- Email: [email protected]
Differences in Major Histocompatibility Complex Frequencies after Multitrait, Divergent Selection for Immunocompetence1
Differences in Major Histocompatibility Complex Frequencies After Multitrait, Divergent Selection for Immunocompetence1 R. P. KEAN Department of Animal Science, Iowa State University, Ames, Iowa 50011 W. E. BRILES
A. CAHANER Department of Genetics, Faculty of Agriculture, The Hebrew University of Jerusalem, P.O.B. 12, Rehovot 76100, Israel A. E. FREEMAN and S. J. LAMONT7 Department of Animal Science, Iowa State University, Ames, Iowa 50011 ABSTRACT White Leghorn chickens from lines selected for four immuneresponse traits (IR lines) were serotyped for B system alloantigens characterizing the haplotypes and genotypes to examine the effect of divergent selection for multitrait immunocompetence on MHC haplotype and genotype frequencies. The selected lines were derived from the Ottawa Strain 7. The selection index included four immunocompetence traits: antibody production against Mycoplasma gallisepticum (MG) and Pasteurella multocida, inflammatory response to phytohemagglutinin, and reticuloendothelial carbon clearance. The four lines include two replicates of high and low mulntrait-immunocompetence lines. After four cycles of selection, significant differences (P < .05) in several B system haplotype frequencies were observed, both among IR lines and between the IR lines and the Ottawa Strain 7. The B2 haplotype frequency was greater in all IR lines than in the Ottawa Strain 7. The B21 frequency was less in both high lines than in the Ottawa Strain 7. In comparisons among lines, frequencies of B21 were greater in both replicates of the low lines and the B12 and B19 frequencies were significantly greater (P < .05) in the high lines. A gene substitution model showed effects (P < .10) of specific haplotypes on MG and on the index. The B2 haplotype had a positive effect associated with MG. Haplotype B21 was positively associated with the multitrait index. Haplotype B13 had a negative effect on both MG and the index. Significant differences (P < .01) in genotype frequencies were also noted among the IR lines. Associations between specific MHC haplotypes or genotypes and immune-response traits may offer insight into MHC-mediated mechanisms of disease resistance. (Key words: chicken, selection, multitrait immune response, immunocompetence, major histocompatibility complex) 1994 Poultry Science 73:7-17
Received for publication June 14, 1993. Accepted for publication September 16, 1993. journal Paper Number J-14989 of the Iowa Agriculture and Home Economics Experiment Sta-
tion, Ames, LA 50011. Project Numbers 2237 and 2351. 2 To whom correspondence should be addressed.
7
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Department of Biological Sciences, Northern Illinois University, DeKalb, Illinois 60115
8
KEAN ET AL.
INTRODUCTION
MATERIALS AND METHODS Chickens
White Leghorn chickens from the Iowa State University IR lines were used. These
I = .522 MG + .546 PM + 1.521 CCA + 1.141 PHA. Determination of B Haplotypes
The alloantigen typing for this project was performed at Northern Illinois University by a tube hemagglutination procedure (Briles and Briles, 1982). Typing was conducted in two groups (Replicate 1 and Replicate 2) when the birds were 4 and 10 wk old, respectively. Blood specimens of a
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Specific haplotypes of the B blood group system in chickens (Briles et al, 1950) have been associated with varied degrees of susceptibility to many diseases (reviewed by Bacon, 1987) and, hence, this complex gene system has been suggested as an appropriate marker for indirect selection for improved disease resistance (van der Zijpp, 1983). Selection for immune response also has been suggested as a method to improve disease resistance (Gavora and Spencer, 1983) and has been carried out in a number of experiments (e.g., Siegel and Gross, 1980; Pevzner et al, 1981; Pitcovski et al, 1987; van der Zijpp and Nieuwland, 1989; Leitner et al, 1992). Each of these experiments was based on some individual facet of immune response. Gavora and Spencer (1978) outlined the merits of selection for general disease resistance rather than resistance to any specific disease agent. The selection program that generated the lines used in studies described in this paper used an index of measures of immunoresponsiveness. The index combines three different facets of immune response (humoral, cellular, and reticuloendothelial) to examine multitrait immunocompetence (Cheng et al, 1991; Kean et al, 1994). The purpose of this project was to determine B haplotype and B genotype frequencies in replicated lines of chickens divergently selected for general immunocompetence and to estimate the gene substitution effect of specific MHC haplotypes. Because of the strong histocompatibility effects of the B system of alloantigens (Schierman and Nordskog, 1961) and the prominent role of the MHC in specific immune responses (Lamont, 1989, 1991, 1993), an association between the B system and general immunocompetence might be expected. An association between specific B haplotypes or genotypes and multitrait immunocompetence also might aid in understanding underlying mechanisms of immune response and disease resistance and identification of beneficial genotypes.
four lines were derived from Ottawa Strain 7. Two replicated pairs of lines were divergently selected for high or low general immunocompetence as previously described (Cheng et al, 1991; Kean et al, 1994). The Ottawa Strain 7 is maintained as a random-mated, pedigreed population at Agriculture Canada, Ottawa, ON, Canada K1A 0C6. Sib matings are avoided and genetic drift is minimized (Gowe and Fairfull, 1980). In the first generation, half-sib groups were divided by index performance in immunocompetence assays, with approximately equal numbers of half-sibs in high (H) and low (L) lines. Progeny of each replicate (divergent line combination) represented 10 different sires. No dams were in common between replicates. In succeeding generations, all matings were within line (1H, 1L, 2H, 2L) with avoidance of sib matings. Index selection was carried out by using three measurements of immune response: 1) antibody production to Mycoplasma gallisepticum (MG) and Pasteurella tnultocida (PM), measured as a ratio to a commercially supplied positive serum; 2) hypersensitivity response to phytohemagglutinin (PHA), in millimeters; and 3) reticuloendothelial clearance of colloidal carbon (CCA), measured in the natural log of absorbance value (at 650 ran) relative to the individual's normal serum. Cheng and Lamont (1988) have previously described these tests. After three generations of index selection, a generation of relaxed selection (nonsib matings within line), then one more generation of selection, all chickens in Generation 6 were included in this study. Details of the selection experiment and direct responses are given elsewhere (Kean et al, 1994). Individual index values for chickens in the generation tested in this study were calculated as
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
sample of Ottawa Strain 7 were taken in 1982, and alloantigen typing was performed at the University of Copenhagen, Denmark (M. Simonsen, Institute for Experimental Immunology, University of Copenhagen, DK-2100, Copenhagen, Denmark). The results of Ottawa Strain 7 typing were reported previously (Gavora et al, 1986). Statistical Analyses
The GLM procedure of SAS® (SAS Institute, 1985) was used to test the effects of MHC genotype on the immunocompetence index and on the four individual immuneresponse traits. Because of the low frequencies of individual MHC genotypes, effects of specific MHC genotypes on immuneresponse traits were not analyzed individually. The model used to analyze overall MHC genotype effect was Yijkim = /x + Rj + Dj + R x Dy + sijk + G] + ejjkta,
where Y, /*, R, D, R x D, s, and e are the same as above, and G is the fixed effect of the 1th MHC genotype. RESULTS
B Haplotype frequencies in the sixth generation of replicated selected lines are reported (Table 1), along with haplotype frequencies previously reported (Gavora et al, 1986) in the Ottawa Strain 7. The means of individuals' breeding values Yijkim = M + Ri + Dj + R x Dij + sijk + frnijH (calculated by using the index) for each + &Oijki +ftsPijki+foqijid+ /Vijid selected line are given in Table 1 and show the differences in composite im+ 06sijkl + e ijklm , munophysiology among the lines. where Y is a dependent trait (Index, MG, Detailed presentation of direct and inPM, CCA, PHA) of the m* individual; /* is direct effects of multitrait, divergent selecthe population mean; R is the fixed effect of tion for immunocompetence are presented the i* replicate (1,2); D is the fixed effect of elsewhere (Kean et al, 1994). the j * selection direction (H, L); s is the Differences in B haplotype frequencies random effect of the k* sire within the ij* exist between the lines of the IR populareplicate and direction; e is a random tion and the Ottawa Strain 7 (Table 1). The residual effect; and j3i... /36 are partial frequency of B2 was greater in lines regression coefficients, which correspond selected for either high or low immune to B2, Bi2, B13, Bis, B™, and m, respectively, response than in the Ottawa Strain 7. The and n . . . s are the number of copies of each frequency of B12 was significantly higher allele present in an individual (absent = 0, in Line 2H and was significantly lower in heterozygote = 1, homozygote = 2). A Line 1L than in the Ottawa Strain 7. generalized inverse was necessary to solve Haplotype B13 was not present in the first these equations. The assumption that all replicate. The frequency of B13 in the gene substitution effects summed to zero second replicate was not significantly was included. The gene substitution effects different from the Ottawa Strain 7. Results of each haplotype refer to the difference in varied by replicate for B15 frequency. In response for each trait associated with Replicate 1, the high-response line had a exchanging one copy of a particular allele significantly greater and the low line had for a hypothetical allele expressing the a significantly lower frequency of B15 than mean effect of all haplotypes tested. Weigel did the Ottawa Strain 7. In the second et al. (1991) suggested including sire effects replicate, however, both lines had less B15 in the model as an adjustment for back- than the Ottawa Strain 7. Haplotype B " ground gene effects. was present in greater frequency in the
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Chickens with B genotypes that were inconsistent with pedigree expectations (parents were also bloodtyped) were removed from all analyses (approximately 2%). Haplotype and MHC genotype frequency differences were analyzed by using 2 x 2 contingency tables (Snedecor and Cochran, 1989). Additive effects of specific MHC haplotypes were analyzed by using a gene substitution model (Batra et al, 1989). The General Linear Models (GLM) procedure of SAS® (SAS Institute, 1985) was used for this analysis. The model used was:
9
10
KEAN ET AL.
