Interaction between the endangered freshwater pearl mussel Margaritifera margaritifera, the duck mussel Anodonta anatina and the fish host (Salmo): acquired and cross-immunity

The common duck mussel Anodonta anatina can live in sympatry with—and use the same host, brown trout (Salmo trutta)—as the endangered freshwater pearl mussel Margaritifera margaritifera. Since the glochidia release of A. anatina takes place seasonally earlier than that of M. margaritifera, brown trout can be sequentially exposed first to A. anatina and then to M. margaritifera. Cross-immunity, an immune reaction induced in fish host against glochidia after the infection with glochidia of another mussel species, is possible. Thus, it was studied experimentally if brown trout can be cross immunized against M. margaritifera by earlier infection with A. anatina. In addition, the hypothesis that consecutive exposures of same glochidial species in different years in the same host may create acquired immunity was tested in brown trout against M. margaritifera. Furthermore, the dose dependence of acquired immunity against M. margaritifera glochidia in the Atlantic salmon (S. salar) was also studied. Cross-immunity was not found; suggesting that occurrence of A. anatina does not pose a threat to M. margaritifera. Instead, acquired immunity and its dose dependence were evident, emphasizing the significance of availability of 0+ age group immunologically naïve Atlantic salmon/brown trout for efficient conservation of M. margaritifera.


Introduction
The freshwater pearl mussel (Margaritifera margaritifera) is a long-lived (Helama & Valovirta 2008), river dwelling bivalve mollusc, which occurs in Europe and North-East North America, but is now critically endangered throughout its range of occurrence (Geist 2010, Lopes-Lima et al. 2016).M. margaritifera has a larval stage, glochidium, which is parasitic on the Atlantic salmon (Salmo salar) and/or the brown trout (S. trutta) (e.g., Young & Williams 1984, Salonen et al. 2016).Some pearl mussel populations exclusively develop on Atlantic salmon (e.g., Ieshko et al. 2016) and others exclusively on brown trout (e.g.Geist et al. 2006).M. margaritifera fulfils the criteria of indicator, flagship, key stone and umbrella species, and can thus be considered an ideal target species for the conservation of aquatic ecosystem functioning (Geist 2010).
The duck mussel, Anodonta anatina, occurs commonly in lakes and rivers of Europe (Lopes-Lima et al. 2016).
The glochidia of A. anatina are known to be able to complete their development on 15 fish species, including the brown trout (Bauer et al. 1991).Although the results by Bauer et al. (1991) suggest that the co-occurrence of both M. margaritifera and A. anatina in a river is not frequent, it is still possible.For example, the River Mustionjoki/Svartå (Finland) and the River Wye (UK) inhabit both mussel species (Lopes-Lima et al. 2016).As the glochidia shedding of A. anatina takes place seasonally earlier (winter-spring, as late as May-June, Taskinen et al. 1997) than that of M. margaritifera (summer-autumn, e.g., Salonen & Taskinen 2016), it is likely that brown trout can be sequentially exposed to the glochidia of A. anatina and M. margaritifera.The immune defense system in vertebrates includes innate and acquired (adaptive, specific) components so that the acquired immunity is based on antibodies that bind to a specific antigen.In a repeated contact with the same parasite or pathogen, the immune reaction is quicker and stronger (memory) due to a faster antibody production and the antibody reserves left from the previous infection (Mutoloki et al. 2014).For example, fish hosts can develop an acquired immunity against the glochidia of unionoid mussels (Bauer and Vogel 1987, Rogers and Dimock 2003, Dodd et al. 2006, Treasurer et al. 2006).Sequential infection with parasites belonging to different (often closely related) species can also provide protection, which is called cross-immunity or cross-resistance.Thus, cross-immunity is a special type of acquired immunity.It has been shown, for instance, that infection with glochidia of one unionid mussel can result in immunity against glochidia of another unionid species (Dodd et al. 2005).As the parasitic stage can largely contribute to the reproduction success of the endangered M. margaritifera, the role of acquired immunity and cross immunity in conservation of M. margaritifera requires attention.Bauer & Vogel (1987) showed that brown trout can eliminate M. margaritifera glochidia by both tissue and humoral reaction so that in repeated exposures to M. margaritifera glochidia the immunologic responses by brown trout against glochidia are strengthened, indicating an acquired immunity against M. margaritifera.In the case of M. margaritifera, lower infection rates have been found in older brown trout in the field or after second infection in the laboratory, suggesting acquired immunity (e.g.Bauer 1987, Ziuganov et al. 1994, Hastie & Young 2001).Nevertheless, even a higher infection success in the second infection has been recorded (see Wächtler et al. 2001).In spite of these contradictory results, the acquired immunity in salmonid hosts against M. margaritifera glochidia has not received much attention by experimental studies.For example, an acquired immune response would mean that the success of M. margaritifera would largely depend on immunologically naïve 0+ host fish in the environmentwith important consequences for the conservation of the species.
The question remains whether the infection by A. anatina can cross-immunize brown trout against M. margaritifera glochidia such a phenomenon could pose a threat to declined M. margaritifera populations in rivers where these unionids live in sympatry.Nevertheless, this is still a widely under researched issue.Bauer et al. (1991) conducted an experiment by infecting brown trout first with A. piscinalis (= A. anatina) glochidia, and subsequently with M. margaritifera glochidia.No evidence for cross-immunity was achieved in this short-term (35 days) experiment.However, as the length of parasitic period of M. margaritifera can be even more than 300 days (Young & Williams 1984), a thorough evaluation of the likelihood and strength of this phenomenon should cover the whole parasitic period.
The aim of the present study is to investigate cross-immunity, i.e. whether the infection with glochidia of the duck mussel, A. anatina, will induce immunity against M. margaritifera in an experiment covering the whole parasitic period of M. margaritifera, as well as acquired immunity in salmonids hosts by M. margaritifera.The aim of the acquired immunity experiment is to examine the magnitude of acquired immunity in brown trout and the dose-dependent acquired immunity in Atlantic salmon against the glochidia of M. margaritifera, i.e. whether the intensity of immunity depends on the number of glochidia to which fish are exposed.Our hypotheses are that (i) Anodonta infection induces cross immunity against M. margaritifera glochidia, (ii) Atlantic salmon and brown trout develop acquired immunity against M. margaritifera and that (iii) the acquired immunity is dose dependent.

