- Email: [email protected]
How specific MHC class I and class II combinations affect disease resistance against infectious salmon anaemia in Atlantic salmon (Salmo salar)
Fish & Shellfish Immunology 21 (2006) 431e441 www.elsevier.com/locate/fsi
How specific MHC class I and class II combinations affect disease resistance against infectious salmon anaemia in Atlantic salmon (Salmo salar) Sissel Kjøglum a,b,*, Stig Larsen c, Hege G. Bakke b, Unni Grimholt b b
a Aqua Gen AS, P.O. Box 1240, Pirsenteret, 7462 Trondheim, Norway Department of Morphology, Genetics and Aquatic Biology, Norwegian School of Veterinary Science, P.O. Box 8146 Dep., 0033 Oslo, Norway c Department of Production Animal Clinical Sciences, Norwegian School of Veterinary Science, P.O. Box 8146 Dep., 0033 Oslo, Norway
Received 9 December 2005; revised 31 January 2006; accepted 3 February 2006 Available online 17 April 2006
Abstract The aim was to evaluate the performance of selected individual MHC class I and class II alpha (A) alleles, and combinations of these on disease resistance against infectious salmon anaemia (ISA). The material consisting of 1966 fish from seven families, contained five MHC class I alleles and four MHC class II A alleles. Which representing given class II A and class II beta (B) haplotypes, totalling 19 MHC class I and class II A genotypes. The fish were challenged with infectious salmon anaemia virus (ISAV), the virus causing ISA. Dead fish were collected daily during the challenge experiment and the survivors were collected at termination. All fish were genotyped for MHC class I and class II A. The total mortality in the material was 85.14%. For MHC class I, UBA*0201 and UBA*0301 were significantly the most resistant alleles, while UBA*0601 for class I and DAA*0301 for class II A were the significantly most susceptible alleles. The analysis of combined MHC class I and class II A genotypes detected that fish with the genotype UBA*0201/*0301;DAA*0201/*0201 were the most resistant fish with a hazard ratio (HR) at 0.750, while the fish with the genotypes UBA*0601/*0801;DAA*0501/*0501 and UBA*0201/*0301;DAA*0301/*0501 were the most susceptible fish with HR of 1.334 and 1.425. In addition, Cox regression analysis within family detected combined MHC class I and class II A genotypes that contributed significantly to resistance and susceptibility. The study confirmed the expectation of performance of individual MHC class I and class II A alleles, and also detected an effect of MHC class I and class II A in combinations. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Atlantic salmon; Association; Infectious salmon anaemia; Major histocompatibility complex; Resistance; Susceptibility
1. Introduction Infectious salmon anaemia (ISA) is an emerging disease causing severe damage to the farming industry in an increasing number of countries. The disease, first reported in Norway in 1984, has since been reported in * Corresponding author. Tel.: þ47 72 45 05 00; fax: þ47 72 45 05 25. E-mail address: [email protected] (S. Kjøglum). 1050-4648/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.fsi.2006.02.001
432
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
Canada (1996), Scotland (1998), the Faroe Islands and possibly Chile (1999) and the USA (2002) [1e6]. The accumulated mortality in an ISA disease outbreak ranges from 15% to 100%. The disease is caused by the infectious salmon anaemia virus (ISAV), first detected by electron microscopy in 1994 and isolated in cell culture in 1995 [7e9]. ISAV is classified as a member of the Orthomyxoviridae [10]. In Europe, the controlling strategy has been to stamp out fish diagnosed with ISA. Vaccination has not been allowed in Europe, while two commercial vaccines are available in Canada/USA [11e13]. The vaccines have been used during the last two years in some areas of Canada/USA with severe ISA problems, but the results so far have varied and significant protection in the field has not been documented [13]. The ISA problem in the fish farming industry can alternatively be controlled by selection for increased resistance. Resistance towards ISA have been found to be heritable with h2 ¼ 0.13 [14]. But the heritability for ISA was characterised as medium to low compared to other bacterial diseases such as V. salmonicida causing cold water vibriosis and V. anguillarum causing vibriosis [14]. Despite the low heritability, ISA was included in the Aqua Gen breeding scheme in 1995 and challenge tests have been applied to measure the fish’s resistance to the disease [15]. The heritability and the additive genetic variance detected using challenge tests suggest that there is at least one gene that contributes to disease resistance [16]. Genes of major histocompatibility complex (MHC) are candidate genes for disease resistance. MHC are crucial elements of adaptive immunity and have been linked to numerous diseases [17e24]. The MHC contain some of the most polymorphic genes known to date and encode polypeptides which recognise and bind both self and foreign peptides and present them to T cells. Generally, foreign peptides produced by the degradation of intracellular pathogens are bound by MHC class I molecules and presented to cytotoxic T cells. Alternately, foreign peptides derived from extracellular pathogens are bound by MHC class II molecules and presented to helper T cells [25]. The peptide binding ability depends on the peptide binding grooves of MHC molecules [23]. Hence, some pathogens may escape recognition by certain MHC molecules since their peptides are not presentable by the MHC. This can lead to increased susceptibility for pathogens. Alternatively, resistance to pathogens may be derived through high affinity binding of certain peptides by specific MHC alleles. In Atlantic salmon, MHC consists of only one expressed classical MHC class I locus (Sasa-UBA) and one classical class II alpha (A) and class II beta (B) haplotype (SasaDAA/DAB). MHC class I and class II reside on different linkage groups [26]. For MHC class II, Atlantic salmon has distinct class II haplotypes consisting of a combination of unique class II A and class II B alleles [27]. Because of the major expressed classical class I and class II locus, the number of peptide-presenting alleles per animal is limited to two within each class. In Atlantic salmon, genotypes of MHC class I and class II have been found to provide resistance to infectious diseases [28e30]. Associations between MHC class II B and resistance to Aeromonas salmonicida [28], and associations between MHC class I and class II A and resistance against Aeromonas salmonicida and ISAV have previously been demonstrated [29]. MHC class I and class II loci associated with susceptibility to infectious haematopoietic necrosis virus (IHNV) have also been reported [30]. The total project was performed sequentially in two steps. In the first step, published in Grimholt et al. [29], associations between MHC and resistance towards ISA was found in a group of fish which included a large number of families. The study included the total number of 1 031 fish from 25 families, presenting 11 alleles for MHC class I and 7 alleles for MHC class II A. Resistant genotypes for both MHC class I and class II A were found. Compared to the total mortality at 42.68% the resistant genotypes provided mortality at 21.13% for class I UBA*0201/*0301 and 28.23% for class I UBA*0701/*0801, and 31.82% for class II DAA*0201/*0501, 32.14% for class II DAA*0501/*0501 and 38.52% for class II DAA*0201/*0201. However, due to the large number of alleles and high polymorphism in both MHC class I and class II A only three groups were large enough to be included in the study with combinations of MHC class I and class II A genotypes. In the second step, i.e. in this study, genotypes of MHC were produced and evaluated in the next generation with regard to expectations from findings in Grimholt et al. [29]. By sampling a large number of animals per family, the study population would provide groups large enough to be included in the analysis of combined MHC class I and class II A genotypes. The aim was to evaluate the performance of selected individual MHC class I and class II A alleles and combinations of these on resistance against ISA.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
433
2. Material and method 2.1. Experimental design Seven full-sib families were reared at the Aqua Gen breeding station at Kyrksæterøra and kept in separate tanks (Fig. 1). The test fish were randomly sampled from each family six months after first feeding and individually tagged before they were included in the challenge experiment. The virus ISAV was used in the challenge. Dead fish were registered and collected daily.
2.2. Families Based on a previous study [29], sire and dams in the F1 generation were selected as parental fish to produce offspring in F2 (Fig. 1). The selected brood fish in F1 were siblings of the fish included in the previous study. The aim was to produce MHC class I and class II A genotypes with expected high and low mortality. Based on expectations from Grimholt et al. [29], UBA*0201, UBA*0301, UBA*0701, UBA*0801, DAA*0201 and DAA*0501 were the best performing alleles while UBA*0401, DAA*0101, DAA*0401 and DAA*0601 were the worst performing alleles. With these alleles, genotypes of interest would be an allele in combination with different alleles to investigate if the allele is well performing as a single allele or increases in performance with combination to a different allele. The sample of 104 breeding candidates from generation F1 could not provide all genotypes of interest. The chosen matings resulted in seven families with 19 MHC class I and class II A genotypes containing the alleles UBA*0201, UBA*0301, UBA*0501, UBA*0601, UBA*0801, DAA*0201, DAA*0301, DAA*0501 and DAA*0701. Families 1, 2 and 3 were offspring of one sire crossed with three dams, hence they were half-siblings (Fig. 1). The other four families were partly related. The number of fish from each family was chosen depending on the number of MHC genotypes in the family. The large number of fish sampled from each family, provided a sufficient number of fish per given MHC class I and class II A genotype to allow statistical analysis with a significant level of 5% and a power of 90%, and resulted in 46e383 animals per each of the 19 genotypes. The total number of fish included in the ISA challenge experiment was 1966.
