) the success of subsequent infections by other parasitic species (see Cox, 2001; Johnson, de Roode, & Fenton, 2015; Kotob, MenanteauLedouble, Kumar, Abdelzaher, & ElMatboulim, 2016; Pedersen & Fenton, 2007; Vaumourin et al., 2015). This influence may lead to a positive interaction, i.e., the presence of one parasite may facilitate subsequent infections by other parasites, or negative interaction between parasitic species (Hoverman, Hoye, & Johnson, 2013; Jackson, Pleass, Cable, Bradley, & Tinsley, 2006; Klemme & Karvonen, 2017; Vaumourin et al., 2015). In general, mechanical damage and immunity are likely to be among the most important factors influencing the success of the infection in preinfected hosts (Pedersen & Fenton, 2007). Thus, in humans preinfection with herpes simplex virus type 2 causes damage to mucous membranes paving the way to HIV infection (Vaumourin et al., 2015 and references therein). In fish, similar mechanical lesions caused by the ectoparasitic crustacean Argulus coregoni increase host susceptibility to the pathogenic bacterium Flavobacterium columnare (Bandilla, Valtonen, Suomalainen, Aphalo, & Hakalahti, 2006). By contrast, the infection with one myxozoan species can prevent the subsequent invasion of other myxozoan species presumably due to crossimmunity (Kotob et al., 2016). Likewise, preinfection also can have no impact on subsequent infection as demonstrated by Karvonen, Seppälä, and Valtonen (2009) in their study of two species of trematodes from the genus Diplostomum infecting rainbow trout and by Chowdhury et al. (2018) in a freshwater pearl mussel study. Although the impact of preinfection by one parasite on the infection success of other parasites has been studied in fish (Kotob et al., 2016), these studies, mainly concerned interactions between microparasites. Preinfection studies with macroparasites have received far less attention. Kotob et al. (2016), for example, reviewed data on parasitic interactions in fish, but did not present any examples of macroparasite–macroparasite interactions (but see Karvonen et al., 2009). Parasitic associations, i.e., positive or negative correlations between infection intensities, are rather commonly reported from wild hosts (Booth, 2006; Johnson & Buller, 2011; Karvonen et al., 2009; Pedersen & Fenton, 2007; Rigaud et al., 2010). In natural conditions, however, true interparasitic interactions are often masked by positively or negatively correlated coinfection (Johnson & Buller, 2011). Consequently, associations between two species can disappear or change from positive to negative, when a controlled experiment is performed (e.g., Johnson & Buller, 2011; Karvonen et al., 2009). However, it is interactions, rather than associations, among parasites that play an important role in structuring populations and communities of both hosts and parasites (Rigaud et al., 2010; Vaumourin et al., 2015). In addition, the order in which different parasitic species or genotypes attack and enter the host can influence the interaction between parasites within the host individual (Hoverman et al., 2013; Klemme & Karvonen, 2017; Read & Taylor, 2001; Telfer et al., 2010). Therefore, an experimental approach is needed to reveal, whether there is real interaction between parasites or, at least, evaluate the consequences of such interactions. The freshwater pearl mussel, Margaritifera margaritifera is a freshwater bivalve critically endangered through its native range (Geist, 2010; LopesLima et al., 2017). This bivalve has an obligate parasitic larval stage, the glochidium, which attaches to gills of salmonid fishes, such as Atlantic salmon (Salmo salar) and brown trout (Salmo trutta) (Salonen, Marjomäki, & Taskinen, 2016; Salonen et al., 2017; Taeubert & Geist, 2017). The latter fish is the exclusive host for freshwater pearl mussel glochidia in many Central European populations (Geist, Porkka, & Kuehn, 2006; Taeubert & Geist, 2013). Often freshwater pearl mussel glochidia are considered to have a positive or, at most, a very weak negative effect on the host’s health (Ziuganov, Zotin, Nezlin, & Tretiakov, 1994; Ziuganov, 2005; but see Taeubert & Geist, 2013). The latter is perhaps more likely to be true since freshwater pearl mussel glochidia demonstrate more than sevenfold growth in the course of 8–10month development on the fish gills with a substantial nutrient transfer from fish to glochidia and a shift in the stable isotope composition during the mussel’s development (Denic, Taeubert, & Geist, 2015). In addition, M. margaritifera glochidia cause a pronounced immune response in their hosts, increase metabolic rates and hematocrit levels and hamper swimming and respiratory performance in brown trout (Chowdhury, Salonen, Marjomäki, & Taskinen, 2018; Filipsson et al., 2017; Österling, Ferm, & Piccilo, 2014; Taeubert & Geist, 