TABLE 1. Frequency i}f B haplotypes in Ottawa Strain 7 and all individuals of lines selected for multitrait immune response (with index mean breeding values included) Line 2 Ottawa 7 n = 80 3
B2 B12 B13 B15
.22 d .05 b .08a .14b .24*" .21» .06
B2i Mean breeding value
IH n = 144
IL n = 141
.48b .06b Ob .21* .25* .01 c 0 2.72"
2H n = 135
.65" Ob .04<* .11' .21» .01 2.65"
2L n = 124
.27c .20= .10* .01 d .32* .10b
.32c .03 b .13" .06c .19b .28"
0 3.19b
2.87"
0
a_d
Values within rows with no common superscript differ significantly (P < .05). !From Gavora et al. (1986). 2 Lines are: I H = Replicate 1, high immune response; IL = Replicate 1, low immune response; 2H Replicate 2, high immune response; 2L = Replicate 2, low immune response. % = number of birds serotyped in each group. 4 Y B = undetermined B haplotype(s).
high-response line in Replicate 2 than in cant for haplotypes 6*5 and B2i. The the Ottawa Strain 7. In Replicate 1, the interaction with respect to haplotype Bis frequency of B19 was less in the low- was due to a directional difference, but the response line than in the Ottawa Strain 7. interaction with respect to haplotype B2i The frequency of B2i was significantly was due only to a difference of magnitude lower in both high IR lines than in the between frequency differences. Frequencies were significantly different between Ottawa Strain 7. Differences in haplotype frequencies high and low groups for all haplotypes existed among the IR lines (Table 2). except Bis. When pooled by direction of selection Although differences between replicates were significant (P < .05) for all haplo- (Table 3), several differences in genotype types, the interaction between replicate frequencies are significant. Genotypes and direction of selection was only signifi- B2B12, B2B15, B2B19, B12B19, and Bi?Bi9 were
TABLE 2. Analysis of B haplotype frequencies (high vs low, Replicate 1 vs Replicate 2, interaction of replicate vs direction of selection) in lines of chickens selected for multitrait immune response (with index mean breeding values included) Replicate
Selection direction
Haplotype B2 B12 B13
Bis B» B21
Index mean *P < .05. »»*P < .001.
High
Low
.36 .12 .05 .11 .27
.49 .01 .06 .05 .14
.24 .05 2.76 2.96 (P = .064)
X
1
2
16.8*** 50.3***
.7 15.2*** 29.1*** 74.9***
.56 .03 0 .12 .18
2 .28 .11 .11 .03 .25
.11 .18 2.69 3.03 (P = .002)
X2
85.1*** 29.0*** 65.5*** 30.5*** 7.8* 11.5***
(X2)
.9 2.2 31.2***
1.1 14.1***
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Haplotype
1
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
CN co
• oo o o\
r-i j - 4
• \© T-5 c 4
o •od
11
CM co i \o CN
coovvocsocMinavf'CincMcNinmsooooovo r-l r-l
CO , :
CM 3 " i - l 3 > CO CO c i :
!oooooo!?P
OS <-< CN CO CO ^
CO
C N O O O ^ H C N
CM CN
0
0
0
?o9
0 0 0 0
H N N H O i n m
r H r - l
•^i/5rsit^r-ico
r-t CN r H CM
0
r - I C s j p p p p p
!P
CM CM oo m
o o o o
S! !fi £ !fi fi 2S 0 0
p p p p p p
0
o
GJ .5
•a
CO
CO
E0
U <
< <
ujt^:H_c-4a;£2co'-'S
<
S"-1
CO
CQ
' i ^ ' i ' i ' i . co o co o co o d\ o co o
•
<
E!
<
<
5 co co
CO <
•
l & o l ^ o S & S o o S SS O CO^H
rtOoNMOoO
fefeS!
" H o n ^ S o H o d o o o o
_
«sifta*i-jrttno\rHinij1i-;o\i-;i-!
w M IN M « n rH d . c i n . r H '-''-* «-• "-1 i i c e tt.
upers eplica
£
GO
B.C0J0<
CO
.P. r J
0 S
tl
1 ered
jncotNi-iirivocOTtirHcO'tcscN
!P
t dif low
CM CN. CM
19o'1'1rlo
<" « § "
a 8 a
is
s-
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
N J
rH CO
12
KEAN ET AL.
TABLE 4. Gene substitution effects1 on individual immune response traits and multitrait index in lines of chickens selected for multitrait immune response Trait 2
Haplotype B
2
B12
B» Bis B« B"
MG
SEM
PM
SEM
CCA
SEM
PHA
SEM
INDEX
SEM
.240+ -.002 -.954" .275 .146 .296
.137 .259 .310 .251 .154 .198
.023 .087 .075 -.089 -.068 -.028
.043 .081 .098 .079 .048 .062
-.020 .056 -.011 -.018 -.033 .026
.020 .038 .045 .037 .022 .029
.013 -.024 .070 -.084 -.019 .044
.031 .059 .071 .058 .035 .045
.123 .104 -.394* -.029 -.032 .228+
.087 .165 .197 .160 .098 .126
!Gene substitution effects are the difference in response, in the units of the trait, for each trait associated with exchanging one copy of a particular allele for a hypothetical allele expressing the mean effect of all haplotypes tested. 2 Traits are: MG = Mycoplasma gallisepticum antibody; PM = Pasteurella multocida antibody; CCA = carbon clearance; PHA = wing web response to phytohemagglutinin; Index = multitrait immune response value. +P < .10. *P < .05. **P < .01.
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
all more frequent in the high group and low line than in the high line in Replicate B2B2, B2B21, B19B21, and B21B21 were present 2, but there was no difference in the in greater frequencies in the low group. frequency of B21B2i between the high and Significant interactions were also present low lines in Replicate 2. with respect to the B2B15 and B12B15 Gene substitution effects of specific genotype frequencies (Table 3). haplotypes (Table 4) show the difference A strong replicate effect is apparent in response for each trait, in the units of within these lines (Table 3). Frequencies of that particular trait (MG and PM are ratios B2B2 and B2B^ were greater in Replicate 1 of sample to a positive serum, PHA is in and frequencies of B2Bi2, B^B™, B12B21, millimeters, and CCA ln[absorbance units B13BW, and Bm* were greater in Repli- (650 nm)] relative to the individual's cate 2. normal serum associated with exchanging The B2B2 genotype frequency was one copy of a particular allele for a greatest in Line 1L and was greater in 1H hypothetical allele expressing the mean than in either 2H or 2L (Table 3). Within effect of all haplotypes tested. The B2 each replicate, B2B12 was significantly (P < haplotype was positively associated with .01) greater in the high lines than in the MG. Haplotype B21 was positively aslow lines. For Genotype B2Bis and B2B", sociated with the multitrait index. Haplofrequencies were greater in Line 1H than type B13 was negatively associated with in 1L but no difference was found between both MG and the index. Lines 2H and 2L. The frequency of Frequencies of several B genotypes genotype B2B2i was greater in both low were less than .10 (Table 3). Because of lines than in the high lines. The frequency these low frequencies and the absence of of B^B" was greater in Line 2H than in 2L specific MHC genotypes in some lines, but genotype B12B™ was absent from substitution effects of specific MHC genoReplicate 1. Genotype B15B19 was present types were not analyzed individually. in a greater frequency in Line 1H than in Analysis of variance for the overall effect 1L, although the frequency was not of genotype on the index and individual greater than 10% in any of the lines. Both immune response traits was conducted. 519521 and B21B21 genotypes were present The results of that analysis (Table 5) show in greater frequencies in the low line than that line (which includes replicate, selecthe high line in Replicate 1. The genotype tion direction, and the interaction) and sire B19B21 frequency was also greater in the effects were significant for MG, PM, PHA,
13
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY TABLE 5. Analysis of variance for MHC genotype, line and sire effects for immunocompetence measures (mean squares) Trait1 Effect
MG
PM
CCA
PHA
Index
Line Sire (line) Genotype Error
28.84*** 8.97*** 7.25** 2.88
2.03** .86*** .22 .31
.124 .076 .068 .065
.700** .262* .153 .164
8.28** 2.27** 2.01* 1.21
and the index. The MHC genotype effect was significant for antibody production to MG and for the index. DISCUSSION
The marked differences in haplotype frequencies that have occurred after selection for multitrait immunocompetence may be due to any of several factors or a combination of factors. Founder effect, random genetic drift, and actual association with the selection criteria are all likely to be involved (Falconer, 1989). Certain aspects of the experimental design, namely replication of the divergent selection and the number of parents, were used to minimize the influence of the effects of founder and drift on the study. Changes in gene frequencies may have also occurred in the Ottawa Strain 7 between 1982 when they were sampled for B haplotypes (Gavora et ah, 1986) and 1988 when eggs were imported to serve as the base population for the lines analyzed in the present study. This is not likely, however, because the Ottawa Strain 7 has been maintained using 80 sires and 240 dams with sires and dams selected to represent sires and dams in the previous generations. The B haplotyping of the Ottawa Strain 7 was conducted by using a small sample (n = 80) of the population, which may result in sampling errors, but the present study includes all individuals in the population. Because of the differences in time and involvement of two laboratories associated with obtaining the
B typing, the comparisons of the selected lines and Ottawa Strain 7 gene frequencies should be conservatively interpreted. The base generation of the IR lines had a natural outbreak of Marek's disease, which resulted in approximately 40% mortality (equal in all four lines) despite routine vaccination with herpes virus of turkeys (HVT). Therefore, substantial natural selection for resistance to Marek's disease probably occurred. From the same Ottawa Strain 7 base population, direct experimental selection for Marek's disease resistance resulted in higher frequency of B2 and B2i (Gavora et al.r 1986). The frequency of B2 has increased in all four lines in the current study of selection for multitrait immunocompetence. The B21 frequency in the selected lines, however, is not significantly greater than in the base population. A strong case can be made for a substantial founder effect between replicates. The complete absence of haplotype B13 from Replicate 1 and virtually no difference between its frequency in 2H and 2L (suggesting neutrality of this allele) would especially support this argument. In addition, frequencies are significantly different between replicates for all six haplotypes (Table 2). The difference in mean index values between replicates supports this as well. Using progeny of 10 sires to establish both 1H and 1L and using progeny of 10 different sires to establish 2H and 2L helps to distinguish founder effect from the effects of divergent selection.
