Fish & Shellfish Immunology 33 (2012) 277e285

Contents lists available at SciVerse ScienceDirect

Fish & Shellfish Immunology

journal homepage: www.elsevier.com/locate/fsi

Pigment-producing granulomatous myopathy in Atlantic salmon: A novel

inflammatory response
Hilde A.S. Larsen a, Lars Austbø b, Turid Mørkøre c, Jim Thorsen d, Ivar Hordvik e, Uwe Fischer f,
Emilio Jirillo g, Espen Rimstad h, Erling O. Koppang a, *
a
Section of Anatomy and Pathology, Institute of Basic Science and Aquatic Medicine, Norwegian School of Veterinary Science, Ullevålsveien 72, PO Box 8146 Dep., 0033 Oslo, Norway
b
Section of Genetics, Department of Basic Science and Aquatic Medicine, Norwegian School of Veterinary Science, Oslo, Norway
c
Nofima Marin AS, Ås, Norway
d
Section of Cancer Cytogenetics, Institute for Medical Informatics, Oslo University Hospital HF, Oslo, Norway
e
Department of Biology, University of Bergen, Bergen, Norway
f
Friedrich-Loeffler-Institut, Federal Research Institute for Animal Health, Institute of Infectology, Greifswald-Insel Riems, Germany
g
Department of Immunology, Faculty of Medicine, University of Bari, Bari, Italy
h
Section of Microbiology, Immunology and Parasitology, Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, Oslo, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Melanin comprises a complex group of pigmented polymers whose primary function is ascribed to
Received 23 December 2011 dermal solar protection, but may also have an interesting role in innate immunity. In ectothermic
Received in revised form vertebrates, melanogenesis is reported in leukocyte populations, but it is not known if this occurs in
29 March 2012
connection with inflammatory reactions. Melanin accumulations in ectopic locations, in particular
Accepted 7 May 2012
Available online 14 May 2012
muscle, represent a serious quality problem in salmon production. Here, we investigated such changes
for the expression of dopachrome tautomerase and tyrosinase as well as some important immune genes
and pathogens. Furthermore, the nature of the pathological changes was addressed by morphological
Keywords:
Granulomatous inflammation
methods. Gene transcripts encoding key enzymes in melanogenesis, suggesting a de novo melanin
Melanin synthesis in pigmented muscle, were found. MHC class II transcripts were up-regulated and there was no
Melanogenesis indication of bacterial or viral infection. The histological examination revealed granulomatous inflam-
Melanin-producing leukocyte mation with distribution of MHC class II positive cells and T cells, analogous to the pattern found in
Pigmentation mammals. Importantly, in contrast to mammals pigmented cells were contributing in the inflammation.
We demonstrate that melanin production occurs in granulomatous inflammation in salmon, revealing
a close and hitherto unreported link between the pigmentary and immune systems.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction shown to produce several inflammatory mediators, suggesting

Melanocytes are melanin-producing cells, and in mammals,
they derive from ectoderm [1]. The functions of melanin synthesis
and deposition in dermal tissues have generally been attributed to
the need for solar protection. Nevertheless, there is an increasing
interest to investigate the relationship between the pigmentary
and the immune systems [2]. There are indications that melanin
plays a role in immune functions such as antimicrobial defense,
suggesting that immune modulation exerted by the pigmentary
system might be an important and underestimated entity [2,3].
Melanocytes respond to cytokines, including interferons, interleu-
kins and tumor necrosis factor [3e5]. Furthermore, they have been
Abbrevations: Dct, dopachrome tautomerase; Tyr, tyrosinase.
* Corresponding author. Tel.: þ47 22 96 45 46; fax: þ47 22 96 47 64.
E-mail address: erling.o.koppang@nvh.no (E.O. Koppang).
participation in the inflammatory response [4,6]. Interestingly, in
humans skin coloration correlates with susceptibility to certain
infectious diseases. This connection was indicated as early as in the
beginning of the 19th century by Dr. William Charles Wells [7].
During the Vietnam War, Allen et al. [8] found that the incidence of
bacterial pyoderma among American troops was significantly
higher in Caucasians than African Americans. Melanin and mela-
nocytes are not only confined to the skin, but are also found in
ectoderm-derived cells in extracutaneous locations, i.e. the brain,
meninges, choroid of the eye and in the inner ear, where protective
roles have been suggested [2]. Increased focus on the immuno-
logical role of mammalian melanocytes invites comparison to non-
mammalian species and their extracutaneous melanocytes, which
seem to be closely linked to the immune system [6,9e11].
In contrast to mammals and birds, ectothermic vertebrates
possess a second group of melanin-producing cells originating from

1050-4648/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fsi.2012.05.012
278 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285

mesenchymal, hematopoietic stem-cells [12]. These cells are

commonly termed melanomacrophages in the literature of ecto-
thermic vertebrates. A well-established Atlantic salmon (Salmo
salar) cell-line (SHK-1 cells), classified as headekidney derived
macrophages, expresses transcripts of the tyrosinase gene family
and produces melanosomes [13,14]. The principal genes related to
melanogenesis include the tyrosinase gene family, and the
expression of tyrosinase-related protein 2/dopachrome tautomer-
ase (TYRP-2/Dct) is considered to be specific for melanocytes and
their precursors [15]. Dct-expression is present in important
lymphatic organs of fish including the head, kidney and the spleen,
where melanomacrophages are found dispersed throughout the
tissue [14,16]. These cells are thought to be part of the antigen-
presenting cell system [16,17].
Inflammatory responses in fish show several similarities to
those of mammals [9]. However, it has been noted that, in contrast
to mammals, chronic inflammation in fish may appear with
abundant pigmentation [18,19]. The presence of melanin in ectopic
locations of farmed salmon, especially in the muscle fillet, repre-
sents a considerable quality problem for the salmon industry
affecting up to 20% of the fillets at the processing plants [20]. The
presence of possible melanin-producing leukocytes in salmon
indicates that melanin may play an active role in inflammation in
fish, and establishes a collaborative relationship between the
pigmentary and immune systems. This would be a unique finding,
as the leukocytes of mammals do not use melanin production in
their defense responses and such an inflammatory response has not
been described previously. The purpose of this work was to
investigate the possible de novo production of melanin at the site of
inflammation, and further to characterize pathological changes and
underlying molecular mechanisms associated with pigmentation of
the fillet, using and combining both morphological and gene-
transcriptional approaches.