2013; Thomas, Taylor, & Garcia de Leaniz, 2014). Impairment of the respiratory capacity of host fish is especially important since gills can be a place where other infectious agents enter the host. It has been demonstrated that exposure of fish to pathogens and parasites increases with the ventilation rate (Mikheev, Pasternak, Valtonen, & Taskinen, 2014). Therefore, freshwater pearl mussel glochidia are likely to be a suitable predecessor for parasitic species infecting brown 9988 | MIKHAIL et AL. trout through gills (i.e., increase their infection success). However, this assumption has never been tested experimentally. Moreover, to our knowledge, there were no previous studies, where the interaction between glochidia and other parasites was tested experimentally. The eye fluke trematode Diplostomum pseudospathaceum is a ubiquitous species (Klemme & Karvonen, 2017) which is a common parasite of a wide array of fishes, such as many cyprinid and salmonid fishes including brown trout in natural environments (Betterton, 1974; Rolbiecki, Sciazko, & Schütz, 2009; Valtonen & Gibson, 1997). D. pseudospathaceum has three hosts in its life cycle (Seppälä, Karvonen, & Valtonen, 2004). Freshwater snails (often Lymnaea stagnalis) are the first intermediate hosts, freshwater fishes are the second intermediate hosts and fisheating birds are the definitive hosts. Infected snails produce thousands of cercariae (the trematode’s dispersal stage) which penetrate fish epithelium primarily through gills (Mikheev et al., 2014). Therefore, the fluke’s infection success strongly depends on the amount of water pumped through fish gill chambers (Mikheev et al., 2014). After penetrating fish epithelium, D. pseudospathaceum moves toward the eye, the place of its final development into the metacercarial stage within the eye lens. In contrast to M. margaritifera, D. pseudospathaceum is likely to be an unsuitable “predecessor” for other parasitic species. In the eye lens, flukes are unprocurable for the host immune system (Höglund & Thuvander, 1990; Wegner, Kalbe, & Reusch, 2007). Therefore, there is no need for them to suppress the host’s immune system to avoid its negative influence on the parasite. Although such immunity regulation is known for numerous other parasitic species (Cox, 2001; Maizels & McSorley, 2016; McSorley, Hewitson, & Maizels, 2013), it is unlikely to be the case for D. pseudospathaceum. Moreover, Höglund and Thuvander (1990) suggested that D. pseudospathaceum nonspecifically enhances host immunity. Although specific antibody production was not recorded, they found that preinfection with D. pseudospathaceum induces some degree of protective immunity against subsequent infections with these parasites, suggesting some kind of nonspecific immunity enhancement caused by eye flukes (Höglund & Thuvander, 1990; see also Karvonen, Paukku, Seppälä, & Valtonen, 2005). In our study, we sequentially infected brown trout with glochidia of the freshwater pearl mussel and cercariae of the eye fluke and vice versa. We hypothesized that fish preinfected with freshwater pearl mussel glochidia are more vulnerable to the cercarial infection by D. pseudospathaceum because glochidia can impair respiratory gas exchange and elevate resting respiratory rate of infected fish (Thomas et al., 2014), thereby increasing the amount of water (and cercariae) pumped through the gill chamber. To test this hypothesis we repeated the experiments twice using different (high and low) doses of cercariae during the infection. We also hypothesized that brown trout preinfected with D. pseudospathaceum would be less or similarly vulnerable to the subsequent freshwater pearl mussel glochidia infection because D. pseudospathaceum does not change or even enhances the performance of the host immune system. In this experiment we also manipulated the timing of preinfection with trematode (i.e., 20 hr vs. 2 weeks before the subsequent infection by glochidia). This took into account the possible host mediated effects because since an immune response may occur in fish soon after the infection D. pseudospathaceum is likely to cause in fish, whereas once it has migrated to the eye lens, it is “invisible” to the host’s immune system (Höglund & Thuvander, 1990; Wegner et al., 2007). TABLE1 Experimental design. (a) Preinfection with M. margaritifera glochidia and (b) Preinfection with D. pseudospathaceum. Fish numbers (n) obtained after the exclusion of individuals with unclear infection state (see the text) Preinfection with glochidia Treatment Exposure to cercariae Fish mass, g, mean ± SD (a) M. margaritifera → D. pseudospathaceum experiment Exposure dose = 5,000 glochidia/fish Infected (n = 36) High dose (300 cercariae/fish) 7.03 ± 2.17 Control (n = 29) Infected (n = 28) Low dose (200 cercariae/fish) 20.10 ± 4.66 Control (n = 55) September, 2014 July, 2015 Preinfection