- Email: [email protected]
Divergent Selection in Japanese Quail for the Plasma Cholesterol Response to ACTH1
Divergent Selection in Japanese Quail for the Plasma Cholesterol Response to ACTH1 H. L. MARKS US Department of Agriculture, Science and Education Administration, Agricultural Research, Southern Regional Poultry Breeding Project, 107 Livestock-Poultry Building, University of Georgia, Athens, GA 30602 H. S. SIEGEL US Department of Agriculture, Science and Education Administration, Agricultural Research, Southeast Poultry Research Laboratory, 934 College Station Road, Athens, GA 30604 (Received for publication September 17, 1979)
1980 Poultry Science 59:1700-1705 INTRODUCTION Adrenal responsiveness to adrenocorticotropin (ACTH) is influenced by genetic selection in the chicken (Edens and Siegel, 1975) and in the t u r k e y ( B r o w n and Nestor, 1973). Although t h e role of the adrenal h o r m o n e s on homeostatic mechanisms in mammals during stress is well d o c u m e n t e d (Selye and Heuser, 1 9 5 6 ; Bransome, 1 9 6 8 ; Mulrow, 1 9 7 3 ; Gale, 1973), less is k n o w n of t h e role of t h e avian adrenal response during stress (Frankel, 1 9 7 0 ; Siegel, 1971). Siegel ( 1 9 7 1 ) reported t h a t a diurnal cycle in ambient t e m p e r a t u r e from 4 C t o 37 C resulted in a significant rise in plasma corticosteriods in chickens, and Brown and Nestor ( 1 9 7 3 ) reported t h a t excessively high or low environmental
t e m p e r a t u r e s resulted in a significant rise in plasma corticosteriods in turkeys. Edens and Siegel ( 1 9 7 5 ) suggested t h a t t h e adrenal cortical response m a y precede the adrenal medullary response in chickens t h a t are subjected t o acutely high environmental t e m p e r a t u r e . These authors suggested t h a t differences between lines selected for high and low responsiveness t o ACTH may have been due to differences in their abilities to maintain an e x t e n d e d responsiveness to the imposed stress. T h e purposes of this paper are t o report the d e v e l o p m e n t of divergent lines of Japanese quail on t h e basis of their high and low plasma cholesterol responses t o ACTH and to estimate t h e genetic parameters of various traits in these lines. MATERIALS AND METHODS
1
Mention of trade name, proprietary product, vendor, or specific equipment does not constitute a guarantee or warranty by the U. S. Department of Agriculture and does not imply its approval to the exclusion of other products or vendors that may be suitable.
Quail from a r a n d o m b r e d p o p u l a t i o n maintained at t h e Southern Regional Poultry Genetics L a b o r a t o r y were the base p o p u l a t i o n for this s t u d y . A p p r o x i m a t e l y 4 0 0 progeny from 120 paired matings were reared in two hatches in five-deck quail battery brooders.
1700
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
ABSTRACT Two lines of Japanese quail were divergently selected for high and low plasma cholesterol levels after three daily injections with 2 IU of adrenocorticotropen (ACTH)/100 g of body weight. Plasma cholesterols were obtained at 31 days of age, just before the first injection, and at 34 days of age, after the last injection. After eight generations of selection, the plasma cholesterol level after ACTH injections in the high line was 34% greater than the level in the control line, whereas the low line level was approximately 14% less than that in the control. Selection differentials were significantly greater in the high line than in the low line. The realized heritability for plasma cholesterol after ACTH stimulation was approximately .15 in both Hnes. The heritability of plasma cholesterol before ACTH treatment calculated by regression of progeny on mid-parent was .25 in the high line and .16 in the low line. (Key words: divergent selection, plasma cholesterol, ACTH, heritability)
D I V E R G E N T S E L E C T I O N IN J A P A N E S E Q U A I L
2 / C T H in 16% gelatin — Acthor, A r m o u r Pharmaceutical C o m p a n y , Phoenix, A Z .