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
t r a i t s are: MG = Mycoplasma gallisepticum antibody; PM = Pasteurella multocida antibody; CCA = carbon clearance; PHA = wing web response to phytohemagglutinin; Index = multitrait immune response value. *P < .05. **P < .01. ***P < .0001.
14
KEAN ET AL. 15
Few associations of B haplotype with in vivo phagocytic ability have been noted. Lamont (1986) reported no association between B haplotype and carbon clearance in a partially inbred line of chickens; however, in two inbred congenic lines differing in B haplotype, those with B15 showed significantly greater reticuloendothelial ability than those with B12. In addition, differences in phagocytic ability, measured in vitro, have been associated with B haplotypes. Puzzi et al. (1990) reported greater phagocytosis of sheep erythrocytes by macrophages from lines with B2 than by macrophages from lines with B5. Lines with the B1^ haplotype had greater phagocytic ability than lines with the B21 haplotype in the same study. The striking decrease in B21 frequency (an allele usually thought to be beneficial, at least in specific reference to resistance to Marek's disease) in both high lines, combined with the high frequency of B19 in the high lines, evokes some interesting hypotheses. These points lend credence to the theory (Cohen et al, 1985) that a low immune response may, in some instances, be beneficial to disease resistance, rather than being detrimental. With reference to Marek's disease, it is possible that a decreased cell-mediated immune (CMI) response may be associated with improved resistance to viral transformation, as has been suggested by Calnek et al (1989). Although their results did not show the hypothesized association for all lines tested, several reasons for this discrepancy were given. Notably, the CMI tests in their study were done in vitro, so some important facets of in vivo response may not have been observable. In the populations used in the present study, the natural selection for resistance to Marek's disease may have had special impact on the B2* haplotype. Challenge experiments and molecular analyses of the MHC with the populations of the present study may clarify the status of Marek's resistance in these lines. Some serologically defined B haplotypes have been reported to consist of several distinct RFLP subtypes when examined with molecular probes for B-L and B-F classes of MHC genes (Chausse et al, 1990);
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
The frequencies of only the B haplotype would support an argument for random genetic drift. Haplotype B^ frequency was significantly different, both between high and low pools and between pooled replicates, and the interaction effect was also significant. Frequency differences in the same direction of haplotypes B", B^, and B2* existed in both replicates, offering evidence for a true association with selection for multitrait immune response. Significant changes in similar directions occurred between high and low immune-response groups, within replicate and between pooled replicates, for each of these haplotypes. Additionally, B2 frequencies differed with selection between pooled replicates. This consistent association of gene frequency differences with selection direction in the two replicates that had completely different founder individuals, and despite a large founder effect, strengthens the argument for a true association of MHC with the selection criteria. Although little is known about associations between general, multitrait immunocompetence and the MHC gene complex, many B haplotypes have been associated with specific facets of immune response as well as with actual resistance to specific diseases. One of the best known associations is the increased frequency of B2i in those lines selected for improved resistance to Marek's disease (Bacon, 1987). In addition, Hansen et al (1967) noted increased susceptibility to Marek's disease associated with the B19 haplotype. Briles et al (1980) reported improved resistance to Marek's disease associated with the B21 haplotype within full-sib families, both with natural and artificial exposure. They found haplotypes B2 and B6 to confer moderate resistance and haplotypes B3, B5, B", Bis, and B™ to be associated with increased susceptibility to Marek's disease. Also, B21 has been associated with increased antibody production to SRBC (Martin et al, 1990). Chickens with the B13 haplotype were shown to have greater hypersensitivity to purified protein derivative of tuberculin than those with Bi (Karakoz et al, 1974). No other haplotypes were compared in the tuberculin study, however, so the relative merit of other haplotypes is not known.
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
haplotype B i was very low in both high lines but was not significantly different between the low lines and the Ottawa Strain 7 (Table 1). The positive effect of haplotype B21 might then be postulated to be associated with a genetic background for low immune response. Because the lines were selected for a combination of traits to evaluate general immunocompetence, it can only be speculated which specific facet, or combination of facets, of immune function may be most strongly associated with the observed B haplotype and genotype frequency differences. Associations with selection for antibody to MG (Tables 4 and 5) may have been responsible for changes in the frequencies of some haplotypes. The presence of several different genotypic combinations in low frequencies may simply be due to low selection pressure and few generations of selection. Alternately, it may be because the selection was for multitrait immunocompetence. Biozzi et al. (1979) suggested that one reason for polymorphism at the MHC locus is that some alleles were favorable for resistance to one disease agent or type of agent and other alleles were favorable for others. Heterozygosity at the B locus has been shown to improve economic traits in chickens, including survivor egg production, egg weight, and viability to 40 wk of age (Sato et al, 1992). Thus, the current selection for multiple traits might favor persistent MHC polymorphism. Multiple studies have illustrated that it is possible to alter the population genotype of chickens by multitrait selection for immunocompetence, as evaluated by endogenous viral (ev) genes (Lamont et al, 1992), DNA fingerprinting (Plotsky, Kaiser, and Lamont, unpublished data), and B haplotype and genotype frequencies (this study). The next hypothesis to be tested is that these alterations in genotype and immunophysiology have resulted in altered disease resistance traits.
Dunnington et al. (1992) reported a significant interaction between the effect of the B21B21 genotype and genetic background on antibody production to Brucella abortus bacterin in lines divergently selected for antibody production to SRBC. Chickens with the B2lB21 genotype had a lower mean antibody titer than chickens with B13B21 or B^B™ genotypes in the genetic background selected for high immune response to SRBC but did not have significantly different titers in the genetic background selected for low immune ACKNOWLEDGMENTS response to SRBC. A similar situation may have occurred in the lines in this study. The determination of alloantigen genoHaplotype B21 may have been negatively types was supported by the National associated with genes that produced a Science Foundation under Grant Number high immune response. The frequency of DCB 9020138 (to W. E. Briles). The authors
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
however, Briles et al. (1993) have recently found that B-F and B-L probes may also detect RFLP due to a locus segregating independently of the B complex. Analysis of effects of specific haplotypes on the index and on the individual traits included in the index (Table 4) gives some insight as to which traits may be associated with haplotype differences. Again, the gene substitution effects of each haplotype refer to the difference in response for each trait, in the units of that particular trait, associated with exchanging one copy of a particular allele for a hypothetical allele expressing the mean effect of all haplotypes tested. Using similar methodology in dairy cattle, Weigel et al. (1991) reported effects of several alleles of the bovine lymphocyte antigen (BoLA) on important health and production traits. In the present study, the B13 haplotype was not different between high and low lines in Replicate 2 and was completely absent in Replicate 1, but it was negatively associated with both MG and the index. Haplotype B2 was positively associated (P < .10) with MG but was present in a greater frequency in both low lines than in the high lines. Likewise, haplotype B21 was present in a significantly greater frequency in the low group than in the high group when pooled across replicates but it was positively associated (P < .10) with the index. Further selection may better differentiate these lines so that these associations can be studied in greater detail.
