Immune parameters correlating with reduced susceptibility to pancreas disease in experimentally challenged Atlantic salmon (Salmo salar)

Fish & Shellfish Immunology 34 (2013) 789e798

Contents lists available at SciVerse ScienceDirect

Fish & Shellfish Immunology journal homepage: www.elsevier.com/locate/fsi

Immune parameters correlating with reduced susceptibility to pancreas disease in experimentally challenged Atlantic salmon (Salmo salar) Søren Grove a,1, Lars Austbø b,1, Kjartan Hodneland c, Petter Frost c, Marie Løvoll a, Marian McLoughlin c, Hanna L. Thim d, Stine Braaen e, Melanie König b, Mohasina Syed b, Jorunn B. Jørgensen d, Espen Rimstad e, * a

Section for Immunology, Norwegian Veterinary Institute, Oslo, Norway Department of Basic Science and Aquatic Medicine, Norwegian School of Veterinary Science, Oslo, Norway c MSD Animal Health, Thormøhlensgate 55, 5008 Bergen, Norway d Norwegian College of Fishery Science, University of Tromsø, 9037 Tromsø, Norway e Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, P.O. Box 8146 Dep, N-0033 Oslo, Norway b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 December 2011 Received in revised form 23 November 2012 Accepted 10 December 2012 Available online 7 January 2013

Two strains of Atlantic salmon (Salmon salar) with different susceptibility to infectious salmon anaemia (ISA) were challenged with salmon pancreas disease virus (SPDV), the etiological agent of salmon pancreas disease (PD), by cohabitation. Serum and tissues were sampled at 0, 1, 3, 6 and 8 weeks postchallenge. Experimental challenge with SAV did not cause mortality, but virus loads and assessment of histopathology indicated that the fish more resistant to ISAV (ISAHi) also was more resistant to PD. Eight weeks post-challenge, the ISAHi strain had higher titres of SAV-neutralising antibodies than the less resistant strain (ISALo). Transcript levels of four adaptive and six innate immune parameters were analysed by real-time RT-PCR in heart, head kidney (HK) and gills of both strains. Secretory IgM (sIgM) and CD8 levels differed most between the two salmon strains. The ISAHi strain had significantly higher levels of sIgM in HK at all samplings, and significantly higher CD8 levels in gills at most samplings. In heart, both sIgM and CD8 levels increased significantly during the challenge, but the increase appeared earlier for the ISALo strain. By hierarchical clustering analysis of mRNA levels, a clear segregation was observed between the two strains prior to the virus challenge. As the viral infection developed, the clustering divide between fish strains disappeared, first for innate and later for adaptive parameters. At eight weeks post-challenge, the divide had however reformed for adaptive parameters. Possible pairwise correlation between transcript levels of immune parameters was evaluated by a non-parametric statistical test. For innate parameters, the extent of correlation peaked at 3 wpc in all tissues; this came rapidly for ISALo and more gradual for ISAHi. The ISAHi strain tended to show higher correlation for innate parameters in heart and gill than ISALo at early sampling times. For adaptive immune parameters, little correlation was observed in general, except for ISAHi in heart at 6 wpc. Overall, the observed differences in immune parameters may provide important clues to the causes underlying the observed difference in susceptibility to PD. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Salmonid alphavirus Disease resistance Adaptive Innate Neutralising antibodies

1. Introduction

Abbreviations: ISAHi, fish strain with relative high resistance to infection with ISAV; ISALo, fish strain with relative high resistance to infection with ISAV. * Corresponding author. Tel.: þ47 22 96 47 66. E-mail address: [email protected] (E. Rimstad). 1 These authors contributed equally to this work. 1050-4648/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fsi.2012.12.014

Salmon pancreas disease (PD) is caused by an alphavirus called Salmon pancreas disease virus [1]. Due to its taxonomic placement, the name Salmonid alphavirus (SAV) has been proposed and is now commonly used [2]. Different isolates of SAV can only be discriminated by genetic analysis. Based on genotypic variations, it is currently proposed that SAV should be divided into six genogroups (subtypes) [3], each correlating with geographical separation.

790

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798

Whereas PD refers to the disease in seawater farmed Atlantic salmon (Salmo salar) and rainbow trout (Oncorhynchus mykiss) in the British Isles/Norway, a related disease called sleeping disease (SD) is mainly found in rainbow trout reared in freshwater (Central Europe). SD in trout caused by subtype SAV2 and SPD in the British Isles is most commonly caused by SAV1, while the Norwegian outbreaks of SAV have been restricted to the subtype 3 [4]. PD is an emerging disease in the Atlantic salmon farming industry in Europe. In Norway a significant increase of disease outbreaks in the southern part of the country has occurred since 2003 [5]. In field outbreaks of PD, clinical signs and severity of pathological changes in the key target organs pancreas, heart and skeletal muscle, will vary between individuals [6]. Clinical signs include lethargy and impaired swimming performance; the latter interpreted as the “sleeping” behaviour typical for SD. Depending on temperature and route of infection, infected fish within days develop viraemia that normally lasts for 2 weeks [7,8]. In SAV cohabitant challenge experiments, SAV infection and PD pathology is found from 3 wpc after shedding [9], but generally experimental challenges do not induce mortality. Waning of viraemia has been shown to coincide in time, but not overlap, with the emergence of neutralizing antibodies (NAb) in the bloodstream [10], which may happen as early as 10 days after challenge by intraperitoneal injection [11]. Tissues [7,8] and blood cells [7] become virus positive within few days, and virus may remain in these cellular compartments weeks after virus have been cleared from the serum. During SAV infection, lesions develop in a definite temporal manner, affecting first pancreas and heart followed by skeletal musculature and possibly CNS [6,10]. Atlantic salmon surviving experimental PD develops protective immunity that can last at least nine months [12]. As reoccurrence of PD has not been reported from previously infected field populations, the protection obtained from natural PD is likely long-lasting, at least on population level [6]. This protective immunity is likely conferred by NAb, as indicated by the protection gained in fish receiving injection of serum from PD convalescent fish [13]. Susceptibility to PD has been shown to vary between commercial strains of Atlantic salmon [10], but the underlying reason is not well known. In response to experimental infection, leukocyte phagocytic activity was shown to increase, as were levels of lysozyme and complement [14]. Reduced PD specific pathological lesions and SAV3 levels in experimentally infected Atlantic salmon injected with the toll-like receptor (TLR)-ligands polyI:C and CpG have recently been reported, and it was demonstrated that this correlated with transcriptional up-regulation of IFNa1, IFNg, Mx and the chemokine CXCL10 [15]. The study found that the IFN system participates in the host defence against SAV, as also indicated by the increased levels of IFNa1 and Mx in SAV challenged salmon [16]. Mammalian alphaviruses are also strong inducers of IFN-I resulting in transcriptional up-regulation of genes with antiviral activities [17]. Studies of alphavirus infection in mice deficient in IFN-I signalling have indicated that this pathway is a primary protective response [17,18]. The viral non-structural protein nsP2 of mammalian alphaviruses is an important regulator of virusehost cell interactions and plays a significant role in suppressing the antiviral response [19]. Experimental vaccines using formalin inactivated [6,20] or attenuated SAV [21] induce protection, but commercial vaccines have not been able to fully control the disease in the field. Description of SAV-specific induction of innate and acquired immune responses, and in particular the nature of protective response may aid the development of efficient prophylaxis. The possibility that increased resistance to a specific pathogen in a bred fish strain may affect susceptibility to other pathogens has been given little attention. In the present study, two strains of Atlantic salmon with differing susceptibility to ISA were challenged with SAV by cohabitation. Based on morbidity parameters and

