Ocular Manifestations of Infectious Salmon Anaemia Virus Infection in Atlantic Salmon ( Salmo salar )

ABSTRACT

Eyes are essential sensory organs needed by teleost Atlantic salmon for high visual acuity and survival in both the wild and in aquaculture settings. In this work, we assessed the ocular manifestations of Infectious Salmon Anaemia Virus (ISAV) infection in Atlantic salmon by a cohabitation‐mediated infection assay and histological and immunohistochemical approaches. The findings reveal that Atlantic salmon with a systemic ISAV infection displayed ISAV antigen accumulation in specific ocular tissues and significant ocular morphological changes that could compromise visual acuity. Immunohistochemical analyses showed that ISAV‐related ocular pathological changes correlate with changes in expression levels and/or spatial localizations of IgM, Sox6, Sox9, collagen type 1, Gata1, and CD10 in specific compartments of the eye. The cornea is likely one of the first ocular tissues to be exposed to the cohabitation‐mediated ISAV infection. The ISAV‐infected Atlantic salmon showed corneal dysplasia, which correlated with increased corneal stromal ISAV antigen expression, dysregulated IgM, Sox6, and collagen type 1 expression, and an induction of Sox9 expression at the corneal surface. The choroid rete mirabile of ISAV‐infected eyes showed a decreased complement of erythrocytes, consistent with anaemia, while Gata1 expression, a key mediator of erythropoiesis, was upregulated compared to non‐infected eyes, suggesting a potential erythropoietic compensatory response. These findings provide a basis for further study of ISAV‐mediated ocular pathological changes and new insight into the pathogenesis of orthomyxoviruses in the eye.

1. Introduction

Infectious salmon anaemia virus (ISAV) is a frequent pathogenic virus that causes a fatal disease called Infectious Salmon Anaemia (ISA), characterized by acute anaemia and internal organ haemorrhage in Atlantic salmon [ Salmo salar ] (Dannevig et al. 1995). In salmon farms, ISA can cause high mortality rates which can lead to high social‐economic impacts (Dannevig et al. 2008). This virus belongs to the family Orthomyxoviridae and the genus Isavirus, sharing many structural, morphological, and physiochemical properties with other viruses in this family (Falk et al. 1995). The ISAV particles are 100–120 nm, containing an external envelope and granules (Dannevig et al. 1995). The genome consists of eight segments of single‐stranded RNA with 8 segments and a tentative negative polarity (Mjaaland et al. 1997).

Two phenotypes of ISA infections exist and are detectable using molecular approaches (Rounsville Jr et al. 2024). Virulent variants are generated by inducing deletions in the highly polymorphic region (HPR) of coding segment 6, and specific modifications in other coding segments such as coding segment 5 can also contribute to the pathogenicity (Cárdenas et al. 2019). HPR0 is a non‐virulent strain causing infection without disease where the main target is epithelial cells in gills and skin while the virulent HPR‐deleted ISAV strain causes a systemic endothelial infection (Aamelfot et al. 2016). Additionally, the levels of virulence in the HPR‐deleted ISAV can vary. Low‐virulence ISAV replicates and is transcribed more rapidly in the gills while the high‐virulence ISAV induces a higher viral load and mortality (McBeath et al. 2014). Evidence suggests that the HPR0 variant can evolve into low‐virulence strains, underscoring its significance as a risk factor in the emergence of ISA (Christiansen et al. 2017, 2011; Madhun et al. 2019).

ISAV has various clinical signs including severe anaemia, leukopenia, congestion of the spleen, foregut, and liver and it also affects both erythrocytes and leukocytes (Falk et al. 1998, 1995). It has been suggested that gills may be an important portal of entry and exit, with the virus entering cells via clathrin‐mediated endocytosis and macropinocytosis (Rimstad et al. 1999; Cárdenas et al. 2024). Further studies have shown that mucosal surfaces such as the pectoral fins and skin also function as entry points for the virus (Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015). The ability of ISAV to invade the salmon's immune system and progress is heavily dependent on temperature (Groves et al. 2023). Salmon were found to have a higher viral load and mortality at lower temperatures (10°C) compared to higher temperatures (20°C) (Groves et al. 2023). This is potentially due to a greater response by the T helper 1 cells and antigen presenting cells at 20°C, which may play a role in viral clearance, as well as the virus's difficulty to replicate at 20°C (Groves et al. 2023). Moreover, ISAV‐HPR0 was found to be at the highest prevalence during the winter (Christiansen et al. 2011). Parasite infestation, such as sea lice, was also found to significantly increase mortality and death rates associated with ISAV in Atlantic salmon (Zhong et al. 2023). Although ISAV has been detected in other salmon species, these species act as carriers of the virus without developing clinical disease (Snow et al. 2001).

While other orthomyxoviruses have been known to infect the eyes, to our knowledge there is limited literature on whether ISAV infects the eyes (Belser et al. 2018; Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015). For instance, the influenza virus H1N1, a member of the Orthomyxoviridae family, can infect and replicate within human retinal microvascular endothelial cells (Yang et al. 2024). H1N1 causes cell death, cell rounding and detachment, leading to ocular manifestations including retinal lesions, vasculitis, retinitis, and conjunctivitis (Yang et al. 2024). While ISAV has been detected in the eye, it was only found following the infection of other organs, suggesting the eyes were infected through the cardiovascular system as part of a systemic infection rather than acting as a point of entry (Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015).

Eyes are essential for visual acuity and survival in both wild and farmed finfish, emphasizing the importance of our understanding of Atlantic salmon ocular health. In the present study, we assessed Atlantic salmon moribund from a systemic ISAV infection as a first step toward determining its influence on the eye. We believe that our findings will allow a better understanding of ISAV pathobiology in the eyes, while also increasing the knowledge of ocular pathogenesis of orthomyxoviruses in general. We intended this study to serve as a basis for further exploration of the ocular manifestations of ISAV disease.

2. Materials and Methods

2.1. Fish Cohabitation‐Mediated ISAV Infection

To assess the ocular pathological changes associated with ISAV infection of Atlantic salmon, a cohabitation‐mediated infection paradigm was used. Viral preparations of infectious salmon anaemia virus were obtained from the Department of Fisheries and Oceans (Moncton, NB), originally provided by the Research and Productivity Council (RPC, Fredericton, NB). The high‐virulence ISAV isolate (ISAV‐HPR4; RPC/NB 04‐085‐1) was suspended in L‐15 culture medium following isolation from Atlantic salmon head kidney tissue, as described by Ritchie et al. (2009). Viral titre (TCID₅₀ = 1 × 10⁵/mL) and virulence were confirmed in vitro using an Atlantic salmon kidney (ASK) cell line monolayer and the Spearman–Kärber method.

For challenge, viral stocks were thawed on ice and aliquoted into 1 mL syringes. Atlantic salmon maintained in a quarantine module were lightly anaesthetised with tricaine methanesulfonate (MS‐222; Syndel, Canada; 100 mg/L) and intraperitoneally injected with 0.1 mL of the ISAV‐HPR4 suspension. Following recovery in aerated auxiliary tanks, Atlantic salmon were identified via PIT tags and transferred to their designated tanks. Donor Atlantic salmon remained in quarantine for 6 days post‐injection before being introduced to experimental cohabitation tanks at a 1:5 donor‐to‐cohabitant ratio. Fish were monitored several times daily for morbidity and mortality.

Figure 1 shows total mortality per tank observed in our study. Overall mortality reached 83.5%, which was slightly higher than expected. The cohabitation challenge began on May 19, 2024, and the first mortality in cohabitant fish was recorded on June 7, corresponding to 20 days post‐infection. Consistent with this progression, clinical and pathological signs were observed within the first 30 days, including pale gills, darkened skin, exophthalmia, haemorrhages at the base of fins and skin, ascites, as well as congested and enlarged liver and splenomegaly. Fish selected for sampling were in a moribund state. Our analyses often include the detection of ISAV in additional tissues, such as head kidney and spleen. In this study, viral load in these tissues was quantified following the protocol described by Snow et al. (2006). However, these data fall outside the scope of the present study and will be reported separately in a dedicated manuscript.

FIGURE 1.

Total mortality per tank observed in the ISAV infection study. Overall mortality reached 83.5%. Consistent with this progression, clinical and pathological signs were observed within the first 30 days, including pale gills, darkened skin, exophthalmia, haemorrhages at the base of fins and skin, ascites, as well as congested and enlarged liver and splenomegaly. Fish selected for sampling were in a moribund state.

