Assessing the impact of swimming exercise and the relative susceptibility of rainbow trout Oncorhynchus mykiss (Walbaum) and Atlantic salmon Salmo salar L. following injection challenge with Weissella ceti

Weissellosis is an emerging disease of cultured rainbow trout Oncorhynchus mykiss (Walbaum) and has been associated with significant mortality events on Chinese, Brazilian and American rainbow trout farms since 2007 (Liu et al. 2009; Figueiredo et al. 2012; Welch & Good 2013). Water temperature elevation during the summer months appears to be instrumental in promoting clinical disease outbreaks, and larger fish, especially those approaching market weight (~1 kg), are most affected (Marancik et al. 2013; Welch & Good 2013). The causative agent of weissellosis is the Gram-positive bacterium, Weissella ceti, and during systemic infection, the agent can be readily isolated from brain and splenic tissues from affected fish (Marancik et al. 2013). Vasculitis is often observed through histopathological evaluation of ocular and cardiac tissues, suggesting that inflammation of blood vessels in multiple organs could be an important component of the pathological process (Welch, Marancik & Good 2014). Granulomatous inflammation of the epicardium and superficial myocardium has also been observed (Welch et al. 2014). To date, the only species demonstrated to be susceptible to weissellosis has been rainbow trout (although unpublished data indicate that brook trout, Salvelinus fontinalis, are also susceptible), and therefore, further research is required to investigate whether other commercially important species can be affected. With sustained swimming exercise, the cardiovascular system of fish undergoes various advantageous developments (Takle & Castro 2013). Among other beneficial outcomes, increased cardiac growth and maximum cardiac output (Farrell et al. 1990; Farrell, Johansen & Suarez 1991), improved capillarization of muscle tissue (Sanger & Potscher 2000) and enhanced cardiac oxidative capacity (Farrell et al. 1990) are observed in exercised fish. Takle & Castro (2013) describe the anti-inflammatory effects of exercise on fish cardiac muscle, which may be a component of the relationship between exercise and increased infectious disease resistance; Gleeson (2007) has reported on this association in humans. Indeed, exercised Atlantic salmon Salmo salar L. have exhibited increased resistance to infectious pancreatic necrosis virus (Castro et al. 2011), a pathogen known to cause cardiac pathologies (Ellis et al. 2010). Further research is required to determine whether exercise assists in reducing susceptibility to other fish pathogens associated with cardiac damage, such as infectious salmon anaemia virus (Jorgensen et al. 2007), piscine reovirus (Palacios et al. 2010) and W. ceti (Welch et al. 2014). In the light of the cardiovascular pathologies observed in fish affected by weissellosis, and in response to the aforementioned lack of data regarding the relative susceptibility of other fish species to this disease, we sought to investigate the impact of swimming exercise on weissellosis susceptibility of both rainbow trout and Atlantic salmon. The results of our study are presented in this short communication. All-female rainbow trout and mixed-sex Atlantic salmon (approximately 200 g and 120 g, respectively), originally procured as eyed eggs from commercial suppliers, were stocked into four 1.3-m3 circular tanks in a flow-through system (50 fish per tank) and maintained under study conditions for a period of five months. The four (i-iv) treatment populations consisted of the following: rainbow trout exposed to (i) sustained rotational water flow to provide swimming exercise or (ii) minimal rotational water flow to provide non-exercise conditions (RBT-E and RBT-N, respectively), and Atlantic salmon held under the same conditions, that is (iii) exercised or (iv) non-exercised conditions (ATL-E and ATL-N, respectively). Targeted rotational water velocities in the exercise tanks provided 2.0 body lengths per second (BL s−1) swimming speed, while swimming speeds in the non-exercise tanks were <0.5 BL s−1. To provide increased swimming speed, the two exercise tanks were retrofitted with magnetic pumps connected to an inlet and a discharge, or velocity, manifold. These pumps created rotational water currents by removing water from the tank and pumping it back into the same tank through the velocity manifold. All pumped water was injected beneath the surface to not influence concentrations of dissolved gasses. Velocity manifolds were composed of 5-cm schedule 40 PVC pipe with holes drilled and spaced approximately 2 cm apart; manifolds were positioned approximately parallel to the tank wall to create the rotational current. Water velocities were measured throughout the study period using a floating sphere timed around specific points of the tanks’ perimeters, as well as using an underwater flow meter. As the fish grew over time, bi-monthly length and weight data were collected, and exercise treatment water velocities were adjusted (via the angle of the velocity manifold) to maintain exercise at 2.0 BL s−1. All fish were kept under 24-h lighting and were fed a standard commercial pelleted feed via automatic feeders, with feeding events occurring every two hours. Dissolved oxygen was maintained at or near 100% saturation, and water temperature ranged between 11.9 and 13.8 °C throughout the study period, with no differences in water temperature existing among tanks at any particular time during the study. Following exercise treatment, all study fish (200) were moved to a separate building, where they were stocked into 0.8-m3 fibreglass circular tanks in a flow-through system and allowed to acclimate for a period of 19 days. During this acclimation period, fish were cohabitated by exercise group, i.e. exercised trout and salmon (100 fish total) were acclimated in one tank, while non-exercised trout and salmon (100 fish total) were acclimated in a separate tank. None of the trout or salmon displayed any signs of morbidity during either the exercise treatment or acclimation periods, and as such, these populations were considered to be appropriately healthy to commence the challenge phase. Immediately prior to pathogen challenge, fish were anesthetized with 90 mg L−1tricaine methanesulfonate (MS-222; Western Chemical, Inc.), and