15
2
16
KEAN ET AL.
thank Ruth Briles and Louise Said, Northern Illinois University, DeKalb, IL 60115, for technical assistance in alloantigen typing, and Agriculture Canada, Centre for Food and Animal Research, Ottawa, ON, Canada K1A 0C6 (R. W. Fairfull) for providing the eggs for the base population. REFERENCES
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Bacon, L. D., 1987. Influence of the major histocompatibility complex on disease resistance and productivity. Poultry Sri. 66:802-811. Batra, T. R., A. J. Lee, J. S. Gavora, and M. J. Stear, 1989. Class I alleles of the bovine major histocompatibility system and their association with economic traits. J. Dairy Sci. 72:2115-2124. Biozzi, G., D. Mouton, O. A. Sant Anna, H. C. Passons, M. Gennari, M. H. Reis, V.C.A. Ferreira, A. M. Heumann, Y. Boothillier, O. M. Ibanez, C. Stiffel, and M. Siqueria, 1979. Genetics of immunoresponsiveness to natural antigens in the mouse. Curr. Top. Microbiol. Immunol. 85:31-98. Briles, W. E., and R. W. Briles, 1982. Identification of haplotypes of the chicken major histocompatibility complex (B). Immunogenetics 15:449-459. Briles, W. E., R. W. Briles, W. H. McGibbon, and H. A. Stone, 1980. Identification of B alloalleles associated with resistance to Marek's disease. Pages 395-416 in: Resistance and Immunity to Marek's Disease, CEC Publication EUR 6470, Luxembourg. Briles, W. E., R. M. Goto, C. Auffray, and M. M. Miller, 1993. A polymorphic system related to but genetically independent of the chicken major histocompatibility complex. Immunogenetics 37:408-414. Briles, W. E., W. H. McGibbon, and M. R. Irwin, 1950. On multiple alleles affecting cellular antigens in the chicken. Genetics 35:633-652. Calnek, B. W., D. F. Adene, K. A. Schat, and H. Abplanalp, 1989. Immune response versus susceptibility to Marek's disease. Poultry Sci. 68: 17-26. Chausse, A.-M., P. Thoraval, F. Coudert, C. Auffray, and G. Dambrine, 1990. Analysis of B complex polymorphism in Rous Sarcoma progressor and regressor chickens with B-G, B-V, and B-L/3 probes. Avian Dis. 34:934-940. Cheng, S., and S. J. Lamont, 1988. Genetic analysis of immunocompetence measures in a White Leghorn chicken line. Poultry Sci. 67:989-995. Cheng, S., M. F. Rothschild, and S. J. Lamont, 1991. Estimates of quantitative genetic parameters of immunological traits in the chicken. Poultry Sci. 70:2023-2027. Cohen, I. R., D. M. Altmann, and A. Friedman, 1985. The advantage of being a low responder. Immunol. Today 6:147-148. Dunnington, E. A., C. T. Larsen, W. B. Gross, and P. B. Siegel, 1992. Antibody responses to combinations of antigens in White Leghorn chickens of different background genomes and major
histocompatibility complex genotypes. Poultry Sci. 71:1801-1806. Falconer, D. S., 1989. Introduction to Quantitative Genetics. 3rd ed. John Wiley and Sons, Inc., New York, NY. Gavora, J. S., M. Simonsen, J. L. Spencer, R. W. Fairfull, and R. S. Gowe, 1986. Changes in the frequency of major histocompatibility haplotypes in chickens under selection for both high egg production and resistance to Marek's disease. J. Anim. Breed. Genet. 103:218-226. Gavora, J. S., and J. L. Spencer, 1978. Breeding for genetic resistance to disease: specific or general? World's Poult. Sci. J. 34:137-148. Gavora, J. S., and J. L. Spencer, 1983. Breeding for immune responsiveness and disease resistance. Anim. Blood Groups Biochem. Genet. 14: 159-180. Gowe, R. S., and R. W. Fairfull, 1980. Performance of six long-term multi-trait selected Leghorn strains and three control strains, and a strain cross evaluation of the selected strains. Pages 141-162 in: Proceedings of the South Pacific Poultry Science Convention, Auckland, New Zealand. Hansen, M. P., J. N. Van Zandt, and G.R.J. Law, 1967. Differences in susceptibility to Marek's disease in chickens carrying two different B locus blood group alleles. Poultry Sci. 46:1268.(Abstr.) Karakoz, I., J. Krejci, K. Hala, B. Blaszczyk, T. Hraba, and J. Pekarek, 1974. Genetic determination of tuberculin hypersensitivity in chicken inbred lines. Eur. J. Immunol. 4:545-548. Kean, R. P., A. Cahaner, A. E. Freeman, and S. J. Lamont, 1994. Direct and correlated responses to multitrait, divergent selection for immunocompetence. Poultry Sci. 73:18-32. Lamont, S. J., 1986. Genetic associations of reticuloendothelial activity in chickens. Vol. XI, Pages 643-647 in: Proceedings of Third World Congress of Genetics Applied to Livestock Production. G. E. Dickerson, and R. K. Johnson, ed. University of Nebraska, Lincoln, NE. Lamont, S. J., 1989. The chicken major histocompatibility complex in disease resistance and poultry breeding. J. Dairy Sci. 72:1328-1333. Lamont, S. J., 1991. Immunogenetics and the major histocompatibility complex. Vet. Immunol. Immunopathol. 30:121-127. Lamont, S. J., 1993. The major histocompatibility complex in chickens. Pages 185-203 in: Manipulation of the Avian Genome. R. J. Etches and A.M.V. Gibbins, ed. CRC Press, Boca Raton, FL. Lamont, S. J., Y. Chen, H.J.M. Aarts, M. C. van der Hulst-van Arkel, G. Beuving, and F. R. Leenstra, 1992. Endogenous viral genes in thirteen highly inbred chicken lines and in lines selected for immune response traits. Poultry Sci. 71:530-538. Leitner, G., Z. Uni, A. Cahaner, M. Gutman, and E. D. Heller, 1992. Replicated divergent selection of broiler chickens for high or low early antibody response to Escherichia coli vaccination. Poultry Sci. 71:27-37. Martin, A., E. A. Dunnington, W. B. Gross, W. E. Briles, R. W. Briles, and P. B. Siegel, 1990. Production traits and alloantigen systems in lines of chickens selected for high or low
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
Siegel, P. B., and W. B. Gross, 1980. Production and persistance of antibodies in chickens to sheep erythrocytes 1. Directional selection. Poultry Sci. 59:1-5. Snedecor, G. W., and W. G. Cochran, 1989. Statistical Methods. 7th ed. The Iowa State University Press, Ames, IA. van der Zijpp, A. J., 1983. Breeding for immune responsiveness and disease resistance. World's Poultry Sci. J. 39:118-131. van der Zijpp, A. J., and M.G.B. Nieuwland, 1989. The Biozzi model applied to the chicken. Pages 160-168 in: Series Analytic: Improving Genetic Disease Resistance in Farm Animals. A. J. van der Zijpp and W. Sybesma, ed. Dordrecht, The Netherlands. Weigel, K. A, M. E. Kehrli, Jr., A. E. Freeman, J. R. Thurston, M. J. Stear, and D. H. Kelley, 1991. Association of class I bovine lymphocyte antigen complex alleles with in vitro blood neutrophil functions, lymphocyte blastogenesis, serum complement and conglutinin levels in dairy cattle. Vet. Immunol. Immunopathol. 27: 321-335.
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
antibody responses to sheep erythrocytes. Poultry Sci. 69:871-878. Pevzner, I. Y., H. A. Stone, and A. W. Nordskog, 1981. Immune response and disease resistance in chickens I. Selection for high and low titer to Salmonella pulhrum antigen. Poultry Sci. 60: 920-926. Pitcovski, J., D. E. Heller, A. Cahaner, and B. A. Peleg, 1987. Selection for early responsiveness of chicks to Escherichia coli and Newcastle Disease Virus. Poultry Sci. 66:1276-1282. Puzzi, J. V., L. D. Bacon, and R. R. Dietert, 1990. A gene controlling macrophage functional activation is linked to the chicken B-complex. Arum. Biotechnol. 1:33-45. SAS Institute, 1985. SAS® User's Guide: Statistics. Version 5 Edition. SAS Institute Inc., Cary, NC. Sato, K., H. Abplanalp, D. Napolitano, and J. Reid, 1992. Effects of heterozygosity of major histocompatibility complex haplotypes on performance of Leghorn hens sharing a common inbred background. Poultry Sci. 71:18-26. Schierman, L. W., and A. W. Nordskog, 1961. Relationship of blood type to histocompatibility in chickens. Science 134:1008.