with cercariae Treatment Exposure to glochidia Fish mass, g, mean ± SD (b) D. pseudospathaceum → M. margaritifera experiment Exposure dose = 250 cercariae/fish Preinfected 20 hr before glochidial infection (n = 15) Exposure dose = 1,500 glochidia/ fish 21.58 ± 4.65 Preinfected 2 weeks before glochidial infection (n = 48) Control (n = 56) August, 2015 | 9989 MIKHAIL et AL. 2 | MATERIAL AND METHODS 2.1 | Experimental infection procedures 2.1.1 | Fish preinfected with freshwater pearl mussel glochidia Youngoftheyear brown trout (Salmo trutta) from the Iijoki (Finland) searun and Rautalampi (Finland) lakerun stock were collected from the Natural Resources Institute Finland (Luke) in Taivalkoski and the Luke in Laukaa (Finland), respectively, in the end of August 2014 and transported to the Konnevesi Research Station of the University of Jyväskylä, Finland. Fish from each stock were placed into four (altogether eight) 163L flowthrough tanks. Fish were acclimated in the laboratory for 3 weeks and then some were massinfected with freshwater pearl mussel glochidia (exposure dose = 5,000 glochidia per fish) that originated from the Livojoki river (Table 1a). The procedure of the glochidia collection and exposure to fish was similar to Chowdhury et al. (2018). To collect glochidia we placed adult freshwater pearl mussels in plastic buckets filled with 5 L of river water for 30 min on the day of infection. The mussels were returned to the river after incubation and the glochidial suspension was transported to the Konnevesi research station. Brown trout were massexposed to glochidia in maintenance tanks, where the water volume was reduced to 70 L and waterflow turnedoff. A similar procedure (i.e. massexposure in 70 L of still water) was used in all experimental infections mentioned throughout the paper. Though individual exposure is sometimes recommended (Douda, 2013 and references therein), simultaneous exposure of the group of fish is commonly used in experimental parasitological practice (e.g., Gopko, Mikheev, & Taskinen, 2015, 2017; Seppälä, Karvonen, & Valtonen, 2008; Seppälä et al., 2004; Taeubert, Gum, & Geist, 2013; Taeubert, Martinez, Gum, & Geist, 2012). Such “massexposure” approach is especially logical when studying fish, which spend a substantial amount of time in shoals (Taeubert et al., 2013) as juvenile salmonids do (Brännäs, Jonsson, & Lundqvist, 2003; Hicks & Watson, 1985). The exposure time was 1 hr, average exposure dose ~5,000 glochidia per fish, and water temperature was 14.2°C. Control animals were treated identically to preinfected fish with the exception that instead of cercariae they were exposed only to water. These methods were used during all infection procedures. After exposure, brown trout were marked using fin clipping and randomly reallocated according to their original stock so that all eight tanks received a similar amount of infected and control fish. In four tanks the right fin was notched in control fish, while the left fin was notched in fish exposed to glochidia. In four other tanks, opposite fins were notched in control and infected fish. Though notching can cause an increased immune response in fish, this effect is likely to be relatively short term. Henrich, Hafer, and Mobley (2014) found out that the effect of spine clipping on host immunity became indistinguishable 2 weeks after the procedure. In our study, we marked fish 3 weeks after the exposure to glochidia (i.e., 9 months before the exposure to D. pseudospathaceum cercariae). Therefore, it is unlikely that fin notching could have had a substantial effect on the success of trematode infection. In addition, fish were finnotched both in infected and control groups. Following Taeubert et al. (2012, 2013) five fish from each tank were sacrificed randomly 3 days postexposure to check for infection success. We found successful infection (ranging from 19 to 782 glochidia per fish with mean ± SD = 313.0 ± 239.3) in glochidia exposed fish, while fish from the control group had no glochidia. Fish were subsequently maintained for 10 months in four similar 163L round flowthrough tanks. Throughout the experiment control fish were treated identically to infected fish. In June–July 2015 brown trout from two randomly chosen tanks with fish from each stock were exposed to D. pseudospathaceum. The fish were exposed in summer to make experimental conditions synchronous with their natural seasonal occurrence. In Finland detachment of M. margaritifera glochidia from fish gills happens in June–July (Salonen & Taskinen, 2017). Therefore, damage caused to fish gills by glochidial excystment and the increase in host vulnerability to subsequent infections is likely to be most pronounced in this period of the year. In addition, snails infected with D. pseudospathaceum are starting to produce cercariae actively when water