2. Materials and methods Fig. 1. Gross pathological changes of investigated Atlantic salmon muscle. A. An area
with brown-black pigmentation, or a pigment spot. B. Section through a typical pig-
2.1. Samples from fish mented area. Bar ¼ 5 mm.

The fish sampled were vaccinated Atlantic salmon with

abnormal pigmentation of the muscle fillet. They originated in four
commercial farms and had routinely been intraperitoneally injec-
ted with oil-adjuvant vaccines before transfer to sea water,
following standard procedures [21]. The fillets were collected from
slaughter-sized salmon harvested at four different locations -
Hjelmeland (8 fillets), Eggebønes (10 fillets), Romsdalsfjord (4
fillets) and Frøya (8 fillets).
Tissue samples were collected from the areas of brownish
discoloration (Fig. 1A,B). In addition to the pigmented sample, non-
pigmented control samples of apparently-normal muscle were
obtained from the same fillet at various distances from the pig-
mented areas. Pigmented and non-pigmented tissue samples were
processed for histology and immunohistochemistry (IHC). Macro-
dissected tissues from the second and third sampling, i.e. a total
of 28 samples from 14 fillets, were analyzed by quantitative real-
time reverse transcriptase polymerase chain reaction (RT-qPCR).
In addition, tissues from the last sampling were analyzed using
laser-assisted micro-dissection.
2.2. Gene expression analysis
Macro-dissected tissue samples (3  3  3 mm) intended for
RT-qPCR were immediately transferred to RNAlaterÒ after sampling
and kept at 4  C until processed. RNA was extracted using
an RNeasy MiniÒ kit (QIAGEN, Hilden, Germany), and concentra-
tions and purity were measured spectrophotometrically with
a Biospec-Nano (Shimadzu Corporation, Kyoto, Japan). The cDNA
synthesis was performed immediately after extraction using
w600 ng RNA/reaction, Oligo(dt)-primer and OmniscriptÒ kit
(QIAGEN). Real-time PCR reactions were set up using PlatinumÒ
Quantitative PCR Supermix-UDG (Invitrogen AS, Oslo, Norway) and
run on a Cromo 4Ô (Bio Rad Laboratories AB, Oslo, Norway). The
following genes were analyzed by real-time RT-PCR e tyrosinase
(Tyr), dopachrome tautomerase (Dct), major histocompability
complex class II (MHC class II), interferon a (IFNa), cluster of
differentiation 3 (CD3z), membrane bound and secretory immu-
noglobulin M (mIgM, sIgM). When possible, primers and probe
were designed to span across intron sections [Table 1]. The level of
expression was normalized using elongation factor 1AA (EF1AA) as
the reference gene [22].
For laser-capture micro-dissection, tissue samples
(10  10  3 mm) of pigmented muscle from eight individuals were
immediately transferred to liquid nitrogen and stored at 70  C
until further processing. Cryo-sections, 14 mm in thickness, were cut
with a cryostat and mounted on special membrane slides, as
described previously [23]. The sections were air dried for 1 h,
washed twice in xylene to remove fat and stained with RNase-free
hematoxylin before air-drying for 1 h. Subsequently, a representa-
tive part of melanomacrophage accumulations (Fig. 2A,B) were
dissected [23]. Up to eight successive dissected sections, with
a total area ranging between 0.3 and 0.5 mm2, were cut and
collected in tubes. From each individual, neighboring non-
 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285 279

Table 1
Gene assays used for quantitative real-time RT-PCR.

Genes GeneBank Assay concentration Sequence (50 e30 ) Ct value mean  SD

accession
EF1Aa X70165 Primers 900 mM F e CCCCTCCAGGACGTTTACAAA 26.33
Probe 200 mM R e CACACGGCCCACAGGTACA 1.50
P e FAM-ATCGGTGGTATTGGAAC-MGB
Tyra DQ408549 Custom TaqMan F e CTGGGCTACGAATACTCCTACCT 32.80
R e TGCATGTAGGGAGTCAGGAACT 2.05
p e FAM-CTCGTGCCTGGGTCAAG-MGB
Dcta,b DQ339140 Custom TaqMan F e TCTCACTCTGCAGCCAATGAC 36.54
R e CAGACTTCCTCATCCACTCATCAAA 1.14
P e FAM-CATGGATCACCAGAAAGA-MGB
MHC II X70165 Primers 900 mM F e CTCACTGAGCCCATGGTGTAT 25.32
Probe 200 mM R e GAGTCCTGCCAAGGCTAAGATG 1.95
P e FAM-CTGGGACCCGTCCCTG-BHQ1
CD3zeta BT060238 Primers 900 mM F e AACAGGGATCCAGAGAGTGCTG 32.35
Probe 200 mM R e AAGGGACGTGTAAGTGTCGTCA 1.87
P e FAM-ACGGCACGCGATAATCGCAGGA-BHQ1
IFNa1 AY216594 Custom TaqMan F e TCAACATTCTAGAACGCCAATTGGA 35.77
R e CGTTTCAGTCTCTTCTCAGGTTTCA 1.76
Pe FAM-CAGGTGATACACAGGATTTA-MGB
Mx U66475 Primers 900 mM F e GATGCTGCACCTCAAGTCCTATTA 29.21
Probe 200 mM R e CACCAGGTAGCGGATCACCAT 2.04
P e FAM-CTGATCAGCCAAACGTTGACTGGATATCCT-TAMRA
mIgM S48658 Primers 900 mM F e TGCGCTGTAGATCACTTGGAA 28.45
Probe 200 mM R e ATGGTGTTGCTGCATGGACA 3.02
P e FAM-AGGGAGACCGAGTGTCTCGTGTTGACTG-MGB
sIgM S48652 Primers 900 mM F e TGCGCTGTAGATCACTTGGAA 22.92
Probe 200 mM R e GCCAGCAGAAAGACAGATGGA 2.83
P e FAM-AGGGAGACCGGAGGAGATCCACAGC-MGB
RIG1 NM_001163699 Primers 900 mM F e ACAGCTGTTACACAGACGACATCA 29.82
Probe 200 mM R e TTTAGGGTGAGGTTCTGTCCGA 1.75
P e FAM-TCGTGTTGGACCCCACTCTGTTCTCTC-BHQ1
Vig-1 BT047610 Primers 900 mM F e AGCAATGGCAGCATGATCAG 34.95
Probe 200 mM R e TGGTTGGTGTCCTCGTCAAAG 3.22
P e FAM-AGTGGTTCCAAACGTATGGCGAATACCTG-BHQ1