At ~30 days post‐infection, fish showing clinical signs of ISAV infection were euthanized with a lethal dose of MS‐222 (250 mg/L) followed by cervical separation. Fish were necropsied and examined for internal and external signs of disease. Both eyes were collected, fixed in 4% paraformaldehyde for 48 h, and processed for histological analysis. Both eyes from three non‐infected Atlantic salmon from the same cohort served as controls. All husbandry and experimental procedures followed the guidelines of the Canadian Council on Animal Care and were approved by the University of Prince Edward Island and Atlantic Veterinary College Animal Care Committee (Protocol #21‐029).

2.2. Sequence Analysis

Sequence comparison of human (AAK26115.1) and Atlantic salmon (XP 045545203.1) Sox6 protein revealed a high degree of conservation between the two species. The BLAST alignment for human and Atlantic salmon Sox6 resulted in 74% identity and 84% similarity at the amino acid level. The mouse monoclonal anti‐Sox6 Ab we used for IHC in our studies was raised against amino acids 506–600 of the human Sox6, a region in which several stretches of sequence of 4–9 consecutive amino acids are 100% identical between the two species (Figure 2). Comparison of human (CAA86598) and Atlantic salmon (ACN10975.1) Sox9 protein also revealed high conservation; BLAST alignment search resulted in 73% identical and 80% similar amino acids (Figure 3A). The mouse monoclonal anti‐Sox9 Ab we used for IHC in our studies was raised against amino acids near the amino terminus of the human Sox9, in which several stretches of sequence of 4–17 consecutive amino acids are identical between the two species (Figure 3A). Sequence comparison of human (AAH09797.1) and Atlantic salmon (NP_001133239.1) Gata1 protein revealed a significant percentage of conservation of this protein between the two species (47% identical and 56% similar amino acids). Of the full length human Gata1 protein, Finger 2 domain sequences contributing to nuclear localization, DNA binding and protein stability (human amino acids 181–289 in Figure 3B) share the highest level of sequence homology with the Atlantic salmon Gata1 ortholog and this is consistent with the high evolutionary conservation known for this region (Yang and Evans 1992). The sequences against which the anti‐Gata1 mouse monoclonal Ab we used here are proprietary; but, outside of the Finger 2 domain, the level of sequence homology between several other regions of human and Atlantic salmon Gata1 are also sufficiently high for potential cross‐reactivity (Figure 3B). Sequence comparison of rat collagen type 1 protein (AAI33729.1) with that of Atlantic salmon (XP 014059932.1) revealed many regions of 4–26 consecutive amino acids which are identical between the two species. The BLAST alignment for rat ( Rattus norvegicus ) and Atlantic salmon collagen type 1 resulted in 77% identical and 83% similar amino acids (Figure 4). The two mouse monoclonal anti‐collagen type 1 Ab we used for IHC in these studies were both raised against full length rat collagen type 1.

FIGURE 2.

Sequence comparison for Sox6 proteins of Homo sapiens versus Salmo salar . COBALT alignment of Sox6 protein. Atlantic salmon (XP 045545203.1), human (AAK26115.1). Red indicates highly conserved regions, and blue indicates less conserved regions. Boxed regions are amino acid sequences against which the mouse monoclonal anti‐human Sox6 Ab used for IHC was raised.

FIGURE 3.

Sequence comparison of the Sox9 and Gata1 proteins of Homo sapiens versus Salmo salar (A), COBALT alignment of Sox9 protein. Human (CAA86598.1), Atlantic salmon (ACN10975.1). Boxed regions are the approximate amino acid sequences against which the rabbit monoclonal anti‐human Sox9 Ab used for IHC was raised. (B) COBALT alignment of Gata1 protein. Human (AAH09797.1), Atlantic salmon (NP_001133239.1). Red indicates highly conserved regions, and blue indicates less conserved regions.

FIGURE 4.

Sequence comparison of the collagen type 1 proteins of Rattus norvegicus versus Salmo salar . COBALT alignment of collagen type 1 protein. Rat (AAI33729.1), Atlantic salmon (XP 014059932.1). Red indicates highly conserved regions, and blue indicates less conserved regions.

Protein Basic Local Alignment Search Tool (BLAST 2.17.0) and COBALT analyses were performed for human (AAK26115) versus Atlantic salmon (XP_045545226) Sox6 protein, human (CAA86598) and Atlantic salmon (XP_014034661) Sox9 protein, human (AAH09797.1) versus Atlantic salmon (NP_001133239.1) Gata1 protein, and rat (AAI33729.1) versus Atlantic salmon (014059932.1) collagen 1 alpha (I) chain protein (COL1A1; also known as collagen type 1).

2.3. Histology, Immunohistochemistry and Immunofluorescence

Both histological and immunohistochemistry (IHC) analyses of ocular tissues were used to characterize the impact of the infection.

Paraffin blocks of Atlantic salmon eye tissues were sectioned onto positively charged slides (Superfrost Plus [VWR]) and processed for IHC as previously described (Gendron et al. 2020, 2022) or for standard Giemsa, Prussian blue, Herovici, RGB‐trichrome and Verhoeff/elastin histological stains. Tissue sections were deparaffinized in xylene and rehydrated in decreasing ethanol steps. Following deparaffinization, sections were processed for antigen retrieval for all IHC using a 10 mM citrate buffer pH 6 with 0.05% Tween 20, at 97°C for 40 min. Antigen retrieval was performed consistently and accurately using 125 μL volume Gene Frame manifolds (Thermo Fisher Scientific Canada) on a PCR instrument (Eppendorf Master Cycler gradient) fitted with a 4‐slide adapter (Eppendorf in situ Adapter). Slides were blocked for 1 h in 2% ECL Advance Blocking Reagent (Fisher Scientific) in Tris buffered saline pH 7.6 (TBS) at room temperature. Antibodies were applied in a vehicle of 2% ECL Advance Blocking reagent in TBS (overnight for primary Ab and 2 h for secondary Ab) at room temperature. Alkaline phosphatase (AP) colour reactions were developed using a Vector Red AP Substrate Kit (Vector Laboratories).

Expression of markers including ISAV, Sox6, Sox9, CD10, Gata1, and collagen type 1 was explored using specific Ab directed against the mammalian counterpart of these proteins. The selection of a panel of marker Ab reagents to analyse these markers was based on both the literature and previous work in our group. Previous work has shown an induction of an anti‐ISAV IgM response upon vaccination against ISAV (Caruffo et al. 2016; Tobar et al. 2015). Our previous work also suggests that IgM positive cells mitigate damage to marine teleost eye structures (Gendron et al. 2020). We previously showed that Sox9 positive cells are involved in an intraocular response to temperature stress in spotted wolffish [ Anarhichas minor ] (Kwabiah et al. 2024). Since ISAV infection causes anaemia, we targeted two key proteins, Sox6 and Gata1, known to orchestrate vertebrate erythropoiesis (Gutiérrez et al. 2008; Chennupati et al. 2018). Sox6 also harbours other multifaceted functions in tissue development and thus could be impacted by ISAV pathogenesis (Hagiwara 2011). We previously showed that CD10 positive cells are involved in an intraocular response to infection in lumpfish [ Cyclopterus lumpus ] (Gendron et al. 2020). Finally, regulation of expression of corneal collagen type 1 has been implicated in congenital eye disease (Li et al. 2025).

