Heart and skeletal muscle inflammation of farmed salmon is associated with infection with a novel reovirus

Abstract

Atlantic salmon (Salmo salar L.) mariculture has been associated with epidemics of infectious diseases that threaten not only local production, but also wild fish coming into close proximity to marine pens and fish escaping from them. Heart and skeletal muscle inflammation (HSMI) is a frequently fatal disease of farmed Atlantic salmon. First recognized in one farm in Norway in 1999, HSMI was subsequently implicated in outbreaks in other farms in Norway and the United Kingdom. Although pathology and disease transmission studies indicated an infectious basis, efforts to identify an agent were unsuccessful. Here we provide evidence that HSMI is associated with infection with piscine reovirus (PRV). PRV is a novel reovirus identified by unbiased high throughput DNA sequencing and a bioinformatics program focused on nucleotide frequency as well as sequence alignment and motif analyses. Formal implication of PRV in HSMI will require isolation in cell culture and fulfillment of Koch's postulates, or prevention or modification of disease through use of specific drugs or vaccines. Nonetheless, as our data indicate that a causal relationship is plausible, measures must be taken to control PRV not only because it threatens domestic salmon production but also due to the potential for transmission to wild salmon populations.

Figure 1. Piscine reovirus (PRV) sequence obtained by pyrosequencing.

Assembled sequence data mapped against the concatenated sequences of PRV. Genomic regions identified by BLASTN, BLASTX, FASTX, and FASD are shown in red, blue, green, and orange respectively.

Figure 2. Phylogenetic analysis of the RNA-dependent RNA-polymerase of Reoviridae.

Full length amino acid sequences were aligned using the ClustalX implemented on MEGA software and refined using T-Coffee to incorporate protein structure data. Phylogenetic analysis was performed using p-distance as model of amino acid substitution as implemented by ICTV for analysis of the Reoviridae family. MEGA was used to produce phylogenetic trees, reconstructed through the Neighbor Joining (NJ) method. The statistical significance of a particular tree topology was evaluated by bootstrap re-sampling of the sequences 1000 times.

Figure 3. Graphical representation of group differences in the log ratio of virus load normalized to a salmon host gene.

Nonparametric approaches were used to determine statistical significance for comparisons of the relative viral load among healthy and HSMI-affected farmed fish. Log transformations, which did not normalize log ratio distributions, were nonetheless performed for all samples after calculating L1 (virus)/EF1A (housekeeping) ratios to aid in graphical representation. (a) Comparison of adjusted log ratio in mixed heart and kidney samples from healthy farmed fish and farmed fish with HSMI; *, p<0.0001 (Mann-Whitney U); (b) comparison of adjusted log ratios in farmed fish without HSMI (healthy farmed fish), in the early phase of an HSMI outbreak, in the middle of an HSMI outbreak, and during the peak of an HSMI outbreak; **, p<0.0005; *, p<0.01 (individual Mann-Whitney U). Adjusted log ratios also differed significantly across all four farmed fish groups (p<0.0001; Kruskal-Wallis).

Figure 4. In situ hybridization was performed using locked nucleic acid (LNA) probes targeting the L2 segment of the Piscine reovirus.

Sections were permeabilized using proteinase K followed by hybridisation with digoxigenin (DIG)-labeled LNA probes. Sections were incubated with a mouse monoclonal anti-DIG-horse radish peroxidase and stained using a Tyramide Signal Amplification System. Sections were counterstained with Meyer's hematoxylin solution. (a) Heart from HSMI-infected fish (10×); (b) heart from HSMI-infected fish (40×); (c) heart from non-infected fish (40×); (d) heart from a fish infected with salmon pancreas disease virus.