Influence of introduced vs. native parasites on the body condition of migrant silver eels

Because parasitism is among the reasons invoked to explain the collapse of Anguilla anguilla, we evaluated the parasitic constraint on body condition (BC) of migrant silver eels as a proxy of fitness with inter-site comparisons. Metazoan parasites were studied in 149 silver eels from five sites (northern Europe). In total, 89% were infected by 13 species including Myxozoa, Monogenea, Cestoda, Nematoda, and Acanthocephala. Anguillicoloides crassus was most common (56%), then Acanthocephalus clavula (30%), and Pseudodactylogyrus sp. (17%). BC, calculated for 58 females, was negatively correlated by abundance of the introduced Pseudodactylogyrus sp. but not by other parasite taxa. Nevertheless, the introduced A. crassus was considered as a severe pathogen based on previous data, whereas the native A. clavula was supposed to have limited impact. Parasite component communities and BC were different between sites. Silver eels from Stockholm Archipelago (Sweden) were the least parasitized (40% vs. 90–95% for other sites) with no parasites on the gills. Burrishoole (Ireland) differed by the absence of A. crassus and high prevalence of A. clavula (84%) but without consequences on BC. Gudenaa (Denmark), Corrib (Ireland), and Frémur (France) were close due to high prevalence of A. crassus (89–93%). Gudenaa and Corrib were the most similar because Pseudodactylogyrus sp. was also highly prevalent (respectively 71% and 60%) whereas absent in Frémur. Our results suggest that the fitness loss induced by the introduced parasites could affect the spawning success of migrant silver eels from Gudenaa and Corrib, and to a lesser extent from Frémur, but probably not those from Stockholm Archipelago and Burrishoole.


Introduction
In spite of the perennial scientific mysteries about its biology and population genetic structure, the European eel Anguilla anguilla is considered as panmictic, therefore a breeder emigrating from a given river theoretically contributes to subsequent glass eel recruitment along the whole continental distribution range [5,16,21]. Depending on subpopulations and habitat characteristics, eels spend about 3-30 years in fresh and brackish waters of Europe and North Africa, growing and accumulating fat reserves for an active swimming migration as silver eel stage across the Atlantic Ocean and reproduction once in the Sargasso Sea [36]. The probability that silver eels reach the spawning grounds and reproduce successfully is likely to vary greatly among continental growing sites [4,18]. A pan-European methodology to estimate the overall breeding potential of silver eels according to relevant criteria (e.g., fat composition, contamination by chemicals, parasitic load) is not ready for use because of the complexity in implementing such an approach. The use of silver eel condition indices may constitute a first step in this direction. Body condition has been demonstrated to indicate energy reserves of Salmo salar [59], Coregonus artedi [54], and Gadus morhua [25,40]. Considering the importance of energy reserves during both transoceanic migration and reproduction of the European eel [7], we hypothesize that body condition may represent a good proxy of silver eel fitness [4,31].
Because most parasitized organisms are generally in poorer condition than unparasitized ones [15,22,61 for reviews] and because parasites interact with natural and anthropogenic stressors to increase mortality and reduce animal health in myriad ways [47 for review], parasitism is among the reasons invoked to explain the European eel collapse [9,20,28,35]. Up to now, 161 species or taxa have been described in A. anguilla from fresh, brackish, and marine waters in 30 countries of Europe and North Africa [32 for review]. Among them, the introduced Anguillicoloides crassus (swimbladder nematode) and Pseudodactylogyrus sp. (gill monogenean) are considered as an important factor able to induce a stress that probably decreases the host fitness and seriously hampers the recovery of the European eel [for reviews : 9, 35]. The scientific research program EEL-IAD (2008-2012) aimed to resolve some of the mysteries of eel biology in order to help conserving European eel stocks. EELIAD provided us the opportunity to investigate the parasitism patterns of some organs (gills, heart, intestine, and swimbladder) of silver eels sampled at the start of their transoceanic migration across five far-distant sites from northern Europe, i.e., Sweden, Denmark, Ireland (two sites), and France. Our objectives were (i) to describe the metazoan parasite community of these five silver eel subpopulations, and (ii) to evaluate in each site the potential parasitic constraint (on the whole and depending on parasite taxa) on the body condition of silver eels. Because a greater pathogenicity is often observed for recent (vs. long-term coevolved) host-parasite associations as for Pseudodactylogyrus sp. and A. crassus in A. anguilla [33,34,48], we suppose that the least parasitized silver eels by these introduced species could be the most susceptible to reach the spawning grounds and to reproduce in the Sargasso Sea compared to the most heavily infected ones.