Anti‐ISAV antigen IHC was performed using anti‐ISAV P10 IgG2b monoclonal Ab (Aquatic Diagnostics Ltd.) at a dilution of 1:10 of the original P10 stock which was prepared by adding 1 mL of phosphate buffered saline containing 0.01% thimerosal to the dried P10 Ab. The ISAV antigen IHCs were performed using goat anti‐mouse‐IgG2b‐AP‐conjugated secondary Ab (Jackson Immunoresearch Laboratories) and was applied at a dilution of 1:250. Following incubation and washing, AP substrate solution (Vector Red) was applied to the slides and red reaction colour was monitored on an inverted microscope. Anti‐salmon IgM IHC was performed using a chicken anti‐salmon IgM (Ghasemieshkaftaki et al. 2024), IgY isotype (Somru BioScience), at a dilution of 1:500. A goat anti‐IgY(H + L)‐AP conjugated secondary Ab (Southern Biotech) was used at a dilution of 1:250. Anti‐Sox6 IHC was performed using a mouse monoclonal anti‐Sox6 primary Ab (Santa Cruz Biotechnology Inc.) at a dilution of 1:50. A goat anti‐mouse‐IgG2b‐AP conjugated secondary Ab was used at a dilution of 1:250. Anti‐Sox9 IHC was performed using a rabbit monoclonal primary Ab D8G8H (Cell Signaling Technology), at a dilution of 1:50. A goat anti‐rabbit IgG‐AP conjugated secondary Ab (Promega, Thermo Fisher Scientific Canada) was used at a dilution of 1:250. Anti‐Gata1 IHC was performed using a mouse monoclonal anti‐Gata1 primary Ab (Proteintech Inc.) at a dilution of 1:50. A goat anti‐mouse‐IgG1‐AP‐conjugated secondary Ab (Jackson Immunoresearch Laboratories) was used at a dilution of 1:250. Anti‐collagen type 1 IHC was performed using a mouse anti‐rat‐collagen 1 monoclonal primary Ab cocktail of Ab 8‐3A5 and 7‐2C12, both raised against rat collagen 1, from Developmental Studies Hybridoma Bank, both at a dilution of 1:10. A goat anti‐mouse IgG1‐AP‐conjugated secondary Ab was used at a dilution of 1:250. Anti‐CD10 IHC was performed using rabbit monoclonal anti‐mammalian CD10 (Abcam SP67) at 1:25 dilution as previously described for lumpfish [ Cyclopterus lumpus ] (Gendron et al. 2020). An anti‐rabbit IgG‐AP‐conjugated secondary Ab was used at a dilution of 1:250. Information for all Ab used herein is summarized in Table 1.

TABLE 1.

Antibody reagents used in this work.

Target antigen	Target species/Strain	Host species/Isotype	Conjugate	Source	Catalogue number
ISAV	HPR0	Mouse IgG2b Monoclonal	NA	Aquatic Diagnostics	P10
IgM	Atlantic salmon	Chicken IgY Polyclonal	NA	Somru BioScience	NA
Sox6	Human	Mouse IgG2b Monoclonal	NA	Santa Cruz	sc‐393314
Sox9	Human	Rabbit IgG Polyclonal	NA	Cell Signaling	D8G8H
Gata1	Human	Mouse IgG1 Monoclonal	NA	Proteintech	6011‐1‐Ig
Collagen 1	Rat	Mouse IgG1 Monoclonal	NA	Developmental Studies Hybridoma Bank	7‐2C12‐5; 8‐3A5‐S
CD10	Human	Rabbit IgG Monoclonal	NA	Abcam	ab227640
Mouse IgG2b	Mouse	Goat Polyclonal	Alkaline Phosphatase	Jackson Immunoresearch	115‐055‐207
Chicken IgY	Chicken	Goat Polyclonal	Alkaline Phosphatase	Invitrogen Thermo Fisher Canada	A16057
Rabbit IgG	Rabbit	Goat Polyclonal	Alkaline Phosphatase	Invitrogen Thermo Fisher Canada	31340
Mouse IgG1	Mouse	Goat Polyclonal	Alkaline Phosphatase	Jackson Immunoresearch	115‐055‐205
Mouse IgG1	Mouse	Goat Polyclonal	Alexafluor647	Invitrogen Thermo Fisher Canada	A21240

Note: Target antigens, target species and/or strain, host species/isotype, sources and vendor catalogue numbers (where appropriate) of Ab reagents used are listed in the order they appear in the results section.

The IHC procedures were performed as outlined in Table 2. Since multiple IHC experiments were required, IHC development reactions for each experiment consisting of the same Ab were performed for an equal amount of time and always included at least one slide specimen from a non‐infected control eye in addition to the slides from the ISAV‐infected animals. Negative controls for the IHC procedure consisted of IHCs done identically except omitting the primary Ab and images of them are included as insets in the IHC figures. Due to antigen retrieval constraints (i.e., only four slides can be processed at a time and the edges of the retrieval chamber manifolds sometimes damaged or destroyed sections of fellow eyes on each slide), not all experiment runs could contain an IHC negative control and not all runs included both eyes from individual animals. However, once the negativity of negative controls was clearly established as being completely and consistently negative (as shown in figures), a priority was placed on marker expression testing on as many individual specimens and fish as possible, as outlined in Table 2. Atlantic salmon tissue infected with ISAV used herein served as positive control for ISAV IHC. Atlantic salmon head kidney and/or blood vessel lumens in the Atlantic salmon tissue used herein served as positive control for IgM IHC. Sox6 and Gata1 western blotting was used to validate positive reaction of the Sox6 and Gata‐1 antibodies against Atlantic salmon eye tissue. Positive controls for Sox9 were non‐infected Atlantic salmon scleral cartilage (data not shown) since we previously described scleral cartilage Sox9 expression in the marine teleost spotted wolffish (Anarachis minor) (Kwabiah et al. 2024). Atlantic salmon corneal tissue, used herein, served as positive control for Collagen 1 IHC. Atlantic salmon head kidney tissue served as positive control for CD10 IHC (data not shown). Sections were examined using a LEICA DM4000 B light microscope. Images were taken using an Infinity 8 camera mounted to the microscope along with the Infinity 8 software.

TABLE 2.

Tissue harvesting and experimental outline.

Marker	Number of specimens used per IHC experiment				Number of non‐infected versus infected specimens per marker		Total number of specimens per marker	Name of specimen used per marker
	1	2	3	4	Non‐infected	Infected
ISAV	4	4	3	ND	3	8	11	A117, B19, B110, B211, B36, C112, D111, E19, M127, N126, O127
IgM	3	3	3	4	4	9	13	A116, B18, B29, B210, B35, C111, C214, D110, E18, M128, N123, N124, O128
Sox6	4	4	4	ND	4	8	12	A115, B17, B28, B34, C110, C215, D19, E17, HK, M126, N122, O126
Sox9	4	4	4	ND	4	8	12	A114, B116, B27, B33, C19, C213, D18, E16, HK, M125, N116, O125
Gata1	3	4	3	3	5	8	13	A111, B11, B12, B21, B22, B23, B24, D15, M121, N11, O120, O121, O122
Collagen 1	3	4	4	3	5	9	14	A112, B14, B25, B322, B325, C17, C211, D16, E14, M122, M123, N119, N120, O123
CD10	4	4	3	ND	4	8	12	A113, B15, B26, B321, C18, C212, D17, E15, HK, M124, N121, O124

Note: Eye tissue harvesting and experiments performed upon these tissues are provided to outline the methodological plan toward the results.

In order to increase resolution and magnification, immunofluorescence for Gata1 expression was performed on some slides containing adjacent sections of the same non‐infected or ISAV‐infected Atlantic salmon choroid rete mirabile specimens using the same processing methods described above for AP colorimetric development except that Alexafluor 647 conjugated anti‐mouse IgG1 specific second Ab (Invitrogen, Thermo Fisher Scientific Canada) was used at a dilution of 1:200 for confocal microscopic imaging. Some experiments included pre‐treatment with TrueBlack lipofuscin autofluorescence quencher (Cell Signaling Technology). Slides were mounted in aqueous fluorescence mounting medium with or without 4′,6‐diamidino‐2‐phenylindole (DAPI) as indicated in the results section. Negative controls consisted of omitting the primary Ab but not the fluorescence‐conjugated secondary Ab. Immunofluorescence sections were examined using a Zeiss LSM 900 Airyscan Confocal microscope and imaging was done using Airyscan 2. Nine Z‐stack slices of 2 μm were taken.