Primers and probes used for realntime RT-PCR on pigmented and un-pigmented muscle tissue of Atlantic salmon. Primers and probe were design to span intron section and
exon-junction sites located in the probes, indicated in bold.
a
Primers and probe were custom designed by Assays-by-DesignSM service consisting of a 20 mix of unlabeled PCR primers and TaqManÒ MGB probes (FAMÔ dye labeled)
and used in a concentration of 900 nM (primer) and 200 nM (probe).
b
Samples with no detectable expression within 45 cycles could not be included in the calculation as this would result in an erroneous average value.

pigmented inflamed tissue was prepared for comparison
(Fig. 2C,D). Total RNA was isolated from the laser-captured material
by using a Nucleo Spin RNA XS (Macherey & Nagel, Düren,
Germany). This kit allows rapid purification of high-quality total
RNA from extremely small samples. The RNA was eluted into 15 ml
RNase-free H2O and directly used for cDNA synthesis without
freezing the RNA samples. The manufacturer’s manual was fol-
lowed including an integrated DNase step. Reverse transcription of
13 mL total RNA solution was performed using M-MLV Reverse
transcriptase (Promega, Madison, WI 53711 USA). The cDNA
synthesis was prepared with oligo(dT) and random hexamer
primers, in addition RNase inhibitor was added to prevent RNA
degradation. Quantitative real-time PCR was performed using
TaqMan Gene Expression Master Mix (Applied Biosystems, Carls-
bad, CA 92008 USA) and carried out in a 7900HT fast real-time PCR
system (Applied Biosystems) according to the producer’s instruc-
tions and running 45 cycles. Tyr and Dct were analyzed in the
micro-dissected material and the level of expression was normal-
ized using EF1AA as the reference gene.
To elucidate of the viral infection status, a selection of virus-
induced genes and sequences from viruses relevant to the local
environment were investigated. Retinoid-inducible gene (RIG1),
a pattern recognizing cytosolic molecule primarily detecting viral
nucleic acids based on structure and localization [24], virus
inducible gene (vig-1) and antiviral IFN-induced gene Mx, the latter
two coding for interferon-inducible, virus inhibitory proteins
[25,26], were investigated. The samples were also tested for the
presence of pancreas disease virus (SPDV), infectious pancreatic
necrosis virus (IPNV) and piscine reovirus (PRV) by RT-qPCR.
In these expression studies, each sample was run in triplicate
and a control lacking the template for each master mix was always
included in the experiments. Sequences and final concentration of
selected probes and primers are displayed in Table 1.
2.3. Statistical analysis
Gene expression in pigmented and non-pigmented tissue
samples was evaluated using the paired t-test, comparing pig-
mented to non-pigmented tissue within each fillet. The data were
tested to examine whether they were sampled from a Gaussian
distribution using the KolmogoroveSmirnov normality test. Differ-
ences in expression between pigmented and non-pigmented tissue
were considered to be significant at p < 0.05. The statistical calcu-
lations and graphical presentation in Figs. 3 and 4 were performed
using Prism 5.0 software (GraphPad Software, San Diego, CA, USA).
2.4. Histology
Tissue samples (10  10  3 mm) were fixed in 10% buffered
formalin for 96 h, paraffin-embedded and cut in 4 mm sections,
mounted and dried at 37  C, before staining with hematoxylin and
eosin (H&E), van Giesson, Sudan black, rubeanic acid or Pearl’s Berlin
blue. To confirm the presence of melanin, sections were stained with
Masson/Fontana. For detection of prokaryotic and eukaryotic
280 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285

Fig. 2. Laser-capture micro-dissection of areas with accumulations of melanomacrophages (A,B) and neighboring non-pigmented inflamed tissue (C,D). Hematoxylin stain.
Bar ¼ 50 mm.

Fig. 3. Relative transcription levels of Tyr (A), Dct (B), MHC class II (C) and mIgM (D) to EF1AA in pigmented lesions and non-pigmented tissue samples of muscle fillets of Atlantic
salmon. LCMD represents samples collected by laser-capture micro-dissection. The scatter plots show the transcript levels for each individual with mean for each group as obtained
by RT-qPCR analysis. Expression levels are presented relative to the mean of the control group that was given value 1. Error bars represent standard error of the mean (SEM).
p-values << 0,01 (significant).
 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285 281