temperatures are above 10°C (Lyholt & Buchmann, 1996; Voutilainen, Taskinen, & Huuskonen, 2010). Therefore, in boreal ecosystems of the northern hemisphere the probability of interactions between M. margaritifera glochidia and D. pseudospathaceum eye flukes within hosts is especially high in summer. The exposure procedure was generally similar to Gopko et al. (2015). In brief, brown trout were exposed to the mixture of cercariae shed by 10 freshwater snails Lymnaea stagnalis. In our study, we did not identify parasites using molecular methods. However, previous studies showed that L. stagnalis snails are typically infected with D. pseudospathaceum (e.g. Rellstab, Louhi, Karvonen, & Jokela, 2011; Selbach, Soldánová, Georgieva, Kostadinova, & Sures, 2015). For instance, in Finland (including Lake Konnevesi) L. stagnalis is infected with D. pseudospathaceum (Louhi, Karvonen, Rellstab, & Jokela, 2010; Rellstab et al., 2011). Cercariae used in our experiments were no older than 5 hr. The exposure time was 30 min which is similar to previous studies (e.g., Gopko et al., 2015, 2017; Klemme & Karvonen, 2016, 2017; Mikheev, Pasternak, Taskinen, & Valtonen, 2010; Mikheev et al., 2014; Seppälä et al., 2004). During the exposure water levels in each tank were decreased to 70 L and water flow was turned off. After the exposure, water flow was turned on and water volumes in the tanks were again increased to 163 L. We repeated the experiment twice with a twoweek interval. In first two (out of four) randomly chosen tanks, fish from Iijoki stock were exposed to the low infection dose (200 cercariae per fish) at 12.5°C. Such a dose, however, resulted in lower metacercariae numbers in fish eye lenses than we planned (some fish were uninfected) (see Section 3). Therefore, we repeated the experiment with a higher doses of cercariae during the exposure (300 cercariae per fish, 12.8°C). In this high dose experiment we used fish from the Rautalampi stock. 9990 | MIKHAIL et AL. After the D. pseudospathaceum infection (on 7th and 12th days in low and high dose infection experiment, respectively), fish were killed with an overdose of 0.01% MS 222 (Sigma Chemical Co., St Louis, U.S.A.), weighed and dissected. It is improbable that a small difference in the periods between infections and dissections influenced the results of dissections. This is because the Diplostomum metacercarial mortality rate within host lenses is likely to be extremely low, while the life span is very long, with a maximum of 4 years (Klemme & Karvonen, 2017; Shigin, 1980). The number of D. pseudospathaceum metacercariae and freshwater pearl mussel glochidia were counted using a dissection microscope (32× magnification). 2.1.2 | Fish preinfected with D. pseudospathaceum A reciprocal experiment (brown trout preinfected with eye flukes and then infected with glochidia at 14.2°C) was conducted in August 2015. One hundred and twenty youngoftheyear brown trout, “Rautalampi” stock, were obtained from the Natural Resources Research Institute (Luke) fish farm at Laukaa (Table 1b). Exposure procedures were similar to the described above. The exposure dose was 250 cercariae per fish. Fish were divided into three groups: (a) exposed to cercariae 2 weeks prior to the glochidia infection, (b) exposed 20 hr prior to the glochidia infection and (c) control fish exposed only to lake water). The 20 hr group was added because eye flukes usually need 24–48 hr to migrate to the host eye lens. Withinhost migration is the only period when the parasite is vulnerable to the fish’s immune system (Höglund & Thuvander, 1990). Then all fish were infected with the glochidia of the freshwater pearl mussel. The exposure dose of glochidia infection was 1,500 glochidia per fish and an exposure time of 1.5 hr. Glochidia were collected from Jukuanoja, a tributary of the Iijoki River and were transferred to Konnevesi research station on 31 of August 2015. Fish were dissected on the 13th day after the infection with trematodes. Metacercariae and glochidia were counted using a dissection microscope (32× magnification). Throughout the experiments, fish were daily fed with commercial food pellets (1.5 mm size, Nutra Parr LB, Norway). The experiments were conducted with the permission of the Centre for Economic Development, Transport and Environment of South Finland (license number ESAVI/6759/04.10.03/2011). 