2.4. Western Blotting

To generate control Atlantic salmon protein lysates, head kidney and/or ocular tissue were harvested post‐mortem from 3 normal healthy adult Atlantic salmon being euthanized for other research purposes (Memorial University Institutional Animal Care Committee protocols 17‐03‐RG; 18‐01‐JS; 18‐02‐JS). Protein lysates of human lung epithelial cell line A549 (American Type Culture Collection, Manassas, VA) were prepared and were used as a positive control tissue. Tissue samples and cells were homogenized in Triton X‐100 lysis buffer (50 mM Tris–HCl pH 7.6, 150 mM NaCl, 0.5% [w/v] sodium deoxycholate, 0.1% [w/v] SDS, 1% [v/v] Triton X‐100 and 10% [v/v] glycerol) supplemented with 1 mM dithiothreitol, 10 μg/mL leupeptin, 0.3 U/mL aprotinin, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM sodium orthovanadate, 50 mM sodium fluoride and 25 mM β‐glycerophosphate. Whole cell lysates were clarified by centrifugation and protein concentration was determined by Bradford assay using bovine serum albumin as a standard. Protein extracts were analysed for Sox6, Gata1 and actin by western blot as described previously (Ho et al. 2015; Gendron et al. 2022; Kwabiah et al. 2024) using the same primary Ab used for IHC followed by an anti‐mouse IgG‐horse radish peroxidase conjugated secondary Ab (Promega, Madison, Wisconsin). Western blot membranes were reprobed using an anti‐actin monoclonal Ab (C4, Santa Cruz Biotechnology, CA) which has previously been used for detecting lumpfish actin (Gendron et al. 2022), to ensure protein integrity and loading control. Western blot signal was revealed by chemiluminescence using LumiGLO substrate kit (KPL, Gaithersburg, MD) and images were captured using a Fuji LAS 4000 digital imaging system.

3. Results

Western blotting for Sox6 and Gata1 demonstrated specific discrete bands in Atlantic salmon tissues, the observed molecular weights of which were similar to molecular weights for these proteins in the human A549 positive control (Zhou et al. 2019; Wang et al. 2017) cell line (Figure 5). Sox6 western blots revealed a main band of 112 kDa in salmon choroid rete mirabile tissues. However, this main 112 kDa band was not detected in salmon head kidney tissues. Gata1 western blots revealed a main band of 44 kDa in salmon tissues co‐migrating with Gata1 expressed in A549 cells. Additional anti‐Gata1 reactive bands of 36 and 38 kDa were observed in salmon choroid rete mirabile and head kidney, respectively. The sequence analysis results for Sox6, Sox9, collagen type 1, and Gata1, taken together with our western blotting results for Sox6 and Gata1, support cross reactivity between mammal and Atlantic salmon for the Ab we used for the IHC studies herein.

FIGURE 5.

Western blot analysis of Sox6 and Gata1 protein expression in Atlantic salmon tissues. Representative western blots for Sox6 (top panel), Gata1 (middle panel), and Actin (bottom panel) of protein lysates prepared from Atlantic salmon head kidney (HK, 200 μg) and choroid rete mirabile (CRM, 200 μg) tissue from 2 separate Atlantic salmon (1, 2), in comparison to 50 μg and 25 μg (50, 25) of protein lysate prepared from the human lung epithelial control cell line A549, as indicated. Arrow in Sox6 blot indicates the main observed band at approximately 112 kDa. Arrows in Gata1 blot indicate the main ~44 kDa band in A549 cells and the two main bands at approximately 44 and 38 kDa in Atlantic salmon tissues. Actin was also probed for loading and protein integrity control.

Immunohistochemistry for IgM revealed IgM positive cells in the central corneal epithelium of both non‐infected (Figure 6A) and ISAV‐infected (Figure 6B) Atlantic salmon, the highest staining intensity of which was found at the outer (more anterior) portion of the epithelium. However, the typically layered cellular morphological structure of the central corneal epithelium in the ISAV‐infected Atlantic salmon was found to be more disorganized, lacking the clear layering of cells (Figure 6B) observed in the non‐infected animals (Figure 6A). Furthermore, the IgM staining observed in the substantia propria (SP) region of the central corneal stroma in the ISAV‐infected specimens (Figure 6B, boxed) appeared increased compared to the non‐infected fish.

FIGURE 6.

Marker expression and morphological analysis of the central cornea region of ISAV‐infected Atlantic salmon. Immunohistochemistry for IgM expression showing the presence of IgM (red colour) in the central corneal epithelium (brackets) of non‐infected (A) and ISAV‐infected Atlantic salmon (B). Increased levels of IgM staining in the central corneal stroma of ISAV‐infected animals are shown in the boxed area. The characteristic morphological cellular layering of the central corneal epithelium indicated with black arrows in non‐infected Atlantic salmon (A) does not appear as prominant in the ISAV‐infected animals (B). A negative control for anti‐IgY‐AP is shown in the inset in (B). Immunohistochemistry for Sox6 showing negative results in non‐infected Atlantic salmon (C). Cells in the central corneal epithelium of ISAV‐infected Atlantic salmon were positive for Sox6 (D), with the highest signal being detected near Bowman's membrane (Bm). A negative control for anti‐IgG2b‐AP is shown in the inset in (D). Sox9 was negative within non‐infected Atlantic salmon (E) but positive in some cells in ISAV‐infected Atlantic salmon (F) and was localized near the outer central corneal epithelium (arrows in F). The upper right inset in (F) shows a magnified view of the cell boxed in black in (F). The bottom left inset in (F) shows a negative control on the same tissue for anti‐rabbit IgG‐AP. Immunohistochemistry for CD10 showed central corneal stromal staining in what appeared to be keratocytes in both non‐infected (G) and ISAV‐infected (H) salmon. A negative control on the same tissue for anti‐rabbit IgG‐AP is shown in the inset in (H). Giemsa and Herovichi histological stains highlighted the change in organization of the central corneal epithelium in healthy (I and K, bracket) versus ISAV‐infected (J, L) salmon. 400X. Scale bar subdivisions equal 10 μm.

Immunohistochemistry for Sox6 revealed the presence of Sox6 in the central corneal epithelium of the ISAV‐infected Atlantic salmon (Figure 6D) while no staining was detected in the non‐infected fish (Figure 6C). Sox6 staining was observed primarily in the basal region of the central corneal epithelium in the ISAV‐infected Atlantic salmon, particularly near Bowman's membrane (Bm). In contrast to Sox6, Sox9 staining was observed in the outer (more anterior) region of the central corneal epithelium in ISAV‐infected Atlantic salmon (Figure 6F, arrows), while no staining was detected in the non‐infected fish (Figure 6E).

The central corneal stroma of both ISAV‐infected (Figure 6H) and non‐infected Atlantic salmon (Figure 6G) revealed CD10 staining in a pattern that was similar to the IgM central corneal stromal staining. This staining pattern was consistent with the distribution of keratocytes in the cornea (Jester et al. 1994). Low levels of basal epithelial central corneal CD10 staining were observed in some samples (Figure 6H).

Giemsa (Figure 6I,J) and Herovici (Figure 6K,L) histological stains were used to reveal the changes in structure, organization, and thickness of the central corneal epithelium and stroma associated with ISAV infection. Both Giemsa and Herovici stainings confirmed a change in the stratification of the cells within the central corneal epithelium of the ISAV‐infected Atlantic salmon, indicating disorganization and the loss of the defined layers (Figure 6J,L) relative to the non‐infected fish (Figure 6I,K).

The cornea limbic region of ISAV‐infected Atlantic salmon showed upregulation of both ISAV antigen immunoreactivity and collagen type 1 in cellular elements of the limbic corneal stroma relative to non‐infected fish (Figure 7B versus Figure 7A and Figure 7D versus Figure 7C). A panel of various histological stains (Giemsa, Prussian blue, Herovici, RGB‐trichrome and Verhoeff elastin staining) indicated cellular infiltration of the limbic corneal stroma, possible extracellular haemoglobin deposition, changes in collagen distribution patterns and limbic corneal dysplasia of ISAV‐infected Atlantic salmon relative to the non‐infected animals (Figure 7A,C,E,G,I,K,M vs. Figure 7B,D,F,H,J,L,N). In particular, increases were observable in cellularity (apparent increases in cellular infiltration or proliferation) and distribution of histological stain colour reflecting collagen distribution changes in Figure 7J,L, compared to Figure 7I,K.

FIGURE 7.