quantification of virus RNA, the two fish strains were shown to differ in susceptibility to PD. In an attempt to resolve the nature of immune factors underlying this differential disease susceptibility, an array of adaptive and innate immune parameters were analysed by real-time reverse transcription quantitative PCR (RT-qPCR) and by serological methods. 2. Materials and methods 2.1. Fish Atlantic salmon (S. salar) of two different strains were obtained from SalmoBreed AS (Bergen, Norway) and kept at the ILAB facility (Bergen, Norway). One strain (ISAHi) was characterised by relative high resistance to infection with ISAV (estimated breeding value of 26.3%) whereas the other (ISALo) had relative low resistance (estimated breeding value of 18%). A third strain of A. salmon (ILAB/08/004) was obtained from ILAB and used as virus shedder fish in cohabitation challenge. Prior to use, the fish were tested and found negative for presence of infectious pancreas necrosis virus (IPNV) and SAV by real-time RT-PCR. The fish were kept at 10e13  C and fed ad libitum during the challenge. 2.2. Cohabitation challenge and sampling For the cohabitation challenge, a Norwegian isolate of SPDV (PD03 13p2) passed twice in CHSE-214 cells, was used to inject the shedder fish. The challenge was performed in two tanks, tank A containing 100 salmon of strain ISAHi (mean weight 35 g) and tank B containing 100 salmon of strain ISALo (mean weight 35 g). At day 0 (0 wpc), each of tanks A and B received 25 shedders of strain ILAB/ 08/004 that had got an intraperitoneal (ip) injection of 0.2 ml virus suspension the same day. Although the exact timing of infection is not possible to pin-point in a cohabitant challenge model this was partly compensated for by using a relatively large number of shedder fish. At day 0, prior to infection, blood and tissue (head kidney (HK), heart and gill) were sampled from 10 fish from each of strain ISAHi and ISALo. Blood samples were taken from anaesthetized fish using heparin-containing Vacutainers and serum was prepared by standard centrifugation protocol and stored at 20  C. Tissue samples were immediately transferred to ice-cold RNAlater (Ambion), then incubated at 4  C for 24 h and finally stored at 20  C. During the subsequent challenge, identical samplings were performed at 1, 3, 6 and 8 wpc. At separate samplings, at 4 and 6 wpc, additional heart tissue samples were taken from 15 fish from each of strain ISAHi and ISALo and fixed in 3.5% formaldehyde in buffered saline (pH 7.0). 2.3. Histopathological examination Formalin fixed heart tissue was processed for haematoxyline eosin staining using standard protocol. A score system was used to evaluate the severity of SAV induced heart lesions, i.e. no lesion: minimal: 1, mild: 2, moderate: 3 and severe: 4 [22]. Scores 2 are considered to be specifically to clinical PD infection. Moribund fish and fish that have died from PD are typically scored as severe (4). The scoring was done as a blinded experiment. 2.4. RNA extraction and cDNA synthesis Tissue samples in RNAlater were distributed between four contributing laboratories (AeD), where total RNA was isolated by use of RNeasy Mini Kit (Qiagen) and eluted in 50 ml of RNase free H2O. The RNA output was checked by gel electrophoresis for the

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798

absence of degradation and quantified spectrophotometrically using Nanodrop ND-1000 (Thermo Scientific) before storage at 80  C. cDNA was synthesised from 1 mg of RNA using QuantiTect Reverse Transcription Kit (Qiagen) following the manufacturer’s recommendations and stored at 20  C until use. Prepared cDNA were exchanged so that each laboratory was provided with cDNA from all samples and time-points. 2.5. Quantitative two-step RT-PCR The RT-qPCR was performed using TaqMan Gene Expression Master Mix (Applied Biosystems). In laboratory A, samples were analysed for transcripts of CD8 and secreted immunoglobulin M (sIgM). In laboratory B, samples were analysed for Viperin (Vip), MHCI and MHCII. In laboratory C, samples were analysed for IFNa1, IFNg, TLR8, Mx protein (Mx) and CXC ligand 10 (CXCL10). Laboratory D performed RT-PCR based detection and semi-quantification of SAV. The primers and probe are as listed in Table 1. Elongation factor 1ab (EF1ab) [23] was used as the reference gene (RG). A previous study has established that the EF1ab gene was satisfactorily stable expressed during PD infection and that inclusion of more reference genes did not improve overall stability [24]. The RTqPCR was carried out with cDNA corresponding to 15 ng of RNA. Each quantification target was amplified in triplicate with negative control lacking the template, for each master mix. The transcript levels were normalized to EF1ab expression using the method of Pfaffl [25]. Further, the EF1ab-normalized transcript levels were