2.2 | Statistical analysis During dissections we found that not all fish marked as preinfected with freshwater pearl mussel really had glochidia on their gills. It could be a result of unsuccessful infection or active dropping off the host gill by the mature parasites. The latter is more likely because the freshwater pearl mussel glochidia prevalence in experimental infections usually is close to 100% (e.g., Taeubert & Geist, 2013). By contrast, the duration of glochidial development on the fish gills is highly variable (from 3 weeks to 10 months) and depends on several factors including parasites’ compatibility with the host (Geist et al., 2006; Marwaha, Jensen, Jakobsen, & Geist, 2017). Since we could not estimate how much time passed since glochidia dropped off the gills, those fish that were marked as infected, but did not have any glochidia on their gills were considered as being of unknown status and were excluded from the subsequent analysis (23 and 44 fish in low and high dose trematode infection treatment, respectively). However, when these fish are included in the analyses, statistical effects remain significant. R software was used for all statistical analyses (R Core Team, 2017). Plots were drawn using the “ggplot2” package (Wickham, 2009). In fish preinfected with glochidia, we started with a general linear model (GLM) where the Box–Cox transformed D. pseudospathaceum infection intensity was a response variable, infection state (preinfected with glochidia or not) and experiment (low vs. high exposure dose) were factors and fish mass was a covariate. Response variable transformation was needed since the residuals’ distribution of our model strongly violated the normality assumption (Shapiro– Wilk (S–W) test, W = 0.928, p < 0.001). After the transformation, data distribution became close to normal (S–W test, W = 0.984, p = 0.09). Since experiments with low and high exposure doses were conducted with a time lag and can be considered independent, we also fitted separate models for each experiment. In the high dose treatment, we started with a general linear model (GLM) where the D. pseudospathaceum infection intensity was the response variable, infection state was a factor and fish mass as a covariate. However, mass did not have a significant influence on the eye fluke infection intensity and was excluded from the final model. A residual values distribution did not differ from normal (S–W test, W = 0.971, p = 0.13). In the low dose treatment, the Shapiro–Wilk test showed that even after transformation there was a violation of the normality assumption in our data due to several obvious outliers seen on the Q–Q plot (S–W test, W = 0.944, p = 0.001). Therefore, we had to fit a robust regression based on Mestimator with Huber’s weights, with the tuning constant k = 1.345σ using MASS package in R programming software (Venables & Ripley, 2002). The f.robtest function in the sfsmisc package was used to compute robust Ftest and get pvalues (Maechler, 2017). The square root transformed intensity of the D. pseudospathaceum infection was a response variable. The Fisher’s exact test was used to estimate whether the probability of getting infected in preinfected fish was higher than in the control group. In the reciprocal experiment (where fish were preinfected with D. pseudospathaceum), we used a generalized linear model with a log link function and Gaussian error structure, where the glochidial infection intensity was the dependent variable, infection state was a factor and fish mass was a continuous predictor. Post hoc comparisons were done using a glht function in the “multcomp” package (Hothorn, Bretz, & Westfall, 2008). pValues were adjusted using Bonferroni corrections. | 9991 MIKHAIL et AL. 3 | RESULTS 3.1 | Fish preinfected with freshwater pearl mussel glochidia There was a significant effect of the preinfection with glochidia on the intensity of the subsequent eyefluke infection (Table 2; Figure 1). As hypothesized, fish preinfected with glochidia were more vulnerable to the subsequent infection with trematodes than control fish (Figure 1). The effect of the covariate fish mass was statistically significant (Table 2), indicating that the intensity of the trematode infection decreased with host size. Finally, as hypothesized, higher exposure dose of cercariae resulted in higher infection intensities in brown trout (Table 2, Figure 1). The effects of interactions were nonsignificant and therefore interactions were excluded from the final model. If low and high dose treatments were analyzed separately, the effect of the glochidial preinfection on the success on subsequent trematode infection was still significant. In the high dose treatment (fish infected with a high dose of cercariae) all brown trout (N = 64, fish mass mean ± SD = 7.03 ± 2.17 g) were successfully infected with D. pseudospathaceum metacercariae (range 2–25). In the high dose treatment, the mean intensity of D. pseudospathaceum infection in fish preinfected with freshwater pearl mussel glochidia was significantly higher compared with the control group (GLM ANOVA, t = 2.7, p = 0.008; Figure 1). When included in the model, fish mass did not significantly influence the D. pseudospathaceum infection intensity (p > 0.6). Weights of control and preinfected brown trout were similar (GLM ANOVA, t = 0.03; p = 