Base population x
x
x
x
x
x
x
UBA*0301/*0301 UBA*0301/*0501 UBA*0201/*0201 UBA*0201/*0501 UBA*0601/*0801 UBA*0701/*0801 UBA*0701/*0801 UBA*0701/*0801 UBA*0601/*0901 UBA*0301/*0501 UBA*0201/*0601 UBA*0601/*1201 UBA*0201/*0601 UBA*0501/*0701 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0501 DAA*0201/*0301 DAA*0201/*0501 DAA*0301/*0701 DAA*0501/*0501 DAA*0501/*0501 DAA*0301/*0301 DAA*0201/*0201 DAA*0201/*0301 DAA*0201/*0301 DAA*0201/*0301 DAA*0301/*0601
F1
UBA*0301/*0301 UBA*0301/*0301 UBA*0301/*0301 UBA*0201/*0201 UBA*0201/*0201 UBA*0301/*0801 UBA*0601/*0801 UBA*0801/*0801 UBA*0501/*0601 UBA*0601/*0601 UBA*0501/*0601 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0501 DAA*0301/*0701 DAA*0501/*0701 DAA*0501/*0501 DAA*0201/*0301 DAA*0201/*0301 DAA*0201/*0301
F2
Fam 1
Fam 2
Fam 3
Fam 5
Fam 6
Fam 7
Fam 8
Fig. 1. The pedigree, showing the families (Fam) in generation F2 with parents in F1 and grandparents in the base population.
434
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
2.3. Challenge experiment The fish, with a mean weight of 20 g, were transported to the test station, VESO Vikan (http://www.veso.no). At the test station fish were put into one tank with freshwater at 12.7 C, with a diameter of 3 m, and a water flow producing an oxygen level of 6.0e11.0 mgO2/L. The fish were infected using a standardised cohabitation challenge after an acclimatisation period of 33 days [31,32]. The cohabitants were inoculated intraperitoneally with the ISAV isolate Glesvaer/2/90, developed the disease and transmitted ISA to the fish sampled from the seven families. The main outbreak caused by ISA started after 22 days and the study terminated 51 days after infection with a total mortality of 85.14%.
2.4. MHC genotyping The base population (Fig. 1) was thoroughly MHC class I and class II A genotyped using cDNA [26], identifying linkage between the polymorphic markers in the 30 UTRs and expressed MHC class I and class II alleles. The class II marker represents class II A and B haplotypes [27]. Finclippings from a parental material of 104 dams and sires from the F1 population (Fig.1) were collected and used for isolation of genomic DNA using a DNeasy Tissue kit (Qiagen, Germany) according to manufacturer’s protocol. The polymorphic repeats located in the 30 UTRs of Sasa-UBA and Sasa-DAA genes [26,27] were amplified using PCR with primers flanking the repeats. The sense primer was fluorescently labelled in both cases. The resultant fragments were analysed on an automated sequencing machine (ABI377) using the Genescan software (Applied Biosystems, USA). For MHC class II, the class II A DAA*0201 marker represents the DAA*0201;DAB*0201 haplotype, the DAA*0301 marker represents the DAA*0301;DAB*0401 haplotype, the DAA*0501 marker represents the DAA*0501;DAB*0301 haplotype and the DAA*0701 marker represents the DAA*0701;DAB*0501 haplotype. All fish in families 1, 2 and 3 had identical MHC genotypes, hence genotyping was not needed. Genomic DNA from 25 mg muscle tissue from 1583 animals of the F2 population challenged with ISAV were isolated using the Qiagen DNeasy 96 Tissue kit according to standard protocol. MHC class I and class II A genotypes were determined using marker typing with the primer sets and analyses described above. For family 5, both the segregating class I alleles UBA*0301 and UBA*0801 have identical markers. Thus, allele specific primers covering the individual alpha 1 domains were designed. For UBA*0301 the forward primer sequence was 50 ttcccagagtttgtgactgtt while the reverse primer sequence was 50 caatggagacctgagtctga. For UBA*0801 the forward primer sequence was 50 ttcccagagtttgtcaatct while the reverse primer sequence was 50 ctgattgaagcgctgcttgac. Standard PCR conditions using 100 ng genomic DNA with 30 cycles and annealing temperatures of 59 C and 64 C amplified fragments of 148 bp and 204 bp for UBA*0301 and UBA*0801 respectively.
2.5. Statistical methods The results are presented by proportional Hazard Ratio (HR), with 95% confidence intervals (CI) [33]. The time until death is graphically expressed using KaplaneMeier survival plots [33]. Additionally the mortality with 95% confidence interval was calculated using the theory of simple Bernoulli sequences [34]. The hazard ratio compares mortality in the group having the allele or genotype versus the group not having the allele or genotype. The proportional HR was calculated using the proportional hazards regression (PHREG) model of Cox for comparison of two groups; hðt; xÞ ¼ ebx h0 ðtÞ [35], where x ¼ 0 for control and x ¼ 1 for allele or genotype. In the Cox model no assumptions are made about the parametric form of h0(t), which is the base ^ ^ is the maximum likelihood estimates of line for the control group. The hazard ratio eb is estimated by eb, where b the b parameter [35]. The confidence interval for the proportional HR is estimated using the Student procedure ^ exp b ^ za=2 seðbÞ ^ [35]. The PHREG model was also used to detect significant effects for the estimated b, on the survival data, using selection procedures. The significance level for entering the regression model was set to 15%. All comparisons were performed two-tailed and differences were considered significant if the p values were found to be less than or equal to a significance level (a) of 5%.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
435
3. Results 3.1. Single MHC class I analysis The MHC class I alleles UBA*0201 and UBA*0301, both present in four families, performed as the most resistant alleles (Table 1). The alleles UBA*0501, UBA*0601 and UBA*0801, present in two to three families performed as the most susceptible alleles. The group of fish with UBA*0201 or UBA*0301 performed significantly better than the groups not having UBA*0201 or UBA*0301, and the group of fish with UBA*0601 performed significantly worse than the group not having UBA*0601. In the single allele analysis for class I, the differences in mortality for the five alleles were within 10.3%.
3.2. Combined MHC class I analysis Of the six MHC class I allele combinations, two combinations performed significantly. The combination UBA*0201/*0301 present in four families, performed best as a significant resistant genotype (Table 2). The combination UBA*0601/*0801 present in one family performed worst as a significant susceptible genotype (Table 2). The range between the group with highest and lowest mortality in the MHC class I groups was 12.4%.
3.3. Single MHC class II A analysis The MHC class II A alleles DAA*0701 and DAA*0201, present in two and six families, were the most resistant alleles (Table 1). The alleles DAA*0501 and DAA*0301, present in two and three families, were the most susceptible alleles. The only significantly performing class II A allele was DAA*0301. The group of fish with DAA*0301 performed significantly worse than the group not having DAA*0301. In the single allele analysis for class II A, the differences in mortality for the four alleles were within 6.0%.
Table 1 Hazard ratio and mortality in the single allele groups with at least one of the given class I (UBA) or class II A (DAA) alleles present Allele
UBA*0201 UBA*0301 UBA*0801 UBA*0601 UBA*0501 DAA*0701 DAA*0201 DAA*0501 DAA*0301
Hazard ratio
Mortality
(95% CI)
(95% CI)
0.837 (0.761, 0.867 (0.783, 1.025 (0.929, 1.202 (1.089, 1.125 (0.987, 0.992 (0.886, 0.92 (0.834, 1.089 (0.985, 1.125 (1.020,
0.922) 0.959) 1.133) 1.327) 1.283) 1.109) 1.014) 1.204) 1.240)
No. of fish
Present in family
80.76% (78.41, 83.11) 81.40% (78.52, 84.29) 83.24% (80.55, 85.93) 90.82% (88.67, 92.96) 91.10% (87.83, 94.36)
1083
1, 2, 3,5
700
1, 2, 3, 5
741
5, 6
697
6, 7, 8
292
7, 8
82.24% (78.86, 85.63) 83.94% (81.85, 86.03) 85.52% (82.89, 88.16) 88.28% (86.03, 90.53)
490
5, 6
1183
1, 2, 3, 5, 7, 8
684
5, 6
785
5, 7, 8
The results are expressed with 95% confidence intervals in parentheses. The family number names, in which the groups are present, are listed in the last column. Significant HR are presented in bold.
436
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
Table 2 Hazard ratio and mortality in groups with class I (UBA) and class II A (DAA) genotypes Allele
UBA*0201/*0801 UBA*0201/*0301 UBA*0801/*0801 UBA*0601/*0801 UBA*0501/*0601 UBA*0601/*0601 DAA*0201/*0201 DAA*0501/*0701 DAA*0201/*0701 DAA*0301/*0501 DAA*0201/*0301 DAA*0301/*0301 DAA*0501/*0501
Hazard ratio
Mortality
(95% CI)
(95% CI)
0.931 (0.822, 0.867 (0.784, 1.016 (0.863, 1.223 (1.034, 1.123 (0.985, 1.119 (0.970, 0.804 (0.718, 0.964 (0.845, 1.048 (0.882, 1.117 (0.946, 1.100 (0.987, 1.021 (0.824, 1.194 (1.020,
1.054) 0.960) 1.195) 1.195) 1.280) 1.291) 0.899) 1.100) 1.246) 1.319) 1.225) 1.263) 1.399)
No. of fish
Present in family
79.58% (75.54, 83.39) 81.40% (78.52, 85.86) 85.64% (80.63, 89.04) 88.82% (84.09, 92.77) 91.1% (87.83, 90.89) 91.95% (88.43, 95.84)
382
5
699
1, 2, 3, 5
188
6
170
6
292
7, 8
235
7, 8
79.88% (76.37, 83.39) 81.62% (77.38, 85.86) 83.43% (77.83, 89.04) 87.93% (83.09, 92.77) 88.11% (85.28, 90.89) 89.9% (83.96, 95.85) 90.05% (85.66, 94,23)
502
1, 2, 3, 7, 8
321
5, 6
169
5
174
5
512
5, 7, 8
99 189
7, 8 6
The results are expressed with 95% confidence intervals in parentheses. The family number names, in which the groups are present, are listed in the last column. Significant HR are presented in bold.