Materials and Methods
In all the experiments, an effort was made to fulfil the key requirements for unbiased procedures for priming and challenge infections of fish by mussel glochidia (see Taeubert et al. 2013).These included e.g., maintenance of experimental fish groups in identical conditions throughout the experiments, identical exposure of fish to glochidia, randomization and sufficient number of replicate fish individuals.

Cross immunity experiment
A total of 300 brown trout fry (age group 1+, River Iijoki stock) were transported from the Taivalkoski fish farm of the Natural Resources Institute Finland (Luke) to Konnevesi research station, University of Jyväskylä, on May 23, 2012.Fish had not been exposed (hereafter 'exposed' and 'infected' are synonymous) to M. margaritifera glochidia in the fish farm.Dissection and examination of the gills of five individuals verified that the trout were not previously infected by glochidia.Fish were randomly allocated into four 163 L flow-through tanks with 100, 100, 50 and 50 individuals per tank.Two-hour exposure of trout with glochidia dissected from A. anatina (collected from Lake Koijärvi, eastern Finland), was performed on May 24, 2012, by decreasing water volume to 70 L and adding 12.3 x 10 4 and 7.1 x 10 4 glochidia to the two tanks with 100 and 50 trout, respectively.The two control tanks holding another 100 and 50 fish per tank received A. anatina gill extract suspension without glochidia.Water temperature during the priming infection with A. anatina was 7.7 °C.
Five brown trout from the 100-fish-tanks and three from the 50-fish-tanks from both primed and control groups were examined for glochidia five days post infection.All primed fish were infected and the number of A.
anatina in primed fish varied from 90 to 232 glochidia fish -1 , indicating a successful priming infection.No glochidia were found from the control fish.On 15 August 2012, when the water temperature had increased to 15.7 °C (2.5 months post infection), one brown trout from each tank was examined and found uninfected, indicating that A. anatina glochidia had already excysted (see also Douda et al. 2013).
On 15 August 2012, fish were marked using fin clipping, and randomly re-allocated into four new 163 L flowthrough tanks so that all tanks received both primed and control fish.In every other tank the primed fish were fin-clipped and control fish unclipped while in every other tank the primed fish were unclipped and control fish fin-clipped.The number of primed fish per tank varied from 21 to 36 whereas the number of control fish per tank varied from 21 to 62.Both the fin-clipped and unclipped fish were anesthetized using MS-222 before marking and handled similarly, except for the clipping.
Challenge infection with M. margaritifera glochidia was done two weeks after marking, on 28 August, 2012, with glochidia from the River Jukuanoja (the River Iijoki catchment), northern Finland.The 2-hour exposure was performed technically as in the priming infection above, by adding 3.0 x 10 5 glochidia to all the four tanks.
Water temperature was 16.7 °C.Glochidia collection was performed by placing 30 adult M. margaritifera in plastic buckets in 5 L of river water for 30 min on the day of infection.The mussels were returned to the river after incubation.Timing of challenge infection was based on the previous knowledge that the River Jukuanoja Margaritifera release glochidia in the end of August (Salonen & Taskinen 2016).
Data were collected at four time points; September 2012 (3 weeks post infection), December 2012 (3 months), May 2013 (9 months) and June 2013 (10 months) (Table 1).Primed and control fish were randomly collected, killed with a sharp blow on the head, and measured for the total length and fresh mass.The gills were cut off and glochidia were examined microscopically for the number and size (length from a subsample of 10 random larvae), except for September sampling when only the right side gills were examined.Therefore, only the data for the right side gills were used in statistical analyses.Throughout the experiment, fish were daily fed with commercial food pellets.During this phase of the experiment the minimum and maximum temperatures were 1.1 °C and 16.8 °C, being the highest in September 2012 and June 2013.

Acquired immunity experiment
Testing of acquired immunity was performed for both of the salmonid host species of M. margaritifera, Atlantic salmon and brown trout.Both host species individuals (age group 0+) originated from the River Iijoki stock reared at Taivalkoski fish farm of the Natural Resources Institute Finland (Luke), from where they were moved to Konnevesi research station on August 21, 2012.
In the brown trout experiment, fish were first randomly allocated to primed vs. control groups in two separate 163 L flow-through tanks with 50 fish per tank.Priming infection of trout was performed on August 28, 2012, using similar methods and origin of M. margaritifera glochidia as in the cross immunity experiment, with 2.9 x 10 5 glochidia tank -1 .Control fish were not exposed to glochidia, but experienced otherwise the same treatment as the primed fish.Next year, in August 26, 2013, when age of the fish was 1+, adipose fins of the primed fish were cut, after which both the primed and control fish were put in one tank.On August 28, 2013, the fish were infected with M. margaritifera glochidia (2.3 x 10 5 glochidia tank -1 ) collected from the River Koivuoja (the River Iijoki catchment), northern Finland.In November 25 (3 months post infection), all the fish were examined with the methods described above (Table 1).
Atlantic salmon were primed at the same time (August 28, 2012), with same origin of glochidia and with the same methods as brown trout mentioned above.Salmon were allocated to two tanks with three treatment groups in each, (1) primed with a high dose (8.8 x 10 5 M. margaritifera glochidia, tip of the right pectoral fin clipped), (2) primed with a low dose (1.7 x 10 5 M. margaritifera glochidia, tip of the left pectoral fin clipped), and (3) control group (not prime infected, adipose fin clipped) with 17, 14 and 13 fish per tank, respectively.After one year, in August 28, 2013, all salmon were infected with M. margaritifera glochidia collected from the River Luttojoki (the River Tuuloma catchment), northern Finland.The challenge infection was performed with 6.0 x 10 5 glochidia tank -1 .As brown trout above, all salmon were examined on November 25, 2013 (3 months post infection) (Table 1).Throughout the experiments, fish were daily fed with commercial food pellets.Due to lack of logistic supports, the dose-dependence experiment was not performed for brown trout.