TABLE 1. Divergent
percentage of change in plasma cholesterol were therefore pooled to form a base population for the current study. From this population 30 males and 30 females with the lowest plasma cholesterol values after ACTH injections were selected to initiate a low line and 30 males and 30 females with the highest plasma cholesterol values were selected to initiate a high line. Subsequent generations were produced in two hatches of approximately 100 birds/line/hatch with 30 paired matings/line. Quail from the control population were reared intermingled with the low and high lines each generation to measure environmental variation across generations. Heritabilities were obtained by dividing the cumulative responses by the cumulative selection differentials (h 2 = R/S) in the high and low lines with the controls used to correct for environmental variation across generations. Heritabilities were also estimated by calculating the regression of mean plasma cholesterol levels on the cumulative selection differentials and from the regression of offspring on mid-parent. R E S U L T S A N D DISCUSSION
The means and standard deviations for plasma cholesterol levels in the control line showed considerable variation across generations (Table 2). The regressions of post-ACTH control means on generation number yielded a regression coefficient (b) of 10.44 ± 8.80 mg/100 ml, whereas the comparable regression coefficient for pre-ACTH was 12.22 ± 3.90 mg/100 ml. Although both regressions were positive, neither was significant, a result that suggests the lack of a time trend across generations in the control line. Variation across generations in all lines was not unexpected and may have resulted from such factors as different
selection for percentage of change in plasma after 3 days im injection of ACTH
Generation 0 cholesterol (mg/1.00 ml)
cholesterol
Generation 1 cholest:erol ( m g / 1 0 0 ml)
Trait
X
Low line
High 1 ine
Control
Post-ACTH Pre-ACTH Change % Change
418 310 108 34.8
481 279 202 72.4
423 238 185 77.7
465 249 216 86.7
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
Birds were fed a 28% commercial gamebird diet. Individual body weights were obtained at 14 and 28 days of age; however, starting in generation 3, body weights were obtained at 31 instead of 28 days of age. At 31 days of age approximately 1 ml of blood was drawn from each bird by cardiac stab. The heparinized blood was stored in ice-chilled glass tubes before it was centrifuged for plasma. Plasma cholesterols were determined by the method of Huang et al. (1961) modified for use on a Technicon automated system. Starting at 31 days and continuing for 3 consecutive days, all birds were injected im with 2 IU of ACTH / 100 g of body weight. At 34 days of age, approximately 1 ml of blood was again drawn from each bird by cardiac stab. These samples were analyzed by the same procedure as those for samples taken at 31 days of age. Plasma cholesterol values from samples taken at 31 days (before ACTH injection) were termed "pre-ACTH" whereas those taken at 34 days (after ACTH injection) were termed "postACTH." Selection for the percentage of change in plasma cholesterol after ACTH injections did not appear to provide a satisfactory technique for developing high and low plasma cholesterol lines (Table 1). After a single generation of selection, pre-ACTH and post-ACTH cholesterol values in the low line were higher than corresponding values in the high line (Table 1). We decided therefore to develop high and low lines based on absolute plasma cholesterol levels after ACTH treatment. All birds in both lines after the single generation of selection based on
1701
1702
MARKS AND SIEGEL TABLE 2. Means and standard deviations for post-ACTH and pre-ACTH plasma cholesterol levels by generation and line
Treatment
Generation
High line
Low line
Control
Post-ACTH
1 2' 3 4 5 6 7 8
438 411 512 472 399 508 529 701
+ ± ± + ± + ± ±
89 130 136 116 131 160 181 169
353 337 436 360 317 350 324 446
+ 92 + 82 ± 99 ± 74 ± 87 + 98 + 119 ± 139
396 351 466 406 342 412 396 521
Pre-ACTH
1 2' 3 4 5 6 7 8
235 292 304 257 243 324 379 361
± ± ± ± ± ± ± ±
48 41 60 49 47 73 90 86
205 254 250 189 186 239 238 281
± ± ± ± ± ± ± ±
208 ± 256± 252 ± 236 ± 217± 274 ± 303 ± 315±
38 42 61 64 46 84 71 80
One hatch only.