17
A. CAHANER Department of Genetics, Faculty of Agriculture, The Hebrew University of Jerusalem, P.O.B. 12, Rehovot 76100, Israel A. E. FREEMAN and S. J. LAMONT7 Department of Animal Science, Iowa State University, Ames, Iowa 50011 ABSTRACT White Leghorn chickens from lines selected for four immuneresponse traits (IR lines) were serotyped for B system alloantigens characterizing the haplotypes and genotypes to examine the effect of divergent selection for multitrait immunocompetence on MHC haplotype and genotype frequencies. The selected lines were derived from the Ottawa Strain 7. The selection index included four immunocompetence traits: antibody production against Mycoplasma gallisepticum (MG) and Pasteurella multocida, inflammatory response to phytohemagglutinin, and reticuloendothelial carbon clearance. The four lines include two replicates of high and low mulntrait-immunocompetence lines. After four cycles of selection, significant differences (P < .05) in several B system haplotype frequencies were observed, both among IR lines and between the IR lines and the Ottawa Strain 7. The B2 haplotype frequency was greater in all IR lines than in the Ottawa Strain 7. The B21 frequency was less in both high lines than in the Ottawa Strain 7. In comparisons among lines, frequencies of B21 were greater in both replicates of the low lines and the B12 and B19 frequencies were significantly greater (P < .05) in the high lines. A gene substitution model showed effects (P < .10) of specific haplotypes on MG and on the index. The B2 haplotype had a positive effect associated with MG. Haplotype B21 was positively associated with the multitrait index. Haplotype B13 had a negative effect on both MG and the index. Significant differences (P < .01) in genotype frequencies were also noted among the IR lines. Associations between specific MHC haplotypes or genotypes and immune-response traits may offer insight into MHC-mediated mechanisms of disease resistance. (Key words: chicken, selection, multitrait immune response, immunocompetence, major histocompatibility complex) 1994 Poultry Science 73:7-17
Received for publication June 14, 1993. Accepted for publication September 16, 1993. journal Paper Number J-14989 of the Iowa Agriculture and Home Economics Experiment Sta-
tion, Ames, LA 50011. Project Numbers 2237 and 2351. 2 To whom correspondence should be addressed.
7
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Department of Biological Sciences, Northern Illinois University, DeKalb, Illinois 60115
8
KEAN ET AL.
INTRODUCTION
MATERIALS AND METHODS Chickens
White Leghorn chickens from the Iowa State University IR lines were used. These
I = .522 MG + .546 PM + 1.521 CCA + 1.141 PHA. Determination of B Haplotypes
The alloantigen typing for this project was performed at Northern Illinois University by a tube hemagglutination procedure (Briles and Briles, 1982). Typing was conducted in two groups (Replicate 1 and Replicate 2) when the birds were 4 and 10 wk old, respectively. Blood specimens of a
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Specific haplotypes of the B blood group system in chickens (Briles et al, 1950) have been associated with varied degrees of susceptibility to many diseases (reviewed by Bacon, 1987) and, hence, this complex gene system has been suggested as an appropriate marker for indirect selection for improved disease resistance (van der Zijpp, 1983). Selection for immune response also has been suggested as a method to improve disease resistance (Gavora and Spencer, 1983) and has been carried out in a number of experiments (e.g., Siegel and Gross, 1980; Pevzner et al, 1981; Pitcovski et al, 1987; van der Zijpp and Nieuwland, 1989; Leitner et al, 1992). Each of these experiments was based on some individual facet of immune response. Gavora and Spencer (1978) outlined the merits of selection for general disease resistance rather than resistance to any specific disease agent. The selection program that generated the lines used in studies described in this paper used an index of measures of immunoresponsiveness. The index combines three different facets of immune response (humoral, cellular, and reticuloendothelial) to examine multitrait immunocompetence (Cheng et al, 1991; Kean et al, 1994). The purpose of this project was to determine B haplotype and B genotype frequencies in replicated lines of chickens divergently selected for general immunocompetence and to estimate the gene substitution effect of specific MHC haplotypes. Because of the strong histocompatibility effects of the B system of alloantigens (Schierman and Nordskog, 1961) and the prominent role of the MHC in specific immune responses (Lamont, 1989, 1991, 1993), an association between the B system and general immunocompetence might be expected. An association between specific B haplotypes or genotypes and multitrait immunocompetence also might aid in understanding underlying mechanisms of immune response and disease resistance and identification of beneficial genotypes.
four lines were derived from Ottawa Strain 7. Two replicated pairs of lines were divergently selected for high or low general immunocompetence as previously described (Cheng et al, 1991; Kean et al, 1994). The Ottawa Strain 7 is maintained as a random-mated, pedigreed population at Agriculture Canada, Ottawa, ON, Canada K1A 0C6. Sib matings are avoided and genetic drift is minimized (Gowe and Fairfull, 1980). In the first generation, half-sib groups were divided by index performance in immunocompetence assays, with approximately equal numbers of half-sibs in high (H) and low (L) lines. Progeny of each replicate (divergent line combination) represented 10 different sires. No dams were in common between replicates. In succeeding generations, all matings were within line (1H, 1L, 2H, 2L) with avoidance of sib matings. Index selection was carried out by using three measurements of immune response: 1) antibody production to Mycoplasma gallisepticum (MG) and Pasteurella tnultocida (PM), measured as a ratio to a commercially supplied positive serum; 2) hypersensitivity response to phytohemagglutinin (PHA), in millimeters; and 3) reticuloendothelial clearance of colloidal carbon (CCA), measured in the natural log of absorbance value (at 650 ran) relative to the individual's normal serum. Cheng and Lamont (1988) have previously described these tests. After three generations of index selection, a generation of relaxed selection (nonsib matings within line), then one more generation of selection, all chickens in Generation 6 were included in this study. Details of the selection experiment and direct responses are given elsewhere (Kean et al, 1994). Individual index values for chickens in the generation tested in this study were calculated as
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
sample of Ottawa Strain 7 were taken in 1982, and alloantigen typing was performed at the University of Copenhagen, Denmark (M. Simonsen, Institute for Experimental Immunology, University of Copenhagen, DK-2100, Copenhagen, Denmark). The results of Ottawa Strain 7 typing were reported previously (Gavora et al, 1986). Statistical Analyses
The GLM procedure of SAS® (SAS Institute, 1985) was used to test the effects of MHC genotype on the immunocompetence index and on the four individual immuneresponse traits. Because of the low frequencies of individual MHC genotypes, effects of specific MHC genotypes on immuneresponse traits were not analyzed individually. The model used to analyze overall MHC genotype effect was Yijkim = /x + Rj + Dj + R x Dy + sijk + G] + ejjkta,
where Y, /*, R, D, R x D, s, and e are the same as above, and G is the fixed effect of the 1th MHC genotype. RESULTS
B Haplotype frequencies in the sixth generation of replicated selected lines are reported (Table 1), along with haplotype frequencies previously reported (Gavora et al, 1986) in the Ottawa Strain 7. The means of individuals' breeding values Yijkim = M + Ri + Dj + R x Dij + sijk + frnijH (calculated by using the index) for each + &Oijki +ftsPijki+foqijid+ /Vijid selected line are given in Table 1 and show the differences in composite im+ 06sijkl + e ijklm , munophysiology among the lines. where Y is a dependent trait (Index, MG, Detailed presentation of direct and inPM, CCA, PHA) of the m* individual; /* is direct effects of multitrait, divergent selecthe population mean; R is the fixed effect of tion for immunocompetence are presented the i* replicate (1,2); D is the fixed effect of elsewhere (Kean et al, 1994). the j * selection direction (H, L); s is the Differences in B haplotype frequencies random effect of the k* sire within the ij* exist between the lines of the IR populareplicate and direction; e is a random tion and the Ottawa Strain 7 (Table 1). The residual effect; and j3i... /36 are partial frequency of B2 was greater in lines regression coefficients, which correspond selected for either high or low immune to B2, Bi2, B13, Bis, B™, and m, respectively, response than in the Ottawa Strain 7. The and n . . . s are the number of copies of each frequency of B12 was significantly higher allele present in an individual (absent = 0, in Line 2H and was significantly lower in heterozygote = 1, homozygote = 2). A Line 1L than in the Ottawa Strain 7. generalized inverse was necessary to solve Haplotype B13 was not present in the first these equations. The assumption that all replicate. The frequency of B13 in the gene substitution effects summed to zero second replicate was not significantly was included. The gene substitution effects different from the Ottawa Strain 7. Results of each haplotype refer to the difference in varied by replicate for B15 frequency. In response for each trait associated with Replicate 1, the high-response line had a exchanging one copy of a particular allele significantly greater and the low line had for a hypothetical allele expressing the a significantly lower frequency of B15 than mean effect of all haplotypes tested. Weigel did the Ottawa Strain 7. In the second et al. (1991) suggested including sire effects replicate, however, both lines had less B15 in the model as an adjustment for back- than the Ottawa Strain 7. Haplotype B " ground gene effects. was present in greater frequency in the
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Chickens with B genotypes that were inconsistent with pedigree expectations (parents were also bloodtyped) were removed from all analyses (approximately 2%). Haplotype and MHC genotype frequency differences were analyzed by using 2 x 2 contingency tables (Snedecor and Cochran, 1989). Additive effects of specific MHC haplotypes were analyzed by using a gene substitution model (Batra et al, 1989). The General Linear Models (GLM) procedure of SAS® (SAS Institute, 1985) was used for this analysis. The model used was:
9
10
KEAN ET AL.