3.4. Combined MHC class II A analysis Of the seven MHC class II A allele combinations, one combination performed significantly better and one significantly worse (Table 2). The combination DAA*0201/*0201 present in 5 families performed best and was a significant resistant class II genotype (Table 2). The combination DAA*0501/*0501, present in one family performed as a significant susceptible genotype (Table 2). The range between the group with highest and lowest mortality in the MHC class II A groups was 10.2%. 3.5. Combined MHC class I and class II A analysis One combination performed significantly better and two significantly worse (Fig. 2a). The fish with genotype UBA*0201/*0301;DAA*0201/*0201, present in three families performed best and was the significantly most resistant genotype (Table 3). The genotypes performing as the significantly susceptible genotypes were UBA*0601/ *0801;DAA*0501/*0501 and UBA*0201/*0301;DAA*0301/*0501. The range between the group with highest and lowest mortality in MHC class I and class II A groups was 23.7%. 3.6. Family and MHC A significant difference in performance between families was detected (Fig. 2b). Families 1 and 2 were significantly resistant families, while family 7 was a significantly susceptible family. The range between the families with highest and lowest mortality was 22.8%.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
437
(a) 1.0 0.9 0.8
Survival
0.7 0.6 0.5
UBA*0201/*0301;DAA*0201/*0201 UBA*0201/*0301;DAA*0301/*0501
0.4
UBA*0601/*0801;DAA*0501/*0501
0.3 0.2 0.1 0.0 5
10
15
20
25
30
35
40
45
50
30
35
40
45
50
day
(b) 1.0 0.9 0.8
Survival
0.7 0.6
Family 1 Family 2
0.5
Family 3 Family 5 Family 6 Family 7
0.4 0.3
Family 8
0.2 0.1 0.0 5
10
15
20
25
day Fig. 2. The number of days from challenge to death expressed by KaplaneMeier survival plots. Only the groups with significantly increased or reduced mortality in the combined class I and class II alpha (a) are included in the figures. KaplaneMeier survival plots for all families are presented (b).
In family 5 the mortality varied by 23.7% between the MHC genotypes performing best and worst (Table 3). The genotype performing best in family 5 was UBA*0201/*0801; DAA*0501/*0701, while the genotype performing worst was UBA*0201/*0301; DAA*0301/*0501. In the Cox regression analysis, both these MHC genotypes contributed significantly to explaining variation in survival. In family 6 the mortality varied by 9.7% between the MHC genotypes performing best and worst (Table 3). The genotype performing best in family 6 was UBA*0801/*0801; DAA*0501/*0701, while the genotype performing worst was UBA*0601/*0801; DAA*0501/*0501. In the Cox regression analysis the MHC genotype UBA*0601/*0801; DAA*0501/*0501 contributed significantly to explaining variation in survival. Families 7 and 8 contained the same MHC genotypes (Table 3). The mortality varied by 5.6% between the genotypes performing best and worst in this group. The genotype performing best was UBA*0501/*0601; DAA*0201/ *0201, while UBA*0501/*0601; DAA*0201/*0301 performed worst. None of these MHC genotypes contributed significantly in explaining the variation in survival.
438
Table 3 Hazard ratio and mortality in groups with combined MHC class I and class II A allele groups MHC genotype Class I
UBA*0201/*0801
Frequencies in family groups Class II alpha
Fam 1
Fam 2
Survival data MHC Fam 3
DAA*0501/*0701
Fam 5
Fam 6
Fam 7
Fam 8
72
DAA*0201/*0201
UBA*0201/*0301
DAA*0201/*0301
77
UBA*0201/*0801
DAA*0301/*0501
98
UBA*0201/*0801
DAA*0201/*0301
126
UBA*0201/*0801
DAA*0201/*0701
86
UBA*0801/*0801
DAA*0501/*0701
UBA*0201/*0301
DAA*0501/*0701
80
UBA*0201/*0301
DAA*0201/*0701
83
UBA*0601/*0801
DAA*0501/*0701
UBA*0501/*0601
DAA*0201/*0201
UBA*0801/*0801
DAA*0501/*0501
UBA*0501/*0601
DAA*0301/*0301
31
22
UBA*0601/*0601
DAA*0201/*0201
34
30
UBA*0601/*0601
DAA*0301/*0301
25
21
UBA*0601/*0801
DAA*0501/*0501
UBA*0601/*0601
DAA*0201/*0301
64
61
UBA*0501/*0601
DAA*0201/*0301
89
95
UBA*0201/*0301
DAA*0301/*0501
Survival data Families
130
128
125
83
86 26
29
105
84
76
Hazard ratio (95% CI)
0.769 (0.62, 0.94)
0.631 (0.51, 0.78)
1.005 (0.82, 1.22)
1.006 (0.91, 1.11)
1.120 (0.99, 1.27)
1.316 (1.15, 1.50)
0.962 (0.84, 1.11)
Mortality (95% CI)
74.61% (67.1, 82.1)
72.66% (64.9, 80.)
84.00% (77.6, 90.4)
82.80% (80.0, 85.6)
87.15% (83.7, 90.6)
95.48% (92.6, 97.7)
87.60% (83.6, 91.6)
The results are expressed with 95% confidence intervals in parentheses. Significant HR are presented in bold.
Mortality
(95% CI)
(95% CI)
0.769 (0.585, 0.750 (0.661, 0.907 (0.704, 0.918 (0.733, 0.981 (0.803, 1.083 (0.853, 0.928 (0.728, 1.087 (0.851, 1.011 (0.797, 1.107 (0.877, 1.174 (0.881, 1.090 (0.884, 0.962 (0.720, 0.984 (0.757, 1.094 (0.805, 1.334 (1.062, 1.205 (0.998, 1.146 (0.978, 1.425 (1.128,
73.61% (61.90, 83.30) 77.02% (72.81, 81.24) 80.51% (69.91, 88.67) 80.61% (71.39, 87.90) 80.95% (74.10, 87.81) 81.40% (71.55, 88.98) 81.93% (71.95, 89.52) 83.75% (73.82, 91.05) 85.54% (76.11, 92.3) 86.05% (76.89, 92.58) 87.27% (75.52, 94.73) 88.57% (82.49, 94.66) 88.68% (76.97, 95.73) 90.62% (80.70, 96.48) 91.30% (79.21, 97.58) 91.67% (83.58, 96.58) 92.8% (88.27, 97.33) 92.93% (89.23, 96.64) 97.37% (90.82, 99.68)
1.011) 0.851) 1.169) 1.151) 1.198) 1.376) 1.183) 1.387) 1.282) 1.397) 1.565) 1.343) 1.286) 1.279) 1.486) 1.677) 1.455) 1.342) 1.800)
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
UBA*0201/*0301
Hazard ratio
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
439
4. Discussion A main difference between the previous study and the follow-up study is the challenge method used. In the previous study by Grimholt et al. [29], all fish were infected by i.p. injection of ISAV, while in this study a cohabitation method was used to infect the fish. The cohabitation method in ISA challenge was not available in 1998 when the previous challenge test was carried out. But at present this method is used as a standard, because it is assumed to represent a more natural infection that also involves innate defence mechanisms. The different challenge methods could cause some deviation from the expectation of MHC performance, but also resulted in different challenge pressures. The mortality in the challenge experiment using i.p. injection was moderate, while the mortality in the challenge experiment using cohabitations was high. For survival analysis information on more individuals is preferred, and therefore the present study with a high challenge pressure and high mortality is more appropriate for such analysis. The present study detected one genotype of combined MHC class I and class II A alleles, which performed significantly better and were more resistant against ISA. Two MHC class I and class II A genotypes were also found to be significantly more susceptible to ISA. The single allele analysis detected a single allele for MHC class I being more resistant, and one MHC class I allele and one class II A allele was found to be significantly more susceptible. The parental material of 104 breeding candidates provided us with 7 families with sufficient number of animals per selected genotype to enable statistical analysis. These 7 families are not representative for the entire population partly due to limitation in the number of alleles, as some of the most susceptible alleles were not present in the breeding candidates [29]. However, the material provided 19 combined MHC class I and class II A genotypes. In the single allele analysis the results were partly in accordance with expectations. For MHC class I, the alleles UBA*0501 and UBA*0601 were included in the study as the susceptible alleles based on expectations from the previous study [29]. These alleles performed in accordance with expectations, and were also the most susceptible alleles in the present study. The class I allele UBA*0801 was included in the study as the most resistant allele, but did not perform as expected. UBA*0201 and UBA*0301, however, performed as expected both in the single allele analysis as well as in the combined class I analysis. The class II A allele DAA*0301 was the susceptible allele included in the study, and in the single allele analysis this allele performed as expected. The most resistant allele for MHC class II A was expected to be DAA*0501, but did not perform as a resistant allele in this study. Hence, some of the MHC class I and class II A the alleles expected to be resistant did not show resistance in the single allele analysis, but the alleles expected to be susceptible performed in accordance with expectations. The small differences in performance between the alleles in the single allele analysis could cause some of the problems in detecting alleles being most resistant. In addition, the family effect including non-MHC genetic effects influence the results [36]. The alleles were present in a limited number of families, hence other genetic effects in these families could strongly influence performance and thus reduce the observed effect of MHC. Also, in the analysis of combined genotypes only small differences in performance between the genotypes appeared. The difference between the best and worse performing genotypes was 23.7%. However, genotypes