Statistical analyses
The effect of the previous infection with either A. anatina (cross immunity) or M. margaritifera (acquired immunity) and other factors (month, dose, tank) on glochidia number in gills and the size of glochidia was analysed by ANOVA.If the tank effect was not significant, the analysis was reduced to the effect of other variables.If necessary, the response variables studied were transferred by Box-Cox-transformation {BCSN = (N  -1)/} to yield as normally distributed variable as possible within each treatment cell.In some cases the distribution within a treatment cell still deviated significantly from normal, which induced a tendency for incorrect rejection of H 0 -hypothesis (bias for too low p-value).Therefore, if the H 0 was rejected (p < 0.05) the hypothesis was also tested using more conservative non-parametric tests (e.g.Kruskall-Wallis).Fish were not measured for length and weight at the time of exposure but when examined.Therefore, the number of M. margaritifera in the gills was not standardised based on the size of the fish as the individual growth rate after exposure, and consequently the size during exposure, was not known.

Cross immunity experiment
Previous infection with A. anatina glochidia had no statistically significant effect (ANOVA) on the number of M. margaritifera glochidia (Fig. 1) or glochidium size (Fig. 2).Thus, brown trout does not develop any nonspecific immunity that would decrease the success of M. margaritifera to parasitize them.Month had a significant effect (ANOVA and Kruskall-Wallis p < 0.001) on the response variables: the number of glochidia declined during the incubation period (Fig. 1) and their size (Fig. 2) increased.

Acquired immunity experiment
The previous infection of brown trout with M. margaritifera had a significant effect on the number of glochidia (Mann-Whitney U, p < 0.001) when re-infected with M. margaritifera glochidia (Fig. 3).Brown trout individuals earlier exposed to glochidia had significantly (Tukey and non-parametric pairwise test p < 0.01) less glochidia than the control (no exposure) group.No significant difference (p> 0.05) in the size of glochidia between the control and exposed group was found (Fig. 4).
Previous infection of Atlantic salmon with M. margaritifera had a significant dose-dependent effect on the number of glochidia (ANOVA and Kruskall-Wallis p < 0.001) when re-infected with M. margaritifera glochidia (Fig. 5).Salmon individuals exposed to high dose of glochidia had significantly (ANOVA with Tukey and Kruskall-Wallis with non-parametric pairwise test p < 0.01) less glochidia than the low dose group or the control group.The difference in the number of glochidia between the control and low dose treatment groups was not significant.
In addition, previous infection with M. margaritifera had a significant effect (p < 0.05) on the size of M. margaritifera glochidia when re-infected (Fig. 6).Glochidium size was significantly smaller in the high dose treatment than in the control group (p < 0.05).The tank effect was also significant with one tank having larger glochidia than the other of the two replicate tanks, presumably due to higher water temperature.Thus, previous infection with M. margaritifera led to both lower intensity of infection and poorer growth of M. margaritifera larvae when re-infected.