Fig. 4. Histopathological investigations of formalin-fixed, paraffin-embedded sections of pigmented skeletal muscle tissue including immunohistochemistry. A. Degeneration and
repair, side by side. An empty vacuole encircled by macrophages and melanomacrophages in an area with fibrosis and fat-infiltration (asterisk). Acute degeneration of a muscle-fiber
(arrow). H&E stain. Bar ¼ 200 mm. B. A vacuole containing material stained for fat (black). Black droplets are also seen in the surrounding macrophages (arrowhead). Sudan black
stain. Bar ¼ 50 mm. C. Melanomacrophages encircling the vacuoles, creating a nearly dense rim of pigmentation. The larger vacuoles are surrounded by sparse amounts of leucocytes,
mainly consisting of pigmented and non-pigmented cells. Hematoxylin stain. Bar ¼ 50 mm. D. A large vacuole in an area with severe fibrosis and fat-infiltration (arrowhead).
Proliferating myoblasts with ineffective regeneration attempts; so-called muscle giant cells (arrow). H&E stain. Bar ¼ 200 mm. E. Cell-rich granulomas with various amounts of
melanin-containing cells displaying a heterogeneous morphology. H&E stain. Bar ¼ 200 mm. F. Transverse section of a necrotic myofibre encircled by MHC class IIþ macrophages
(red) (arrowhead). MHC class IIþ macrophages (red) entering the lumen and cleaning up the debris of a necrotic myofibre (arrow). MHC class II immune stain, hematoxylin counter
stain. Bar ¼ 200 mm. G. MHC class IIþ macrophages (red) in an area with abundant leukocyte infiltration. A vacuole encircled by macrophages and melanomacrophages is seen in the
upper left corner. Macrophages with only sparse amount of pigment-granules (arrowhead) are staining stronger than the heavily-pigmented cells that are encircling the vacuoles.
MHC class II immune stain, hematoxylin counter stain. Bar ¼ 50 mm. H. A cell-rich granuloma with central vacuoles and abundant MHC class IIþ macrophages (red). A few mel-
anomacrophages are present within the cell accumulation. MHC class II immune stain, hematoxylin counter stain. Bar ¼ 200 mm. I. T cells (red staining) encircling a cell-rich
granuloma. CD3 immune stain, hematoxylin counter stain. Bar ¼ 200 mm. J. IgMþ plasma in the vessel lumen staining intensely red. Only sparse single cells are IgMþ in the
pigmented areas. IgM immune stain, hematoxylin counter stain. Bar ¼ 50 mm. K. a-actinþ cells (red) interpreted as activated satellite cells in an area with repair, displayed by the
formation of myotubes with an internal chain of nuclei (arrowhead). a-actin immune stain, hematoxylin counter stain. Bar ¼ 100 mm. L. Negative control for the immunohisto-
chemistry lacking primary antisera. Hematoxylin counter stain. Bar ¼ 200 mm. H, I and L were produced in serial sections. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

infectious agents, sections were stained with Ziel-Nielsen
(Mycobacteria), periodic acid-Shiff (mucus of fungus) and Gram
(Gram bacteria). All stainings followed standard procedures [27].
2.5. Immunohistochemistry
Formalin-fixed, paraffin-embedded sections were subjected to
immunohistochemistry to elucidate the morphological distribution
pattern of the leukocytes in the inflammatory response. Polyclonal
rabbit antisera recognizing MHC class II b-chain (dilution 1:1000)
[16], CD33 (dilution 1:600) [28], immunoglobulin M (IgM) (dilution
1:25,000; kindly provided by Dr. Knut Falk, The National Veterinary
Institute, Oslo, Norway) [29], and monoclonal antibody against a-
actin (dilution 1:1000) (Thermo Fisher Scientific, Fremont, CA
94538 USA) were used. Tissue sections (4 mm) were placed on
polysine-coated slides (Menzel-Gläser, Braunschweig, Germany)
282 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285

and dried for at least 12 h at 37  C. After a standard de-waxing and
rehydration procedure, sections were autoclaved in citrate buffer
(10 mM citric acid monohydrate, pH 6.0) at 121  C for 15 min,
transferred to PBS and treated with 0.05% phenylhydrazine
(SigmaeAldrich, St. Louis, MI, USA) at 37  C for 40 min to inhibit
endogenous peroxidase, and washed overnight in PBS at 4  C. The
next day, the sections were incubated for 20 min in normal caprine
serum (SigmaeAldrich) diluted 1:50 in 5% bovine serum albumin
(BSA) (SigmaeAldrich) in Tris-buffer solution (TBS). For immune
labeling, an indirect method (EnVisionÔ þ System HRP, Dako, CA
93013, USA) was used according to the manufacturer’s instructions.
The primary antibodies were diluted in 1% BSA. All incubations
were performed in a humid chamber on a rocking table at room
temperature. The sections were washed in distilled water and
either left unstained, or counterstained with hematoxylin, washed
in distilled water and mounted with polyvinyl alcohol mounting
medium, pH 8.2, and examined with a light microscope. Sections
treated with pre-immune rabbit and mouse sera served as negative
controls in the experiment. Microphotographs were captured with
a SpotRT Slider digital camera (Diagnostic Instruments, Sterling
Heights, MI, USA).
3. Results
3.1. Macroscopic pathology
The body condition of the affected fish ranged from adipose to
moderately reduced, evaluated from the fillet size, shape and
texture. Areas of brown-black discoloration were seen on the
muscle fillet after removal of the peritoneum and spine (Fig. 1A).
The pigmentation occurred mostly sub-peritoneally in the
abdominal wall, but caudal and dorsal localizations remote from
the abdominal cavity were also observed. Some of the pigmented
areas were found deep in the musculature, with no connection to
the skin or peritoneum. Each affected fillet had mostly only one, but
occasionally up to three areas with discoloration. Pigmented areas
varied strongly in size and shape, from a few millimeters to several
centimeters in diameter (Fig. 1B). Occasionally, the pigmented areas
penetrated the entire muscle throughout to the skin. The larger
areas showed a firmness and contraction characteristic of scar
tissue.
3.2. RT-qPCR
3.3. Histopathology
The changes in the pigmented samples showed the character-
istic features of a chronic polyphasic necrotizing myopathy, with
the formation of granulomas and large extracellular vacuoles
(Fig. 4A). Pigmentation by melanin-containing cells, interpreted as
teleost melanomacrophages, occurred abundantly in well-
organized granulomas and chronically inflamed areas with signs
of repair, whereas more acute lesions, indicated by the presence of
destruction and lack of organization, showed only sparse infiltra-
tion by such cells. Vacuoles of various sizes and shapes were
observed throughout the affected areas. The vacuoles seemed
empty, although a few contained an amorphous material staining
positive for fat, 4 (Fig. 4B). Abundant melanomacrophages encircled
the vacuoles, creating a dense rim of pigmentation, apparently
creating a barrier between the content of the vacuoles and the
surrounding tissue. The larger vacuoles were surrounded by more
modest numbers of leucocytes, mainly consisting of pigmented and
non-pigmented macrophages (Fig. 4C). High levels of fibrosis and
fatty infiltrates were present in areas with large vacuoles, con-
firming previous tissue destruction and subsequent repair by
scarring, rather than by restoration (Fig. 4D). Cell-rich granulomas
contained varying numbers of melanomacrophages (Fig. 4E) dis-
playing differing cell-morphology. The melanomacrophages
surrounding the vacuoles were strongly elongated with densely-
packed melanin compartments. However, the appearance of pig-
mented cells varied strongly throughout the granulomatous tissue,
from sparsely-pigmented, rounded macrophage-like cells to volu-
minous, dendritic-shaped cells containing abundant amounts of
melanin (Fig. 5). This morphological variation of the melanoma-
crophages was also observed in less organized areas of the inflamed
tissue, like areas with abundant leukocyte infiltration and without
vacuoles and granulomas.
Staining with Masson/Fontana technique confirmed the pres-
ence of melanin in the pigmented cells. Staining for iron and copper
proved negative, as did the search for infectious agents (staining
with Ziel-Nielsen, PAS and Gram) (data not shown). This is
consistent with previous reports [18].
3.4. Immunohistochemistry
Based on examinations of the inflamed and pigmented areas in
all fillets, the immunostaining intensity was labeled as abundant,