0.98). In the low dose treatment (fish infected with a low dose of cercariae) only 28 brown trout out of 83 (N = 83, fish mass mean ± SD = 20.10 ± 4.66 g) were infected with the eye fluke. In the low dose treatment, the probability of being infected was significantly higher in preinfected fish compared with the control group (Fisher’s exact test, p = 0.016). The robust regression based on Mestimation showed that in the low dose experiment both preinfection with glochidia and fish mass significantly influence the D. pseudospathaceum infection intensity. Fish preinfected with M. margaritifera glochidia were significantly more vulnerable to the subsequent infection with D. pseudospathaceum trematode compared with control fish (Robust F = 7.113, p = 0.009). The infection intensity of D. pseudospathaceum decreased with the fish size in the low dose experiment (Robust F = 10.47, p = 0.002). Fish mass (mean ± SD = 21.58 ± 4.65 g) was similar in control and preinfected fish (GLM ANOVA, t = 1.25; p = 0.22). Surprisingly, in both treatments there was no significant relationship between the intensity of glochidia preinfection and the intensity of D. pseudospathaceum subsequent infection, when only preinfected fish were taken into consideration (Pearson’s correlation coefficient, r = −0.13, p = 0.53, N = 28 and r = 0.30, p = 0.12, N = 28, high and low dose treatment, respectively). There was also no significant correlation between the fish mass and the glochidia infection intensity (r = 0.19, p > 0.33, N = 28 in the high dose treatment and r = 0.15, p = 0.47, N = 28 in the low dose treatment). When the two datasets (glochidiainfected fish from high and low dose treatment) were merged and possible confounding variables (fish mass and treatment) were taken into account, the result was similar (see Supporting Information Table S1). Intensities of the glochidial infection were (mean ± SD) 376.5 ± 275.3 and 438.7 ± 286.7 glochidia/ fish in low and high dose treatment respectively. TABLE2 The GLM demonstrated that there were significant effects of treatment (preinfection with glochidia), fish mass and exposure dose on the success of the subsequent D. pseudospathaceum infection (Box–Cox transformed values). Preinfected fish were more vulnerable to the infection compared with controls. In addition, bigger brown trout had lower eye fluke infection intensity than smaller ones. As expected lower exposure dose of trematode’s cercariae result in lower infection intensities in brown trout Source Estimate SE tValue pValue Treatment (infected) 0.323 0.082 3.913 0.0001 Mass −0.031 0.011 −2.874 0.0047 Experiment (low exposure dose) −1.586 0.160 −9.944 <0.0001 FIGURE1 Both in the low and the high cercaria dose treatment, brown trout preinfected with freshwater pearl mussel glochidia were more vulnerable to infestation by D. pseudospathaceum trematode cercariae than control fish. The “box” represents the interquartile range (IQR) of the D. pseudospathaceum metacercaria infection intensities within groups with median (black line). Whiskers extend from the highest to lowest values within 1.5*IQR. Suspected outliers, i.e., all observations lying outside 1.5*IQR, are shown as dots 0 5 10 15 20 25 High infection Low infection Number of metace rcrariae Treatment Control Infected Trematode infection in fish pre−infected with glochidia 9992 | MIKHAIL et AL. 3.2 | Fish preinfected with D. pseudospathaceum The GLM followed by post hoc comparisons showed that fish preinfected with D. pseudospathaceum 20 hr before the subsequent infection with M. margaritifera glochidia were significantly more vulnerable to the glochidial infection compared with control fish (Figure 2, Table 3). On the other hand, fish preinfected with D. pseudospathaceum cercariae14 days prior to the infection with glochidia had glochidial loads that did not differ significantly from those of the control fish. Glochidial loads significantly increase with the increase of the fish weight (Table 3). Fish weights did not differ between the treatments (ANOVA F2,116 = 0.097; p = 0.91). The intensities of glochidial infection were moderate (mean ± SD glochidia/fish was 138.9 ± 41.5 in 20 hr earlier preinfected, 119.4 ± 36.8 in 14 days earlier preinfected and 114.6 ± 38.9 in control fish respectively, Figure 2, see also Supporting Information Figure S1). 