Marker expression and morphological analysis of the limbic region of cornea of ISAV‐infected Atlantic salmon. A, C, E, G, I, K, M, Corneal limbic region of non‐infected Atlantic salmon. B, D, F, H, J, L, N, Corneal limbic region of ISAV‐infected Atlantic salmon. A, B, Anti‐ISAV antigen IHC (red) indicating increased ISAV staining in cellular elements of the limbic corneal stroma associated with ISAV infection. C, D, Anti‐collagen type 1 IHC (red) indicating increased collagen type 1 staining in cellular elements of the limbic corneal stroma associated with ISAV infection. E, F, Giemsa, suggesting cellular infiltration in limbic corneal stroma of ISAV‐infected Atlantic salmon. G, H, Prussian blue, suggesting extracellular haemoglobin deposition (darker blue staining in black boxed area in H compared to G) in limbic corneal stroma and limbic corneal epithelial dysplasia in ISAV‐infected fish (white boxed area in H compared to G). I, J, Herovici, suggesting changes in collagen distribution patterns in limbic corneal stroma in ISAV‐infected Atlantic salmon. K, L, RGB‐trichrome, suggesting cellular infiltration and changes in extracellular matrix deposition patterns in limbic corneal stroma and limbic corneal epithelial dysplasia of ISAV‐infected Atlantic salmon (L), compared to uninfected (K). (M, N) Verhoeff elastin stain suggesting cellular infiltration, changes in extracellular matrix deposition patterns and limbic corneal epithelial dysplasia of ISAV‐infected salmon (N), compared to uninfected (M). Ce, corneal epithelium. Cs, corneal stroma. Cd, corneal dysplasia. Representative negative control for ISAV staining is shown in inset in B (cornea negative control for anti‐IgG2b‐AP). Representative negative control for Collagen 1 staining is shown in inset in D (cornea negative control for anti‐mouse IgG1‐AP). 400×. Scale bar subdivisions in M are 10 μm and representative for all panels.

Immunohistochemistry for Sox9 revealed nuclear expression of Sox9 within cells of the ISAV‐infected salmon eyelid skin (Figure 8B), while no staining was observed in the non‐infected fish (Figure 8A). In the ISAV‐infected specimens, Sox9 expression was localized near the outer edge of the eyelid skin (boxed in Figure 8B). The histological stain Giemsa was used to further assess the changes between the eyelid skin of non‐infected and ISAV‐infected Atlantic salmon. While the non‐infected specimens showed normal stratified epithelial morphology, such morphology was lost in the ISAV‐infected specimens, in which the epithelium appeared dysplastic (Figure 8C versus Figure 8D, respectively).

FIGURE 8.

Sox9 expression in the eyelid skin of ISAV‐infected salmon. Representative IHC for Sox9 expression in eyelid skin adjacent to the corneal limbic area showing a non‐infected specimen testing negative for Sox9 (A) and an ISAV‐infected Atlantic salmon (B) with regions of Sox9 expression boxed. Giemsa staining in non‐infected (C) and ISAV‐infected (D) Atlantic salmon are shown. The boxed regions in (C) versus (D) show the loss of the normal stratified epithelial morphology in the ISAV‐infected specimen. 400×. Scale bar subdivisions equal 10 μm.

Immunohistochemistry for ISAV antigen expression indicated more intense staining within the vasculature walls of the choroid rete mirabile in the ISAV‐infected Atlantic salmon compared to healthy fish (box in Figure 9B versus Figure 9A), consistent with infection of vascular walls. Giemsa staining was used to reveal the changes in the level of organization of the vasculature within the choroid rete mirabile of ISAV‐infected Atlantic salmon (Figure 9C,D). In contrast to the well organized and aligned choroid rete mirabile vasculature of the non‐infected Atlantic salmon (Figure 9C), the vascular tissue of the ISAV‐infected fish displayed disorganized regions with an increase in clustering of cells (Figure 9D). The RGB‐trichrome staining method was used to further analyse the impact of ISAV infection on the choroid rete mirabile. Although this stain was developed for the histological assessment of skeletal tissues, fast‐green is a component of RGB‐trichrome stain and is known for labeling erythrocytes bright green (Gaytan et al. 2020; Xuereb and Beesley 1954). RGB‐trichrome staining of choroid rete mirabile in non‐infected salmon (Figure 9E) showed robust presence of bright green‐staining erythrocytes, while salmon infected with pathogenic ISAV (Figure 9F) showed a decreased presence of erythrocytes within the tissue accompanied by an increased density of small red/orange staining cells across the whole section. In non‐infected specimens, Gata1 expression was mainly confined to the walls of fine capillaries (Figure 9G), while ISAV‐infected Atlantic salmon choroid rete mirabile displayed increased Gata1 expression in both vascular wall structures and in numerous round cells (Figure 9H). The increase and expansion of Gata1 expression pattern in the ISAV‐infected specimens was even more apparent using confocal immunofluorescence Z stack imaging at higher magnification under oil immersion (Figure 10). The fluorescence imaging approach also allows imaging of teleost choroid rete mirabile cellular autofluorescence (Murphy et al. 2024) marking mature erythrocytes (bright blue) and DAPI (darker blue) marking nuclei. In these images, Gata1 immunoreactivity in ISAV‐infected Atlantic salmon choroid rete mirabile appeared in tightly packed clusters of numerous Gata1‐positive leukocytic‐like round cells within areas surrounded by Gata1‐positive vascular wall‐like structures (Figure 10D).

FIGURE 9.

Marker expression and morphological analysis of the choroid rete mirabile in ISAV‐infected Atlantic salmon. (A, B) Representative ISAV antigen IHC showing (A), Vascular staining (red) for ISAV in the rete mirabile of a non‐infected Atlantic salmon and (B) Intense ISAV antigen staining in the vasculature (boxed) of an ISAV‐infected fish. Inset in (B) shows negative control staining using anti‐mouse IgG2b‐AP. (C) Giemsa staining of a non‐infected and (D) ISAV‐infected Atlantic salmon. Regions of cellular pooling are boxed. (E) RGB‐trichrome staining of a non‐infected and F, ISAV‐infected Atlantic salmon. Paucity of mature erythrocytes (green) and increased density of small round red/orange staining cells are evident across the whole image of the ISAV‐infected specimen. (G) Gata1 immunoreactivity of a non‐infected Atlantic salmon showing Gata1 staining (red) in fine capillary walls (arrows). (H) Gata1 immunoreactivity in ISAV‐infected fish showing staining (red) for both the vascular wall and numerous small round cells (exemplified by arrows in zoomed‐in area in left inset in H in which Gata1‐positive nucleoli are also visible). Inset on the right side in H shows negative control staining using anti‐mouse IgG1‐AP. All panels 400× except zoomed inset in (H). Scale bar subdivisions equal 10 μm.

FIGURE 10.

Gata1 expression changes in the choroid rete mirabile in ISAV‐infected Atlantic salmon. (A) Gata1 immunoreactivity (red) of a non‐infected Atlantic salmon showing light Gata1 expression in fine capillary walls (arrows). (B) Gata1 immunoreactivity (red) in ISAV‐infected Atlantic salmon showing increased vascular wall staining (compare arrowed areas in B versus A) and strong cellular staining of numerous round cells. (C) Negative control with primary Ab omitted from an adjacent section of ISAV‐infected Atlantic salmon showing no immunofluorescence signal. (D) Confocal Z stack with red signal indicating Gata1 immunoreactivity, bright blue signals showing erythrocyte autofluorescence (yellow arrow) and darker blue DAPI signal showing nuclei in ISAV‐infected Atlantic salmon. The white boxed areas indicate the presence of tightly packed clusters of numerous Gata1 positive round cells within areas surrounded by Gata1 positive vascular wall‐like structures. Panels A–C show images in which TrueBlack autofluorescence blocking procedure was performed to eliminate tissue autofluorescence. Panel D shows an image on which TrueBlack autofluorescence blocking procedure was omitted. All panels 630×. Oil immersion. Scale bar in (A) equals 10 μm for all panels. Representative images shown.

Immunostaining for ISAV antigen confirmed the presence of ISAV within the neural retina of Atlantic salmon, with more intense staining observed in discrete areas of the inner retina of ISAV‐infected fish compared to non‐infected fish (Figure 11A–C). Histological comparison of the retina between non‐infected and ISAV‐infected Atlantic salmon revealed substantial changes in neural retinal structure (Figure 11). Giemsa staining revealed severe rosetting throughout the retina in all ISAV‐infected Atlantic salmon (black box in Figure 11E), while the retina of non‐infected Atlantic salmon was arranged in straight layers (Figure 11D). A change in the individual layers of the retina from ISAV‐infected Atlantic salmon was also observed, such as the scattering of the inner nuclear layer (white box in Figure 11D) and a thickening of the rods and cones layer.