In the macro-dissected tissue samples with pigmented lesions,

transcription of both Tyr and Dct was significantly up-regulated
compared to non-pigmented areas within the same fillet
(Fig. 3A,B). Transcription of Dct was detected in 13 out of 14
samples from pigmented areas, whereas only 1 out of 14 from non-
pigmented areas (Fig. 3B).Transcription of MHC class II (Fig. 3C) and
mIgM (Fig. 3D) was significantly higher in pigmented samples
compared to non-pigmented controls, while the opposite was
noted for RIG1 (not shown). No significant differences in the tran-
scription levels of CD3, sIgM, vig-1, IFNa or Mx were detected. Mean
ct-values are given in Table 1.
SAV, IPNV and PRV were not detected, which is consistent with
the transcription level of the virus-induced genes.
In the micro-dissected tissue samples transcription of both Tyr
and Dct was significantly up-regulated in areas containing accu-
mulations of melanomacrophages compared to non-pigmented
parts at the same site of inflammation (Fig. 3A,B). Additionally,
transcription level of Dct as measured relative to EF1AA was
significantly higher in the micro-dissected tissue compared to the Fig. 5. Voluminous melanomacrophage with dendritic shape. Hematoxylin stain.
macro-dissected. Bar ¼ 20 mm.
 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285
moderate, sparse or negative (no labeled cells detectable). Variation
in labeling intensity and frequency between individuals was
present in all groups.
Anti-MHC class II reactivity was found in macrophages encir-
cling necrotic muscle fibers (Fig. 4F), in sites with abundant
leukocyte infiltration (Fig. 4G) and in cell-rich granulomas (Fig. 4H).
Multifocal positive staining was also observed within melanoma-
crophages, mainly in cells with sparse pigmentation and low den-
dricity (Fig. 4G). Only sparse staining was observed in the
macrophages bordering the vacuoles (Fig. 4G).
A broad rim of CD33þ cells was located in the periphery of cell-
rich granulomas (Fig. 4I). Clusters of such cells were also seen
around necrotic muscle fibers and throughout areas with abundant
leukocyte infiltration. Immunopositive cells were often located in
close association with melanomacrophages.
Anti-IgM reactivity in the form of abundant positive staining
was seen in the lumen of vessels and capillaries containing blood
(Fig. 4J) and in sparse single cells randomly distributed throughout
the inflamed areas.
In most lesions, focal to multifocal foci with a-actin-immuno-
positive cells interpreted as myofibroblasts or satellite cells differ-
entiating into myoblasts were observed (Fig. 4K). The smooth
musculature of the vessel walls stained in a similar manner, serving
as an internal positive control. Myotubes and mature skeletal
muscle fibers were negative.
4. Discussion
In this study, we have investigated abnormal pigmented
changes in muscles of Atlantic salmon. Affected muscle showed
expression of genes confined to melanin-synthesizing cells. There
was a consistent up-regulation of Tyr-transcription in pigmented
versus non-pigmented samples. Additionally, Dct-transcription,
exclusively occurring in cells capable of pigment production, was
detected in 13 of 14 pigmented tissue samples, and undetected in
13 of 14 control samples. The pathological changes of the fillets
contained areas of inflammation with and without pigmentation.
Most control tissue was collected from healthy-appearing muscle
far distant from the lesion, but within the same fillet. However, in
sampling nr 3 (Romsdalsfjord, 4 fillets) control material was har-
vested in close vicinity to visible pigmentation. Nevertheless, as the
samples selected for RT-qPCR cannot simultaneously be investi-
gated by histology, there is no guarantee that the control samples
were free from sparse melanomacrophages. The single sample with
a clear divergent result, gave room for suspicion concerning this
matter. By applying the laser-dissection method we were able to
make a more precise comparison, avoiding the risk of misinter-
preting the results. Furthermore, by comparing the expression
pattern of Dct in the macro-dissected tissue with the more precise
micro-dissected, we found the relative expression to be signifi-
cantly higher in the latter, strongly connecting the transcription of
Dct to the pigmented cells. Histological analysis revealed a granu-
lomatous inflammation in the affected tissue with differently-
shaped melanin-containing cells, interpreted as melanomacroph-
ages, a specialized type of leukocytes found in ectothermic verte-
brates [12,17]. Seen together, these results strongly indicate
enzymatic de novo production of melanin in the inflamed tissue,
excluding a non-enzymatic production as has been suggested to
occur in the brain during synthesis of neuromelanin [30].
The changes in the pigmented muscle were characterized as
chronic polyphasic necrotizing myopathy with melanin production
and granuloma formation, indicating an ongoing degenerative
process. New lesions have been forming simultaneously with
regeneration, displaying signs of repair side by side with acute
necrosis [31]. Based on the morphological appearance, the a-actinþ
283
cells were interpreted as being activated satellite cells capable of
replacing degenerated myofibers [31]. The absence of detectable
infectious agents, the lack of granulocytes, the vacuoles staining
positive for fat and the presence of granulomas, are indications of
a non-infectious, foreign-body type of inflammation, i.e. granulo-
matous inflammation.
Granulomatous inflammation is dominated by cells of the
monocyte-macrophage line, consisting of macrophages, epitheloid
cells and multinucleated giant cells [32]. The morphological
appearance in combination with abundant immunopositive stain