Table 1 Primers and probes used in Atlantic salmon quantitative RT-PCR. Primers and probe were design to span intron section and exon-junction sites. FAM: 6carboxyfluorescein, MGB: minor groove binder, BHQ: black hole quencher, NFQ: non-fluorescent quencher (Applied Biosystems). Genes

Primer

Sequence (50 e30 )

EF1b

Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe Forward Reverse Probe

TGCCCCTCCAGGATGTCTAC CACGGCCCACAGGTACTG FAM-AAATCGGCGGTATTGG-MGB-BHQ CCTTTCCCTGCTGGACCA TGTCTGTAAAGGGATGTTGGGAAAA FAM-CTTTGTGATATCTCCTCCCATC-MGB-BHQ GATGCTGCACCTCAAGTCCTATTA CGGATCACCATGGGAATCTGA FAM-CAGGATATCCAGTCAACGTT-MGB-BHQ AAGGGCTGTGATGTGTTTCTG TGTACTGAGCGGCATTACTCC FAM-TTGATGGGCTGGATGACTTTAGGA-MGB-BHQ AGGAGTGTGCAGTAAATCTGTGAAC CTCATGGTGCTCTCTGTTCCA FAM-CAATTCCACTAAGAACTTG-MGB-BHQ ACCAAAACCACTAATGACATCATCTTCA TGGTGATGCCATCAGGTATGTTT FAM-CTCAGTCGACGCTCCTC-MGB-BHQ AGCAATGGCAGCATGATCAG TGGTTGGTGTCCTCGTCAAAG FAM-AGTGGTTCCAAACGTATGGCGAATACCTG-BHQ GGAAGAGCACTCTGATGAGGACAG CACCATGACTCCACTGGGGTAG FAM-TCAGTGTCTCTGCTCCAGAAGACCCCCT-BHQ CCACCTGGAGTACACACCCAG TTCCTCTCAGCCTCAGGCAG FAM-TCCTGCATGGTGGAGCACATCAGC-BHQ CTACAAGAGGGAGACCGGAG AGGGTCACCGTATTATCACTAGTTT FAM-TCCACAGCGTCCATCTGTCTTTC-BHQ CGTCTACAGCTGTGCATCAATCAA GGCTGTGGTCATTGGTGTAGTC FAM-CTGGGCCAGCCCCTAC-MGB-NFQ CCGGCCCTGAACCAGTT GTAGCCAAGTGGGAGAAAGCT VIC-CTGGCCACCACTTCGA-MGB-BHQ

IFN a/b

Mx1/2

IFNg

CXCL10

TLR8

Viperin

MHC I

MHCII

sIgM

CD8a

SAV

791

normalized against a calibrator which was the mean of ISAHi and ISALo transcript levels at 0 wpc. Detection of the SAV3 nsP1 gene was used for quantitative realtime RT-PCR analysis, and the PCR parameters used have been described earlier [26]. Each sample was tested in triplicate and considered as positive when the Cq [27] value was 37.5. A fish was determined SAV positive if a sample from that fish was found positive. To assess PCR efficiency, template dilutions were used to generate standard curve. The load of SAV RNA was estimated in tissue samples at 0, 1, 3, 6 and 8 wpc and in serum at 3 wpc. 2.6. Neutralization assay Five serum samples from each sampling (0, 1, 3, 6 and 8 wpc) were analysed for their SAV-neutralizing activity. Serum was heat inactivated at 43  C for 45 min and then two-fold dilutions of the serum in L-15 cell culture medium were mixed with an equal volume of Irish SAV1 (F93-125) [1] containing 5  102 TCID50 ml1, passed 9 times in CHSE-214 cells. Following incubation at 20  C for 2 h, 50 ml of each serumevirus mixture was added in three parallels to Chinook salmon embryo cells (CHSE) grown in 96-well plates. After 10 days of incubation at 20  C, the cells were fixed in 80% acetone and stained for SAV using an a-SAV1 mAb (5A5) as earlier described [28]. Neutralizing activity was expressed as the reciprocal value of the serum dilution that inhibited virus detection in 50% of the inoculated cultures. 2.7. Statistical analysis Statistical analyses and hierarchical clustering were performed using the JMPÔ software (SAS Institute Inc., North Carolina, USA). Both RT-PCR based virus detection data and histopathology score data were analysed using the non-parametric ManneWhitney U test. Data obtained by quantitative RT-PCR showed a significant logarithmic distribution. These data were log2 transformed and this resulted in a satisfactory normal distribution as judged by use of Normal Quantile plotting. The log2 transformed data were analysed by ANOVA and a possible pair-wise correlation between mRNA levels of immune parameters was evaluated by the nonparametric Kendall’s Tau test. Hierarchical clustering analysis was performed using JMP 10 (SAS institute) and the Ward’s minimum variance method on data standardised by mean and standard deviation. 3. Results 3.1. Course of infection Histopathologic evaluation of heart tissue sampled 4 and 6 wpc showed little variation in the mean histology score for the ISAHi strain. In contrast, the ISALo group had a significant higher score at 4 wpc (ManneWhitney U test, p < 0.01) (Fig. 1). Comparing the strains, ISALo had a significant higher mean histology score at 4 wpc but not at 6 wpc (ManneWhitney U test, p < 0.01). The prevalence of fish having a histology score 2 (i.e. with clinical PD) was also significantly higher in ISALo (ManneWhitney U test, p < 0.05) as compared to the ISAHi strain (data not shown). No mortality was observed during the experiment. 3.2. SAV in serum, heart, gills and head kidney The load of SAV RNA was estimated by real-time PCR in tissue samples from both ISALo and ISAHi fish at 0, 1, 3, 6 and 8 wpc and in serum at 3 wpc (Table 2). All fish sampled at 0 wpc and 1 wpc were negative for SAV. Tissue samples from ISALo fish were more

792

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798 Table 3 SAV-neutralising capacity of serum samples from groups ISAHi (Hi) and ISALo (Lo). Data is given as number of sera that had no/little neutralising activity (<20), intermediate neutralising activity (<50) or high neutralising activity (80). (N ¼ 5 for each group at each sampling). 0 wpc