TABLE 1. Frequency i}f B haplotypes in Ottawa Strain 7 and all individuals of lines selected for multitrait immune response (with index mean breeding values included) Line 2 Ottawa 7 n = 80 3
B2 B12 B13 B15
.22 d .05 b .08a .14b .24*" .21» .06
B2i Mean breeding value
IH n = 144
IL n = 141
.48b .06b Ob .21* .25* .01 c 0 2.72"
2H n = 135
.65" Ob .04<* .11' .21» .01 2.65"
2L n = 124
.27c .20= .10* .01 d .32* .10b
.32c .03 b .13" .06c .19b .28"
0 3.19b
2.87"
0
a_d
Values within rows with no common superscript differ significantly (P < .05). !From Gavora et al. (1986). 2 Lines are: I H = Replicate 1, high immune response; IL = Replicate 1, low immune response; 2H Replicate 2, high immune response; 2L = Replicate 2, low immune response. % = number of birds serotyped in each group. 4 Y B = undetermined B haplotype(s).
high-response line in Replicate 2 than in cant for haplotypes 6*5 and B2i. The the Ottawa Strain 7. In Replicate 1, the interaction with respect to haplotype Bis frequency of B19 was less in the low- was due to a directional difference, but the response line than in the Ottawa Strain 7. interaction with respect to haplotype B2i The frequency of B2i was significantly was due only to a difference of magnitude lower in both high IR lines than in the between frequency differences. Frequencies were significantly different between Ottawa Strain 7. Differences in haplotype frequencies high and low groups for all haplotypes existed among the IR lines (Table 2). except Bis. When pooled by direction of selection Although differences between replicates were significant (P < .05) for all haplo- (Table 3), several differences in genotype types, the interaction between replicate frequencies are significant. Genotypes and direction of selection was only signifi- B2B12, B2B15, B2B19, B12B19, and Bi?Bi9 were
TABLE 2. Analysis of B haplotype frequencies (high vs low, Replicate 1 vs Replicate 2, interaction of replicate vs direction of selection) in lines of chickens selected for multitrait immune response (with index mean breeding values included) Replicate
Selection direction
Haplotype B2 B12 B13
Bis B» B21
Index mean *P < .05. »»*P < .001.
High
Low
.36 .12 .05 .11 .27
.49 .01 .06 .05 .14
.24 .05 2.76 2.96 (P = .064)
X
1
2
16.8*** 50.3***
.7 15.2*** 29.1*** 74.9***
.56 .03 0 .12 .18
2 .28 .11 .11 .03 .25
.11 .18 2.69 3.03 (P = .002)
X2
85.1*** 29.0*** 65.5*** 30.5*** 7.8* 11.5***
(X2)
.9 2.2 31.2***
1.1 14.1***
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Haplotype
1
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
CN co
• oo o o\
r-i j - 4
• \© T-5 c 4
o •od
11
CM co i \o CN
coovvocsocMinavf'CincMcNinmsooooovo r-l r-l
CO , :
CM 3 " i - l 3 > CO CO c i :
!oooooo!?P
OS <-< CN CO CO ^
CO
C N O O O ^ H C N
CM CN
0
0
0
?o9
0 0 0 0
H N N H O i n m
r H r - l
•^i/5rsit^r-ico
r-t CN r H CM
0
r - I C s j p p p p p
!P
CM CM oo m
o o o o
S! !fi £ !fi fi 2S 0 0
p p p p p p
0
o
GJ .5
•a
CO
CO
E0
U <
< <
ujt^:H_c-4a;£2co'-'S
<
S"-1
CO
CQ
' i ^ ' i ' i ' i . co o co o co o d\ o co o
•
<
E!
<
<
5 co co
CO <
•
l & o l ^ o S & S o o S SS O CO^H
rtOoNMOoO
fefeS!
" H o n ^ S o H o d o o o o
_
«sifta*i-jrttno\rHinij1i-;o\i-;i-!
w M IN M « n rH d . c i n . r H '-''-* «-• "-1 i i c e tt.
upers eplica
£
GO
B.C0J0<
CO
.P. r J
0 S
tl
1 ered
jncotNi-iirivocOTtirHcO'tcscN
!P
t dif low
CM CN. CM
19o'1'1rlo
<" « § "
a 8 a
is
s-
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
N J
rH CO
12
KEAN ET AL.
TABLE 4. Gene substitution effects1 on individual immune response traits and multitrait index in lines of chickens selected for multitrait immune response Trait 2
Haplotype B
2
B12
B» Bis B« B"
MG
SEM
PM
SEM
CCA
SEM
PHA
SEM
INDEX
SEM
.240+ -.002 -.954" .275 .146 .296
.137 .259 .310 .251 .154 .198
.023 .087 .075 -.089 -.068 -.028
.043 .081 .098 .079 .048 .062
-.020 .056 -.011 -.018 -.033 .026
.020 .038 .045 .037 .022 .029
.013 -.024 .070 -.084 -.019 .044
.031 .059 .071 .058 .035 .045
.123 .104 -.394* -.029 -.032 .228+
.087 .165 .197 .160 .098 .126
!Gene substitution effects are the difference in response, in the units of the trait, for each trait associated with exchanging one copy of a particular allele for a hypothetical allele expressing the mean effect of all haplotypes tested. 2 Traits are: MG = Mycoplasma gallisepticum antibody; PM = Pasteurella multocida antibody; CCA = carbon clearance; PHA = wing web response to phytohemagglutinin; Index = multitrait immune response value. +P < .10. *P < .05. **P < .01.
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
all more frequent in the high group and low line than in the high line in Replicate B2B2, B2B21, B19B21, and B21B21 were present 2, but there was no difference in the in greater frequencies in the low group. frequency of B21B2i between the high and Significant interactions were also present low lines in Replicate 2. with respect to the B2B15 and B12B15 Gene substitution effects of specific genotype frequencies (Table 3). haplotypes (Table 4) show the difference A strong replicate effect is apparent in response for each trait, in the units of within these lines (Table 3). Frequencies of that particular trait (MG and PM are ratios B2B2 and B2B^ were greater in Replicate 1 of sample to a positive serum, PHA is in and frequencies of B2Bi2, B^B™, B12B21, millimeters, and CCA ln[absorbance units B13BW, and Bm* were greater in Repli- (650 nm)] relative to the individual's cate 2. normal serum associated with exchanging The B2B2 genotype frequency was one copy of a particular allele for a greatest in Line 1L and was greater in 1H hypothetical allele expressing the mean than in either 2H or 2L (Table 3). Within effect of all haplotypes tested. The B2 each replicate, B2B12 was significantly (P < haplotype was positively associated with .01) greater in the high lines than in the MG. Haplotype B21 was positively aslow lines. For Genotype B2Bis and B2B", sociated with the multitrait index. Haplofrequencies were greater in Line 1H than type B13 was negatively associated with in 1L but no difference was found between both MG and the index. Lines 2H and 2L. The frequency of Frequencies of several B genotypes genotype B2B2i was greater in both low were less than .10 (Table 3). Because of lines than in the high lines. The frequency these low frequencies and the absence of of B^B" was greater in Line 2H than in 2L specific MHC genotypes in some lines, but genotype B12B™ was absent from substitution effects of specific MHC genoReplicate 1. Genotype B15B19 was present types were not analyzed individually. in a greater frequency in Line 1H than in Analysis of variance for the overall effect 1L, although the frequency was not of genotype on the index and individual greater than 10% in any of the lines. Both immune response traits was conducted. 519521 and B21B21 genotypes were present The results of that analysis (Table 5) show in greater frequencies in the low line than that line (which includes replicate, selecthe high line in Replicate 1. The genotype tion direction, and the interaction) and sire B19B21 frequency was also greater in the effects were significant for MG, PM, PHA,
13
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY TABLE 5. Analysis of variance for MHC genotype, line and sire effects for immunocompetence measures (mean squares) Trait1 Effect
MG
PM
CCA
PHA
Index
Line Sire (line) Genotype Error
28.84*** 8.97*** 7.25** 2.88
2.03** .86*** .22 .31
.124 .076 .068 .065
.700** .262* .153 .164
8.28** 2.27** 2.01* 1.21
and the index. The MHC genotype effect was significant for antibody production to MG and for the index. DISCUSSION
The marked differences in haplotype frequencies that have occurred after selection for multitrait immunocompetence may be due to any of several factors or a combination of factors. Founder effect, random genetic drift, and actual association with the selection criteria are all likely to be involved (Falconer, 1989). Certain aspects of the experimental design, namely replication of the divergent selection and the number of parents, were used to minimize the influence of the effects of founder and drift on the study. Changes in gene frequencies may have also occurred in the Ottawa Strain 7 between 1982 when they were sampled for B haplotypes (Gavora et ah, 1986) and 1988 when eggs were imported to serve as the base population for the lines analyzed in the present study. This is not likely, however, because the Ottawa Strain 7 has been maintained using 80 sires and 240 dams with sires and dams selected to represent sires and dams in the previous generations. The B haplotyping of the Ottawa Strain 7 was conducted by using a small sample (n = 80) of the population, which may result in sampling errors, but the present study includes all individuals in the population. Because of the differences in time and involvement of two laboratories associated with obtaining the
B typing, the comparisons of the selected lines and Ottawa Strain 7 gene frequencies should be conservatively interpreted. The base generation of the IR lines had a natural outbreak of Marek's disease, which resulted in approximately 40% mortality (equal in all four lines) despite routine vaccination with herpes virus of turkeys (HVT). Therefore, substantial natural selection for resistance to Marek's disease probably occurred. From the same Ottawa Strain 7 base population, direct experimental selection for Marek's disease resistance resulted in higher frequency of B2 and B2i (Gavora et al.r 1986). The frequency of B2 has increased in all four lines in the current study of selection for multitrait immunocompetence. The B21 frequency in the selected lines, however, is not significantly greater than in the base population. A strong case can be made for a substantial founder effect between replicates. The complete absence of haplotype B13 from Replicate 1 and virtually no difference between its frequency in 2H and 2L (suggesting neutrality of this allele) would especially support this argument. In addition, frequencies are significantly different between replicates for all six haplotypes (Table 2). The difference in mean index values between replicates supports this as well. Using progeny of 10 sires to establish both 1H and 1L and using progeny of 10 different sires to establish 2H and 2L helps to distinguish founder effect from the effects of divergent selection.