lots of ACTH, season, temperature, and other environmental factors. There was a tendency for male plasma cholesterol values to be larger than female values; differences between sexes approached significance at the P<.05 level. The ratherlarge standard errors also provide evidence of considerable variation for both pre- and post-ACTH levels. Increased variation associated with increased population means in the high line is more apparent than decreased variation associated with decreased means in the low lines. These trends do, however, give rise to the suggestion of scaling effects. Genetic response appears to have been greater in selecting for high post-ACTH levels than in selecting for low (Fig. 1). The percentage of deviation from control in generation 8 was approximately 34 in the high line and 14 in the low line. The coefficients of regression of deviations from control by generation yielded b values of 3.77 ± .57 mg/100 ml in the high line and - 1 . 7 0 ± .48 mg/100 ml in the low. Both of these regression coefficients were significant at the P«.01 level. These data indicate an asymmetrical response to selection; progress in the upward direction was approximately twice as great as progress in the downward direction. The difference between the coefficients approached significance at the P<.05 level. Divergent selection often gives rise to asymmetrical response and is often associated with scaling effects (Falconer, 1960). Plotting the data after transformation to logarithmic values
resulted in the removal of a great deal of the asymmetry in Figure 1. This removal does not negate the fact that the response was greater in the upward direction than in the downward, but it does account for the reason for asymmetry in this study. These data are in agreement
35
High
30
Jy
25
c o O E
sf
20 y
15
//'
o
10 5
/ /
// // -/// /- / / / /
/
/v
0 -5 -10
£
\ 7*^^ - v Low
-15 20
-
25
0
1
2
3
4
5
6
7
8
Generation FIG. 1. Post-ACTH plasma cholesterol percentage of deviation from control by line and generation.
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
1
41 43 49 38 45 49 58 74
± 91 ± 91 ±119 ± 84 + 99 ± 105 ± 136 + 132
DIVERGENT SELECTION IN JAPANESE QUAIL
c o O High Line
E o c o ro
"> 01
Q c
4) O
a.
0
1 2
3
4
5
6
Generation
FIG. 2. Pre-ACTH plasma cholesterol percentage of deviation from control by line and generation.
response to artificial selection on these lines. Evidence showed that the magnitude of the selection differential declined more rapidly across generations in the low line than in the high. Regression of generation number on weighted selection differentials yielded nonsignificant b values of —5.63 ± 4.61 in the low line and - . 2 3 ± 5.3 5 mg/100 ml in the high.
TABLE 3. Selection differentials by post-ACTH line and generation High line (mg/100 ml)
Low line (mg/100 ml) Generation
Expected
Effective
i1
1 2 3 4 5 6 7 8
105.2 74.5 34.2 85.5 34.1 56.7 90.5 57.7
111.0 80.9 35.3 86.9 37.3 57.6 95.9 30.2
1.21
X
67.3 a
66.9 a
.99 .36
1.17 .43 .59 .81 .22 .72
Expected
Effective
i>
138.3 81.0 89.3 138.1 57.2 121.8 136.4 85.7 106.0b
146.0 70.6 81.5 140.0 64.0 117.4 130.5 95.9 105.7b
1.65
1 i = intensity of selection. a ' b Means bearing different superscripts are significantly different (P<.01).