<20 <50 80

1 wpc

3 wpc

6 wpc

8 wpc

Hi

Lo

Hi

Lo

Hi

Lo

Hi

Lo

Hi

Lo

5 0 0

5 0 0

5 0 0

5 0 0

4a 1 0

4b 1 0

1 0 4

1 0 4

0 0 5

4 0 1

a One of the tested sera was virus positive, i.e. neutralizing effect could not be estimated in this serum. b Two of the tested sera were virus positive, i.e. neutralizing effect could not be estimated in these sera.

each strain had a moderate neutralising antibody titre (50 < X < 80). At 6 wpc, 4 out of 5 fish from each strain had developed high neutralising activity (>80). The two strains differed substantially only at 8 wpc, where 5 out of 5 ISAHi fish showed high neutralising activity (>80) compared to only 1 out of 5 fish for the ISALo strain. Fig. 1. Histology score for heart pathology at 4 and 6 weeks post-challenge for ISAHi (Hi) and ISALo (Lo) strains. Y-axis: histology score. Error bars represent SEM.

frequently positive for SAV RNA (SAVþ) at samplings 3, 6 and 8 wpc, as compared to ISAHi fish. A single exception to this was seen in HK samples at 6 wpc. This difference was however, only statistically significant for heart tissue sampled at 8 wpc (ManneWhitney U test, p < 0.01). There was no statistical significant difference in the prevalence or load of SAV in serum samples at 3 wpc for the two groups. Compared to ISAHi group, the Cq values in SAV positive ISALo fish were lower, suggesting a higher virus load in the latter strain. Within the ISAHi and ISALo strains, the number of SAVþ samples was relatively equal between the tissue types at 3 wpc. The load of virus in the two groups, estimated by combining the number of RT-qPCR positive and the individual virus loads, peaked at 3 wpc. At 3 wpc, 65 and 75% of the ISAHi and ISALo fish respectively, were viraemic which clearly indicated that the infection was well-established in the groups. At 6 wpc, the number of positive head kidney samples had decreased in both strains but increased for heart samples. At 8 wpc the virus was no longer found in head kidney and only in a minor extent in the gills, but was still present in 9 out of 10 heart samples from the ISALo group. 3.3. Neutralizing antibodies No SAV-neutralising antibodies were detected at the time of challenge (0 wpc) or at 1 wpc (Table 3). At 3 wpc, 1 out of 5 fish from

3.4. Innate immune parameters 3.4.1. mRNA levels With few and mostly modest exceptions, the mRNA levels of the innate immune parameters IFNab, CXCL10, Mx, Vip, TLR8 and IFNg, differed little between ISAHi and ISALo strains irrespective of tissue type and sampling time (Fig. 2). At 0 wpc however, the level of CXCL10 was approximately 16 times higher in the heart of ISAHi than ISALo, but the difference disappeared in the course of the experiment. A similar but much less pronounced trend was observed in heart and HK for Vip, Mx, TLR8, IFN-I and IFNg. For IFNg, the levels in gills of both strains tended to increase slightly towards 3 wpc. At 3 wpc Vip, Mx, CXCL10 and IFNg mRNA levels increased more abruptly, but notably the increase arose exclusively from fish in which SAV had been demonstrated in all three analysed tissues (SAV triple-positive). 3.4.2. Clustering The hierarchical clustering analysis of innate genes showed that both the ISALo and ISAHi strains formed distinct clusters at 0 wpc (Fig. 3A). The ISAHi cluster, which was particular distinct, was characterised by relative high levels of innate transcripts in heart and head kidney. At 1 wpc, the distinction between the fish strains was less pronounced and the strains now formed a looser cluster that primarily was distinguished by relative high transcript levels of all innate genes in the gills. Analysis of the samplings at 3 and 6 wpc (Fig. 3B) showed a very distinct cluster formed by SAV triple-

Table 2 RT-PCR based detection of SAV RNA in gill, head kidney (HK) and heart tissue of ISAHi (Hi) and ISALo (Lo) strains at 3, 6 and 8 wpc. Ratio in bold indicate the number of SAV RNA positive samples over the total number analysed. The Cq range indicates minimum and maximum Cq values. The numbers in the lower row of the sub-table for each sampling time indicate from which individual fish the SAV positive sample was taken. All fish sampled at 0 wpc and 1 wpc were negative for SAV and therefore not included in the table. Gill-Lo No Pos 3 wpc 9/10 1e8, 10 6 wpc 6/8 2e6, 8 8 wpc 4/10 4, 6, 8, 9

Gill-Hi

HK-Lo Cq range

No Pos

HK-Hi

Cq range

No Pos

Cq range

30.7e37.5

5/10 29.4e37.7 1, 5, 6, 9, 10

6/10 30.1e38.0 1e3, 5, 6, 10

31.9e39.1

6/10 3e5, 8e10

36.2e38.5

1/10 4

36.6e38.1

1/10 3

38.2

0/10 e

No Pos

Heart-Lo Cq range

Heart-Hi

No Pos

Cq range

No Pos

5/10 31.7e35.9 1, 5, 6, 9, 10

8/10 1e7, 10

28.5e37.9

5/10 32.2e38.9 1, 5, 6, 9, 10

39.3

3/10 4,5,8

35.1e37.3

9/10 1e6, 8e10

28.4e34.3

9/10 1e6, 8e10

32.7e39.1

e

0/10 e

e

9/10 1, 2, 4e10

28.5e37.9

2/10 2, 3

32.2e38.9

Ratio shown in bold numbers indicates the number of SAV RNA positive samples over the total number of samples analysed.

Ct range

Fig. 2. RT-PCR quantification of Mx, TLR, Vip, CXCL10, IFNab and IFNg mRNA in tissue samples from Atlantic salmon challenged with SAV. X-axis: sampling time wpc (number) and fish strain (Hi/Lo). Y-axis: log2 transformed relative ratios (relative to mean of ISAHi and ISALo at 0 wpc). Due to the log2 transformation, the Y-value zero does not represent the mean of ISAHi and ISALo at 0 wpc. Results for individual fish are indicated by diamonds; open diamonds indicate fish that are positive for SAV in all analysed tissues (triple-positive). For each organ, the horizontal line indicates the grand mean (of all samples). Mean diamonds show the means at each sampling time (midmost horizontal line) and the 95% confidence intervals (upper and lower corner of diamond). Sampling means differ significantly if the Overlap marks (upper and lower horizontal line in mean diamonds) do not overlap in the vertical dimension.