contributing significantly to the resistance and susceptibility were detected. The variation between the mortality in families was also small. The present results detected a family effect expressed by significant variation in performance between three of the families. The major part of the differences in prevalence between the families was caused by genetic effects and not due to environmental factors because the fish were challenged in one tank. Minor parts of the variation could be caused by different environment prior to the challenge experiment when the fish were kept in separate family tanks. The genetic effect consisted of MHC, analysed in this study, and non-MHC genetic effects that also affect performance [36]. The material did not give the opportunity to evaluate MHC versus family effect, but analysis applied in four families that contained more than one genotype showed variation in performance of MHC genotypes within the families. Families 5 and 6 did not perform significantly different from expected, but in both families MHC genotypes contributed significantly to resistance and susceptibility and MHC had an effect on resistance. Family 7 was a significantly more susceptible family. In this family no MHC genotypes contributed to resistance, hence in family 7 there was the non-MHC genetic effect that contributed to susceptibility. Resistance towards ISA has been found to be heritable, but with low heritability compared with other viral diseases and also bacterial diseases. This implies that the observed genetic variation for resistance is small, which is in accordance with the small variation found in the present study for both MHC genotypes and families. This might also explain why MHC effects are stronger for furunculosis resistance than for ISA resistance ([28,29]; Kjøglum et al.,
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
440
unpublished). The relatively low MHC influence on ISA resistance could be caused by a viral host battle, where the ISA virus avoids the defence mechanisms of the host. The ISA virus has been shown to down regulate expression of MHC class I when infecting a salmon cell line [37]. In conclusion, the present study confirmed the expectations in performance of selected individual MHC class I and class II A alleles, and combinations of these were found to significantly influence disease resistance against ISA. Acknowledgement This study was funded by Aqua Gen AS and the Norwegian Research Council. References [1] Jarp J, Karlsen E. Infectious salmon anaemia (ISA) risk factors in sea-cultured Atlantic salmon Salmo salar. Dis Aquat Org 1997;28:79e86. [2] Mullins JE, Groman D, Wadowska D. Infectious salmon anaemia in salt water Atlantic salmon (Salmo salar L.) in New Brunswick, Canada. Bull Eur Assoc Fish Pathol 1998;11:110e4. [3] Rodger HD, Turnbull T, Muir F, Millar S, Richards RH. Infectious salmon anaemia (ISA) in the United Kingdom. Bull Eur Assoc Fish Pathol 1998;18:115e6. [4] Kibenge FSB, Ga´rate ON, Johnson G, Arriagada R, Kibenge MJT, Wadowska D. Isolation and identification of infectious salmon anaemia virus (ISAV) from Coho salmon in Chile. Dis Aquat Org 2001;45:9e18. [5] Bouchard DA, Brockway K, Giary C, Keleher W, Merrill PL. First report of infectious salmon anaemia (ISA) in the United States. Bull Eur Assoc Fish Pathol 2001;21:86e8. [6] Murray AG, Smith RJ, Stagg RM. Shipping and the spread of infectious salmon anemia in Scottish aquaculture. Emerging Infect Dis 2002;8:1e5. [7] Hovland T, Nylund A, Watanabe K, Endresen C. Observation of infectious salmon anaemia virus in Atlantic salmon, Salmo salar L. J Fish Dis 1994;17:291e6. [8] Dannevig BH, Falk K, Namork E. Isolation of the causal virus of infectious salmon anaemia (ISA) in long-term cell line from Atlantic salmon head kidney. J Gen Virol 1995;76:1353e9. [9] Nylund A, Hovland T, Watanabe K, Endresen C. Presence of infectious salmon anaemia virus (ISAV) in tissues of Atlantic salmon, Salmo salar L., collected during three separate outbreaks of the disease. J Fish Dis 1995;18:135e45. [10] Falk K, Namork E, Rimstad E, Mjaaland S, Dannevig BH. Characterization of infectious salmon anemia virus, an orthomyxo-like virus isolated from Atlantic salmon (Salmo salar L.). J Virol 1997;71:9016e23. [11] Jones SRM, MacKinnon AM, Salonius K. Vaccination of fresh-water-reared Atlantic salmon reduces mortality associated with infectious salmon anaemia virus. Bull Eur Assoc Fish Pathol 1999;19:98e101. [12] Brown LL, Sperker SA, Clouthier S, Thornton JC. Development of a vaccine against infectious salmon anaemia virus (ISAV). Bulletin of the Aquaculture Association of Canada 2000;100:4e7. [13] Biering E, Villoing S, Sommerset I, Christie KE. Update on viral vaccines for fish. In: Midtlyng P, editor. Progress in fish vaccinology. 1st ed. Basel: S. Krager; 2005. p. 97e113. [14] Gjoen HM, Refstie T, Ulla O, Gjerde B. Genetic correlations between survival of Atlantic salmon in challenge and field tests. Aquaculture 1997;158:277e88. [15] Midtlyng PJ, Storset A, Michel C, Slierendrecht WJ, Okamoto N. Breeding for disease resistance in fish. Bull Eur Assoc Fish Pathol 2002;22:166e72. [16] Falconer DS. Quantitative genetics. 3rd ed. England: Longman Scientific and Technical; 1989. [17] Lamont SJ. The chicken major histocompatibility complex and disease. Rev Off Int Epizoot 1998;17:128e42. [18] Paterson S, Wilson K, Pemberton JM. Major histocompatibility complex variation associated with juvenile survival and parasite resistance in a large unmanaged ungulate population (Ovis aries L.). Proc Natl Acad Sci USA 1998;95:3714e9. [19] Palti Y, Nichols KM, Waller KI, Parsons JE, Thorgaard GH. Association between DNA polymorphisms tightly linked to MHC class II genes and IHN virus resistance in backcrosses of rainbow and cutthroat trout. Aquaculture 2001;194:283e9. [20] Hedrick PN. Pathogen resistance and genetic variation at MHC loci. Evolution 2002;56:1902e8. [21] Arkush KD, Giese AR, Mendonca HL, McBride AM, Marty GD, Hedrick PW. Resistance to three pathogens in the endangered winter-run chinook salmon (Oncorhynchus tshawytscha): effects of inbreeding and major histocompatibility complex genotypes. Can J Fish Aquat Sci 2002;59:966e75. [22] Penn DJ, Damjanovich K, Potts WK. MHC heterozygosity confers a selective advantage against multiple-strain infections. Proc Natl Acad Sci USA 2002;99:11260e4. [23] Bernatchez L, Landry C. MHC studies in nonmodel vertebrates: what have we learned about natural selection in 15 years? J Evol Biol 2003;16:363e77. [24] Wedekind C, Walker M, Portmann J, Cenni B, Mu¨ller R, Binz T. MHC-linked susceptibility to a bacterial infection, but no MHC cryptic female choice in whitefish. J Evol Biol 2004;17:11e8. [25] Marieb EN. Human anatomy and physiology. 6th ed. San Francisco: Pearson Benjamin Cummings; 2004.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
441
[26] Grimholt U, Drablos F, Jorgensen SM, Hoyheim B, Stet RJM. The major histocompatibility class I locus in Atlantic salmon (Salmo salar L.): polymorphism, linkage analysis and protein modelling. Immunogenetics 2002;54:570e81. [27] Stet RJM, de Vries B, Mudde K, Hermsen T, van Heerwaarden J, Shum BP, et al. Unique haplotypes of co-segregating major histocompatibility class II A and class II B alleles in Atlantic salmon (Salmo salar) give rise to diverse class II genotypes. Immunogenetics 2002;54:320e31. [28] Langefors A, Lohm J, Grahn M, Andersen O, von Schantz T. Association between major histocompatibility complex class IIB alleles and resistance to Aeromonas salmonicida in Atlantic salmon. Proc R Soc Lond B Biol Sci 2001;268:479e85. [29] Grimholt U, Larsen S, Nordmo R, Midtlyng P, Kjoeglum S, Storset A, et al. MHC polymorphism and disease resistance in Atlantic salmon (Salmo salar); facing pathogens with single expressed major histocompatibility class I and class II loci. Immunogenetics 2003;55:210e9. [30] Miller KM, Winton JR, Schulze AD, Purcell MK, Ming TJ. Major histocompatibility complex loci are associated with susceptibility of Atlantic salmon to infectious hematopoietic necrosis virus. Environ Biol Fishes 2004;69:307e16. [31] Jones SRM, Groman DB. Cohabitation transmission of infectious salmon anemia virus among freshwater-reared Atlantic salmon. J Aquat Anim Health 2001;13:340e6. [32] Raynard RS, Snow M, Bruno DW. Experimental infection models and susceptibility of Atlantic salmon Salmo salar to a Scottish isolate of infectious salmon anaemia virus. Dis Aquat Org 2001;47:169e74. [33] Kleinbaum DG. Survival analysis: a self-learning text. New York: Springer-Verlag; 2000. [34] Altman DG. Practical statistics for medical research. London: Chapman and Hall; 2001. [35] Collett D. Modelling survival data in medical research. 1st ed. Boca Raton: Chapman and Hall; 1994. [36] Kjøglum S, Grimholt U, Larsen S. Non-MHC genetic and tank effects influence disease challenge tests in Atlantic salmon (Salmo salar). Aquaculture 2005;250:102e9. [37] McBeath AJA, Collet B, Paley R, Duraffour S, Aspehaug V, Biering E, et al. Identification of an interferon antagonist protein encoded by segment 7 of infectious salmon anaemia virus. Virus Res 2006;115:176e84.