Discussion
In natural populations, individuals are usually infected not only by one but also with multiple parasitic species.
Interaction between the co-infecting parasitic species within a host individual can be negative (antagonistic), leadingin extreme casesto competitive exclusion (Holmes 1961).However, species can also be independent of each other or the interaction can be even positive (co-operation, facilitation) (Poulin 2001, Lello et al. 2004).
Thus, the interaction between A. anatina and M. margaritifera could also lead to one of these three possible outcomes.Because the earlier study by Dodd et al. (2005) showed that the previous infection of the host fish with the glochidia of Lampsilis reeveiana lowered the infection success of other unionid mussels, L. abrupta, Villosa iris and Utterbackia imbecillis, we hypothesised that the effect of A. anatina infection on M. margaritifera would be negative, or at most insignificant.We also hypothesised that if the effect of A. anatina infection on M. margaritifera would be negative, it would be due to cross-immunityrepresenting so-called immune-mediated 'apparent competition' where one parasite species elicits an immune response which harms its competitors (see Read & Taylor 2001).In natural conditions, the possible negative impact of A. anatina on M. margaritifera could be also due to direct interference competition.M. margaritifera glochidia that occupy brown trout gills in autumn could be interfered by the glochidia of A. anatina in the spring when the glochidia shedding of the latter species takes place.However, in the present study, the exposure of brown trout was sequential so that A. anatina infection occurred earlier (in spring/early summer) and that of M. margaritifera started in autumn; glochidia of only one species was present in brown trout at a time.
The acquired (adaptive) immune system of vertebrates activates slowly, but brings a specific and long-lasting immunity against subsequent infections.However, the acquired immune response developed against one parasite genotype may be cross-reactive and provide protection against other genotypes of the same species (e.g., Rellstab et al. 2013), or even to those of different species (e.g., Dodd et al. 2005, Karvonen et al. 2009).
From our point of view, in the parasite, two factors determine the importance of cross immunity in the case of multiple infections.First, probability of cross immunity decreases with the genetic distance between the infecting parasites strains/species (Read & Taylor 2001).Second, the sequence of infections influences on the relative benefits and costs of cross immunity.In sequential exposure, only the first parasite enjoys the slow activation of the adaptive immune system whereas the later arrival bears the full costs of the acquired immunity (see Jackson et al. 2006, Hoverman et al. 2013, Klemme et al. 2016).
The previous experimental study on the acquired immunity in brown trout fish host against the glochidia of M. margaritifera (Bauer & Vogel 1987) suggested acquired resistance; brown trout developed humoral immunological response and the infection success decreased in repeated infection.Our results verified this finding.Three months after the challenge infection the number of M. margaritifera glochidia was lower in individuals infected 1 y earlier with M. margaritifera glochidia than in the control group that were not previously exposed to M. margaritifera.Furthermore, evidence for acquired immunity was obtained not only in brown trout but also in the Atlantic salmon, indicating that both of the two suitable host fishes of M. margaritifera (see e.g., Salonen et al. 2016) are able to mount an acquired immune reaction against M. margaritifera.
In addition to the number of glochidia, the negative impact of previous infection on success of M. margaritifera was seen also in the growth rate of glochidia.When measured 3 months after the challenge infection, glochidia in previously infected fish were smaller than in control fish in the Atlantic salmon.The size of glochidium at the time of excystment from the fish host correlates with the survival rate of the juvenile M. margaritifera (Eybe et al. 2015).Thus, previously infected hosts produce less and lower quality juveniles than immunologically naïve hosts.