Fig. 3. Hierarchical clustering analysis of innate and adaptive immune parameters in tissue samples from Atlantic salmon challenged with SAV. A. Innate genes at 0 and 1 wpc. B. Innate genes at 3 and 6 wpc. C. Adaptive genes at 0 and 1 wpc. D. Adaptive genes at 3 and 6 wpc. X-axis shows the clustering of the immune parameter mRNA levels in heart (He), head kidney (HK) and gills (Gi), respectively. Y-axis shows the clustering of individual fish with their sampling time (number) and strain (Hi/Lo). Each combination of sampling time and fish strain is indicated by a specific colour. Symbols þ and  indicate SAV detection by RT-PCR in He, HK and Gi tissues, respectively. The þ indicates that SAV was detected by RT-PCR while the , that SAV was not detected.

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798

positive fish from both strains sampled 3 wpc. Fish in this cluster generally had a high level of all innate transcripts in all tissues. A second cluster was formed by fish sampled at 6 wpc that were SAV positive in at least one tissue. While a certain distinction between ISAHi and ISALo strains could still be made at 3 and 6 wpc, these differences were now reduced to the sub-cluster level (Supplementary Fig. 1). 3.4.3. Correlation For the innate parameters, the tendency to correlation varied greatly with the sampling time, revealing a peak at 3 wpc for all tissues and both fish strains (Supplementary Fig. 2). In heart, the correlation peak appeared abruptly at 3 wpc for ISALo, whereas it had a more gradual development in the ISAHi strain. In gill, the ISAHi strain tended to have more pronounced correlation at early samplings than ISALo. Comparison of SAV negative with SAV positive fish, irrespective of fish strain and sampling time, clearly revealed that correlation of innate immune parameters were much stronger in the SAV positive fish in all tissues (data not shown). 3.5. Adaptive immune parameters 3.5.1. mRNA levels For MHCI and MHCII, transcript levels remained relatively stable throughout the experiment, with little differences between fish strains and sampling times (Supplementary Fig. 3). At 0 wpc however, mRNA levels for both MHCI and MHCII were significantly higher for the ISAHi strain in all analysed tissues. By 1 wpc these differences were less pronounced and only significant for MHCI in heart and for MHCII in gills. For sIgM and CD8, changes in transcript levels were more prominent (Fig. 4). Both ISAHi and ISALo strains

795

showed an increase in sIgM transcripts levels in HK from 0 wpc to 1 wpc, but the ISAHi strain had consistently a significant higher transcript level at all samplings. While both strains had a similar early increase in sIgM mRNA in gills from 0 wpc to 1 wpc, there was no general difference in the transcript levels between the two strains. In heart tissue, sIgM levels in ISAHi and ISALo were similar at early time points but increased significantly at 3 wpc for ISALo and 6 wpc for ISAHi strains. For CD8, the transcript levels in HK showed a slight and gradual decrease towards 6 wpc/8 wpc for both strains. In gills, levels of CD8 mRNA were generally higher in the ISAHi strain except for 6 wpc, where this strain experienced a significant drop. In heart tissue, CD8 transcript levels were similar in the strains at early time point, but experienced an increase that coincided with the increase in sIgM mRNA, i.e. at 3 wpc for ISALo and at 6 wpc for ISAHi. In gill tissue, however, the minimum transcript level for CD8 for ISAHi was at 6 wpc, i.e. the same time point where maximum was found in heart tissue. 3.5.2. Clustering Hierarchical clustering analysis of the adaptive genes at 0 and 1 wpc showed a clustering that primarily followed the fish strain and secondarily the sampling time, producing four distinct clusters (Fig. 3C). Fish sampled 3 and 6 wpc showed less distinct clustering where the strains formed clusters at 3 wpc but not at 6 wpc (Fig. 3D). Clustering analysis of the individual sampling times, i.e. 3 and 6 wpc respectively, confirmed the distinct clustering along strain at 3 wpc but not at 6 wpc (Supplementary Fig. 4). At 8 wpc, analysis based on sIgM and CD8 only, showed that the clustering along strains was again formed. In contrast to the innate genes, clusters based on adaptive gene data did at no time point follow any obvious pattern in viral status.

Fig. 4. RT-PCR quantification of sIgM and CD8 mRNA in tissue samples from Atlantic salmon challenged with SAV. (See Fig. 2 for explanatory text).