How specific MHC class I and class II combinations affect disease resistance against infectious salmon anaemia in Atlantic salmon (Salmo salar) Sissel Kjøglum a,b,*, Stig Larsen c, Hege G. Bakke b, Unni Grimholt b b
a Aqua Gen AS, P.O. Box 1240, Pirsenteret, 7462 Trondheim, Norway Department of Morphology, Genetics and Aquatic Biology, Norwegian School of Veterinary Science, P.O. Box 8146 Dep., 0033 Oslo, Norway c Department of Production Animal Clinical Sciences, Norwegian School of Veterinary Science, P.O. Box 8146 Dep., 0033 Oslo, Norway
Received 9 December 2005; revised 31 January 2006; accepted 3 February 2006 Available online 17 April 2006
Abstract The aim was to evaluate the performance of selected individual MHC class I and class II alpha (A) alleles, and combinations of these on disease resistance against infectious salmon anaemia (ISA). The material consisting of 1966 fish from seven families, contained five MHC class I alleles and four MHC class II A alleles. Which representing given class II A and class II beta (B) haplotypes, totalling 19 MHC class I and class II A genotypes. The fish were challenged with infectious salmon anaemia virus (ISAV), the virus causing ISA. Dead fish were collected daily during the challenge experiment and the survivors were collected at termination. All fish were genotyped for MHC class I and class II A. The total mortality in the material was 85.14%. For MHC class I, UBA*0201 and UBA*0301 were significantly the most resistant alleles, while UBA*0601 for class I and DAA*0301 for class II A were the significantly most susceptible alleles. The analysis of combined MHC class I and class II A genotypes detected that fish with the genotype UBA*0201/*0301;DAA*0201/*0201 were the most resistant fish with a hazard ratio (HR) at 0.750, while the fish with the genotypes UBA*0601/*0801;DAA*0501/*0501 and UBA*0201/*0301;DAA*0301/*0501 were the most susceptible fish with HR of 1.334 and 1.425. In addition, Cox regression analysis within family detected combined MHC class I and class II A genotypes that contributed significantly to resistance and susceptibility. The study confirmed the expectation of performance of individual MHC class I and class II A alleles, and also detected an effect of MHC class I and class II A in combinations. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Atlantic salmon; Association; Infectious salmon anaemia; Major histocompatibility complex; Resistance; Susceptibility
1. Introduction Infectious salmon anaemia (ISA) is an emerging disease causing severe damage to the farming industry in an increasing number of countries. The disease, first reported in Norway in 1984, has since been reported in * Corresponding author. Tel.: þ47 72 45 05 00; fax: þ47 72 45 05 25. E-mail address: [email protected] (S. Kjøglum). 1050-4648/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.fsi.2006.02.001
432
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
Canada (1996), Scotland (1998), the Faroe Islands and possibly Chile (1999) and the USA (2002) [1e6]. The accumulated mortality in an ISA disease outbreak ranges from 15% to 100%. The disease is caused by the infectious salmon anaemia virus (ISAV), first detected by electron microscopy in 1994 and isolated in cell culture in 1995 [7e9]. ISAV is classified as a member of the Orthomyxoviridae [10]. In Europe, the controlling strategy has been to stamp out fish diagnosed with ISA. Vaccination has not been allowed in Europe, while two commercial vaccines are available in Canada/USA [11e13]. The vaccines have been used during the last two years in some areas of Canada/USA with severe ISA problems, but the results so far have varied and significant protection in the field has not been documented [13]. The ISA problem in the fish farming industry can alternatively be controlled by selection for increased resistance. Resistance towards ISA have been found to be heritable with h2 ¼ 0.13 [14]. But the heritability for ISA was characterised as medium to low compared to other bacterial diseases such as V. salmonicida causing cold water vibriosis and V. anguillarum causing vibriosis [14]. Despite the low heritability, ISA was included in the Aqua Gen breeding scheme in 1995 and challenge tests have been applied to measure the fish’s resistance to the disease [15]. The heritability and the additive genetic variance detected using challenge tests suggest that there is at least one gene that contributes to disease resistance [16]. Genes of major histocompatibility complex (MHC) are candidate genes for disease resistance. MHC are crucial elements of adaptive immunity and have been linked to numerous diseases [17e24]. The MHC contain some of the most polymorphic genes known to date and encode polypeptides which recognise and bind both self and foreign peptides and present them to T cells. Generally, foreign peptides produced by the degradation of intracellular pathogens are bound by MHC class I molecules and presented to cytotoxic T cells. Alternately, foreign peptides derived from extracellular pathogens are bound by MHC class II molecules and presented to helper T cells [25]. The peptide binding ability depends on the peptide binding grooves of MHC molecules [23]. Hence, some pathogens may escape recognition by certain MHC molecules since their peptides are not presentable by the MHC. This can lead to increased susceptibility for pathogens. Alternatively, resistance to pathogens may be derived through high affinity binding of certain peptides by specific MHC alleles. In Atlantic salmon, MHC consists of only one expressed classical MHC class I locus (Sasa-UBA) and one classical class II alpha (A) and class II beta (B) haplotype (SasaDAA/DAB). MHC class I and class II reside on different linkage groups [26]. For MHC class II, Atlantic salmon has distinct class II haplotypes consisting of a combination of unique class II A and class II B alleles [27]. Because of the major expressed classical class I and class II locus, the number of peptide-presenting alleles per animal is limited to two within each class. In Atlantic salmon, genotypes of MHC class I and class II have been found to provide resistance to infectious diseases [28e30]. Associations between MHC class II B and resistance to Aeromonas salmonicida [28], and associations between MHC class I and class II A and resistance against Aeromonas salmonicida and ISAV have previously been demonstrated [29]. MHC class I and class II loci associated with susceptibility to infectious haematopoietic necrosis virus (IHNV) have also been reported [30]. The total project was performed sequentially in two steps. In the first step, published in Grimholt et al. [29], associations between MHC and resistance towards ISA was found in a group of fish which included a large number of families. The study included the total number of 1 031 fish from 25 families, presenting 11 alleles for MHC class I and 7 alleles for MHC class II A. Resistant genotypes for both MHC class I and class II A were found. Compared to the total mortality at 42.68% the resistant genotypes provided mortality at 21.13% for class I UBA*0201/*0301 and 28.23% for class I UBA*0701/*0801, and 31.82% for class II DAA*0201/*0501, 32.14% for class II DAA*0501/*0501 and 38.52% for class II DAA*0201/*0201. However, due to the large number of alleles and high polymorphism in both MHC class I and class II A only three groups were large enough to be included in the study with combinations of MHC class I and class II A genotypes. In the second step, i.e. in this study, genotypes of MHC were produced and evaluated in the next generation with regard to expectations from findings in Grimholt et al. [29]. By sampling a large number of animals per family, the study population would provide groups large enough to be included in the analysis of combined MHC class I and class II A genotypes. The aim was to evaluate the performance of selected individual MHC class I and class II A alleles and combinations of these on resistance against ISA.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
433
2. Material and method 2.1. Experimental design Seven full-sib families were reared at the Aqua Gen breeding station at Kyrksæterøra and kept in separate tanks (Fig. 1). The test fish were randomly sampled from each family six months after first feeding and individually tagged before they were included in the challenge experiment. The virus ISAV was used in the challenge. Dead fish were registered and collected daily.
2.2. Families Based on a previous study [29], sire and dams in the F1 generation were selected as parental fish to produce offspring in F2 (Fig. 1). The selected brood fish in F1 were siblings of the fish included in the previous study. The aim was to produce MHC class I and class II A genotypes with expected high and low mortality. Based on expectations from Grimholt et al. [29], UBA*0201, UBA*0301, UBA*0701, UBA*0801, DAA*0201 and DAA*0501 were the best performing alleles while UBA*0401, DAA*0101, DAA*0401 and DAA*0601 were the worst performing alleles. With these alleles, genotypes of interest would be an allele in combination with different alleles to investigate if the allele is well performing as a single allele or increases in performance with combination to a different allele. The sample of 104 breeding candidates from generation F1 could not provide all genotypes of interest. The chosen matings resulted in seven families with 19 MHC class I and class II A genotypes containing the alleles UBA*0201, UBA*0301, UBA*0501, UBA*0601, UBA*0801, DAA*0201, DAA*0301, DAA*0501 and DAA*0701. Families 1, 2 and 3 were offspring of one sire crossed with three dams, hence they were half-siblings (Fig. 1). The other four families were partly related. The number of fish from each family was chosen depending on the number of MHC genotypes in the family. The large number of fish sampled from each family, provided a sufficient number of fish per given MHC class I and class II A genotype to allow statistical analysis with a significant level of 5% and a power of 90%, and resulted in 46e383 animals per each of the 19 genotypes. The total number of fish included in the ISA challenge experiment was 1966.