4 | DISCUSSION The results of our study indicate that the infection with glochidia of the freshwater pearl mussel can predispose fish to a concomitant infection with other parasitic species. Brown trout preinfected with freshwater pearl mussel glochidia were infected with more D. pseudospathaceum during the subsequent experimental infection compared with control fish. We repeated the experiment with two different doses of cercariae during the subsequent infection and both times the result was similar. In addition, in the low dose treatment—i.e., when a low dose of cercariae was used in the subsequent infection—fish preinfected with glochidia were more likely to become infected with D. pseudospathaceum than the control fish. The numbers of freshwater pearl mussel glochidia in preinfected fish (~400 glochidia per fish on average) were within the range of glochidia numbers found in natural salmonid populations (e.g., Salonen & Taskinen, 2017; Ziuganov, 2005), thereby providing a close biological relevance to the experimental findings. Previous studies have demonstrated that freshwater pearl mussel glochidia constitute a respiratory burden for the infected fish (Thomas et al., 2014). In turn, D. pseudospathaceum cercariae enter their hosts mainly through gills and increased ventilation rate also increases infection rate of D. pseudospathaceum infection in fish (Mikheev et al., 2014). Therefore, an enhanced ventilation rate in fish infected with glochidia is likely to be a plausible mechanism explaining their higher vulnerability to the D. pseudospathaceum infection. In addition, June–July, when subsequent infections with D. pseudospathaceum cercariae were performed, is the period when freshwater pearl mussel glochidia detach from fish (Salonen et al., 2017)—an event that typically damages gills as glochidial cysts rupture. This should further increase the vulnerability of fish to D. pseudospathaceum cercariae. It has been proposed earlier that the relationship between M. margaritifera glochidia and fish host is neutral or mutualistic FIGURE2 Fish preinfected with the D. pseudospathaceum were more vulnerable to the freshwater pearl mussel glochidia infection compared with control fish, when the preinfection took place 20 hr before the exposure to glochidia. Two weeks after preinfection, however, this pattern did not occur. For an explanation of the box, vertical line and whiskers, see Figure 1. Suspected outliers, i.e., all observations lying outside 1.5*IQR, are shown as dots. These points were included in the analysis, but when a robust regression was used, results of statistical tests remained similar (see the Supporting Information Table S2) 50 100150 200 20 hrs 14 days Control Treatment Number of glochidi a Glochidia infection in fish pre−infected with trematodes TABLE3 Results of the GLM and subsequent post hoc comparisons examining the effect of the preinfection with D. pseudospathaceum on brown trout vulnerability to subsequent freshwater pearl mussel glochidia infection in brown trout Source Estimate SE tValue pValue Results of generalized linear model Treatment 1 day 0.202 0.075 2.69 0.008 Treatment 14 days 0.030 0.058 0.52 0.602 Mass 0.116 0.023 5.10 <0.0001 Comparison Estimate SE zValue pValue Multiple comparisons 20 hr–14 days −0.172 0.076 −2.25 0.073 Control–14 days −0.030 0.058 −0.52 1.000 Control–20 hr 0.202 −0.075 −2.69 0.022 | 9993 MIKHAIL et AL. (e.g., Ziuganov, 2005; Ziuganov et al., 1994), but our results provide evidence that the relationship at least moderately antagonistic. Therefore, our data is consistent with previous studies, which demonstrated that brown trout infected with freshwater pearl mussel glochidia perform generally worse compared to uninfected fish (Filipsson et al., 2017; Österling et al., 2014; Taeubert & Geist, 2013; Thomas et al., 2014). This is in line with the fact that freshwater pearl mussel glochidia grow intensively on the fish gills and obtain nutrients from their hosts (Denic et al., 2015). However, Ziuganov (2005) found lower prevalence of saprolegniosis in Atlantic salmon infected by freshwater pearl mussel glochidia than in noninfected salmon and suggested an increase in nonspecific resistance in fish host caused by freshwater pearl mussel glochidia infection. Our results do not rule out this possibility, but indicate that either the proposed immunity enhancement does not include macroparasites, such as trematode cercariae, or that negative effects of gill damage due to detaching glochidia override a possible positive effect of the immune enhancement. Therefore, it would be interesting to perform subsequent infections with fungal, bacterial and viral pathogens in addition to parasites and at different times after the preinfection with freshwater pearl mussel glochidia. Shortly after the preinfection and at the time of glochidia detachment, the negative effect of gill damage should prevail, but in the course of glochidia development on gills or after the detachment period the possible immunity enhancement should be observed. A concomitant infection by trematodes caused by preinfection with freshwater pearl mussel glochidia may hamper the host directly, e.g., by utilizing additional host resources and/or loading its immune system. However, indirect effects of such concomitant infection can be even more severe. Diplostomum eye flukes, for example, can cause