796

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798

3.5.3. Correlation For both ISAHi and ISALo strains, correlation between adaptive parameters tended to be low irrespective of tissue and sampling time (data not shown). One noteworthy exception was observed in heart at 6 wpc, where adaptive immune parameters were both more and statistically stronger correlated in the ISAHi strain than in ISALo (Supplementary Fig. 5). 4. Discussion In the present SAV cohabitant challenge experiment it was found that two strains of Atlantic salmon bred to high and low resistance to ISA, showed a similar difference in susceptibility to PD. The disease was more severe in the ISALo group compared to the ISAHi group, manifested by both higher prevalence and load of SAV in the ISALo group. These differences, however, levelled out from 6 wpc and onwards. It was shown that the two fish strains differed in immune response to the infection, regarding both adaptive immune parameters and the overall level of coordination of both innate and adaptive immune parameters. These immune parameter differences likely contribute to the unequal PD susceptibility observed in the two fish strains. The mean histology score, as well as the prevalence of fish with clinical PD, was significantly higher in the ISALo group at 4 wpc. Regarding presence of SAV RNA, more fish of the ISALo strain were SAV positive at 3 wpc and 8 wpc, but the difference was only statistically significant in heart tissue at 8 wpc. Moreover, the Cq range for the SAV positive fish was lower in ISALo, suggesting higher virus RNA levels in this strain. A similar difference in prevalence, severity of lesions and in serological responses, due to an SAV infection, has previously been shown between commercial strains of Atlantic salmon in Ireland [10]. Neutralising antibodies were detectable from 3 wpc in both fish strains, but a difference in NAb levels was only observed at 8 wpc where ISAHi had the highest prevalence and titres. In contrast, the observed difference in disease susceptibility was already manifested at 3e4 wpc, as observed by histopathological and virological parameters, suggesting that the causal factors leading to this were already at work prior to 3 wpc. It is hence less likely that differences in NAb levels played a primary role for ISAHi being the less susceptible strain. In accordance with this observation, a previous study could not explain demonstrated differences in SAV1 susceptibility in three strains of Atlantic salmon by differences in levels of NAb [10]. Neutralizing antibodies became detectable only after SAV had been cleared from plasma of infected fish [10,14]. The ISAHi fish had high titres of NAb and few SAVþ hearts at 8 wpc, while the opposite was true for ISALo, which may suggest that NAb are important for SAV clearance in tissues. As demonstrated by passive immunisation experiments, fish receiving injection of NAb (serum) from PD convalescent fish are completely protected against subsequent SAV challenge [13,29]. Transcript levels of sIgM showed significant differences in two respects during the challenge experiment. Firstly, the ISAHi strain consistently had significantly higher sIgM transcript levels in head kidney. Secondly, sIgM transcript levels increased relative to 0 wpc in all tissues of both strains, although this increase came at different time points in the heart for the two strains. The difference in head kidney sIgM transcript levels between the strains is interesting because it appeared in a tissue that is central to systemic humoral immunity, but also because it was present prior to the first encounter with SAV and was maintained after the first increase from 0 wpc to 1 wpc. Taken together, a non-transient difference in sIgM transcript levels between the two fish strains was present and could thus be related to the difference in PD susceptibility. At 0 wpc,

fish were naïve to SAV and hence the higher PD resistance observed in ISAHi is not likely based on pre-existing specificity to SAV antigens. A possibility could be that the observed higher level of sIgM also reflects a larger population of B cell lineage cells in head kidney of ISAHi fish. From a larger pool of B cells with more antigen specificities to select from, a resulting specific antibody response may be more efficient, likely improving disease resistance. In a recent study, increased levels of C4 in serum, an important factor for classical complement activation, were found in SAV infected salmon [15]. In the same study, C4 levels were depleted in vaccinated groups which showed high protection against PD. C4 consumption is a sensitive measure of classical activation, suggesting a role of antibody-mediated complement activation in protection against piscine alphaviruses. The initial higher levels of sIgM found in the ISAHi strain may lead to more efficient activation of complement mediated protection against SAV. In heart, but not in head kidney and gills, the increase of sIgM observed coincided closely with a significant increase in CD8 mRNA levels. In ISAHi fish, sIgM and CD8 transcripts were further significantly correlated at both 6 and 8 wpc. The observed increases in heart and gills suggest augmented activity and/or influx of plasmacells or precursors and cytotoxic T cells, respectively, and for sIgM it specifically indicates local antibody production. Similar peripheral antibody production has been demonstrated in Atlantic halibut (Hippoglossus hippoglossus) where it was indicated that target tissues for nodavirus became major sites of antibody production [30]. The late occurrence of the sIgM and CD8 mRNA increase in heart favours the possibility that plasmacells/blasts and cytotoxic T cells migrating to the infected target tissue are indeed SAV antigen specific. Conversely, the early sIgM increase in gills suggests influx or activation of cells from the B cell lineage that are not yet antigen specific. As IgM is the most abundant immunoglobulin type in salmonids [31], the levels of sIgM transcript are likely associated with the observed titres of NAb. The level of CD8 mRNA was significantly higher in gills of ISAHi at most samplings including at 0 wpc. Atlantic salmon gills have recently been shown to harbour a lymphoid-like tissue hosting many T cells [32] and may thus be an important compartment for development of T cell responses. Similarly to sIgM in head kidney, the high levels of CD8 in ISAHi gills may be interpreted as a distinct presence of the cell type indicated by the CD8 cell marker, i.e. cytotoxic T cells. Again, a larger pool of T cells and antigen specificities could endow ISAHi with a better capacity to launch an earlier and/or more efficient T cell response. Compared to previous studies with viral salmon diseases, the innate response observed in the current experiment was modest [33,34]. Transcript levels of several innate genes were at 0 wpc significantly higher in heart of ISAHi compared to ISALo, but the magnitude of these differences was relatively small. The most pronounced response was seen at 3 wpc and both RT-qPCR data and clustering analysis clearly indicate that the most distinct responders were SAV triple-positive fish from both fish strains. While this finding suggests a late innate response associated primarily with systemic infection, it is possible that the sampling frequency (weeks) used in the experiment failed to identify an early (shortlived) innate response. For viperin and Mx, which are induced by IFN type I, a transcript level up-regulation was observed at 3 wpc. This however, was not accompanied by an increase in of IFNa1 expression at this time point, probably reflecting that the IFNa1 response had already waned. This latter would be in accordance with previous results from Skjaeveland et al. [16], where the transcript levels of this gene were higher at day 7 than at 4 wpc. In mammals, alphaviruses are strong inducers of IFN type I, resulting in transcriptional up-regulation of genes with antiviral activities [17], but the virus may also strongly inhibit host synthesis of mRNA and protein. Such a cytotoxic effect and shut-down of protein