Materials and methods
Study areas and silver eel sampling ( Figure 1; Table 1) Silver eels (149) were caught during their seaward migration between October and February (i.e., at the beginning of their transoceanic migration) in five coastal water bodies from northern Europe: Stockholm Archipelago in Sweden The eel collection was focused on silver eel stage at the same maturation degree (see below the three criteria for silver eel assessment). Migrant silver eels were intercepted using pound nets (SWE-ARCH), coghill nets (IRE-COR), or eel-traps installed on dams (DEN-GUD, IRE-BUR, and FRA-FRE) depending on the site, these different methods of capture being adapted to both the size of the eels and the studied systems. Water bodies sampled differed in trophic status and level of anthropogenic pressure (Table 1). Among sites, IRE-BUR is considered as the pristine reference site due to the clean water and null anthropogenic pressure (no fishery, controls on movement of boat, water, or fishing gear).
Silver eels: Development stage and sex determination, aging by otoliths Each fish collected was anesthetized with Benzocaine at a concentration of 0.15 g L À1 . Then, total weight (TW, g) and total length (TL, mm) were measured respectively to the nearest g and mm. Silver eels were identified by three criteria [1,2]: color of the back and belly, presence of a well-defined lateral line, as well as ocular index (OI ! 6.5, Pankhurst's 1982).
The sex was assigned by macroscopic observation of the gonads using the criteria described by Colombo et al. [14]. Because most silver eels sampled were females (132 vs. 17 males), we only selected female eels to study the parasitic impact on the eel body condition.
As a proxy of fitness, we used index developed by Le Cren [41] by calculating the body condition (BC) as the relative total weight TWr = 100 TW/TWstd, where TWstd is the predicted standard total weight of a fish at the same TL, as calculated according to the log 10 TW À log 10 TL regression equation (least squares means fit) for the whole sample of 132 female silver eels.
A sample of 58 female silver eels (SWE-ARCH, n = 4; DEN-GUD, n = 10; IRE-BUR, n = 17; IRE-COR, n = 11; FRA-FRE, n = 16) was aged by examination of the sagittal otoliths. The extracted otoliths were glued dorsal side up with crystal bond, grounded along the longitudinal plane until the nucleus was reached. Following the method described by the EIFAC ICES working group on eel [30], the fish continental age (number of years in fresh water) was determined through a stereomicroscope by counting the number of annuli from the first growth check outside the so-called zero band. This band is commonly assumed as the beginning of eel's continental growth.

Parasitological research in sampled Anguilla anguilla
All the 149 silver eels sampled were frozen before the research of metazoan parasites made on the gills, the heart, the digestive tract, and the swimbladder (the other organs were not available for parasitic investigation). Each organ was dissected using binocular microscope and all the metazoan parasites were numbered per organ and per eel (except for Myxozoa for which the presence or absence of cysts was recorded), identified and preserved in alcohol 70°. The parasitological parameter used to describe the community structure of parasites was species richness (number of parasite species in a sample of hosts); those to describe parasite populations (for a given parasite species) were prevalence (number of hosts infected with a particular parasite species/number of examined hosts), mean abundance (average abundance of a parasite species among all members of a host sample), and mean intensity (total number of parasites of a particular species found in a sample divided by the number of hosts infected with that parasite) [10].

Statistical analyses
All the statistical analyses were made with the R-Cran project free software (http://www.r-project.org/). Differences were considered statistically significant at p < 0.05. Mean values of data are reported as means ± standard error (SE) except for prevalences ± 95% confidence limits (CL).