.54 .60
1.20 .49 .74 .72 .57 .81
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
with the means for high and low plasma cholesterol lines of mice (Weibust, 1973). The percentage of deviation from control in pre-ACTH level in the high line was not as great as that in post-ACTH level in the high line (Fig. 2). The percentage of deviation from control in pre-ACTH level in the low line was, however, not greatly different from that in the post-ACTH levels across generations. The percentage of deviation in pre-ACTH levels in the low and high lines were not greatly different as indicated by regression coefficient values across generations (Fig. 2). Not until the fourth generation was any real movement observed in the downward direction of pre-ACTH levels. Movement in the upper direction was observed in the first three generations. This pattern is similar to that observed in yolk cholesterol level when selection was practiced for milligram of cholesterol per gram of yolk (Marks and Washburn, 1977). Apparently, whatever mechanisms may have been involved in the resistance of the pre-ACTH level to move in a downward direction were overcome with continued selection pressure after the third generation. The greater progress in the high post-ACTH line than in the low was due at least in part to larger selection differentials in the high line than in the low (Table 3). The differences between selection differentials in the high and the low lines were significant (P«.Ul). Differences between the expected (unweighted) and effective (weighted) selection differentials were small and not significant in either the high or the low line. Therefore, natural selection has apparently exerted little influence on the
1703
1704
MARKS AND SIEGEL
The pre-ACTH cholesterol heritability estimates obtained by regression of progeny on mid-parent were generally larger than comparable post-ACTH estimates (Table 5). Estimates were also generally higher in the high line than in the low line. The mean estimates of .25 and .16 for the high and low lines are similar to heritability estimates for serum cholesterol in the domestic fowl (Cherms et al., I960; Hollands « al, 1980). The heritabilities obtained from the high and low lines in the current study cannot be directly compared with serum and plasma cholesterol values reported in the literature. Our estimates were from lines during a stress response injection after ACTH, whereas Dunnington et al. (1977) and Weibust (1973) estimates were from nonstressed lines. One would assume that heritability estimates would be lower in lines selected for plasma cholesterol after ACTH
TABLE 4. Heritability estimates for post-ACTH plasma cholesterol in high and low lines of quail Generation
High line
1 2 3 4 5 6 7 8
.16 .20 .14 .11 .30 .23 -.41 .15
± ± ± ± ± ± ± ±
.08 .23 .14 .18 .10 .18 .15 .08
Low line .00 .11 .08 -.15 .14 .52 .22 -.23
± + + ± ± ± ± ±
.16 .16 .18 .12 .21 .17 .30 .20
X
.16
.13
Realized 1
.13 ± .09
.11 ± .12
Realized 2
.18
.14
1
Regression. Cumulative h2 = R/S.
treatment than before because of greater environmental variation resulting from such a stress. Dunnington et al. (1977) reported a heritability of .31 ± .07 for serum cholesterol by sire-son regression in an unselected randombred line of mice. Realized heritabilities for plasma cholesterol in mice after five generations of selection were .51 ± .05 for males and .50 ± .03 for females (Weibust, 1973); however, actual data were not analyzed in that report. Because the variance was not independent of the mean, data were transformed to logarithms before analysis and calculation of heritabilities. For both sexes, realized h 2 estimates were greater in the low-line mice.
TABLE 5. Mid-parent heritability estimates for pre-ACTH plasma cholesterol in high and low lines of quail Generation
High line
Low line
1 2 3 4 5 6 7 8
.24 -.05 .44 .18 .19 .14 .54 .29
.18 .19 .42 .16 .12 .04 -.06 .18
X
.25
± ± ± ± ± ± ± ±
.05 .15 .14 .07 .09 .07 .12 .11
.16
± ± ± ± ± ± ± ±
.07 .13 .14 .13 .16 .18 .07 .18
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
Intensity of selection (i) values obtained by dividing the selection differential by the phenotypic standard deviation o p of the population (Falconer, 1960) also illustrates the importance of scaling on the magnitude of the selection differentials of the two lines (Table 3). Although i values were higher in the high line than in the low line, differences were smaller than those between actual selection differentials (Table 3). The heritability of plasma cholesterol after ACTH injection in quail appears to be of relatively low magnitude. The mid-parent heritability estimates showed considerable variation across generations in both the high and the low lines. The mean of generation estimates was .16 in the high line and .13 in the low (Table 4). These values are in good agreement with realized estimates from alternate methods of calculation (Table 4). Realized estimates obtained by dividing the cumulative gains by cumulative selection differentials (h 2 = R/S) yielded values of .18 in the high line and .14 in the low. Regression of generation means on cumulative selection differentials resulted in estimates of .13 ± .09 in the high and .11 ± .12 in the low. These data indicate little difference in the heritability of plasma post-ACTH cholesterol in the upward or downward direction. The similarity of heritability estimates between lines indicates that the greater progress in the upward direction than in the downward direction was due primarily to the greater magnitude of the selection differentials in the high line (Table 3).
