Heart and skeletal muscle inflammation (HSMI) disease diagnosed on a British Columbia salmon farm through a longitudinal farm study

Abstract

Heart and skeletal muscle inflammation (HSMI) is an emerging disease of marine-farmed Atlantic Salmon (Salmo salar), first recognized in 1999 in Norway, and later also reported in Scotland and Chile. We undertook a longitudinal study involving health evaluation over an entire marine production cycle on one salmon farm in British Columbia (Canada). In previous production cycles at this farm site and others in the vicinity, cardiac lesions not linked to a specific infectious agent or disease were identified. Histologic assessments of both live and moribund fish samples collected at the farm during the longitudinal study documented at the population level the development, peak, and recovery phases of HSMI. The fish underwent histopathological evaluation of all tissues, Twort's Gram staining, immunohistochemistry, and molecular quantification in heart tissue of 44 agents known or suspected to cause disease in salmon. Our analysis showed evidence of HSMI histopathological lesions over an 11-month timespan, with the prevalence of lesions peaking at 80-100% in sampled fish, despite mild clinical signs with no associated elevation in mortalities reported at the farm level. Diffuse mononuclear inflammation and myodegeneration, consistent with HSMI, was the predominant histologic observation in affected heart and skeletal muscle. Infective agent monitoring identified three agents at high prevalence in salmon heart tissue, including Piscine orthoreovirus (PRV), and parasites Paranucleospora theridion and Kudoa thyrsites. However, PRV alone was statistically correlated with the occurrence and severity of histopathological lesions in the heart. Immunohistochemical staining further localized PRV throughout HSMI development, with the virus found mainly within red blood cells in early cases, moving into the cardiomyocytes within or, more often, on the periphery of the inflammatory reaction during the peak disease, and reducing to low or undetectable levels later in the production cycle. This study represents the first longitudinal assessment of HSMI in a salmon farm in British Columbia, providing new insights on the pathogenesis of the disease.

Fig 1. Timeline of the events occurring at the farm during the period of study.

All sampling events carried out at the farm are shown. The Federal regulatory Fish Health and Surveillance audit dates carried out by the Aquaculture Management Division of Fisheries and Oceans Canada (AMD-DFO FH) are boxed in blue. Dates for Sea lice Abundance Count Data supplied by industry are boxed in peach (indicating average number of sea lice per fish, including a distinct count for Lepeophtheirus salmonis motiles and females, Caligus sp. and chalimus lifestage for all species, when concurrently present in the same fish). Industry reports of medical treatment and Fish Health submissions boxed in purple. Comments from the veterinarian hired for this study as part of the Strategic Salmon Health Initiative (SSHI) program boxed in green. The samples analysed in the study belong to sampling dates colored in green.

Fig 2. Clinical signs, gross lesions, and weekly mortalities reported for the farm in this study.

A) Non-specific clinical signs and gross lesions which have previously been associated with HSMI, as reported by Kongtorp and colleagues [1,9,11]. Any of the following were included: slow-swimmer, off-feed, ascites (e.g. ascites, hemorrhagic ascites, visceral petechiae, etc.), liver (e.g. enlarged, congested, etc.; excluding “green” and “haemorrhagic”), spleen (e.g. enlarged, congested, etc.), and heart (e.g. enlarged, pericardial effusion, etc.). B) All clinical signs and lesions presented in “A” were further aggregated into one category, called “~HSMI”, and compared to additional lesions and clinical signs observed over the course of the production cycle. The latter were aggregated to include five broad categories for: sea lice, mouth rot disease (MMY; myxobacteriosis), poor performers (PP), central nervous system (CNS; e.g. inflammation in red, hemorrhages in purple), and HSMI-related (from Fig 2A). Percentages of fish sampled showing clinical signs or lesions were calculated from the number of fish collected (moribund or dead) at each sampling event (light grey font color located above the sampling date). C) Weekly mortality counts (provided by the company) summarized based on their general classification. The inset demonstrates weekly mortality for each pen of the farm between October 1^st and November 26^th, 2013: during that period, one pen had increased mortalities (red line), with two more pens minorly affected (royal blue and charcoal lines), and the company veterinarian had submitted fish samples as part of routine investigation.

Fig 3. Heart and Skeletal Muscle Inflammation (HSMI) lesions in BC farmed Atlantic salmon (H&E).

A) Severe diffuse/ infiltrative panmyocarditis, involving the epicardium (arrows) and the compact layer of the myocardium (stars). The spongy layer is also involved in widespread inflammatory nodules (arrowheads). Bar scale: 500μm. B) Solid line inset of Fig A. The inflammatory infiltrate comprises mononuclear cells (lymphocytes and macrophages) in both epicardium (arrows) and compact myocardium (stars). Along with the inflammation, myocardial degeneration and necrosis are also visible (arrowheads). Scale bar: 100μm. C) Dashed line inset of Fig A. Multiple foci of endo-myocarditis (arrows) and myocardionecrosis (arrowheads) in the spongy layer of the myocardium Scale bar: 100μm. D) Severe diffuse zonal degeneration and necrosis in liver. Scale bar: 250μm. E) Severe diffuse infiltrative myositis in skeletal muscle. Note the white portion of the muscle (star) completely unaffected by the inflammatory reaction. Scale bar: 500μm. F) Inset of Fig E. Severe myonecrosis (along with some regenerating muscle cells) and infiltrating mononuclear cells inflammation are evident. Scale bar: 100μm.