FIGURE 11.

ISAV antigen expression and morphological analysis of the neural retina in ISAV‐infected Atlantic salmon. Representative IHC for ISAV antigen expression (red staining): No primary Ab negative control stained with anti‐mouse IgG2b‐AP only (A), non‐infected (B) and ISAV‐infected Atlantic salmon retina (C). Black boxed area in (C) indicates intense ISAV immunoreactivity in discrete regions of the inner retinal regions. (D) Giemsa staining of the non‐infected retina showing the retinal pigment epithelium (rpe), rods and cones (rc), outer nuclear layer (on), outer plexiform layer (op), inner nuclear layer (in), inner plexiform layer (ip), and ganglion cell layer/nerve fibre layer (gc). (E) Giemsa staining of Atlantic salmon infected with the pathogenic ISAV showing rosetting (black box) and scattering of the inner nuclear layer (white box) with rc, on, in and ip also indicated. 200×. Scale bar subdivisions equal 10 μm.

Histological analysis combined with IgM IHC highlighted the degeneration of Bruch's membrane in Atlantic salmon infected with ISAV. The changes in this barrier located between the retinal pigment epithelium (RPE) and the choroidal vasculature were observed in all ISAV‐infected Atlantic salmon (Figure 12B) while remaining intact in all the non‐infected Atlantic salmon examined (Figure 12A). In multiple regions within the ISAV‐infected Atlantic salmon, Bruch's membrane appeared thinned (arrows in Figure 12B), while in other regions the membrane was not visible (arrows in Figure 12B). Verhoeff elastin histological staining was used to further visualize the retinal and choroidal changes associated with the infection. A degeneration of elastin fibres (purple), including Bruch's membrane, the RPE (black), and photoreceptors (green), was observed in the ISAV‐infected fish (Figure 12D) compared to that of the non‐infected specimens (Figure 12C).

FIGURE 12.

IgM expression and morphological analysis of retinal‐choroidal region in ISAV‐infected Atlantic salmon. Representative IHC for IgM expression (red) showing in (A), intact Bruch's membrane (arrows), not stained, between RPE (rpe) and choroid (c) in a non‐infected Atlantic salmon and (B), disrupted and thinned Bruch's membrane in a ISAV‐infected specimen (arrows). The anti‐IgM staining can be observed in both blood vessel lumens and on endothelial linings. Inset in (B) shows a negative control stained using anti‐IgY‐AP. Verhoeff elastin staining (purple) highlighting Bruch's membrane (arrows), the RPE (black) and photoreceptors (green) in non‐infected (C) and ISAV‐infected Atlantic salmon (D). 400×. Scale bar subdivisions equal 10 μm.

CD10 staining was observed in the extraocular muscle myofibrils of both ISAV‐infected (Figure 13B) and non‐infected Atlantic salmon and did not appear to be present in extracellular areas of these tissues (Figure 13A). However, the staining intensity of the ISAV‐infected specimens generally appeared higher than the non‐infected fish. Staining patterns exhibited a combination of striated (boxed) and punctate (arrow). Histological Herovici staining of both non‐infected (Figure 13C) and ISAV‐infected Atlantic salmon (Figure 13D) facilitated further visualization of the muscle fibre vasculature and myofibrils. The extraocular muscle myofibrils of ISAV‐infected Atlantic salmon exhibited a more pronounced reddish/brown colour with Herovici staining compared to the more purple/blue colour seen in the non‐infected specimens. Furthermore, degeneration of the muscle myofibrils in ISAV‐infected Atlantic salmon (Figure 13D, arrows) and infiltration of leukocytic‐like cells (Figure 13D, boxed) were observed.

FIGURE 13.

CD10 expression and morphological analysis of extraocular muscle myofibrils of Atlantic salmon. Representative IHC for CD10 expression showing staining (red) in both the non‐infected (A) and ISAV‐infected (B) Atlantic salmon. Both striated (boxed) and punctate (arrows) staining was observed, with more intense staining seen in ISAV‐infected than non‐infected specimens. Inset on the right side of B shows a negative control stained using anti‐rabbit IgG‐AP. Herovici staining in non‐infected (C) and ISAV‐infected fish (D) is shown, with a reddish/brown staining seen in ISAV‐infected specimens. Muscle degeneration (D, arrows) and leukocytic‐like cell infiltration (D, boxed) are also observed in ISAV‐infected specimens. eom, extraocular muscle myofibril. 400×. Scale bar subdivisions equal 10 μm.

Most of the results described herein were consistent in all animals investigated. While some animals served as better examples, either due to having more evident pathological changes or the tissues were more intact for imaging, all pathological changes that we discussed in the manuscript were present in all the ISAV infected specimens and absent or observed at a much lesser extent in the non‐infected specimens. While both eyes from the same individual were examined, the extent of variation of these findings within animals was more difficult to determine since, for example, fellow eyes were sometimes damaged during the retrieval process.

4. Discussion

In this work, we assessed the ocular manifestations of ISAV infection in Atlantic salmon using both histological characterization and marker expression analyses. The findings revealed that moribund salmon with a systemic ISAV infection show specific morphological changes and changes in protein markers as determined by IHC that provide insight into the molecular pathways that might be involved in the ocular pathogenesis. These findings provide a basis for expanding our understanding of ISAV infection ocular pathobiology and contribute to the general knowledge of the ocular pathogenesis of orthomyxoviruses.

Some of the Ab reagents used in these studies to characterize the expression pattern of specific markers were raised against either ISAV‐specific epitopes (Falk et al. 1998) or Atlantic salmon IgM specific, where epitope reactivity is established by complete sequence identity. Other Ab reagents we used herein, raised against mammalian epitopes, such as the anti‐Sox9 and anti‐CD10 Ab, have been previously validated by us as cross‐reactive in North Atlantic marine teleost tissues (Kwabiah et al. 2024; Gendron et al. 2020). The transcription factor Gata1 plays an essential role in steady‐state and stress‐induced erythropoiesis, and several studies have characterized Gata1 orthologs in teleosts (Tian et al. 2021; Belele et al. 2009; Xu et al. 2015; Gallo and Civinini 2003). Although there is no extensive literature on collagen type 1 in teleost tissues, the level of sequence homology we found between mammalian and Atlantic salmon proteins is sufficiently high for expecting cross reactivity. Furthermore, the deduced polypeptide of collagen type 1 of the teleost Chu's croaker ( Nibea coibor ) contains the same signal peptides, C‐propeptide and N‐propeptide domains, and triple helix domains, all of which are characteristics of type I collagen in vertebrates, and further supports the cross reactivity of the anti‐mammalian collagen type 1 Ab reagent we used herein with teleost (Rong et al. 2022). The anti‐Sox6 and anti‐Gata1 Ab we used recognized proteins within Atlantic salmon tissues with molecular weights that are similar to those appearing in a control human cell line. Additional lower molecular weight anti‐Gata1 reactive bands were also detected in the fish tissues. Shorter isoforms of Gata1 protein have been reported and studied in human tissue (Aluri et al. 2025). Species variations in splicing, as we have presented and discussed previously in other marine teleosts (Gendron et al. 2020, 2022; Kwabiah et al. 2024), could produce isoforms of different sizes. However, further work will be required to explore what these shorter anti‐Gata1 reactive bands are in Atlantic salmon tissue. Overall, our sequence and western blotting analyses, taken together with our previously published work and additional information in the literature, strongly supports the existence of Atlantic salmon orthologs of the proteins we assessed herein.

Our results indicating an increase in expression level and/or a redistribution of ISAV, IgM, Sox6, Sox9, and collagen type 1 in central corneal tissues of ISAV‐infected salmon suggest an IgM‐related immune response to ISAV in central corneal tissues accompanied by Sox family protein and extracellular matrix (ECM) regulation. Sox9 and Sox6 play critical roles in ECM‐rich tissues such as cartilage (Lefebvre 2010), but their roles in response to pathogens in marine teleost cornea, which is also an ECM‐rich tissue, have not been reported. In ISAV‐infected salmon, we observed that the outermost layer of central corneal epithelial cells, which are the cells that would be the first in line to receive a waterborne cohabitation‐delivered ISAV insult, upregulates Sox9, while more basal central corneal epithelial cells upregulate Sox6. Our present studies were not performed at the resolution levels necessary to determine if Ig‐related responses in these animals are restricted to B cells, extracellular fluids, or Ig‐binding macrophages. Results shown here in Figure 6B might suggest that the localization of IgM in corneal tissues from ISAV‐infected fish changes noticeably, but further work will be needed to determine where IgM expression is located exactly. We recently described the use of an anti‐CD14 antibody in studies of ocular pathology in lumpfish infected with the Vibrio anguillarum marine bacterium. As infection proceeds in this model, CD14‐like expression was found to be increased in head kidney and the ocular choroid rete mirabile vasculature (Kwabiah et al. 2025). Since CD14 is a well‐established macrophage marker, future multiplex studies applying anti‐CD14 reagents to the Atlantic salmon ISAV ocular infection model should shed light on these questions.