DIVERGENT SELECTION IN JAPANESE QUAIL
ACKNOWLEDGMENTS We thank J. W. Latimer, Joyce Hinesely, Jennifer Windham, and Vivian Patten for their technical assistance. REFERENCES Bransome, E. D., Jr., 1968. Adrenal cortex. Ann. Rev. Physiol. 30:171-212. Brown, K. I., and K. E. Nestor, 1973. Some physiological responses of turkeys selected for high and low adrenal response to cold stress. Poultry Sci. 52:1948-1954. Cherms, F. L., Jr., F. H. Wilcox and C. S. Shaffner,
1960. Genetic studies of serum cholesterol levels in the chicken. Poultry Sci. 39:889-892. Dunnington, E. A., J. M. White, and W. E. Vinson, 1977. Genetic parameters of serum cholesterol levels, activity and growth in mice. Genetics 85:659-668. Edens, F. W., and H. S. Siegel, 1975. Adrenal responses in high and low ACTH response lines of chickens during acute heat stress. Gen. Comp. Endocrinol. 25:67-73. Falconer, D. S., 1960. Introduction to quantitative genetics. Ronald Press Co., NY. Frankel, A. J., 1970. Neurohumoral control of the avian adrenal: A review. Poultry Sci. 49: 869-921. Gale, C. C , 1973. Neuroendocrine aspects of thermoregulation. Ann. Rev. Physiol. 35:391-430. Hollands, K. G., A. A. Grunder, and C. J. Williams, 1980. Response to five generations of selection for blood cholesterol levels in White Leghorns. Poultry Sci. 59: 1316-1323. Huang, T. C , C. P. Chen, V. Wefler, and A. Raftery, 1961. A stable reagent for the LiebermannBurchand Reaction. Application to rapid serum cholesterol determination. Anal. Chem. 33: 1405-1407. Marks, H. L., and K. W. Washburn, 1977. Divergent selection for yolk cholesterol in laying hens. Brit. Poultry Sci. 18:179-188. Mulrow, P. J., 1973. The adrenal cortex. Ann. Rev. Physiol. 34:409-424. Seyle, H., and G. Heuser, 1956. The fifth annual report on stress. M. D. Publications, NY. Siegel, H. S., 1971. Adrenals, stress and the environment. World's Poultry Sci. J. 27:327-349. Weibust, R. S., 1973. Inheritance of plasma cholesterol levels in mice. Genetics 73:303 — 312.
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
However, when the actual data expressed as mg/100 ml were considered the percentage of deviation from control was 13 to 23% greater in the high line than in the low line. Also during the first three generations, the deviation was two to three times greater in the high line than in the low line. This selection response in mice plasma cholesterol agrees with our results (Figs. 1 and 2). The higher heritabilities observed in plasma cholesterol in mice than in quail may be due to the fact that estimates for quail were obtained before maturity, whereas estimates for mice were obtained after maturity (60 to 100 days of age). Possibly the fact that less environmental variation is present after maturity than during the developmental phases would result in higher heritabilities.
1705
1980 Poultry Science 59:1700-1705 INTRODUCTION Adrenal responsiveness to adrenocorticotropin (ACTH) is influenced by genetic selection in the chicken (Edens and Siegel, 1975) and in the t u r k e y ( B r o w n and Nestor, 1973). Although t h e role of the adrenal h o r m o n e s on homeostatic mechanisms in mammals during stress is well d o c u m e n t e d (Selye and Heuser, 1 9 5 6 ; Bransome, 1 9 6 8 ; Mulrow, 1 9 7 3 ; Gale, 1973), less is k n o w n of t h e role of t h e avian adrenal response during stress (Frankel, 1 9 7 0 ; Siegel, 1971). Siegel ( 1 9 7 1 ) reported t h a t a diurnal cycle in ambient t e m p e r a t u r e from 4 C t o 37 C resulted in a significant rise in plasma corticosteriods in chickens, and Brown and Nestor ( 1 9 7 3 ) reported t h a t excessively high or low environmental
t e m p e r a t u r e s resulted in a significant rise in plasma corticosteriods in turkeys. Edens and Siegel ( 1 9 7 5 ) suggested t h a t t h e adrenal cortical response m a y precede the adrenal medullary response in chickens t h a t are subjected t o acutely high environmental t e m p e r a t u r e . These authors suggested t h a t differences between lines selected for high and low responsiveness t o ACTH may have been due to differences in their abilities to maintain an e x t e n d e d responsiveness to the imposed stress. T h e purposes of this paper are t o report the d e v e l o p m e n t of divergent lines of Japanese quail on t h e basis of their high and low plasma cholesterol responses t o ACTH and to estimate t h e genetic parameters of various traits in these lines. MATERIALS AND METHODS
1
Mention of trade name, proprietary product, vendor, or specific equipment does not constitute a guarantee or warranty by the U. S. Department of Agriculture and does not imply its approval to the exclusion of other products or vendors that may be suitable.