A novel finding in the present study was the dose dependence of acquired immunity in M. margaritifera-fish host relationship.The higher the number of M. margaritifera glochidia that the fish were exposed to in priming the lower the number of glochidia, and the smaller their size, after re-infection with M. margaritifera glochidia.Dose dependence was evident in Atlantic salmon but since the dose dependence has been earlier observed in immunization and vaccination of salmonids (Munag'andu et al. 2013, Ballesteros et al. 2015) it is reasonable to assume that the result can be extrapolated also to brown trout.Thus, the negative effect of the previous infection on both the number and the quality of M. margaritifera glochidia (and juveniles) depends on the density of glochidia in the previous exposure.
The acquired immunity could explain the previous contrasting findings of M. margaritifera infection rate with respect to host fish age.In some studies the infection rate has been lower in older host fish (Bauer 1987, Hastie & Young 2001) while in some studies the opposite was found (see Wächtler et al. 2001).For example, if production of glochidia does not take place every year in a particular M. margaritifera population, there can be years in which both the 0+ and 1+ age group fish are immunologically naïve with respect to M. margaritifera glochidia.In such a condition, the larger sized 1+ fish, due to their large gill area, are probably more intensively parasitized by M. margaritifera glochidia than the 0+ fish that is also supported by Geist et al. (2006).During the year that follows production of M. margaritifera glochidia, the negative effect of acquired immunity may override the positive effect of larger size among the 1+ age group fish, resulting in situation where the younger and smaller but immunologically naïve 0+ individuals are more heavily infected by M. margaritifera than the 1+ fish.The dose dependence of acquired immunity can strengthen this process.
Importantly, the acquired immunity emphasizes the importance of the availability of 0+ age group fish for M. margaritiferaand explains the association between the density of 0+ fish hosts and density of young M. margaritifera in the population (Bauer 1987).For the conservation of M. margaritifera, therefore, the availability of 0+ aged (immunologically naïve) hosts is essential.In other words, the acquired immunity would mean that the recruitment success to post-parasitic life stage of M. margaritifera could strongly depend on the abundance of immunologically naïve 0+ host fish in the environment.
Our results suggest that the cross immunity between M. margaritifera and A. anatina is not as important an impediment for the success and conservation of M. margaritifera as the acquired immunity.Brown trout primed with glochidia of the unionid mussel A. anatina did not harbor significantly lower number of glochidia when challenged with M. margaritifera.The pattern was consistent throughout the 9-month parasitic period of M. margaritifera.This is in line with the results of the short term experiment performed earlier: previous infection with A. anatina had no influence on the survival of M. margaritifera glochidia in brown trout within 35 d (Bauer et al. 1991).That study also showed that previous infection with another unionoid species, Unio crassus, had no influence on the survival of M. margaritifera glochidia in brown trout (Bauer et al. 1991).Together these results indicate that the exposure of host fish to glochidia of other mussel species would not pose a threat to the endangered freshwater pearl mussel, M. margaritifera.However, the present study shows that the immunity is related to the dose of exposure.Thus, possibility of such a cross immunity cannot be ruled completely out, for example, if brown trout is heavily exposed to A. anatina glochidia.
As hypothesized, acquired immunity and its dose dependence existed in M. margaritifera-host fish relationship.However, our experiment, as the previous study by Bauer et al. (1991) did not find evidence for cross immunity between M. margaritifera and A. anatinacontrasting the earlier study by Dodd et al. (2005) conducted between two Unionidae species.It is possible that M. margaritifera (family Margaritiferidae) and A. anatina (family Unionidae) are immunologically so distant that the antibodies produced for one species do not protect against the other species.