Inter-site comparison of the communities of parasites
Taxa with low prevalence ( 2%) were excluded of the analysis (Acanthocephalus lucii, Echinorhynchus truttae, Eustrongylides sp., Pomphorhynchus laevis, and Raphidascaris acus) because of their rarity and probable limited impact on silver eel subpopulations.
First, the structure of the component parasite communities in silver eels (n = 149) among the five sites was detected and represented by a Multiple Correspondence Analysis (MCA, Ade4 package) performed on the presence/absence of eight parasite taxa (Myxidium giardi, Pseudodactylogyrus sp., Bothriocephalus claviceps, A. crassus, Paraquimperia tenerrima, Camallanus lacustris, Acanthocephalus clavula, and Acanthocephalus anguillae). Second, similar tests (MCA) were focused on the abundances of the three major parasite taxa (i.e., the two introduced A. crassus and Pseudodactylogyrus sp., and the native A. clavula) that could exert a significant constraint on silver eels, potentially compromising the migration and spawning success [28,35]. In factorial maps (Figures 2 and 3) realized from MCA results, ellipses are only visual summary. The prob-ability to be located in the ellipse is p = 1 À exp(Àk 2 /2) with k = 1.5 showing that around 67% of observations were located in the ellipses.

Factors influencing the body condition of female silver eels
Five variables which could explain BC were chosen: the age of female silver eels, the study-site, and the number of each of the three major parasite taxa (A. crassus, A. clavula, and Pseudodactylogyrus sp.). Their influence was analyzed using generalized linear models (glm, base package). We tested all possible combinations with these variables including interactions. Nineteen different models were tested and compared using classical selection by both step-akaike information criterion corrected from small sample sizes (AICc) and deviance explained (DE) [29]. Then, the effect of parameters was analyzed by analysis of variance (ANOVA) from the selected model. As the sample size (58) used for the glm analyses was small according to the number of study-sites and variables tested, a power test using pwr package (pwr.f2.test) was performed on the selected model in order to give confidence in the acceptation of the null hypothesis (the fitness loss was not affected in some localities).

Component communities of metazoan parasites in each subpopulation of Anguilla anguilla
According to the MCAs (Figures 2 and 3), the structure of the parasite component communities (Table 3)   Second, the intestine was heavily infected with a total prevalence of 90% (CL = 78-97%), mostly by the acanthocephalan A. clavula that was highly occurrent in terms of prevalence (84% with CL = 71-93%), mean intensity (30.51 ± 4.40), and mean abundance (25.53 ± 4.02). A second site, SWE-ARC, was also significantly different from the others (even with the small sample size of 10 fish compared to other sites) due to an overall prevalence less than half (40% with CL = 12-74% vs. 90-95% with CL = 77-100% in other sites) and to a total absence of parasites on the gills ( Table 3). The three other sites: DEN-GUD, IRE-COR, and FRA-FRE, were characterized by high prevalences of A. crassus varying from 89% (CL = 67-104%) to 93% (80-99%). Among all sites, DEN-GUD and IRE-COR were the most similar because of the highest species richness (eight parasite taxa vs. four and five in other sites) and of the highest prevalences of Pseudodactylogyrus sp.; no monogenean was found in silver eels from FRA-FRE (Table 3).

Influence of parasitism on the body condition of female silver eels
Results from the 19 linear models are summarized in Table 4. According to the two criterions of selection (AICc and DE), the model #18 was selected. Results of ANOVA performed on this model are summarized in Table 5. The selected model explained 42.43% of the total deviance, and 33.10% of this explained deviance was explained only by the site (or 14.04% of total deviance, p = 0.013). The combined effect of female age and study-site was responsible of 27.70% of the explained deviance (p = 0.061). Among the three major parasite taxa, the number of Pseudodactylogyrus sp. had a strong negative effect on the BC of female silver eels (p = 0.003) and explained 27.10% of the explained deviance. The number of A. crassus tends to have a positive effect on the BC (p = 0.048). Model selection using step-AICc process showed that A. clavula had no significant effect on the BC. Indeed, AICc and explained deviance were not different between models #18 and #19, this latter being similar to #18 but with A. clavula as additional variable (Table 5).