Quail from a r a n d o m b r e d p o p u l a t i o n maintained at t h e Southern Regional Poultry Genetics L a b o r a t o r y were the base p o p u l a t i o n for this s t u d y . A p p r o x i m a t e l y 4 0 0 progeny from 120 paired matings were reared in two hatches in five-deck quail battery brooders.
1700
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
ABSTRACT Two lines of Japanese quail were divergently selected for high and low plasma cholesterol levels after three daily injections with 2 IU of adrenocorticotropen (ACTH)/100 g of body weight. Plasma cholesterols were obtained at 31 days of age, just before the first injection, and at 34 days of age, after the last injection. After eight generations of selection, the plasma cholesterol level after ACTH injections in the high line was 34% greater than the level in the control line, whereas the low line level was approximately 14% less than that in the control. Selection differentials were significantly greater in the high line than in the low line. The realized heritability for plasma cholesterol after ACTH stimulation was approximately .15 in both Hnes. The heritability of plasma cholesterol before ACTH treatment calculated by regression of progeny on mid-parent was .25 in the high line and .16 in the low line. (Key words: divergent selection, plasma cholesterol, ACTH, heritability)
D I V E R G E N T S E L E C T I O N IN J A P A N E S E Q U A I L
2 / C T H in 16% gelatin — Acthor, A r m o u r Pharmaceutical C o m p a n y , Phoenix, A Z .
TABLE 1. Divergent
percentage of change in plasma cholesterol were therefore pooled to form a base population for the current study. From this population 30 males and 30 females with the lowest plasma cholesterol values after ACTH injections were selected to initiate a low line and 30 males and 30 females with the highest plasma cholesterol values were selected to initiate a high line. Subsequent generations were produced in two hatches of approximately 100 birds/line/hatch with 30 paired matings/line. Quail from the control population were reared intermingled with the low and high lines each generation to measure environmental variation across generations. Heritabilities were obtained by dividing the cumulative responses by the cumulative selection differentials (h 2 = R/S) in the high and low lines with the controls used to correct for environmental variation across generations. Heritabilities were also estimated by calculating the regression of mean plasma cholesterol levels on the cumulative selection differentials and from the regression of offspring on mid-parent. R E S U L T S A N D DISCUSSION
The means and standard deviations for plasma cholesterol levels in the control line showed considerable variation across generations (Table 2). The regressions of post-ACTH control means on generation number yielded a regression coefficient (b) of 10.44 ± 8.80 mg/100 ml, whereas the comparable regression coefficient for pre-ACTH was 12.22 ± 3.90 mg/100 ml. Although both regressions were positive, neither was significant, a result that suggests the lack of a time trend across generations in the control line. Variation across generations in all lines was not unexpected and may have resulted from such factors as different
selection for percentage of change in plasma after 3 days im injection of ACTH
Generation 0 cholesterol (mg/1.00 ml)
cholesterol
Generation 1 cholest:erol ( m g / 1 0 0 ml)
Trait
X
Low line
High 1 ine
Control
Post-ACTH Pre-ACTH Change % Change
418 310 108 34.8
481 279 202 72.4
423 238 185 77.7
465 249 216 86.7
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