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798

synthesis, tentatively exerted by the SAV capsid protein, have been described [1] and might have influenced the induction of immune genes. As opposed to the IFN type I induced genes, salmon CXCL10 is an IFNg-inducible gene, and its expression along with the other innate response genes was correlated with the expression of IFNg of both strains. A recent report shows that salmonid IFNg has a potent antiviral activity against SAV and that this may in part also be linked to up-regulation of other innate parameters such as Mx, ISG15 and viperin [35]. Whereas the individual innate parameters may not convincingly be correlated to the observed difference in disease susceptibility, the demonstrated changes and differences in correlation between these parameters could be of interest. While the individual pairwise correlations have to be interpreted with great caution, the revealed differences in range of correlation may distinguish the two salmon strains regarding their immunological response. For example, the observation of more correlated innate immune parameters in gills of the ISAHi strain at 0 wpc is noteworthy. At 0 wpc the fish had not been exposed to SAV and the correlation disparity may thus reflect a difference in coordination of innate factors in gills. Better coordination of innate immune factors in this tissue, which is a possible port of virus entry, could make the ISAHi strain less susceptible to SAV infection. In heart, correlations tended to appear very abrupt in ISALo and more gradual, and with a less distinct peak in ISAHi. The extensive correlation pattern, generally observed at 3 wpc, most likely reflects a major activation of the innate antiviral genes due to the intruding virus. As the ISAHi group was more resistant, it can be speculated that a more gradual, and perhaps more host-controlled, development of the innate response in the SAV target tissue could be advantageous. While innate immune parameters were highly correlated in both ISAHi and ISALo strains in all tissues at 3 wpc, only the ISAHi strain showed a comparable correlation trend for adaptive parameters and then only in heart at 6 wpc. At 8 wpc, only sIgM and CD8 were analysed, and their levels were highly correlated in ISAHi but not in ISALo. The correlations suggest a coordinated activity of sIgMþ plasmacells/-blasts, CD8þ cytotoxic T cells and MHCIIþ antigen presenting cells on-site in the infected target organ. The clustering analysis showed that ISAHi and ISALo formed separate clusters at 0 wpc for both innate and adaptive immune parameters. Whereas this ISAHi/ISALo clustering was maintained at 1 and 3 wpc for the adaptive parameters, it was only partial at 1 wpc for the innate parameters, and by 3 wpc it was completely absent. The clustering along strains at 0 wpc, prior to the encounter of SAV, suggests a pre-existing difference in immune parameters. The fact that the distinct strain-specific clustering was resumed at 8 wpc (adaptive), implies that the differences seen at 0 wpc were not accidental, but rather was a basic non-transient difference that can be correlated to the observed difference in disease resistance in the two strains. The higher mRNA level of innate parameters seen in the ISAHi cluster at 0 wpc may suggest a generally increased transcription activity in this strain. In a recent QTL study on Atlantic salmon, the transcription factor HIV-EP2/MBP-2 was suggested as one of the strong correlates for ISA resistance [36]. The collapse of the strain-specific clustering at 3 and 6 wpc for innate and at 6 wpc for adaptive parameters most likely reflects the general impact of the viral infection on the immune system. The very distinct cluster formed by SAV triple-positive fish at 3 wpc (innate), is noteworthy as it seem to show that widespread (systemic) infection induce innate parameters (at a late time point) while more confined infection do not. The present study demonstrates that two strains of Atlantic salmon genetically selected for resistance to ISA, also differ in resistance to pancreas disease in an experimental cohabitation

797

challenge. The differences in immune parameters observed between the two salmon strains may provide important clues to the causes underlying this resistance difference. As the ISA and SAV viruses differ in structure, virulence strategy and pathogenehost interaction, the observed resistance increase likely includes factors of a more general antiviral nature. Future comparison of immune responses in the ISAHi and ISALo strains to ISAV and SAV infections will facilitate identification of immune parameters and responses that are important to antiviral protection, both regarding these specific viruses and viruses in general. Acknowledgements We would like to thank Ingebjørg Modahl, Guro Seternes, Randi Faller, Ingebjørg Sævareid and Irene Gabestad for excellent technical assistance. Fish strains were kindly provided by SalmoBreed AS (Bergen, Norway). The work was supported by the Research Council of Norway, grant # 183196/S40, and by MSD Animal Health. Appendix A. Supplementary material Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.fsi.2012.12.014. References [1] Nelson RT, McLoughlin MF, Rowley HM, Platten MA, McCormick JI. Isolation of a Toga-like virus from farmed Atlantic salmon Salmo salar with pancreas disease. Dis Aquat Org 1995;22:25e32. [2] Weston J, Villoing S, Bremont M, Castric J, Pfeffer M, Jewhurst V, et al. Comparison of two aquatic alphaviruses, salmon pancreas disease virus and sleeping disease virus, by using genome sequence analysis, monoclonal reactivity, and cross-infection. J Virol 2002;76:6155e63. [3] Fringuelli E, Rowley HM, Wilson JC, Hunter R, Rodger H, Graham DA. Phylogenetic analyses and molecular epidemiology of European salmonid alphaviruses (SAV) based on partial E2 and nsP3 gene nucleotide sequences. J Fish Dis 2008;31:811e23. [4] Hodneland K, Bratland A, Christie KE, Endresen C, Nylund A. New subtype of salmonid alphavirus (SAV), Togaviridae, from Atlantic salmon Salmo salar and rainbow trout Oncorhynchus mykiss in Norway. Dis Aquat Org 2005;66:113e20. [5] Kristoffersen AB, Viljugrein H, Kongtorp RT, Brun E, Jansen PA. Risk factors for pancreas disease (PD) outbreaks in farmed Atlantic salmon and rainbow trout in Norway during 2003e2007. Prev Vet Med 2009;90:127e36. [6] McLoughlin MF, Graham DA. Alphavirus infections in salmonids e a review. J Fish Dis 2007;30:511e31. [7] Houghton G. Kinetics of infection of plasma, blood leukocytes and lymphoidtissue from Atlantic salmon Salmo salar experimentally infected with pancreas disease. Dis Aquat Org 1995;22:193e8. [8] Villoing S, Castric J, Jeffroy J, Le Ven A, Thiery R, Bremont M. An RT-PCR-based method for the diagnosis of the sleeping disease virus in experimentally and naturally infected salmonids. Dis Aquat Org 2000;40:19e27. [9] Graham DA, Frost P, McLaughlin K, Rowley HM, Gabestad I, Gordon A, et al. A comparative study of marine salmonid alphavirus subtypes 1e6 using an experimental cohabitation challenge model. J Fish Dis 2011;34:273e86. [10] McLoughlin MF, Graham DA, Norris A, Matthews D, Foyle L, Rowley HM, et al. Virological, serological and histopathological evaluation of fish strain susceptibility to experimental infection with salmonid alphavirus. Dis Aquat Org 2006;72:125e33. [11] McLoughlin MF, Nelson RT, Rowley HM, Cox DI, Grant AN. Experimental pancreas disease in Atlantic salmon Salmo salar post-smolts induced by salmon pancreas disease virus (SPDV). Dis Aquat Org 1996;26:117e24. [12] Houghton G. Acquired protection in Atlantic salmon Salmo salar parr and post-smolts against pancreas disease. Dis Aquat Org 1994;18:109e18. [13] Houghton G, Ellis AE. Pancreas disease in Atlantic salmon: serum neutralisation and passive immunisation. Fish Shellfish Immunol 1996;6:465e72. [14] Desvignes L, Quentel C, Lamour F, Le Ven A. Pathogenesis and immune response in Atlantic salmon (Salmo salar L.) parr experimentally infected with salmon pancreas disease virus (SPDV). Fish Shellfish Immunol 2002;12:77e95. [15] Strandskog G, Villoing S, Iliev DB, Thim HL, Christie KE, Jorgensen JB. Formulations combining CpG containing oliogonucleotides and poly I:C enhance the magnitude of immune responses and protection against pancreas disease in Atlantic salmon. Dev Comp Immunol 2011;35:1116e27. [16] Skjaeveland I, Iliev DB, Strandskog G, Jorgensen JB. Identification and characterization of TLR8 and MyD88 homologs in Atlantic salmon (Salmo salar). Dev Comp Immunol 2009;33:1011e7. [17] Ryman KD, Klimstra WB. Host responses to alphavirus infection. Immunol Rev 2008;225:27e45.