Discussion
Whatever the host-parasite combination, parasites and their host compete for resources in such a way that both survival and fecundity of the host could be affected, even if no pathology is obvious and even if this effect may be drowned in the background noise of all other factors that affect survival and fecundity [15,22,52,61 for reviews]. Other environmental stressors (e.g., pollutants, pathogens) that can also influence the health status of the eels [3,23,26,56] could be considered as a potential confounding factor, but in most cases, contaminants and parasites occurring together are shown to exacerbate the detrimental effects on individuals (synergistic effects), suggesting Table 3. Component communities of metazoan parasites in Anguilla anguilla (149) from the five sampling sites: species richness, prevalence (P%), mean intensity (I ± SE) and mean abundance (A ± SE). See abbreviations in Tables 1 and 2.  Myxidium giardii cysts (Mgia) were not numbered in the gills. CL95% was indicated in parentheses for total prevalence per site.
that parasitized fish in polluted environments are in a poorer condition than unparasitized fish [47 for review]. Thus, to consider parasitism and its impact can help to understand the collapse of the European eel. In our study, 89 ± 5% of all silver eels were parasitized and 13 metazoan taxa were identified including Myxozoa, Monogenea, Cestoda, Nematoda, and Acanthocephala commonly recorded in A. anguilla [32 for review]. No Trematoda were found from any site in spite of 39 trematode species described in A. anguilla as definitive host [32]. One possible explanation is that our silver eels were originated from waters with salinity 5 psu and that trematodes infecting eels are shown to be significantly less frequent in fresh vs. marine waters [37,38]. The body condition of silver eels was used as a proxy of fitness that can reveal the impact of parasite infections. Among the three major parasite taxa recorded in the female silver eels studied here (A. crassus, A. clavula, and Pseudodactylogyrus sp.), only Pseudodactylogyrus sp. decreased the host body condition in relation with increasing abundance (up to 80 worms) on the gills. Pseudodactylogyrus sp. is the third more prevalent taxon in this study (17 ± 6%) and a specific gill monogenean of the genus Anguilla transferred from its native host Anguilla japonica to A. anguilla after introduction in Europe in 1977 [33,35,62]. Pseudodactylogyrus sp. is browsing the host gill surface and induces epithelial lesions potentially leading to lethal hypoxia of the highly susceptible European eel and facilitating infections by various opportunist pathogens (virus, bacteria, and fungi) [33,35,62].
In contrast with Pseudodactylogyrus sp., the infection by the introduced hematophagous A. crassus, the most prevalent species (56 ± 8%), had no significant effect on the condition factor of A. anguilla, as also shown by other studies [58 for review], and even tended to have a positive effect with increasing nematode abundance in the swimbladder. But this criterion is not the best reflecting the eel pathogenicity due to the short life cycle of A. crassus, and severely damaged swimbladders are shown to harbor very few or even no living nematodes [42,63]. Nevertheless, using the Swimbladder Degenerative Index, Lefebvre et al. [44] recently demonstrated that the most affected eels had greater body length and mass (+11% and +41% respectively) than unaffected eels of the same age. Despite these surprising counterintuitive results, high virulence and severe impacts of A. crassus are expected because A. anguilla lacks an adaptive immune response, and various pathogenic potentially lethal effects (e.g., anemia, energy drain, swimming performance decrease) have been demonstrated, threatening the success of spawning migration in the Sargasso Sea [35,43,44,53,58].
The body condition of female silver eels was also not influenced by the second most prevalent parasite of our study, A. clavula (30 ± 7%), a native generalist acanthocephalan with A. anguilla as preferred definitive host [8,11]. The apparent unchanged fitness of A. clavula-infected eels is in accordance with the absence of pronounced symptoms of disease generally observed for most fish infected by acanthocephalans (including those with high parasite intensities) [60 for review].
