Labile resistance of Atlantic salmon, Salmo salar L., to infections with Gyrodactylus derjavini Mikailov, 1975: implications for host specificity

A number of investigations have shown a clear host specificity of gyrodactylids parasitizing salmonid hosts. Gyrodactylus derjavini Mikailov, 1975 succesfully infect brown trout, Salmo trutta L., but cannot reproduce satisfactorily on Atlantic salmon, S. salar L. (Buchmann & Uldal 1997; Bakke, Soleng & Harris 1999). In contrast, the congener G. salaris Malmberg, 1957 preferentially infects the latter host, whereas brown trout does not support significant population growth of this parasite (Bakke, Jansen & Hansen 1990; Bakke et al. 1999). The basic mechanisms behind this strict host specificity or differential resistance have remained enigmatic. However, some studies have shown that treatments of hosts with corticosteroids, such as hydrocortisone and dexamethasone, modulate this association significantly. Thus, brown trout injected with the former steroid become susceptible to G. salaris (Harris, Soleng & Bakke 2000) and salmon from the Scottish (Conon) strain of Atlantic salmon treated with the latter hormone support population increases of G. derjavini (Nielsen & Buchmann, in press). It has been shown that the resistance of salmon to G. salaris differs according to their origin. Thus, Neva salmon from the Baltic are less susceptible to this particular parasite compared with Norwegian salmon (Bakke et al. 1990) and Baltic Iijoki salmon are relatively resistant to G. derjavini (Buchmann & Uldal 1997). Whether the resistance of Baltic salmon towards the latter gyrodactylid is sensitive to steroid treatment of the host is an important question. Thus, the present work elucidates the influence of injection with the corticosteroid dexamethasone on the host susceptibility of both Atlantic and Baltic strains of salmon to infection with G. derjavini.

Two strains of Atlantic salmon were used for the experiments. Eyed eggs from the River Lule in Sweden and from the River Conon, Scotland (delivered by the Danish Wild Salmon Centre, Randers, Denmark) were brought to the hatching facility in spring 2001. The Conon salmon stock was previously found susceptible to infections with G. salaris (Bakke & MacKenzie 1993) and relatively resistant to G. derjavini (Buchmann & Uldal 1997). The status of the Baltic Lule salmon in this regard has not been investigated previously. Both egg batches were hatched and reared at the salmon hatchery of Bornholm (Nexø, Denmark) in pathogen-free conditions. Approximately 8 months post-hatching the fish were brought to the laboratory for at least 4 weeks acclimatization before experimentation. The salmon used had a mean body weight of 2.3 g (SD 0.6) and a mean body length of 7.0 cm (SD 0.6) (n=70). As a control for parasite reproductive capacity, rainbow trout, Oncorhynchus mykiss (Walbaum), a susceptible host, were infected and treated as salmon. These trout [mean body weight 3.0 g (SD 0.3) and mean body length 6.5 cm (SD 0.4) (n=10)] were also hatched and reared at the same salmon hatchery.

A strain of G. derjavini originally isolated from the Danish trout farm ‘Mosbjerg’ was maintained on rainbow trout (fresh water at 12 °C) in the laboratory 1 year before experimentation. Municipal water at pH 7.3, calcium carbonate 390 mg L⁻¹ and no detectable nitrite and nitrate was used in the trials. Temperature during the experiments varied between 8 and 12 °C. Fish to be treated were injected intraperitoneally (i.p.) with 200 μg Dexadresone^® (Intervet, Skovlunde, Denmark) (dexamethasone sodium phosphate, 100 μL fish⁻¹). All injections were performed on anaesthetized fish. Fish (for infection, injection or parasite counting) were anaesthetized by immersion in MS222 (tricaine methane sulphonate; Sigma, St Louis, MO, USA) (50 mg L⁻¹).

Two types of infections were performed. Some fish were infected by placing one parasite only on each fish (isolated infections), other fish were infected by cohabitation. In isolated infections anaesthetized fish (Lule salmon) were infected by placing a single specimen of G. derjavini on the tail fin of individual fish. Isolated infected fish were then placed in plastic aquaria (total volume 4 L) containing 3 L water (daily water change). Thus, micropopulations could develop from one founder parasite on susceptible hosts. Ten fish were used in each experimental group. Results were compared with similar previous experiments using Conon salmon of the same size (Nielsen & Buchmann, in press) (Tables 1 & 2). In cohabitation infections heavily infected rainbow trout (three fish harbouring more than 300 parasites) were used as donor fish. They were placed in a 128 L aquarium (aerated with internal Eheim biofilters) with 40 salmon: 10 non-treated Conon salmon, 10 dexamethasone treated Conon salmon, 10 non-treated Lule salmon and 10 dexamethasone treated Lule salmon. The parasites were allowed to spread from trout to salmon for 2 weeks before infected rainbow trout were removed. The number of parasites on all fish was monitored by weekly counting for 5 weeks. Anaesthetized salmon were placed (covered with water) under a dissection microscope (magnification 7–40×) and parasites were counted. The microhabitats of parasites on all body parts were recorded (Buchmann & Uldal 1997). The mean intensity and prevalence (Bush, Lafferty, Lotz & Shostak 1997), and the total parasite population in each group were calculated and the Mann–Whitney U-test was conducted to detect differences of mean with a probability level of 0.05. The percentages of the parasite population infecting the corneal surface and rest of the fish were calculated for each group for each week.