for MHC class II b-chain and the significant up-regulation of MHC
class II expression, a receptor found on antigen-presenting cells
including activated macrophages, strongly indicates that macro-
phages were the dominating inflammatory cell type in the pig-
mented tissue. No definite macrophage cell marker is available for
salmonids. T cells, identified in our study by antisera recognizing
the CD33 chain, participate in granuloma production. T cells here
are thought to be regulatory, preventing widespread tissue damage
by the activated macrophages [32]. In our material, a broad rim of T
cells was seen encircling the cell-rich granulomas. This appearance
is consistent with the architecture of mammalian granuloma
formation [32]. The lack of significant up-regulation of CD3z-
mRNAs compared to the positive immunostaining of the CD33 chain
may due to the dynamics of transcription and translation of the CD3
molecules. However, information is sparse about this matter in
salmon. Few cells stained positive for IgM, suggesting B cells to be
of lesser importance in the localized inflammatory response, also
consistent with granulomatous myopathy of mammals [32].
Granulomas serve to demark pathogens/foreign bodies resisting
destruction. In our material, the well-organized granulomas with
large central vacuoles and sparse leukocyte response were sur-
rounded by a solid rim of elongated melanomacrophages. Such cells
may function as a seal, isolating the self from the supposed foreign
body, preventing repeated triggering of the immune response. This
assumption may be supported by the sparse amount of MHC class II
reactivity in macrophages and melanomacrophages adjacent to the
vacuoles. In mammals, chronic inflammation with macrophage-
lineage cells as a key component, although devoid of pigment
production has been linked to non-disposable materials likely to
accumulate in macrophages. Such material could consist of oil-
based adjuvant amongst several others [33]. Additionally, degen-
erative diseases or trauma causing damaged and dying cells, can
also act as powerful adjuvant, stimulating chronic inflammation
[33]. Hereby, antigen-presenting cells may be activated by alarm
signals from stressed or damaged tissues [34]. The granulomatous
inflammation pattern, together with the results from the immu-
nohistochemical studies, show that the pathological changes in the
salmon share several similarities with homologous reactions in
mammals [32]. However, the presence and production of melanin
represent a major difference. In addition to the flattened melano-
macrophages surrounding vacuoles, large dendritic-shaped mela-
nomacrophages, seemingly very active, were also encountered in
the granulomatous tissue. Such cells may be the ones responsible
for the detection of transcripts of the tyrosinase gene family.
Pigment-producing leukocytes of ectothermic vertebrates (classi-
fied by Sichel and co-workers [12]), are often referred to as mela-
nomacrophages in the literature (reviewed by Agius & Roberts
[17]). They can be distinguished from melanocytes of ectoderm
origin and from mammalian Langerhans cells or macrophages
phagocytizing pigments, the last of which is referred to as mela-
nophages of human skin [35]. We assume that the melanoma-
crophages found in the muscle lesions are constitutively melanin-
producing leukocytes summoned to the site of inflammation
together with other leukocyte populations. In salmonids, the main
concentration of such cells is normally found in the anterior part of
284 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285
the kidney, a tissue considered to be the teleost equivalent to
mammalian bone-marrow [14]. The gross pathology observations
in the present study revealed in most cases a sub-peritoneal or
profound localization of pigment, far distant from cutanous mela-
nocytes and without pathological changes in the skin. In mammals,
visceral melanin deposition caused by migrating pigment-
containing leukocytes (melanophages) is also well-known [36],
but de novo synthesis by such cells has not been reported.
A fundamental question is why granulomatous inflammation in
salmon may involve melanin formation. One of the functions of
melanin, as part of the host antioxidant system is to trap radicals
[37,38]. This effect of melanin may be particularly relevant in long-
term functioning post-mitotic pigment cells, with protracted
exposure to oxidative stress, as would be the case for the mela-
nomacrophages encircling the vacuoles staining for fat. Radicals
produced during melanogenesis are nevertheless believed to exert
strong antimicrobial activity over a short action radius [6]. Melanin
production could thus serve as a compensatory or additional
defense mechanism to the innate immune system in poorly-
vascularized tissue, such as tissue with severe scarring and
mature granulomas. On the other hand, melanin per se seems to be
able to trigger the inflammatory pathway in macrophages also in
the presence of LPS via a TLR-4 mediated mechanism [39]. Even if
TLR-4 presence and function in fish are still a matter of debate, one
can speculate that melanin is able, once produced, to maintain
a sort of vicious circle in the fish via release of proinflammatory
cytokines.
Melanin synthesis occurs both in prokaryotes and in eukaryotes,
in the latter specialized melanocytes are dedicated to this
production. Throughout evolution, melanin has been an important
factor for survival, and appeared long before the molecules of the
classical adaptive immune system, which first developed with the
introduction of jawed vertebrates [40]. Melanin is known to protect
microorganisms, like bacteria and fungus, against cell damage such
as solar exposure and reactive oxygen species (ROS) [41]. In some