Base population x
x
x
x
x
x
x
UBA*0301/*0301 UBA*0301/*0501 UBA*0201/*0201 UBA*0201/*0501 UBA*0601/*0801 UBA*0701/*0801 UBA*0701/*0801 UBA*0701/*0801 UBA*0601/*0901 UBA*0301/*0501 UBA*0201/*0601 UBA*0601/*1201 UBA*0201/*0601 UBA*0501/*0701 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0501 DAA*0201/*0301 DAA*0201/*0501 DAA*0301/*0701 DAA*0501/*0501 DAA*0501/*0501 DAA*0301/*0301 DAA*0201/*0201 DAA*0201/*0301 DAA*0201/*0301 DAA*0201/*0301 DAA*0301/*0601
F1
UBA*0301/*0301 UBA*0301/*0301 UBA*0301/*0301 UBA*0201/*0201 UBA*0201/*0201 UBA*0301/*0801 UBA*0601/*0801 UBA*0801/*0801 UBA*0501/*0601 UBA*0601/*0601 UBA*0501/*0601 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0201 DAA*0201/*0501 DAA*0301/*0701 DAA*0501/*0701 DAA*0501/*0501 DAA*0201/*0301 DAA*0201/*0301 DAA*0201/*0301
F2
Fam 1
Fam 2
Fam 3
Fam 5
Fam 6
Fam 7
Fam 8
Fig. 1. The pedigree, showing the families (Fam) in generation F2 with parents in F1 and grandparents in the base population.
434
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
2.3. Challenge experiment The fish, with a mean weight of 20 g, were transported to the test station, VESO Vikan (http://www.veso.no). At the test station fish were put into one tank with freshwater at 12.7 C, with a diameter of 3 m, and a water flow producing an oxygen level of 6.0e11.0 mgO2/L. The fish were infected using a standardised cohabitation challenge after an acclimatisation period of 33 days [31,32]. The cohabitants were inoculated intraperitoneally with the ISAV isolate Glesvaer/2/90, developed the disease and transmitted ISA to the fish sampled from the seven families. The main outbreak caused by ISA started after 22 days and the study terminated 51 days after infection with a total mortality of 85.14%.
2.4. MHC genotyping The base population (Fig. 1) was thoroughly MHC class I and class II A genotyped using cDNA [26], identifying linkage between the polymorphic markers in the 30 UTRs and expressed MHC class I and class II alleles. The class II marker represents class II A and B haplotypes [27]. Finclippings from a parental material of 104 dams and sires from the F1 population (Fig.1) were collected and used for isolation of genomic DNA using a DNeasy Tissue kit (Qiagen, Germany) according to manufacturer’s protocol. The polymorphic repeats located in the 30 UTRs of Sasa-UBA and Sasa-DAA genes [26,27] were amplified using PCR with primers flanking the repeats. The sense primer was fluorescently labelled in both cases. The resultant fragments were analysed on an automated sequencing machine (ABI377) using the Genescan software (Applied Biosystems, USA). For MHC class II, the class II A DAA*0201 marker represents the DAA*0201;DAB*0201 haplotype, the DAA*0301 marker represents the DAA*0301;DAB*0401 haplotype, the DAA*0501 marker represents the DAA*0501;DAB*0301 haplotype and the DAA*0701 marker represents the DAA*0701;DAB*0501 haplotype. All fish in families 1, 2 and 3 had identical MHC genotypes, hence genotyping was not needed. Genomic DNA from 25 mg muscle tissue from 1583 animals of the F2 population challenged with ISAV were isolated using the Qiagen DNeasy 96 Tissue kit according to standard protocol. MHC class I and class II A genotypes were determined using marker typing with the primer sets and analyses described above. For family 5, both the segregating class I alleles UBA*0301 and UBA*0801 have identical markers. Thus, allele specific primers covering the individual alpha 1 domains were designed. For UBA*0301 the forward primer sequence was 50 ttcccagagtttgtgactgtt while the reverse primer sequence was 50 caatggagacctgagtctga. For UBA*0801 the forward primer sequence was 50 ttcccagagtttgtcaatct while the reverse primer sequence was 50 ctgattgaagcgctgcttgac. Standard PCR conditions using 100 ng genomic DNA with 30 cycles and annealing temperatures of 59 C and 64 C amplified fragments of 148 bp and 204 bp for UBA*0301 and UBA*0801 respectively.
2.5. Statistical methods The results are presented by proportional Hazard Ratio (HR), with 95% confidence intervals (CI) [33]. The time until death is graphically expressed using KaplaneMeier survival plots [33]. Additionally the mortality with 95% confidence interval was calculated using the theory of simple Bernoulli sequences [34]. The hazard ratio compares mortality in the group having the allele or genotype versus the group not having the allele or genotype. The proportional HR was calculated using the proportional hazards regression (PHREG) model of Cox for comparison of two groups; hðt; xÞ ¼ ebx h0 ðtÞ [35], where x ¼ 0 for control and x ¼ 1 for allele or genotype. In the Cox model no assumptions are made about the parametric form of h0(t), which is the base ^ ^ is the maximum likelihood estimates of line for the control group. The hazard ratio eb is estimated by eb, where b the b parameter [35]. The confidence interval for the proportional HR is estimated using the Student procedure ^ exp b ^ za=2 seðbÞ ^ [35]. The PHREG model was also used to detect significant effects for the estimated b, on the survival data, using selection procedures. The significance level for entering the regression model was set to 15%. All comparisons were performed two-tailed and differences were considered significant if the p values were found to be less than or equal to a significance level (a) of 5%.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
435
3. Results 3.1. Single MHC class I analysis The MHC class I alleles UBA*0201 and UBA*0301, both present in four families, performed as the most resistant alleles (Table 1). The alleles UBA*0501, UBA*0601 and UBA*0801, present in two to three families performed as the most susceptible alleles. The group of fish with UBA*0201 or UBA*0301 performed significantly better than the groups not having UBA*0201 or UBA*0301, and the group of fish with UBA*0601 performed significantly worse than the group not having UBA*0601. In the single allele analysis for class I, the differences in mortality for the five alleles were within 10.3%.
3.2. Combined MHC class I analysis Of the six MHC class I allele combinations, two combinations performed significantly. The combination UBA*0201/*0301 present in four families, performed best as a significant resistant genotype (Table 2). The combination UBA*0601/*0801 present in one family performed worst as a significant susceptible genotype (Table 2). The range between the group with highest and lowest mortality in the MHC class I groups was 12.4%.
3.3. Single MHC class II A analysis The MHC class II A alleles DAA*0701 and DAA*0201, present in two and six families, were the most resistant alleles (Table 1). The alleles DAA*0501 and DAA*0301, present in two and three families, were the most susceptible alleles. The only significantly performing class II A allele was DAA*0301. The group of fish with DAA*0301 performed significantly worse than the group not having DAA*0301. In the single allele analysis for class II A, the differences in mortality for the four alleles were within 6.0%.
Table 1 Hazard ratio and mortality in the single allele groups with at least one of the given class I (UBA) or class II A (DAA) alleles present Allele
UBA*0201 UBA*0301 UBA*0801 UBA*0601 UBA*0501 DAA*0701 DAA*0201 DAA*0501 DAA*0301
Hazard ratio
Mortality
(95% CI)
(95% CI)
0.837 (0.761, 0.867 (0.783, 1.025 (0.929, 1.202 (1.089, 1.125 (0.987, 0.992 (0.886, 0.92 (0.834, 1.089 (0.985, 1.125 (1.020,
0.922) 0.959) 1.133) 1.327) 1.283) 1.109) 1.014) 1.204) 1.240)
No. of fish
Present in family
80.76% (78.41, 83.11) 81.40% (78.52, 84.29) 83.24% (80.55, 85.93) 90.82% (88.67, 92.96) 91.10% (87.83, 94.36)
1083
1, 2, 3,5
700
1, 2, 3, 5
741
5, 6
697
6, 7, 8
292
7, 8
82.24% (78.86, 85.63) 83.94% (81.85, 86.03) 85.52% (82.89, 88.16) 88.28% (86.03, 90.53)
490
5, 6
1183
1, 2, 3, 5, 7, 8
684
5, 6
785
5, 7, 8
The results are expressed with 95% confidence intervals in parentheses. The family number names, in which the groups are present, are listed in the last column. Significant HR are presented in bold.
436
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
Table 2 Hazard ratio and mortality in groups with class I (UBA) and class II A (DAA) genotypes Allele
UBA*0201/*0801 UBA*0201/*0301 UBA*0801/*0801 UBA*0601/*0801 UBA*0501/*0601 UBA*0601/*0601 DAA*0201/*0201 DAA*0501/*0701 DAA*0201/*0701 DAA*0301/*0501 DAA*0201/*0301 DAA*0301/*0301 DAA*0501/*0501
Hazard ratio
Mortality
(95% CI)
(95% CI)
0.931 (0.822, 0.867 (0.784, 1.016 (0.863, 1.223 (1.034, 1.123 (0.985, 1.119 (0.970, 0.804 (0.718, 0.964 (0.845, 1.048 (0.882, 1.117 (0.946, 1.100 (0.987, 1.021 (0.824, 1.194 (1.020,
1.054) 0.960) 1.195) 1.195) 1.280) 1.291) 0.899) 1.100) 1.246) 1.319) 1.225) 1.263) 1.399)
No. of fish
Present in family
79.58% (75.54, 83.39) 81.40% (78.52, 85.86) 85.64% (80.63, 89.04) 88.82% (84.09, 92.77) 91.1% (87.83, 90.89) 91.95% (88.43, 95.84)
382
5
699
1, 2, 3, 5
188
6
170
6
292
7, 8
235
7, 8
79.88% (76.37, 83.39) 81.62% (77.38, 85.86) 83.43% (77.83, 89.04) 87.93% (83.09, 92.77) 88.11% (85.28, 90.89) 89.9% (83.96, 95.85) 90.05% (85.66, 94,23)
502
1, 2, 3, 7, 8
321
5, 6
169
5
174
5
512
5, 7, 8
99 189
7, 8 6
The results are expressed with 95% confidence intervals in parentheses. The family number names, in which the groups are present, are listed in the last column. Significant HR are presented in bold.