Previous studies have indicated significant differences in the biology of the central versus limbic corneal regions. In mammals, central versus peripheral corneal regions show differences in ECM, biomechanical and immunochemistry properties (Jan et al. 2018; Hamrah et al. 2003). Mammalian corneal epithelium is renewed by limbal corneal epithelial stem/progenitor cells (LEPC) which are located at the corner‐scleral junction (Mikhailova et al. 2016) at which collagen fibre orientation differs significantly from central cornea (Girard et al. 2011). Human peripheral corneal cell division has been observed to be up‐regulated when the tissue is exposed to differentiation media with collagen substrates (Yoon et al. 2013). Therefore, the significant differences in the biology of central versus limbal corneal regions might impact host susceptibility and response to ISAV. Several of our histological stains did indeed highlight cellular changes (either infiltration or proliferation) and changes in collagen distribution patterns specific to the stroma of the corneal limbic region of ISAV‐infected Atlantic salmon compared to non‐infected fish. On the other hand, corneal epithelial dysplasia found associated with ISAV infection was observed in both the limbic and central regions. A previously published study associated increased nuclear localization of Sox9 with the activation of human LEPC in response to corneal epithelial wound healing (Menzel‐Severing et al. 2018). Interestingly, the Menzel‐Severing et al. (2018) study also showed predominantly cytoplasmic Sox9 expression in basally located LEPCs but showed nuclear expression in the supra‐basal cells (Menzel‐Severing et al. 2018). Our histochemical analyses using special stains indicating cellular changes and dysplasia of the corneal epithelium and stroma in ISAV‐infected salmon further supports a corneal pathological response to ISAV infection. Similar to low pathogenic influenza, epithelial cells are the main target of HPR0, the non‐virulent ISAV strain (Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015; Aamelfot et al. 2016), which does not cause systemic endothelial infection or appreciable immune responses. Whether HPR‐del, the virulent strain of ISAV, definitively targets corneal epithelial cells in vivo has not been previously reported. Our results herein suggest that Atlantic salmon corneal epithelium is indeed a target for pathogenic ISAV.

We also observed an upregulation in expression of nuclear Sox9 within cells near the outer edge of the ISAV‐infected Atlantic salmon eyelid skin. The eyelid skin tissue in ISAV‐infected specimens also showed morphological disorganization of the tissue. Therefore, eyelid skin could also potentially serve as a point of entry for ISAV or high‐virulence ISAV.

Using various histological stains and marker analyses, we observed vascular reorganizational changes that correlated closely with ISAV antigen immunoreactivity and increase in Gata1 expression in choroid rete mirabile of ISAV‐infected salmon eyes. The recently introduced RGB‐trichrome staining method developed by Gaytan et al. (2020) incorporated fast green and was designed for the histological assessment of skeletal tissues but is also useful for labeling erythrocytes bright green (Gaytan et al. 2020; Xuereb and Beesley 1954). Our observations in choroid rete mirabile tissue of a robust presence of bright green stained erythrocytes of non‐infected fish versus a paucity of bright green‐stained erythrocytes in ISAV‐infected fish suggests an ocular manifestation of the systemic anaemia in the ISAV‐infected animals. The increased density of small red/orange stained leukocytic‐like cells across the whole section of choroid rete mirabile in RGB‐stained sections from the ISAV‐infected specimens suggests that additional cellular changes might be occurring in this tissue. Mammalian Gata1 is important for homeostatic and stress‐induced erythropoiesis (Gutiérrez et al. 2008; Chennupati et al. 2018) and Gata1 orthologs have been characterized in teleosts (Tian et al. 2021; Xu et al. 2015; Gallo and Civinini 2003). In light of ISAV‐caused anaemia, which is a stress to erythropoiesis, we hypothesized that Gata1 might be deregulated in choroid rete mirabile tissues of ISAV‐infected Atlantic salmon. The nuclear and cytoplasmic subcellular localizations we observed for Gata1 in the choroid rete mirabile tissues of infected fish is consistent with that previously described in mouse bone marrow derived cells which undergo erythropoiesis (Gutiérrez et al. 2007). The increase we observed in Gata1 immunoreactivity in both vascular wall structures and tightly packed pockets of leukocytic‐like round cells in choroid rete mirabile tissues of ISAV‐infected fish could be the result of a compensatory response for ISAV‐driven anaemia by inducing erythropoiesis. However, as a transcription factor, Gata1 is known as a master regulator in programming haematopoietic progenitors toward both megakaryocyte and erythrocyte development (Ferreira et al. 2005). Ferreira et al. (2005) and Iwasaki et al. (2003) also highlight Gata1's role in development and terminal differentiation of eosinophils and mast cells. Both eosinophils and mast cells are known to participate in the host response to pathogens in mammals (Fulkerson and Rothenberg 2008; Migliaccio et al. 2007). Regardless of phenotype, we do not yet know if these Gata1 positive cells accumulate in the eye after entry from the systemic circulation or if they originate from precursors within the eye. Extramedullary haematopoiesis is rare but can occur in some cases in ocular tissues (Arredondo et al. 2008). In teleosts, it is generally accepted that the barrier between haematopoietic tissue and blood within the circulation is relatively weak, allowing immature erythrocytes into tissues (Witeska 2013). Future studies applying additional phenotyping reagents could address these questions.

Previous studies described ISAV in the eye but following infection of other organs, which was taken to suggest that the eyes were infected through the cardiovascular system rather than acting as an actual point of entry (Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015). Relevant to the vascular route, we observed increased expression of ISAV in the choroid rete mirabile of ISAV‐infected specimens. The choroid rete mirabile receives blood from the pseudobranch which contacts water (Wittenberg and Haedrich 1974). Gills are thought to be a point of entry for ISAV (Rimstad et al. 1999; Cárdenas et al. 2024). If the pseudobranch is an entry point, any incoming retinal ISAV could originate from the choroid rete mirabile either as water or blood‐borne material. Thus, it is possible that the choroid rete mirabile vasculature may be a mode of entry of ISAV into the eye. The other potential route of entry for ISAV is the central cornea or the limbic regions (Gendron et al. 2022). Since both non‐infected and ISAV‐infected fish showed a low level of corneal epithelial ISAV immunoreactivity, it is also possible that high‐virulence ISAV replicates in the corneal limbic stroma, after which it could penetrate the inner ocular tissues such as retina. Our results showing significant molecular changes (ISAV, IgM, Sox6, Sox9, collagen type 1 immunoreactivity) and morphological dysplasia in both central and limbic corneal tissues in the ISAV‐infected samples indicates that the corneal tissues in ISAV‐infected Atlantic salmon might lose the microbiological barrier properties that healthy corneal tissues normally serve in most animals (Kinoshita et al. 2001).

The neural retinal dysplasia and degeneration we observed in the ISAV‐infected Atlantic salmon is consistent with the loss in integrity of Bruch's membrane barrier in these animals. This level of retinal pathological changes, if observed in human subjects, would likely cause significant vision loss. While the most abundant protein component of mammalian Bruch's membrane is collagen (Clark et al. 2025), elastin also plays an important role in neural retinal barrier and Bruch's membrane integrity. Abnormal elastin turnover/degradation is associated with several ocular diseases (Navneet and Rohrer 2022). The Verhoeff elastin histology results herein indicating a degeneration of Bruch's membrane in ISAV‐infected Atlantic salmon suggest that the ISAV infection in these fish can progress to retinal barrier breakdown. This observation is important since retinal barrier breakdown at the level of Bruch's membrane is a hallmark of some human retinal diseases including age‐related macular degeneration (AMD) (Clark et al. 2025; Biasella et al. 2023).