Component communities of metazoan parasites and body condition of silver eels widely varied between sites, suggesting differences in biocenosis (e.g., host species occurrence) and environmental constraints or stressors (including parasitism), with probable consequences on the spawning migration and reproduction success. Based on the inter-site differences and the virulence of Pseudodactylogyrus sp. and A. crassus, the silver eels from Stockholm Archipelago (Sweden) and from Burrishoole (Ireland) could be the most successful, whereas the migration success may be affected by the pathogenic effects of introduced parasites for the silver eels from the other sites studied, i.e., Frémur (France), Corrib (Ireland), and Gudenaa (Denmark). Indeed, only four out of the ten Swedish silver eels were infected and always with a very low parasite intensity. Moreover, they harbored no parasites on their gills and had both low A. crassus intensity (one helminth per swimbladder) and prevalence (three out of ten fish) suggesting low parasite pathogeny. Therefore, despite the low sample size, one can suppose that metazoan parasites have probably a limited influence on the condition of Swedish silver eels and on their migration success.
Burrishoole is considered as a pristine site with oligotrophic acid waters, no anthropogenic influence, and low levels of organic contaminants [46]. The growth rate of eels from this site is known to be extremely low [55], thus we consider that the one-year sooner sampling date (compared to the four other sites) had no significant impact on the eel parasite community and did not introduce a bias in our inter-site comparison. In our study, 90 ± 8% of the silver eels from Burrishoole were parasitized but this high total prevalence was not synonym of a strong parasitic constraint since they were mainly infected by the native intestinal acanthocephalan A. clavula (prevalence of 84 ± 10%, intensity of 31 ± 4) for which we demonstrated no influence on the host body condition. A. crassus was not recorded suggesting that the nematode may not complete its entire life cycle in Burrishoole despite its numerous intermediate crustacean (ostracods, copepods) and paratenic fish (and even snails) hosts [27,49]. Moreover, no parasitic pathology probably occurred in the gills of Burrishoole silver eels because of the rarity of Pseudodactylogyrus sp. (only two infected out of the 49 specimens, harboring only one or three monogeneans).
Despite differences in the anthropogenic influence and the trophic status of Frémur (France), Corrib (Ireland), and Gudenaa (Denmark), A. crassus was omnipresent in A. anguilla from these sites with a prevalence varying from 89% (67/104) to 93% (80/99) probably inducing pathogenic effects [35 for review]. In addition, for 60% (36/81) and 71% (49/87) of the Corrib and Gudenaa silver eels, hypoxia due to the impaired gills was probably a major symptom as they were severely infected by Pseudodactylogyrus sp. with a mean intensity of 23 ± 12 and 19 ± 6 respectively. The parasite constraint appeared more substantial for the silver eels from Corrib and Gudenaa having both damaged swimbladder and gills than for those from Frémur since Pseudodactylogyrus sp. was absent.
Finally, based on the potential fitness loss induced by parasitism, we suppose that the migrant silver eels from Stockholm Archipelago (Sweden) and Burrishoole (Ireland) are able to contribute to the recruitment and gene pool of A. anguilla pop-ulation, whereas those from our other study-sites [in particular from Corrib (Ireland) and Gudenaa (Denmark)] have a lower probability to reach the spawning grounds in Sargasso Sea.
Numerous aspects remain to be explored to explain the decline of the European eel population, probably induced by various interacting abiotic and biotic factors (e.g., habitat loss or fragmentation, changing hydrology, overfishing, pollution, and pathogens [9,20,21,28,36,56]) and resulting in difficult and complex measures of preservation of this species. Enhancing the production of viable silver eels by water system represents the current target of the conservation strategy of the European eel [19]. However, we believe that the question of animal quality among river systems, which is presumed to influence the reproductive success, is also a key issue that must be urgently pursued for European eel conservation. Every potential impact on silver eel body condition warrants examination, more especially as synergistic effects can occur between environmental stressors such as parasites and contaminants, increasing the mortality of exposed organisms [47 for review].