In isolated infections the infection success of untreated Lule salmon was very low. Only three of 10 fish remained infected and the propagation of G. derjavini on these was low. Only one fish was infected at week 5. The infection of Lule salmon treated with dexamethasone was more successful. Thus, 60 % remained infected and the number of parasites increased to significantly higher levels (Tables 1 & 2). In cohabitation infections the infection pressure was high during the first 2 weeks due to the presence of heavily infected rainbow trout in the fish tanks. Following removal of donor fish the infection level stabilized during the following 3 weeks. The dexamethasone treated fish showed a trend for a higher parasite burden in the first weeks (Table 3). In fish infected by cohabitation the parasites aggregated. Thus, up to 31% of the monogeneans were found on the corneal surface of the fish. Non-treated salmon especially experienced a rapid population increase on the cornea compared with dexamethasone treated fish (Table 4).

The present study has shown that salmon (both the Scottish Conon and the Swedish Lule strains) is generally an unsuitable host for G. derjavini. This parasite is known to preferentially infect brown and rainbow trout (Buchmann & Uldal 1997; Bakke et al. 1999). The suitability of rainbow trout for this particular monogenean was also clearly confirmed. However, it was also demonstrated that injections with dexamethasone elicit higher susceptibility in isolated Lule river salmon infected by single parasites. This is in accordance with findings from similar experiments with River Conon salmon (Nielsen & Buchmann, in press) (Tables 1 & 2). Cohabitation experiments involving heavy exposure to parasites also indicated that injections with dexamethasone affect the initial resistance of these two salmon strains to infections with G. derjavini. It thus suggests that host specificity of G. derjavini can be affected by i.p. injection with dexamethasone of an otherwise unsuitable host. In addition, this observation parallels the corticosteroid labile resistance of trout to G. salaris described by Harris et al. (2000). The present experiments did not exclusively demonstrate that the injected dexamethasone was responsible for the altered susceptibility or if endogenous cortisol produced in response to this type of manipulation elicited the changes. However, previous studies did not show changes in susceptibility of trout following injection with inert substances such as saline or protein sonicate (Larsen, Bresciani & Buchmann 2002) and this suggests that the strict host specificity of gyrodactylids, at least in some cases, is sensitive to corticosteroid treatment. The mechanisms involved in these reactions are not clearly understood. However, the corticosteroid drugs are immunosuppressive (Wilkens & De Rijk 1997) and this would imply that immunological factors in the host skin may at least partly determine host specificity. This may well be so (Buchmann 1999; Buchmann & Lindenstrøm 2002) and the predilection of G. derjavini for the corneal surface in unsuitable hosts also points to this. The cornea was suggested by Buchmann & Bresciani (1998) and Sigh & Buchmann (2000) to represent an immuno-privileged site for gyrodactylids parasitizing responding or unsuitable hosts. The cohabitation system using exposure to a high number of gyrodactylids was suitable, in contrast to the isolated infections with only a few parasites, for analysis of this hypothesis. Under these high exposure conditions it was noteworthy that more than 30% of the total parasite population aggregated on the cornea which represents less than 1% of the total fish surface area. Likewise, the predilection of the monogeneans for the cornea was lower in dexamethasone treated salmon (both Conon and Lule strains). Therefore, the immunosuppressive effect of the steroids should not be excluded as an explanation. However, corticosteroids have an additional effect on host cells which can provide the basis for an additional, although speculative, explanation of host specificity of gyrodactylids. These steroids have a stimulatory effect on mammalian epithelial cells (Sakthivel, Hamdan, Yang, Guzman & Nandi 1993; Kanazawa, Enami & Kohmoto 1999) and affect structure and function of rainbow trout epithelia (Iger, Balm & Jenner 1995) which will lead to an increased turnover of such cells. Epithelial cells in salmonids are the primary food source of gyrodactylids browsing the skin and induction of cell division in the outer epidermal layer could provide the parasite with an easily accessible and probably highly digestible food source. This implies that successful parasites are those fed with new (and perhaps more digestible) cells which in turn would make the parasite able to reproduce following corticosteroid treatment of the host. Alternatively, it still cannot be excluded that steroids, such as natural or artificial sex hormones or glucocorticoids, can affect the reproductive rate of gyrodactylids (Buchmann 1997). However, this remains to be demonstrated.