Fig 4. Prevalence of fish with lesions in the heart and skeletal muscle over the saltwater production cycle at the farm of interest.

Bars represent the prevalence of heart lesions for sampling date. Dashed line represents prevalence of skeletal muscle lesions (skeletal muscle score ≥1). The prevalence for both tissues includes all live, moribund and freshly dead fish collected.

Fig 5. Prevalence of fish in developing stage and recovery stage of HSMI, and fish with skeletal muscle lesions over the saltwater production cycle at the farm of interest.

“HSMI Fish” represents the prevalence of fish with a heart score ≥4. “Developing HSMI” represents the prevalence of fish classified in the developing stage of HSMI, with a heart score ratio (epicardium/myocardium) >1. “Recovering HSMI” represents the prevalence of fish classified in the recovery stage of HSMI, with a heart score ratio (epicardium/myocardium) ≤1.The prevalence includes all live, moribund and freshly dead fish collected. Polynomial trend lines are used to show trends on samples distribution over time for each group. Black dots (“Skeletal Muscle lesions”) and relative black dashed polynomial trend line represent prevalence of fish with skeletal muscle lesions (skeletal muscle score ≥1).

Fig 6. P. theridion and K. thyrsites detection in the tissues through Twort’s Gram staining.

A) Cluster of P. theridion (small, dark blue organisms) inside myocardial cells in the spongy layer (arrow). Bar scale: 20μm. B) P. theridion in myocardial cells (arrows) and in an endocardial cell (arrowhead). Scale bar: 10μm. C) K. thyrsites inducing nodular granulomatous myositis in red muscle. The parasites are located in both red (star) and white (pound) portion of the skeletal muscle. One plasmodium inside a white muscle cell is also inducing a nodular granulomatous inflammatory reaction (arrow). Scale bar: 250μm. D) Inset of Fig C. Dark blue parasites are visible inside the nodular inflammation sites: 100μm.

Fig 7. PRV and K. thyrsites detection on the tissues through immunohistochemistry.

A) Red blood cells (RBCs) positive to PRV in a sample showing high PRV load but no heart lesions. The positive marking (brown—DAB) shows as intracytoplasmatic granules (tentatively called “viral factories”, arrow) or involves the whole cytoplasm (arrowheads). Bar scale: 20μm. B) PRV localization in a HSMI fish. The virus is present (red—Novared) in RBCs and leukocyte-like cells in the inflammatory infiltrate of the epicardium (arrowheads) as well as in myocardiocytes of the compact layer of the myocardium just on the border of the myocardial inflammatory infiltrate (arrows). Scale bar: 250μm. C) Solid line inset of Fig B. Leukocyte-like cells (arrowheads) and intracytoplasmatic inclusions in RBCs (arrow) positive to PRV (red—Novared). Bar scale 50 μm. D) Dashed line inset of Fig B. Cardiomyocytes of the compact layer of the myocardium, E) This sample also showed several positive cardiomyocytes to PRV (red—Novared) in the spongy layer of the myocardium. Scale bar: 100μm. F) K. thyrsites localization in the heart. Small plasmodium (red—Novared) in the spongy layer of the myocardium, in absence of inflammatory reaction. Scale bar: 20μm.

Fig 8. Mean pathogen load during the period (August 28^th, 2013 –January 15^th, 2014).

Line plots of mean pathogens loads in heart tissue over time (n = 38–42), as determined by qPCR method from the BioMark^™ platform. Data are presented as copy numbers +1, calculated based on a standard curve and normalized into a logarithmic scale base 10. Error bars represent the standard error of the mean.

Fig 9. Prevalence and load of PRV in association with heart lesions.

Plot of PRV prevalence (light grey plots) and load (dark grey box plots) in the heart over time. PRV load data are presented as copy numbers +1 calculated based on a standard curve and normalized into a base 10 logarithmic scale, ranging between 0 and 6. Black circle points define outliers. Heart score (red box plots) corresponds to the sum of the scores for both myocardial lesions and epicardial lesions, rated between 0 and 3 according to the severity (total score ranging between 0 and 6).

Fig 10. PRV load in association with heart lesions.

Box plots of PRV load in heart tissue of fish analysed on the basis of the presence of heart lesions. Median of PRV load is highlighted in red. PRV is presented as copy numbers + 1 normalized into a base 10 logarithmic scale. Presence of lesions is defined by a histopathological score in myocardium and epicardium equal or higher than 1 in a scale of 0–6.