length and weight data were collected on all fish. Additionally, five fish from each treatment group (20 fish total) were randomly selected, killed with an overdose (200 mg L−1) of MS-222 and assessed for viscerosomatic, hepatosomatic and cardiosomatic indices (VSI, HSI, and CSI, respectively), that is the percentage of total body weight represented by the visceral contents, liver and heart, respectively. Following this initial data collection, all fish were challenged with intracoelomic injection of 9.5 x 106 W. ceti NC36 cells (in 200 μL phosphate-buffered saline [PBS]) per fish. Injections were performed using 1-cm3 syringes equipped with 26-gauge, 1.27-cm (0.5 in) needles. Fish were fed daily and monitored for 30 days post-challenge, and mortalities were removed and recorded daily (all mortalities displayed typical signs of morbidity, including abnormal swimming and exophthalmia, and 20% of mortalities each day were sampled and cultured to confirm the presence of W. ceti). Six fish per group were randomly selected at 24 h, 3, 6, 12 and 27days post-challenge and killed with MS-222; for each selected fish, the spleen and brain were aseptically removed, weighed and homogenized in PBS (10 μL mg−1) using a Bead Beater (BioSpec). Tissue homogenates were serially diluted, and 100 μL from each dilution was plated on tryptic soy–blood agar. Colony-forming units (CFUs) were enumerated after incubation for 24 h at 30 °C. One-way analyses of variance (ANOVAs) were performed in GraphPad Prism 5 (La Jolla) separately for splenic and brain load data [(CFU/mg) x (total mg/tissue)] from each sampling day, with species and exercise groups as independent variables; Bonferroni's multiple comparison post-test was then applied to determine significantly different outcomes between independent variables. Survival curves were also compared in GraphPad Prism 5 using the log-rank (Mantel–Cox) test. Within-species morphometric indices were compared between exercised and unexercised groups using group mean comparison tests in Stata 9 (College Station). Both species demonstrated a significantly (P < 0.05) greater weight gain while under exercised versus unexercised conditions (Table 1). While no significant differences were observed in VSI between treatment groups, visceral weight was generally lower (P = 0.052) in exercised Atlantic salmon, which likely indicates less deposition (or increased mobilization) of visceral fat reserves in response to sustained exercise. Likewise, while no significant differences were noted in HSI among treatment groups, HSI was generally higher (P = 0.053) in unexercised rainbow trout, possibly indicating lipid deposition due to increased fat conversion in the absence of energy utilization through exercise. Atlantic salmon demonstrated significantly greater CSI in exercised versus unexercised groups, which corresponds to previous findings reviewed by Takle & Castro (2013). Whether the relative differences in these metrics between rainbow trout and Atlantic salmon represent genuine species-specific anatomical and physiological responses to exercise requires further investigation incorporating larger sample sizes. Post-challenge bacterial load data (Fig. 1) indicated no major differences between rainbow trout and Atlantic salmon splenic load during the initial post-challenge period, although Atlantic salmon generally demonstrated less splenic bacterial loads in subsequent assessments up to 27 days post-challenge. No significant differences in splenic load were noted between exercised and unexercised groups within the same species. While a few individuals within the Atlantic salmon treatment groups were determined to have culturable bacteria in the brain up to six days post-challenge, the vast majority did not (unlike the rainbow trout, which had high levels of W. ceti isolated from brain tissue throughout the post-challenge period), indicating that the pathogen generally did not establish the typical brain infection observed in rainbow trout (Welch et al. 2014). The Atlantic salmon splenic and brain load data are very similar to previous, unpublished data for this species, indicating that Atlantic salmon are resistant to subsequent dissemination of the pathogen to target tissues and do not develop the clinical disease signs observed in rainbow trout, including mortality (as demonstrated in the survival curves, Fig. 2). Additionally, no significant protective effect of exercise was determined, considering splenic and brain loads, as well as overall survival, in either species. Shortcomings of the present study should be noted, to guide further research in this area. First, there is a possibility that the effects of exercise were reduced to a degree during the acclimation period and/or the pathogen challenge period, and therefore, reducing the period between exercise treatment and pathogen challenge would be beneficial in future studies. Secondly, low statistical power might have prevented detection of significant morphological differences between exercised and non-exercised fish, and further research might benefit from larger sample sizes for these assessments. Finally, due to space and pathogen containment constraints, we were unable to expand our scope to include sham-injected fish and a dose–response component; both of these components would be useful in future studies. The major findings of this study are as follows: (i) Atlantic salmon appear to be resistant to weissellosis and (ii) sustained swimming exercise, while conferring benefits related to growth performance and physical conditioning, did not impart protection against infection and mortality in rainbow trout under our study conditions. Further research is required to examine other commercially important fish species to determine their relative susceptibility to W. ceti. This research was funded by the USDA Agricultural Research Service (ARS) under Agreement No. 59-1930-5-510. Appreciation is extended to Jennifer Lipscomb (USDA-ARS-NCCCWA) for conducting the molecular analyses and to Travis Moreland (USDA-ARS-NCCCWA) for care of the research animals. The authors do not have any conflicts of interest to declare. Experimental protocols described are in compliance with the Animal Welfare Act (9CFR) and were approved by our Institutional Animal Care and Use Committee (IACUC). Use of trade names does not imply endorsement by the US Government.