pathogenic fungi and bacteria, melanin has also been found to
protect the microbe against host immune responses, thereby
contributing to their virulence [42]. A number of plants contain
melanin, where its production may function as a defense mecha-
nism during injury, herbivore or pathogen attack or during expo-
sure to external stresses [43]. In insects, the enzyme cascade of
which melanin is the end product is considered to be the central
antimicrobial defense system [44]. A well-known immune mech-
anism in insects involves using melanin to trap invading microor-
ganisms within humoral capsules [6]. The antimicrobial effect of
radicals would be especially prudent against phagocytized micro-
organisms compartmentalized within melanosomes, as seen in
insects [44].
Given the antimicrobial actions of melanin and melanogenesis
in evolutionary early defense mechanisms, it is possible that
modern bony fishes or teleosts, being one of the first groups to
possess an adaptive immune system based on the receptors of the
immunoglobulin superfamily, still rely on ancient immune
responses related to melanin. Even in mammals, there seems to be
a relic of this system related to inflammatory changes in the skin
[2], but mammalian leukocytes lack melanin-synthesizing abilities.
It is possible that melanin production could serve as an ancient
compensatory mechanism for the innate immune system in tissue
with poor vascularization, as would be the case for fat tissue of
morbidly obese persons as studied by Randhawa et al. [45]. In
humans and in some other mammals, melanin is also found in the
brain. Neuronal melanin is termed neuromelanin and is found
mainly in the substantia nigra of the brain stem. Due to its high
affinity for a variety of organic and inorganic toxins, accumulation
of neuromelanin during aging is linked to the protective role of the
pigment, thereby preventing neuronal damage [30]. The selective
death of the neuromelanin-containing cells causes Parkinson’s
disease, which is associated with oxidative stress [46]. Therefore,
melanin production in fish granulomatous tissue could also serve as
a protective mechanism against oxidation in chronic inflammatory
conditions, which may be important in salmon muscle tissue which
is rich in polyunsaturated fatty acids.
In this study, we have shown that melanogenesis occurs in
muscle-located granulomas in the Atlantic salmon, thus indicating
an association between the immune and pigmentary systems. The
pigment-producing granulomas in salmon are a hitherto non-
described inflammatory reaction. Future studies on this inflam-
matory response could bridge important evolutionary gaps
between the pigmentary and immune systems.
Authors contribution
H.L., study design, collection of material, performing experi-
ments, interpreting results and writing manuscript. L.A., study
design, collection of material, designing probes, statistical analysis,
interpreting results and editing manuscript. T.M., identification and
collection of material. J.T., designing probes and primers for
melanin genes. I.H., sequence analysis, probe design and antisera
development. U.F. and E.J., providing antisera. E.R., study design,
providing probes and primers for viruses and interpreting results.
E.O.K., study design, collection of material, providing antisera,
interpreting results, editing manuscript and supervising the study.
All authors read and commented on the manuscript.
Acknowledgment
Marine Harvest AS, represented by Dr. Olav Breck, VMD Øyvind
Oaland and others, have kindly provided material and assisted
harvesting of material. Valuable help have also been provided by
co-workers of the Norwegian School of Veterinary Science and the
InNoVacc research project (Project RCN 183196/S40). The Research
Council of Norway and The Fishery and Aquaculture Industry
Research Fund supported this project financially (project: RCN
179022/S40 Regulatory processes of melanization in the Atlantic
salmon), and in addition, grants were obtained from the Norwegian
School of Veterinary Science.
References
[1] Hearing VJ, Tsukamoto K. Enzymatic control of pigmentation in mammals.
FASEB J 1991;5:2902e9.
[2] Plonka PM, Passeron T, Brenner M, Tobin DJ, Shibahara S, Thomas A, et al.
What are melanocytes really doing all day long.? Exp Dermatol 2009;18:
799e819.
[3] Burkhart CG, Burkhart CN. The mole theory: primary function of melanocytes
and melanin may be antimicrobial defense and immunomodulation (not solar
protection). Int J Dermatol 2005;44:340e2.
[4] Le Poole IC, Mutis T, van den Wijngaard RM, Westerhof W, Ottenhoff T, de Vries RR,
et al. A novel, antigen-presenting function of melanocytes and its possible rela-
tionship to hypopigmentary disorders. J Immunol 1993;151:7284e92.
[5] Slominski A, Tobin DJ, Shibahara S, Wortsman J. Melanin pigmentation in
mammalian skin and its hormonal regulation. Physiol Rev 2004;84:1155e228.
[6] Mackintosh JA. The antimicrobial properties of melanocytes, melanosomes
and melanin and the evolution of black skin. J Theor Biol 2001;211:101e13.
[7] Wells WC. Some observations on the causes of the differences in colour and
form between the white and negro races of men. Two essays: one: upon
a single vision with two eyes, the other: on dew. London: Constable; 1818.
[8] Allen AM, Taplin D, Twigg L. Cutaneous streptococcal infections in Vietnam.
Arch Dermatol 1971;104:271e80.
[9] Koppang EO, Fischer U, Satoh M, Jirillo E. Inflammation in fish as seen from
a morphological point of view with special reference to the vascular
compartment. Curr Pharm Des 2007;13:3649e55.
[10] Raposo G, Fevrier B, Stoorvogel W, Marks MS. Lysosome-related organelles:
a view from immunity and pigmentation. Cell Struct Funct 2002;27:443e56.
[11] Sichel G, Scalia M, Corsaro C. Amphibia Kupffer cells. Microsc Res Tech 2002;