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
t r a i t s are: MG = Mycoplasma gallisepticum antibody; PM = Pasteurella multocida antibody; CCA = carbon clearance; PHA = wing web response to phytohemagglutinin; Index = multitrait immune response value. *P < .05. **P < .01. ***P < .0001.
14
KEAN ET AL. 15
Few associations of B haplotype with in vivo phagocytic ability have been noted. Lamont (1986) reported no association between B haplotype and carbon clearance in a partially inbred line of chickens; however, in two inbred congenic lines differing in B haplotype, those with B15 showed significantly greater reticuloendothelial ability than those with B12. In addition, differences in phagocytic ability, measured in vitro, have been associated with B haplotypes. Puzzi et al. (1990) reported greater phagocytosis of sheep erythrocytes by macrophages from lines with B2 than by macrophages from lines with B5. Lines with the B1^ haplotype had greater phagocytic ability than lines with the B21 haplotype in the same study. The striking decrease in B21 frequency (an allele usually thought to be beneficial, at least in specific reference to resistance to Marek's disease) in both high lines, combined with the high frequency of B19 in the high lines, evokes some interesting hypotheses. These points lend credence to the theory (Cohen et al, 1985) that a low immune response may, in some instances, be beneficial to disease resistance, rather than being detrimental. With reference to Marek's disease, it is possible that a decreased cell-mediated immune (CMI) response may be associated with improved resistance to viral transformation, as has been suggested by Calnek et al (1989). Although their results did not show the hypothesized association for all lines tested, several reasons for this discrepancy were given. Notably, the CMI tests in their study were done in vitro, so some important facets of in vivo response may not have been observable. In the populations used in the present study, the natural selection for resistance to Marek's disease may have had special impact on the B2* haplotype. Challenge experiments and molecular analyses of the MHC with the populations of the present study may clarify the status of Marek's resistance in these lines. Some serologically defined B haplotypes have been reported to consist of several distinct RFLP subtypes when examined with molecular probes for B-L and B-F classes of MHC genes (Chausse et al, 1990);
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
The frequencies of only the B haplotype would support an argument for random genetic drift. Haplotype B^ frequency was significantly different, both between high and low pools and between pooled replicates, and the interaction effect was also significant. Frequency differences in the same direction of haplotypes B", B^, and B2* existed in both replicates, offering evidence for a true association with selection for multitrait immune response. Significant changes in similar directions occurred between high and low immune-response groups, within replicate and between pooled replicates, for each of these haplotypes. Additionally, B2 frequencies differed with selection between pooled replicates. This consistent association of gene frequency differences with selection direction in the two replicates that had completely different founder individuals, and despite a large founder effect, strengthens the argument for a true association of MHC with the selection criteria. Although little is known about associations between general, multitrait immunocompetence and the MHC gene complex, many B haplotypes have been associated with specific facets of immune response as well as with actual resistance to specific diseases. One of the best known associations is the increased frequency of B2i in those lines selected for improved resistance to Marek's disease (Bacon, 1987). In addition, Hansen et al (1967) noted increased susceptibility to Marek's disease associated with the B19 haplotype. Briles et al (1980) reported improved resistance to Marek's disease associated with the B21 haplotype within full-sib families, both with natural and artificial exposure. They found haplotypes B2 and B6 to confer moderate resistance and haplotypes B3, B5, B", Bis, and B™ to be associated with increased susceptibility to Marek's disease. Also, B21 has been associated with increased antibody production to SRBC (Martin et al, 1990). Chickens with the B13 haplotype were shown to have greater hypersensitivity to purified protein derivative of tuberculin than those with Bi (Karakoz et al, 1974). No other haplotypes were compared in the tuberculin study, however, so the relative merit of other haplotypes is not known.
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
haplotype B i was very low in both high lines but was not significantly different between the low lines and the Ottawa Strain 7 (Table 1). The positive effect of haplotype B21 might then be postulated to be associated with a genetic background for low immune response. Because the lines were selected for a combination of traits to evaluate general immunocompetence, it can only be speculated which specific facet, or combination of facets, of immune function may be most strongly associated with the observed B haplotype and genotype frequency differences. Associations with selection for antibody to MG (Tables 4 and 5) may have been responsible for changes in the frequencies of some haplotypes. The presence of several different genotypic combinations in low frequencies may simply be due to low selection pressure and few generations of selection. Alternately, it may be because the selection was for multitrait immunocompetence. Biozzi et al. (1979) suggested that one reason for polymorphism at the MHC locus is that some alleles were favorable for resistance to one disease agent or type of agent and other alleles were favorable for others. Heterozygosity at the B locus has been shown to improve economic traits in chickens, including survivor egg production, egg weight, and viability to 40 wk of age (Sato et al, 1992). Thus, the current selection for multiple traits might favor persistent MHC polymorphism. Multiple studies have illustrated that it is possible to alter the population genotype of chickens by multitrait selection for immunocompetence, as evaluated by endogenous viral (ev) genes (Lamont et al, 1992), DNA fingerprinting (Plotsky, Kaiser, and Lamont, unpublished data), and B haplotype and genotype frequencies (this study). The next hypothesis to be tested is that these alterations in genotype and immunophysiology have resulted in altered disease resistance traits.
Dunnington et al. (1992) reported a significant interaction between the effect of the B21B21 genotype and genetic background on antibody production to Brucella abortus bacterin in lines divergently selected for antibody production to SRBC. Chickens with the B2lB21 genotype had a lower mean antibody titer than chickens with B13B21 or B^B™ genotypes in the genetic background selected for high immune response to SRBC but did not have significantly different titers in the genetic background selected for low immune ACKNOWLEDGMENTS response to SRBC. A similar situation may have occurred in the lines in this study. The determination of alloantigen genoHaplotype B21 may have been negatively types was supported by the National associated with genes that produced a Science Foundation under Grant Number high immune response. The frequency of DCB 9020138 (to W. E. Briles). The authors
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
however, Briles et al. (1993) have recently found that B-F and B-L probes may also detect RFLP due to a locus segregating independently of the B complex. Analysis of effects of specific haplotypes on the index and on the individual traits included in the index (Table 4) gives some insight as to which traits may be associated with haplotype differences. Again, the gene substitution effects of each haplotype refer to the difference in response for each trait, in the units of that particular trait, associated with exchanging one copy of a particular allele for a hypothetical allele expressing the mean effect of all haplotypes tested. Using similar methodology in dairy cattle, Weigel et al. (1991) reported effects of several alleles of the bovine lymphocyte antigen (BoLA) on important health and production traits. In the present study, the B13 haplotype was not different between high and low lines in Replicate 2 and was completely absent in Replicate 1, but it was negatively associated with both MG and the index. Haplotype B2 was positively associated (P < .10) with MG but was present in a greater frequency in both low lines than in the high lines. Likewise, haplotype B21 was present in a significantly greater frequency in the low group than in the high group when pooled across replicates but it was positively associated (P < .10) with the index. Further selection may better differentiate these lines so that these associations can be studied in greater detail.