3.4. Combined MHC class II A analysis Of the seven MHC class II A allele combinations, one combination performed significantly better and one significantly worse (Table 2). The combination DAA*0201/*0201 present in 5 families performed best and was a significant resistant class II genotype (Table 2). The combination DAA*0501/*0501, present in one family performed as a significant susceptible genotype (Table 2). The range between the group with highest and lowest mortality in the MHC class II A groups was 10.2%. 3.5. Combined MHC class I and class II A analysis One combination performed significantly better and two significantly worse (Fig. 2a). The fish with genotype UBA*0201/*0301;DAA*0201/*0201, present in three families performed best and was the significantly most resistant genotype (Table 3). The genotypes performing as the significantly susceptible genotypes were UBA*0601/ *0801;DAA*0501/*0501 and UBA*0201/*0301;DAA*0301/*0501. The range between the group with highest and lowest mortality in MHC class I and class II A groups was 23.7%. 3.6. Family and MHC A significant difference in performance between families was detected (Fig. 2b). Families 1 and 2 were significantly resistant families, while family 7 was a significantly susceptible family. The range between the families with highest and lowest mortality was 22.8%.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
437
(a) 1.0 0.9 0.8
Survival
0.7 0.6 0.5
UBA*0201/*0301;DAA*0201/*0201 UBA*0201/*0301;DAA*0301/*0501
0.4
UBA*0601/*0801;DAA*0501/*0501
0.3 0.2 0.1 0.0 5
10
15
20
25
30
35
40
45
50
30
35
40
45
50
day
(b) 1.0 0.9 0.8
Survival
0.7 0.6
Family 1 Family 2
0.5
Family 3 Family 5 Family 6 Family 7
0.4 0.3
Family 8
0.2 0.1 0.0 5
10
15
20
25
day Fig. 2. The number of days from challenge to death expressed by KaplaneMeier survival plots. Only the groups with significantly increased or reduced mortality in the combined class I and class II alpha (a) are included in the figures. KaplaneMeier survival plots for all families are presented (b).
In family 5 the mortality varied by 23.7% between the MHC genotypes performing best and worst (Table 3). The genotype performing best in family 5 was UBA*0201/*0801; DAA*0501/*0701, while the genotype performing worst was UBA*0201/*0301; DAA*0301/*0501. In the Cox regression analysis, both these MHC genotypes contributed significantly to explaining variation in survival. In family 6 the mortality varied by 9.7% between the MHC genotypes performing best and worst (Table 3). The genotype performing best in family 6 was UBA*0801/*0801; DAA*0501/*0701, while the genotype performing worst was UBA*0601/*0801; DAA*0501/*0501. In the Cox regression analysis the MHC genotype UBA*0601/*0801; DAA*0501/*0501 contributed significantly to explaining variation in survival. Families 7 and 8 contained the same MHC genotypes (Table 3). The mortality varied by 5.6% between the genotypes performing best and worst in this group. The genotype performing best was UBA*0501/*0601; DAA*0201/ *0201, while UBA*0501/*0601; DAA*0201/*0301 performed worst. None of these MHC genotypes contributed significantly in explaining the variation in survival.
438
Table 3 Hazard ratio and mortality in groups with combined MHC class I and class II A allele groups MHC genotype Class I
UBA*0201/*0801
Frequencies in family groups Class II alpha
Fam 1
Fam 2
Survival data MHC Fam 3
DAA*0501/*0701
Fam 5
Fam 6
Fam 7
Fam 8
72
DAA*0201/*0201
UBA*0201/*0301
DAA*0201/*0301
77
UBA*0201/*0801
DAA*0301/*0501
98
UBA*0201/*0801
DAA*0201/*0301
126
UBA*0201/*0801
DAA*0201/*0701
86
UBA*0801/*0801
DAA*0501/*0701
UBA*0201/*0301
DAA*0501/*0701
80
UBA*0201/*0301
DAA*0201/*0701
83
UBA*0601/*0801
DAA*0501/*0701
UBA*0501/*0601
DAA*0201/*0201
UBA*0801/*0801
DAA*0501/*0501
UBA*0501/*0601
DAA*0301/*0301
31
22
UBA*0601/*0601
DAA*0201/*0201
34
30
UBA*0601/*0601
DAA*0301/*0301
25
21
UBA*0601/*0801
DAA*0501/*0501
UBA*0601/*0601
DAA*0201/*0301
64
61
UBA*0501/*0601
DAA*0201/*0301
89
95
UBA*0201/*0301
DAA*0301/*0501
Survival data Families
130
128
125
83
86 26
29
105
84
76
Hazard ratio (95% CI)
0.769 (0.62, 0.94)
0.631 (0.51, 0.78)
1.005 (0.82, 1.22)
1.006 (0.91, 1.11)
1.120 (0.99, 1.27)
1.316 (1.15, 1.50)
0.962 (0.84, 1.11)
Mortality (95% CI)
74.61% (67.1, 82.1)
72.66% (64.9, 80.)
84.00% (77.6, 90.4)
82.80% (80.0, 85.6)
87.15% (83.7, 90.6)
95.48% (92.6, 97.7)
87.60% (83.6, 91.6)
The results are expressed with 95% confidence intervals in parentheses. Significant HR are presented in bold.
Mortality
(95% CI)
(95% CI)
0.769 (0.585, 0.750 (0.661, 0.907 (0.704, 0.918 (0.733, 0.981 (0.803, 1.083 (0.853, 0.928 (0.728, 1.087 (0.851, 1.011 (0.797, 1.107 (0.877, 1.174 (0.881, 1.090 (0.884, 0.962 (0.720, 0.984 (0.757, 1.094 (0.805, 1.334 (1.062, 1.205 (0.998, 1.146 (0.978, 1.425 (1.128,
73.61% (61.90, 83.30) 77.02% (72.81, 81.24) 80.51% (69.91, 88.67) 80.61% (71.39, 87.90) 80.95% (74.10, 87.81) 81.40% (71.55, 88.98) 81.93% (71.95, 89.52) 83.75% (73.82, 91.05) 85.54% (76.11, 92.3) 86.05% (76.89, 92.58) 87.27% (75.52, 94.73) 88.57% (82.49, 94.66) 88.68% (76.97, 95.73) 90.62% (80.70, 96.48) 91.30% (79.21, 97.58) 91.67% (83.58, 96.58) 92.8% (88.27, 97.33) 92.93% (89.23, 96.64) 97.37% (90.82, 99.68)
1.011) 0.851) 1.169) 1.151) 1.198) 1.376) 1.183) 1.387) 1.282) 1.397) 1.565) 1.343) 1.286) 1.279) 1.486) 1.677) 1.455) 1.342) 1.800)
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
UBA*0201/*0301
Hazard ratio
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
439
4. Discussion A main difference between the previous study and the follow-up study is the challenge method used. In the previous study by Grimholt et al. [29], all fish were infected by i.p. injection of ISAV, while in this study a cohabitation method was used to infect the fish. The cohabitation method in ISA challenge was not available in 1998 when the previous challenge test was carried out. But at present this method is used as a standard, because it is assumed to represent a more natural infection that also involves innate defence mechanisms. The different challenge methods could cause some deviation from the expectation of MHC performance, but also resulted in different challenge pressures. The mortality in the challenge experiment using i.p. injection was moderate, while the mortality in the challenge experiment using cohabitations was high. For survival analysis information on more individuals is preferred, and therefore the present study with a high challenge pressure and high mortality is more appropriate for such analysis. The present study detected one genotype of combined MHC class I and class II A alleles, which performed significantly better and were more resistant against ISA. Two MHC class I and class II A genotypes were also found to be significantly more susceptible to ISA. The single allele analysis detected a single allele for MHC class I being more resistant, and one MHC class I allele and one class II A allele was found to be significantly more susceptible. The parental material of 104 breeding candidates provided us with 7 families with sufficient number of animals per selected genotype to enable statistical analysis. These 7 families are not representative for the entire population partly due to limitation in the number of alleles, as some of the most susceptible alleles were not present in the breeding candidates [29]. However, the material provided 19 combined MHC class I and class II A genotypes. In the single allele analysis the results were partly in accordance with expectations. For MHC class I, the alleles UBA*0501 and UBA*0601 were included in the study as the susceptible alleles based on expectations from the previous study [29]. These alleles performed in accordance with expectations, and were also the most susceptible alleles in the present study. The class I allele UBA*0801 was included in the study as the most resistant allele, but did not perform as expected. UBA*0201 and UBA*0301, however, performed as expected both in the single allele analysis as well as in the combined class I analysis. The class II A allele DAA*0301 was the susceptible allele included in the study, and in the single allele analysis this allele performed as expected. The most resistant allele for MHC class II A was expected to be DAA*0501, but did not perform as a resistant allele in this study. Hence, some of the MHC class I and class II A the alleles expected to be resistant did not show resistance in the single allele analysis, but the alleles expected to be susceptible performed in accordance with expectations. The small differences in performance between the alleles in the single allele analysis could cause some of the problems in detecting alleles being most resistant. In addition, the family effect including non-MHC genetic effects influence the results [36]. The alleles were present in a limited number of families, hence other genetic effects in these families could strongly influence performance and thus reduce the observed effect of MHC. Also, in the analysis of combined genotypes only small differences in performance between the genotypes appeared. The difference between the best and worse performing genotypes was 23.7%. However, genotypes