Surprisingly, we observed CD10 expression in the extraocular muscle myofibrils of both non‐infected and ISAV‐infected Atlantic salmon, with general upregulation of CD10 in the infected specimens. CD10 is a marker for haematopoietic progenitor cells in mammalian bone marrow (Wang et al. 2024), plays a role in B cell development (Torlakovic et al. 2005) and chondrogenesis (Ding et al. 2020). CD10 is an endopeptidase expressed both on cell surfaces of fibroblasts and neutrophils, and as a soluble form observed in serum and cerebrospinal fluid (Bayés‐Genís et al. 2016; Komada et al. 1986). The soluble form of CD10 retains its catalytic activity similar to that of the membrane‐bound form (Bayés‐Genís et al. 2015). Extraocular muscles are specialized skeletal muscles responsible for eye movement. While they share characteristics with other skeletal muscle types, they are unique in that they are typically unaffected by muscular dystrophies as well as resistant to fatigue (Titova et al. 2024). Extraocular muscles are divided into two layers: The outer orbital layer that contains thin muscle fibres, and the global layer that contains thicker muscle fibres (Horn and Straka 2021). After injury, satellite cells and fibro‐adipogenic progenitor cells (FAPs) both play key roles in the repair process. Satellite cells are muscle resident stem cells that differentiate into myocytes, and FAPs are mesenchymal stem cells embedded in muscle connective tissue that maintain structural integrity for regenerating muscles (Collins and Kardon 2021). Human FAPs, as well as satellite cells, express CD10 and CD10 have been shown to be upregulated in cases of rhabdomyolysis (Schutz et al. 2024; Wu et al. 2018). Herovici histological stain revealed possible ocular rhabdomyolysis or degeneration of the ocular muscle myofibrils in ISAV‐infected Atlantic salmon. As CD10 is a marker of FAPs and satellite cells, the increase in expression might be related to an attempted muscle regeneration in response to tissue degradation mediated by ISAV infection but further studies will be needed to explore this.

Important concerns about results interpretation in histopathology and comparative pathobiology in aquatic animal health sciences have been raised and recently reiterated (Wolf et al. 2015; Speare et al. 2025). Several aspects of the work herein address these concerns. Tissue fixation, embedding and processing involved a fish pathology diagnostics laboratory (Aquatic Diagnostic Services, AVC‐UPEI) and was standardized and consistent between the control and experimental groups. Care was taken to embed the eyes at the same orientation to avoid any plane‐of‐section artefacts. Histological staining was done by qualified expert technologists at a Histology Core Facility (Histology Core Facility Faculty of Medicine, Memorial University). Microscopic imaging was performed using state‐of‐the‐art high‐resolution instrumentation at optimized light and colour balance levels. While the observations were qualitative, sufficient numbers of tissue blocks, fellow eyes, tissue sections, histological stainings and IHC were conducted to increase the confidence level of the observations and findings. The IHC approach itself served as further validation of the histological changes observed since in many cases the histopathology correlated with distinctive marker protein expression changes. Western blotting was also used to further validate cross reactivity of Ab reagents used in IHC in cases where such literature in Atlantic salmon is not yet available.

This study has several limitations and caveats. First, ISAV‐moribund animals were assessed at only one time point, 30 days post infection, of the co‐habitation infection process. The study of earlier time points might provide better insight into points of entry and the tissue changes underlying subsequent ocular pathology. Second, IHC and histology are only correlative and cannot prove the function of specific proteins in pathogenesis. However, since there is limited literature in marine teleosts for the role of the protein markers we assessed in ISAV disease, we view these studies as a springboard for future exploration. Third, as we have reported previously, marine teleost orthologs of marker proteins can display interesting divergence from their mammalian counterparts (Gendron et al. 2022; Kwabiah et al. 2024, 2025).

Previous studies on ISAV pathological changes have indicated that different tissues such as head kidney, gills, and spleen respond differently to the infection (Gervais et al. 2022). Our results herein reflect this pattern of diversity in response to ISAV pathology in the Atlantic salmon eye. Previous work on ISAV has shown that early replication of ISAV occurs in various mucosal surfaces such as pectoral fin and skin, with multiple potential entry points for the virus other than the gills (Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015). Aamelfot, Dale, et al. (2015) and Aamelfot, McBeath, et al. (2015) reported that initial infection begins in epithelial cells and progresses to endothelial cells. It was suggested that eye tissue is infected through a systemic route rather than primary viral entry (Aamelfot, Dale, et al. 2015; Aamelfot, McBeath, et al. 2015). Our results raise the new possibilities of a choroid rete mirabile and/or corneal entry route for ISAV. Recent studies have implicated aerosol transmission of ISAV (Krishna et al. 2025). Farmed Atlantic salmon can display above‐ocean‐surface swimming behaviour (Gamperl et al. 2021). Therefore, corneal points of entry of ISAV might be relevant to aerosol‐based ISAV infection. Regardless of the entry route, our results also suggest that ISAV ocular pathological changes might include an ocular erythropoietic or leukocytic response to the infection. Whether such a response originates inside or outside of the eye remains to be explored. Previous work has shown that the nonvirulent ISAV‐HPR0 strain acts as a reservoir and is a risk factor for emergence of ISA disease. There is also evidence of genetic and functional evolution from an ISAV‐HPR0 variant to a low‐virulence ISAV variant (Christiansen et al. 2017). These studies, taken together with our work herein, raise the possibility that the Atlantic salmon eye, which is a barrier organ, might serve as an isolated reservoir for the pathogenesis of virulent ISAV. Orthomyxoviruses such as H1N1 have been recognized to infect eye tissues, including human retinal microvascular endothelial cells (Belser et al. 2018; Yang et al. 2024). H1N1 can lead to ocular manifestations including retinal lesions, vasculitis, retinitis, and conjunctivitis (Yang et al. 2024). Avian and human influenza A and H7 viruses affect eyes, use ocular tissue as a portal of entry, and show ocular tropism (Belser et al. 2018). Human/avian influenza virus is capable of replicating in distinct ocular cell types, and conjunctivitis is the primary complication that occurs in the eye with an influenza infection (Belser et al. 2018). Our current findings on the ISAV orthomyxovirus broaden our basic understanding of how these viruses use and damage ocular tissue and suggest that ISAV might serve as a new model for studying orthomyxovirus‐related ocular disease.

Author Contributions

Rebecca R. Kwabiah: conceptualization, data curation, formal analysis, visualization, writing – original draft, methodology, investigation, supervision, writing – review and editing, validation. Hélène Paradis: conceptualization, data curation, formal analysis, visualization, writing – original draft, methodology, investigation, writing – review and editing, validation, resources. Emily Mahon: conceptualization, data curation, formal analysis, visualization, writing – original draft, methodology, investigation, writing – review and editing, validation. John Law: conceptualization, visualization, methodology, investigation, resources. Ignacio Vasquez: writing – original draft, writing – review and editing. Sathees Duglas: writing – original draft, methodology, writing – review and editing, resources. Shams Al‐Badoosh: writing – original draft, writing – review and editing. Raja Ram Gurung: writing – original draft, methodology, writing – review and editing. Javier Santander: conceptualization, writing – original draft, visualization, methodology, investigation, supervision, project administration, writing – review and editing, funding acquisition, resources. Manuel A. Soto‐Dávila: conceptualization, data curation, visualization, writing – original draft, methodology, investigation, writing – review and editing, resources. Robert L. Gendron: conceptualization, data curation, formal analysis, visualization, writing – original draft, methodology, investigation, supervision, project administration, writing – review and editing, validation, funding acquisition, resources. Thurka Paramanathan: writing – original draft, methodology, writing – review and editing, resources. Mark D. Fast: conceptualization, data curation, formal analysis, visualization, writing – original draft, methodology, investigation, supervision, project administration, writing – review and editing, validation, funding acquisition, resources.

Funding

This work was funded through grants from Medical Research Foundation of the Faculty of Medicine, through leverage funds from School of Graduate Studies, Memorial University, NSERC‐Discovery (RGPIN‐2022‐04171 to R.G.; RGPIN‐2018‐05942 to J.S.) and NSERC Undergraduate Student Research Award to E.M. Some research infrastructure used in this work was funded in part by Canada Foundation for Innovation.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data can be made available upon reasonable request.