cataracts in fish eye lenses hampering fish vision and their ability to forage (Karvonen, Seppälä, & Valtonen, 2004; Owen, Barber, & Hart, 1993). More importantly, D. pseudospathaceum along with many other trophically transmitted parasite species (see Poulin, 2010 for the review) are able to manipulate host behavior predisposing fish to predation by the definitive host—fisheating bird (e.g., Crowden & Broom, 1980; Gopko et al., 2015, 2017; Mikheev et al., 2010; Seppälä et al., 2004, 2008). Therefore, by increasing the probability of concomitant infection freshwater pearl mussel glochidia can hamper their hosts in more subtle ways than it was suggested earlier (Filipsson et al., 2017; Österling et al., 2014; Taeubert & Geist, 2013; Thomas et al., 2014) which are not easy to detect without conducting comprehensive experimentation. Salmonids and freshwater pearl mussels inhabit running waters while the snail host of D. pseudospathaceum, L. stagnalis, prefers standing waters. Thus, cooccurrence of freshwater pearl mussel glochidia and D. pseudospathaceum in a salmonid host is probably infrequent, even though the geographic the distribution of the two species largely overlaps. However, many river systems have lentic sections or a lake as the source of the river. Because trematode cercariae can aggregate as dense clouds in the surface layer (Horák et al., 2015) and potentially can be carried over long distances by the wind, the flow of D. pseudospathaceum cercariae into freshwater pearl mussel rivers is possible. Indeed, today many freshwater pearl mussel rivers have been dammed for hydroelectric power, or modified for other purposes creating standing water reservoirs, ponds or stagnant water microhabitats favoring occurrence of L. stagnalis. Diplostomatids are masterful dispersers owing to their bird definitive host (see Louhi et al., 2010). Consequently, diplostomatid parasites are commonly recorded in brackish water and freshwater systems harboring fish and snails worldwide (BlascoCosta & Locke, 2017; Louhi et al., 2010; Valtonen & Gibson, 1997). In addition, the nutrient input to rivers is also increasing, which should favor L. stagnalis. These anthropogenic changes, together with climate warming, which should also benefit snails, probably increase the probability of the freshwater pearl mussel glochidia—Diplostomum coinfection in the future. When fish preinfected with freshwater pearl mussel glochidia were subsequently infected with a low dose of cercariae, the number of D. pseudospathaceum metacercariae declined with fish mass, whereas in the high dose treatment no significant relationship was found. A possible explanation could be related to fish size. In the low dose treatment fish were about three times bigger compared to the high dose treatment. Although in the place of their final localization (eye lens) D. pseudospathaceum metacercariae are unprocurable for the host immune system, they are attacked by innate immunity while moving to the lens through host tissues (Höglund & Thuvander, 1990; Karvonen et al., 2005; Wegner et al., 2007). The bigger the fish, the longer the distance for the parasite to migrate to the eye lens and the longer the period when the parasite is thus vulnerable to a host immune response. It is necessary to mention, however, that we are unaware about studies, where a correlation between the fish size and time taken by D. pseudospathaceum to reach eye lenses was demonstrated. The difference in fish sizes between experiments may be explained by different origin of brown trout used in the highand lowinfection treatment. Since fish originated from different stocks, their tolerance to laboratory conditions (e.g., feeding regime, temperature, salinity etc.) can be slightly different, which in turn can lead to unequal growth rates during 10month maintenance in the laboratory. This also refers to the different D. pseudospathaceum infection intensities in fish from lowand highexposure experiments. Fish from different stocks (populations) may differ in their susceptibility to parasites (BryanWalker, Leung, & Poulin, 2007; Hasu, Benesh, & Valtonen, 2009; Scharsack & Kalbe, 2014). However such difference is often a result of the evolution in different environmental condition, e.g., in lake (high parasitic pressure) compared to rivers (low parasitic pressure) (Scharsack & Kalbe, 2014), while in our study we were dealing with two river populations. Therefore, difference in parasitic load in two experiments is likely to a result of different exposure doses. However interpopulation differences also may play some role. The intensity of the glochidial infection also did not correlate with the fish size both in the high and the low dose treatment, which is in agreement with previous studies (e.g., Thomas et al., 2014). Though in early stages (few weeks after the infection) bigger fish