15
2
16
KEAN ET AL.
thank Ruth Briles and Louise Said, Northern Illinois University, DeKalb, IL 60115, for technical assistance in alloantigen typing, and Agriculture Canada, Centre for Food and Animal Research, Ottawa, ON, Canada K1A 0C6 (R. W. Fairfull) for providing the eggs for the base population. REFERENCES
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
Bacon, L. D., 1987. Influence of the major histocompatibility complex on disease resistance and productivity. Poultry Sri. 66:802-811. Batra, T. R., A. J. Lee, J. S. Gavora, and M. J. Stear, 1989. Class I alleles of the bovine major histocompatibility system and their association with economic traits. J. Dairy Sci. 72:2115-2124. Biozzi, G., D. Mouton, O. A. Sant Anna, H. C. Passons, M. Gennari, M. H. Reis, V.C.A. Ferreira, A. M. Heumann, Y. Boothillier, O. M. Ibanez, C. Stiffel, and M. Siqueria, 1979. Genetics of immunoresponsiveness to natural antigens in the mouse. Curr. Top. Microbiol. Immunol. 85:31-98. Briles, W. E., and R. W. Briles, 1982. Identification of haplotypes of the chicken major histocompatibility complex (B). Immunogenetics 15:449-459. Briles, W. E., R. W. Briles, W. H. McGibbon, and H. A. Stone, 1980. Identification of B alloalleles associated with resistance to Marek's disease. Pages 395-416 in: Resistance and Immunity to Marek's Disease, CEC Publication EUR 6470, Luxembourg. Briles, W. E., R. M. Goto, C. Auffray, and M. M. Miller, 1993. A polymorphic system related to but genetically independent of the chicken major histocompatibility complex. Immunogenetics 37:408-414. Briles, W. E., W. H. McGibbon, and M. R. Irwin, 1950. On multiple alleles affecting cellular antigens in the chicken. Genetics 35:633-652. Calnek, B. W., D. F. Adene, K. A. Schat, and H. Abplanalp, 1989. Immune response versus susceptibility to Marek's disease. Poultry Sci. 68: 17-26. Chausse, A.-M., P. Thoraval, F. Coudert, C. Auffray, and G. Dambrine, 1990. Analysis of B complex polymorphism in Rous Sarcoma progressor and regressor chickens with B-G, B-V, and B-L/3 probes. Avian Dis. 34:934-940. Cheng, S., and S. J. Lamont, 1988. Genetic analysis of immunocompetence measures in a White Leghorn chicken line. Poultry Sci. 67:989-995. Cheng, S., M. F. Rothschild, and S. J. Lamont, 1991. Estimates of quantitative genetic parameters of immunological traits in the chicken. Poultry Sci. 70:2023-2027. Cohen, I. R., D. M. Altmann, and A. Friedman, 1985. The advantage of being a low responder. Immunol. Today 6:147-148. Dunnington, E. A., C. T. Larsen, W. B. Gross, and P. B. Siegel, 1992. Antibody responses to combinations of antigens in White Leghorn chickens of different background genomes and major
histocompatibility complex genotypes. Poultry Sci. 71:1801-1806. Falconer, D. S., 1989. Introduction to Quantitative Genetics. 3rd ed. John Wiley and Sons, Inc., New York, NY. Gavora, J. S., M. Simonsen, J. L. Spencer, R. W. Fairfull, and R. S. Gowe, 1986. Changes in the frequency of major histocompatibility haplotypes in chickens under selection for both high egg production and resistance to Marek's disease. J. Anim. Breed. Genet. 103:218-226. Gavora, J. S., and J. L. Spencer, 1978. Breeding for genetic resistance to disease: specific or general? World's Poult. Sci. J. 34:137-148. Gavora, J. S., and J. L. Spencer, 1983. Breeding for immune responsiveness and disease resistance. Anim. Blood Groups Biochem. Genet. 14: 159-180. Gowe, R. S., and R. W. Fairfull, 1980. Performance of six long-term multi-trait selected Leghorn strains and three control strains, and a strain cross evaluation of the selected strains. Pages 141-162 in: Proceedings of the South Pacific Poultry Science Convention, Auckland, New Zealand. Hansen, M. P., J. N. Van Zandt, and G.R.J. Law, 1967. Differences in susceptibility to Marek's disease in chickens carrying two different B locus blood group alleles. Poultry Sci. 46:1268.(Abstr.) Karakoz, I., J. Krejci, K. Hala, B. Blaszczyk, T. Hraba, and J. Pekarek, 1974. Genetic determination of tuberculin hypersensitivity in chicken inbred lines. Eur. J. Immunol. 4:545-548. Kean, R. P., A. Cahaner, A. E. Freeman, and S. J. Lamont, 1994. Direct and correlated responses to multitrait, divergent selection for immunocompetence. Poultry Sci. 73:18-32. Lamont, S. J., 1986. Genetic associations of reticuloendothelial activity in chickens. Vol. XI, Pages 643-647 in: Proceedings of Third World Congress of Genetics Applied to Livestock Production. G. E. Dickerson, and R. K. Johnson, ed. University of Nebraska, Lincoln, NE. Lamont, S. J., 1989. The chicken major histocompatibility complex in disease resistance and poultry breeding. J. Dairy Sci. 72:1328-1333. Lamont, S. J., 1991. Immunogenetics and the major histocompatibility complex. Vet. Immunol. Immunopathol. 30:121-127. Lamont, S. J., 1993. The major histocompatibility complex in chickens. Pages 185-203 in: Manipulation of the Avian Genome. R. J. Etches and A.M.V. Gibbins, ed. CRC Press, Boca Raton, FL. Lamont, S. J., Y. Chen, H.J.M. Aarts, M. C. van der Hulst-van Arkel, G. Beuving, and F. R. Leenstra, 1992. Endogenous viral genes in thirteen highly inbred chicken lines and in lines selected for immune response traits. Poultry Sci. 71:530-538. Leitner, G., Z. Uni, A. Cahaner, M. Gutman, and E. D. Heller, 1992. Replicated divergent selection of broiler chickens for high or low early antibody response to Escherichia coli vaccination. Poultry Sci. 71:27-37. Martin, A., E. A. Dunnington, W. B. Gross, W. E. Briles, R. W. Briles, and P. B. Siegel, 1990. Production traits and alloantigen systems in lines of chickens selected for high or low
IMMUNE SELECTION AND B COMPLEX ALLELIC FREQUENCY
Siegel, P. B., and W. B. Gross, 1980. Production and persistance of antibodies in chickens to sheep erythrocytes 1. Directional selection. Poultry Sci. 59:1-5. Snedecor, G. W., and W. G. Cochran, 1989. Statistical Methods. 7th ed. The Iowa State University Press, Ames, IA. van der Zijpp, A. J., 1983. Breeding for immune responsiveness and disease resistance. World's Poultry Sci. J. 39:118-131. van der Zijpp, A. J., and M.G.B. Nieuwland, 1989. The Biozzi model applied to the chicken. Pages 160-168 in: Series Analytic: Improving Genetic Disease Resistance in Farm Animals. A. J. van der Zijpp and W. Sybesma, ed. Dordrecht, The Netherlands. Weigel, K. A, M. E. Kehrli, Jr., A. E. Freeman, J. R. Thurston, M. J. Stear, and D. H. Kelley, 1991. Association of class I bovine lymphocyte antigen complex alleles with in vitro blood neutrophil functions, lymphocyte blastogenesis, serum complement and conglutinin levels in dairy cattle. Vet. Immunol. Immunopathol. 27: 321-335.
Downloaded from http://ps.oxfordjournals.org/ at Karolinska Institutet University Library on May 31, 2015
antibody responses to sheep erythrocytes. Poultry Sci. 69:871-878. Pevzner, I. Y., H. A. Stone, and A. W. Nordskog, 1981. Immune response and disease resistance in chickens I. Selection for high and low titer to Salmonella pulhrum antigen. Poultry Sci. 60: 920-926. Pitcovski, J., D. E. Heller, A. Cahaner, and B. A. Peleg, 1987. Selection for early responsiveness of chicks to Escherichia coli and Newcastle Disease Virus. Poultry Sci. 66:1276-1282. Puzzi, J. V., L. D. Bacon, and R. R. Dietert, 1990. A gene controlling macrophage functional activation is linked to the chicken B-complex. Arum. Biotechnol. 1:33-45. SAS Institute, 1985. SAS® User's Guide: Statistics. Version 5 Edition. SAS Institute Inc., Cary, NC. Sato, K., H. Abplanalp, D. Napolitano, and J. Reid, 1992. Effects of heterozygosity of major histocompatibility complex haplotypes on performance of Leghorn hens sharing a common inbred background. Poultry Sci. 71:18-26. Schierman, L. W., and A. W. Nordskog, 1961. Relationship of blood type to histocompatibility in chickens. Science 134:1008.
17