Birds were fed a 28% commercial gamebird diet. Individual body weights were obtained at 14 and 28 days of age; however, starting in generation 3, body weights were obtained at 31 instead of 28 days of age. At 31 days of age approximately 1 ml of blood was drawn from each bird by cardiac stab. The heparinized blood was stored in ice-chilled glass tubes before it was centrifuged for plasma. Plasma cholesterols were determined by the method of Huang et al. (1961) modified for use on a Technicon automated system. Starting at 31 days and continuing for 3 consecutive days, all birds were injected im with 2 IU of ACTH / 100 g of body weight. At 34 days of age, approximately 1 ml of blood was again drawn from each bird by cardiac stab. These samples were analyzed by the same procedure as those for samples taken at 31 days of age. Plasma cholesterol values from samples taken at 31 days (before ACTH injection) were termed "pre-ACTH" whereas those taken at 34 days (after ACTH injection) were termed "postACTH." Selection for the percentage of change in plasma cholesterol after ACTH injections did not appear to provide a satisfactory technique for developing high and low plasma cholesterol lines (Table 1). After a single generation of selection, pre-ACTH and post-ACTH cholesterol values in the low line were higher than corresponding values in the high line (Table 1). We decided therefore to develop high and low lines based on absolute plasma cholesterol levels after ACTH treatment. All birds in both lines after the single generation of selection based on
1701
1702
MARKS AND SIEGEL TABLE 2. Means and standard deviations for post-ACTH and pre-ACTH plasma cholesterol levels by generation and line
Treatment
Generation
High line
Low line
Control
Post-ACTH
1 2' 3 4 5 6 7 8
438 411 512 472 399 508 529 701
+ ± ± + ± + ± ±
89 130 136 116 131 160 181 169
353 337 436 360 317 350 324 446
+ 92 + 82 ± 99 ± 74 ± 87 + 98 + 119 ± 139
396 351 466 406 342 412 396 521
Pre-ACTH
1 2' 3 4 5 6 7 8
235 292 304 257 243 324 379 361
± ± ± ± ± ± ± ±
48 41 60 49 47 73 90 86
205 254 250 189 186 239 238 281
± ± ± ± ± ± ± ±
208 ± 256± 252 ± 236 ± 217± 274 ± 303 ± 315±
38 42 61 64 46 84 71 80
One hatch only.
lots of ACTH, season, temperature, and other environmental factors. There was a tendency for male plasma cholesterol values to be larger than female values; differences between sexes approached significance at the P<.05 level. The ratherlarge standard errors also provide evidence of considerable variation for both pre- and post-ACTH levels. Increased variation associated with increased population means in the high line is more apparent than decreased variation associated with decreased means in the low lines. These trends do, however, give rise to the suggestion of scaling effects. Genetic response appears to have been greater in selecting for high post-ACTH levels than in selecting for low (Fig. 1). The percentage of deviation from control in generation 8 was approximately 34 in the high line and 14 in the low line. The coefficients of regression of deviations from control by generation yielded b values of 3.77 ± .57 mg/100 ml in the high line and - 1 . 7 0 ± .48 mg/100 ml in the low. Both of these regression coefficients were significant at the P«.01 level. These data indicate an asymmetrical response to selection; progress in the upward direction was approximately twice as great as progress in the downward direction. The difference between the coefficients approached significance at the P<.05 level. Divergent selection often gives rise to asymmetrical response and is often associated with scaling effects (Falconer, 1960). Plotting the data after transformation to logarithmic values
resulted in the removal of a great deal of the asymmetry in Figure 1. This removal does not negate the fact that the response was greater in the upward direction than in the downward, but it does account for the reason for asymmetry in this study. These data are in agreement
35
High
30
Jy
25
c o O E
sf
20 y
15
//'
o
10 5
/ /
// // -/// /- / / / /
/
/v
0 -5 -10
£
\ 7*^^ - v Low
-15 20
-
25
0
1
2
3
4
5
6
7
8
Generation FIG. 1. Post-ACTH plasma cholesterol percentage of deviation from control by line and generation.
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
1
41 43 49 38 45 49 58 74
± 91 ± 91 ±119 ± 84 + 99 ± 105 ± 136 + 132
DIVERGENT SELECTION IN JAPANESE QUAIL
c o O High Line
E o c o ro
"> 01
Q c
4) O
a.