contributing significantly to the resistance and susceptibility were detected. The variation between the mortality in families was also small. The present results detected a family effect expressed by significant variation in performance between three of the families. The major part of the differences in prevalence between the families was caused by genetic effects and not due to environmental factors because the fish were challenged in one tank. Minor parts of the variation could be caused by different environment prior to the challenge experiment when the fish were kept in separate family tanks. The genetic effect consisted of MHC, analysed in this study, and non-MHC genetic effects that also affect performance [36]. The material did not give the opportunity to evaluate MHC versus family effect, but analysis applied in four families that contained more than one genotype showed variation in performance of MHC genotypes within the families. Families 5 and 6 did not perform significantly different from expected, but in both families MHC genotypes contributed significantly to resistance and susceptibility and MHC had an effect on resistance. Family 7 was a significantly more susceptible family. In this family no MHC genotypes contributed to resistance, hence in family 7 there was the non-MHC genetic effect that contributed to susceptibility. Resistance towards ISA has been found to be heritable, but with low heritability compared with other viral diseases and also bacterial diseases. This implies that the observed genetic variation for resistance is small, which is in accordance with the small variation found in the present study for both MHC genotypes and families. This might also explain why MHC effects are stronger for furunculosis resistance than for ISA resistance ([28,29]; Kjøglum et al.,
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
440
unpublished). The relatively low MHC influence on ISA resistance could be caused by a viral host battle, where the ISA virus avoids the defence mechanisms of the host. The ISA virus has been shown to down regulate expression of MHC class I when infecting a salmon cell line [37]. In conclusion, the present study confirmed the expectations in performance of selected individual MHC class I and class II A alleles, and combinations of these were found to significantly influence disease resistance against ISA. Acknowledgement This study was funded by Aqua Gen AS and the Norwegian Research Council. References [1] Jarp J, Karlsen E. Infectious salmon anaemia (ISA) risk factors in sea-cultured Atlantic salmon Salmo salar. Dis Aquat Org 1997;28:79e86. [2] Mullins JE, Groman D, Wadowska D. Infectious salmon anaemia in salt water Atlantic salmon (Salmo salar L.) in New Brunswick, Canada. Bull Eur Assoc Fish Pathol 1998;11:110e4. [3] Rodger HD, Turnbull T, Muir F, Millar S, Richards RH. Infectious salmon anaemia (ISA) in the United Kingdom. Bull Eur Assoc Fish Pathol 1998;18:115e6. [4] Kibenge FSB, Ga´rate ON, Johnson G, Arriagada R, Kibenge MJT, Wadowska D. Isolation and identification of infectious salmon anaemia virus (ISAV) from Coho salmon in Chile. Dis Aquat Org 2001;45:9e18. [5] Bouchard DA, Brockway K, Giary C, Keleher W, Merrill PL. First report of infectious salmon anaemia (ISA) in the United States. Bull Eur Assoc Fish Pathol 2001;21:86e8. [6] Murray AG, Smith RJ, Stagg RM. Shipping and the spread of infectious salmon anemia in Scottish aquaculture. Emerging Infect Dis 2002;8:1e5. [7] Hovland T, Nylund A, Watanabe K, Endresen C. Observation of infectious salmon anaemia virus in Atlantic salmon, Salmo salar L. J Fish Dis 1994;17:291e6. [8] Dannevig BH, Falk K, Namork E. Isolation of the causal virus of infectious salmon anaemia (ISA) in long-term cell line from Atlantic salmon head kidney. J Gen Virol 1995;76:1353e9. [9] Nylund A, Hovland T, Watanabe K, Endresen C. Presence of infectious salmon anaemia virus (ISAV) in tissues of Atlantic salmon, Salmo salar L., collected during three separate outbreaks of the disease. J Fish Dis 1995;18:135e45. [10] Falk K, Namork E, Rimstad E, Mjaaland S, Dannevig BH. Characterization of infectious salmon anemia virus, an orthomyxo-like virus isolated from Atlantic salmon (Salmo salar L.). J Virol 1997;71:9016e23. [11] Jones SRM, MacKinnon AM, Salonius K. Vaccination of fresh-water-reared Atlantic salmon reduces mortality associated with infectious salmon anaemia virus. Bull Eur Assoc Fish Pathol 1999;19:98e101. [12] Brown LL, Sperker SA, Clouthier S, Thornton JC. Development of a vaccine against infectious salmon anaemia virus (ISAV). Bulletin of the Aquaculture Association of Canada 2000;100:4e7. [13] Biering E, Villoing S, Sommerset I, Christie KE. Update on viral vaccines for fish. In: Midtlyng P, editor. Progress in fish vaccinology. 1st ed. Basel: S. Krager; 2005. p. 97e113. [14] Gjoen HM, Refstie T, Ulla O, Gjerde B. Genetic correlations between survival of Atlantic salmon in challenge and field tests. Aquaculture 1997;158:277e88. [15] Midtlyng PJ, Storset A, Michel C, Slierendrecht WJ, Okamoto N. Breeding for disease resistance in fish. Bull Eur Assoc Fish Pathol 2002;22:166e72. [16] Falconer DS. Quantitative genetics. 3rd ed. England: Longman Scientific and Technical; 1989. [17] Lamont SJ. The chicken major histocompatibility complex and disease. Rev Off Int Epizoot 1998;17:128e42. [18] Paterson S, Wilson K, Pemberton JM. Major histocompatibility complex variation associated with juvenile survival and parasite resistance in a large unmanaged ungulate population (Ovis aries L.). Proc Natl Acad Sci USA 1998;95:3714e9. [19] Palti Y, Nichols KM, Waller KI, Parsons JE, Thorgaard GH. Association between DNA polymorphisms tightly linked to MHC class II genes and IHN virus resistance in backcrosses of rainbow and cutthroat trout. Aquaculture 2001;194:283e9. [20] Hedrick PN. Pathogen resistance and genetic variation at MHC loci. Evolution 2002;56:1902e8. [21] Arkush KD, Giese AR, Mendonca HL, McBride AM, Marty GD, Hedrick PW. Resistance to three pathogens in the endangered winter-run chinook salmon (Oncorhynchus tshawytscha): effects of inbreeding and major histocompatibility complex genotypes. Can J Fish Aquat Sci 2002;59:966e75. [22] Penn DJ, Damjanovich K, Potts WK. MHC heterozygosity confers a selective advantage against multiple-strain infections. Proc Natl Acad Sci USA 2002;99:11260e4. [23] Bernatchez L, Landry C. MHC studies in nonmodel vertebrates: what have we learned about natural selection in 15 years? J Evol Biol 2003;16:363e77. [24] Wedekind C, Walker M, Portmann J, Cenni B, Mu¨ller R, Binz T. MHC-linked susceptibility to a bacterial infection, but no MHC cryptic female choice in whitefish. J Evol Biol 2004;17:11e8. [25] Marieb EN. Human anatomy and physiology. 6th ed. San Francisco: Pearson Benjamin Cummings; 2004.
S. Kjøglum et al. / Fish & Shellfish Immunology 21 (2006) 431e441
441
[26] Grimholt U, Drablos F, Jorgensen SM, Hoyheim B, Stet RJM. The major histocompatibility class I locus in Atlantic salmon (Salmo salar L.): polymorphism, linkage analysis and protein modelling. Immunogenetics 2002;54:570e81. [27] Stet RJM, de Vries B, Mudde K, Hermsen T, van Heerwaarden J, Shum BP, et al. Unique haplotypes of co-segregating major histocompatibility class II A and class II B alleles in Atlantic salmon (Salmo salar) give rise to diverse class II genotypes. Immunogenetics 2002;54:320e31. [28] Langefors A, Lohm J, Grahn M, Andersen O, von Schantz T. Association between major histocompatibility complex class IIB alleles and resistance to Aeromonas salmonicida in Atlantic salmon. Proc R Soc Lond B Biol Sci 2001;268:479e85. [29] Grimholt U, Larsen S, Nordmo R, Midtlyng P, Kjoeglum S, Storset A, et al. MHC polymorphism and disease resistance in Atlantic salmon (Salmo salar); facing pathogens with single expressed major histocompatibility class I and class II loci. Immunogenetics 2003;55:210e9. [30] Miller KM, Winton JR, Schulze AD, Purcell MK, Ming TJ. Major histocompatibility complex loci are associated with susceptibility of Atlantic salmon to infectious hematopoietic necrosis virus. Environ Biol Fishes 2004;69:307e16. [31] Jones SRM, Groman DB. Cohabitation transmission of infectious salmon anemia virus among freshwater-reared Atlantic salmon. J Aquat Anim Health 2001;13:340e6. [32] Raynard RS, Snow M, Bruno DW. Experimental infection models and susceptibility of Atlantic salmon Salmo salar to a Scottish isolate of infectious salmon anaemia virus. Dis Aquat Org 2001;47:169e74. [33] Kleinbaum DG. Survival analysis: a self-learning text. New York: Springer-Verlag; 2000. [34] Altman DG. Practical statistics for medical research. London: Chapman and Hall; 2001. [35] Collett D. Modelling survival data in medical research. 1st ed. Boca Raton: Chapman and Hall; 1994. [36] Kjøglum S, Grimholt U, Larsen S. Non-MHC genetic and tank effects influence disease challenge tests in Atlantic salmon (Salmo salar). Aquaculture 2005;250:102e9. [37] McBeath AJA, Collet B, Paley R, Duraffour S, Aspehaug V, Biering E, et al. Identification of an interferon antagonist protein encoded by segment 7 of infectious salmon anaemia virus. Virus Res 2006;115:176e84.