798

S. Grove et al. / Fish & Shellfish Immunology 34 (2013) 789e798

[18] Ryman KD, Meier KC, Gardner CL, Adegboyega PA, Klimstra WB. Non-pathogenic Sindbis virus causes hemorrhagic fever in the absence of alpha/beta and gamma interferons. Virology 2007;368:273e85. [19] Frolova EI, Fayzulin RZ, Cook SH, Griffin DE, Rice CM, Frolov I. Roles of nonstructural protein nsP2 and alpha/beta interferons in determining the outcome of Sindbis virus infection. J Virol 2002;76:11254e64. [20] Lopez-Doriga MV, Smail DA, Smith RJ, Domenech A, Castric J, Smith PD, et al. Isolation of salmon pancreas disease virus (SPDV) in cell culture and its ability to protect against infection by the ‘wild-type’ agent. Fish Shellfish Immunol 2001;11:505e22. [21] Moriette C, LeBerre M, Lamoureux A, Lai TL, Bremont M. Recovery of a recombinant salmonid alphavirus fully attenuated and protective for rainbow trout. J Virol 2006;80:4088e98. [22] Christie KE, Graham DA, McLoughlin MF, Villoing S, Todd D, Knappskog D. Experimental infection of Atlantic salmon Salmo salar pre-smolts by i.p. injection with new Irish and Norwegian salmonid alphavirus (SAV) isolates: a comparative study. Dis Aquat Org 2007;75:13e22. [23] Olsvik PA, Lie KK, Jordal AEO, Nilsen TO, Hordvik I. Evaluation of potential reference genes in real-time RT-PCR studies of Atlantic salmon. BMC Mol Biol 2005;6. [24] Lovoll M, Austbo L, Jorgensen JB, Rimstad E, Frost P. Transcription of reference genes used for quantitative RT-PCR in Atlantic salmon is affected by viral infection. Vet Res 2011;42. [25] Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 2001;29. [26] Hodneland K, Endresen C. Sensitive and specific detection of salmonid alphavirus using real-time PCR (TaqMan (R)). J Virol Methods 2006;131: 184e92. [27] Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: minimum information for publication of quantitative realtime PCR experiments. Clin Chem 2009;55:611e22.

[28] Graham DA, Jewhurst VA, Rowley HM, McLoughlin MF, Todd D. A rapid immunoperoxidase-based virus neutralization assay for salmonid alphavirus used for a serological survey in Northern Ireland. J Fish Dis 2003;26: 407e13. [29] Murphy TM, Drinan EM, Gannon F. Studies with an experimental model for pancreas disease of Atlantic salmon Salmo Salar L. Aquacult Res 1995;26: 861e74. [30] Grove S, Johansen R, Reitan LJ, Press CM, Dannevig BH. Quantitative investigation of antigen and immune response in nervous and lymphoid tissues of Atlantic halibut (Hippoglossus hippoglossus) challenged with nodavirus. Fish Shellfish Immunol 2006;21:525e39. [31] Salinas I, Zhang Y-A, Sunyer JO. Mucosal immunoglobulins and B cells of teleost fish. Dev Comp Immunol 2011;35:1346e65. [32] Koppang EO, Fischer U, Moore L, Tranulis MA, Dijkstra JM, Kollner B, et al. T cells assemble in the thymus, spleen and in novel interbranchial lymphoid tissue. J Anat 2010;217:728e39. [33] Ingerslev HC, Ronneseth A, Pettersen EF, Wergeland HI. Differential expression of immune genes in Atlantic salmon (Salmo salar L.) challenged intraperitoneally or by cohabitation with IPNV. Scand J Immunol 2009;69: 90e8. [34] Jorgensen SM, Hetland DL, Press CM, Grimholt U, Gjoen T. Effect of early infectious salmon anaemia virus (ISAV) infection on expression of MHC pathway genes and type I and II interferon in Atlantic salmon (Salmo salar L.) tissues. Fish Shellfish Immunol 2007;23:576e88. [35] Sun B, Skjaeveland I, Svingerud T, Zou J, Jorgensen J, Robertsen B. Antiviral activity of salmonid gamma interferon against infectious pancreatic necrosis virus and salmonid alphavirus and its dependency on type I interferon. J Virol 2011;85:9188e98. [36] Li J, Boroevich KA, Koop BF, Davidson WS. Comparative genomics identifies candidate genes for infectious salmon anemia (ISA) resistance in Atlantic salmon (Salmo salar). Mar Biotechnol (NY) 2011;13:232e41.