FiguresFigure
Figures Figure 1.Box-plot showing the number of M. margaritifera in the gills of brown trout at different times after challenging with M. margaritifera on 28 August, 2012.No exposure = the fish were not exposed to A. anatina before challenging with M. margaritifera, exposure = the fish were exposed to A. anatina before challenging.The box indicates range between lower and higher quartile, the vertical line in the box is median and the whiskers indicate minimum and maximum values, excluding outliers (values deviating more than 1.5 interquartile ranges from the closest quartile).

Figure 2 .
Figure 2. Box-plot showing the length (m) of M. margaritifera glochidia at different times.No exposure = the fish were not exposed to A. anatina before challenging with M. margaritifera, exposure = the fish were exposed to A. anatina before challenging.

Figure 3 .
Figure 3. Box-plot showing the number (N) of M. margaritifera glochidia in the gills of brown trout in individuals previously not infected with M. margaritifera (no exposure) and infected with M. margaritifera glochidia (exposure).

Figure 4 .
Figure 4. Box-plot showing the length (m) of M. margaritifera glochidia in the gills of brown trout in individuals previously not infected with M. margaritifera (no exposure) and infected with M. margaritifera glochidia (exposure).

Figure 5 .
Figure 5. Box-plot showing the number (N) of M. margaritifera glochidia in the gills of Atlantic salmon in individuals previously not exposed to M. margaritifera (zero dose), exposed to a low number of M. margaritifera glochidia (low dose) and to a high number of M. margaritifera glochidia (high dose).

Figure 6 .
Figure 6.Box-plot showing the length (m) of M. margaritifera glochidia in the gills of Atlantic salmon in individuals previously not infected with M. margaritifera (zero dose), infected with a low number of M.margaritifera glochidia (low dose) and with a high number of M. margaritifera glochidia (high dose).