57:477e90.
 H.A.S. Larsen et al. / Fish & Shellfish Immunology 33 (2012) 277e285
[12] Sichel G, Scalia M, Mondio F, Corsaro C. The amphibian Kupffer cells build and
demolish melanosomes: an ultrastructural point of view. Pigment Cell Res
1997;10:271e87.
[13] Haugarvoll E, Thorsen J, Laane M, Huang Q, Koppang EO. Melanogenesis and
evidence for melanosome transport to the plasma membrane in a CD83þ
teleost leukocyte cell line. Pigment Cell Res 2006;19:214e25.
[14] Thorsen J, Høyheim B, Koppang EO. Isolation of the Atlantic salmon tyrosinase
gene family reveals heterogenous transcripts in a leukocyte cell line. Pigment
Cell Res 2006;19:327e36.
[15] Kelsh RN, Schmid B, Eisen JS. Genetic analysis of melanophore development in
zebrafish embryos. Dev Biol 2000;225:277e93.
[16] Koppang EO, Hordvik I, Bjerkås I, Torvund J, Aune L, Thevarajan J, et al.
Production of rabbit antisera against recombinant MHC class II beta chain and
identification of immunoreactive cells in Atlantic salmon (Salmo salar). Fish
Shellfish Immunol 2003;14:115e32.
[17] Agius C, Roberts RJ. Melano-macrophage centres and their role in fish
pathology. J Fish Dis 2003;26:499e509.
[18] Koppang EO, Haugarvoll E, Hordvik I, Aune L, Poppe TT. Vaccine-associated
granulomatous inflammation and melanin accumulation in Atlantic salmon,
Salmo salar L., white muscle. J Fish Dis 2005;28:13e22.
[19] Poppe TT, Breck O. Pathology of Atlantic salmon Salmo salar intraperitoneally
immunized with oil-adjuvanted vaccine. A case report. Dis Aquat Org 1997;
29:219e26.
[20] Mørkøre T, Heia K. Mørke flekker i laksefilet - omfang og målemetoder. Norsk
Fiskeoppdrett 2012;3:50e3.
[21] Sommerset I, Krossøy B, Biering E, Frost P. Vaccines for fish in aquaculture.
Expert Rev Vaccines 2005;4:89e101.
[22] Olsvik PA, Lie KK, Jordal AE, Nilsen TO, Hordvik I. Evaluation of potential
reference genes in real-time RT-PCR studies of Atlantic salmon. BMC Mol Biol
2005;6:21.
[23] Haugarvoll E, Bjerkås I, Nowak BF, Hordvik I, Koppang EO. Identification and
characterization of a novel intraepithelial lymphoid tissue in the gills of
Atlantic salmon. J Anat 2008;213:202e9.
[24] Biacchesi S, LeBerre M, Lamoureux A, Louise Y, Lauret E, Boudinot P, et al.
Mitochondrial antiviral signaling protein plays a major role in induction of the
fish innate immune response against RNA and DNA viruses. J Virol 2009;83:
7815e27.
[25] Duschene KS, Broderick JB. The antiviral protein viperin is a radical SAM
enzyme. FEBS Lett 2010;584:1263e7.
[26] Robertsen B. Expression of interferon and interferon-induced genes in
salmonids in response to virus infection, interferon-inducing compounds and
vaccination. Fish Shellfish Immunol 2008;25:351e7.
[27] Bancroft JD, Gamble M. Theory and practice of histological techniques. 5th ed.
London: Churchill Livingstone; 2002.
[28] Koppang EO, Fischer U, Moore L, Tranulis MA, Dijkstra JM, Köllner B, et al.
Salmonid T cells assemble in the thymus, spleen and in novel interbranchial
lymphoid tissue. J Anat 2010;217:728e39.
285
[29] Håvarstein LS, Aasjord PM, Ness S, Endresen C. Purification and partial char-
acterization of an IgM-like serum immunoglobulin from Atlantic salmon
(Salmo salar). Dev Comp Immunol 1988;12:773e85.
[30] Zecca L, Tampellini D, Gerlach M, Riederer P, Fariello RG, Sulzer D. Substantia
nigra neuromelanin: structure, synthesis, and molecular behaviour. Mol
Pathol 2001;54:414e8.
[31] Valentine BA, McGavin MD. Skeletal muscle [Chapter 15]. In: McGavin MD,
Zachary JF, editors. Pathologic basis of veterinary disease. 4th ed. St. Louis:
Mosby Elsevier; 2007. p. 973e1039.
[32] Tews DS, Pongratz DE. Immunohistological analysis of sarcoid myopathy.
J Neurol Neurosurg Psychiatry 1995;59:322e5.
[33] Hamilton JA. Nondisposable materials, chronic inflammation, and adjuvant
action. J Leukoc Biol 2003;73:702e12.
[34] Matzinger P. The danger model: a renewed sense of self. Science 2002;296:
301e5.
[35] Weiss JS, James WD, Cooper KD. Melanophages in inflammatory skin
disease demonstrate the surface phenotype of OKM5þ antigen-
presenting cells and activated macrophages. J Am Acad Dermatol
1988;19:633e41.
[36] Plonka PM, Michalczyk D, Popik M, Handjiski B, Slominski A, Paus R. Splenic
eumelanin differs from hair eumelanin in C57BL/6 mice. Acta Biochim Pol
2005;52:433e41.
[37] Scalia M, Geremia E, Corsaro C, Santoro C, Baratta D, Sichel G. Lipid perox-
idation in pigmented and unpigmented liver tissues: protective role of
melanin. Pigment Cell Res 1990;3:115e9.
[38] Sichel G, Corsaro C, Scalia M, Sciuto S, Geremia E. Relationship between
melanin content and superoxide dismutase (SOD) activity in the liver of
various species of animals. Cell Biochem Funct 1987;5:123e8.
[39] Oberg F, Haseeb A, Ahnfelt M, Ponten F, Westermark B, El-Obeid A. Herbal
melanin activates TLR4/NF-kappaB signaling pathway. Phytomedicine 2009;
16:477e84.
[40] Cooper MD, Alder MN. The evolution of adaptive immune systems. Cell 2006;
124:815e22.
[41] Nosanchuk JD, Casadevall A. The contribution of melanin to microbial path-
ogenesis. Cell Microbiol 2003;5:203e23.
[42] Liu GY, Nizet V. Color me bad: microbial pigments as virulence factors. Trends
Microbiol 2009;17:406e13.
[43] Mayer AM. Polyphenol oxidases in plants and fungi: going places? A review.
Phytochemistry 2006;67:2318e31.
[44] Boman HG, Hultmark D. Cell-free immunity in insects. Annu Rev Microbiol
1987;41:103e26.
[45] Randhawa M, Huff T, Valencia JC, Younossi Z, Chandhoke V, Hearing VJ, et al.
Evidence for the ectopic synthesis of melanin in human adipose tissue. FASEB
J 2009;23:835e43.
[46] Zucca FA, Giaveri G, Gallorini M, Albertini A, Toscani M, Pezzoli G, et al. The
neuromelanin of human substantia nigra: physiological and pathogenic
aspects. Pigment Cell Res 2004;17:610e7.