0
1 2
3
4
5
6
Generation
FIG. 2. Pre-ACTH plasma cholesterol percentage of deviation from control by line and generation.
response to artificial selection on these lines. Evidence showed that the magnitude of the selection differential declined more rapidly across generations in the low line than in the high. Regression of generation number on weighted selection differentials yielded nonsignificant b values of —5.63 ± 4.61 in the low line and - . 2 3 ± 5.3 5 mg/100 ml in the high.
TABLE 3. Selection differentials by post-ACTH line and generation High line (mg/100 ml)
Low line (mg/100 ml) Generation
Expected
Effective
i1
1 2 3 4 5 6 7 8
105.2 74.5 34.2 85.5 34.1 56.7 90.5 57.7
111.0 80.9 35.3 86.9 37.3 57.6 95.9 30.2
1.21
X
67.3 a
66.9 a
.99 .36
1.17 .43 .59 .81 .22 .72
Expected
Effective
i>
138.3 81.0 89.3 138.1 57.2 121.8 136.4 85.7 106.0b
146.0 70.6 81.5 140.0 64.0 117.4 130.5 95.9 105.7b
1.65
1 i = intensity of selection. a ' b Means bearing different superscripts are significantly different (P<.01).
.54 .60
1.20 .49 .74 .72 .57 .81
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
with the means for high and low plasma cholesterol lines of mice (Weibust, 1973). The percentage of deviation from control in pre-ACTH level in the high line was not as great as that in post-ACTH level in the high line (Fig. 2). The percentage of deviation from control in pre-ACTH level in the low line was, however, not greatly different from that in the post-ACTH levels across generations. The percentage of deviation in pre-ACTH levels in the low and high lines were not greatly different as indicated by regression coefficient values across generations (Fig. 2). Not until the fourth generation was any real movement observed in the downward direction of pre-ACTH levels. Movement in the upper direction was observed in the first three generations. This pattern is similar to that observed in yolk cholesterol level when selection was practiced for milligram of cholesterol per gram of yolk (Marks and Washburn, 1977). Apparently, whatever mechanisms may have been involved in the resistance of the pre-ACTH level to move in a downward direction were overcome with continued selection pressure after the third generation. The greater progress in the high post-ACTH line than in the low was due at least in part to larger selection differentials in the high line than in the low (Table 3). The differences between selection differentials in the high and the low lines were significant (P«.Ul). Differences between the expected (unweighted) and effective (weighted) selection differentials were small and not significant in either the high or the low line. Therefore, natural selection has apparently exerted little influence on the
1703
1704
MARKS AND SIEGEL
The pre-ACTH cholesterol heritability estimates obtained by regression of progeny on mid-parent were generally larger than comparable post-ACTH estimates (Table 5). Estimates were also generally higher in the high line than in the low line. The mean estimates of .25 and .16 for the high and low lines are similar to heritability estimates for serum cholesterol in the domestic fowl (Cherms et al., I960; Hollands « al, 1980). The heritabilities obtained from the high and low lines in the current study cannot be directly compared with serum and plasma cholesterol values reported in the literature. Our estimates were from lines during a stress response injection after ACTH, whereas Dunnington et al. (1977) and Weibust (1973) estimates were from nonstressed lines. One would assume that heritability estimates would be lower in lines selected for plasma cholesterol after ACTH
TABLE 4. Heritability estimates for post-ACTH plasma cholesterol in high and low lines of quail Generation
High line
1 2 3 4 5 6 7 8
.16 .20 .14 .11 .30 .23 -.41 .15
± ± ± ± ± ± ± ±
.08 .23 .14 .18 .10 .18 .15 .08
Low line .00 .11 .08 -.15 .14 .52 .22 -.23
± + + ± ± ± ± ±
.16 .16 .18 .12 .21 .17 .30 .20
X
.16
.13
Realized 1
.13 ± .09
.11 ± .12
Realized 2
.18
.14
1
Regression. Cumulative h2 = R/S.
treatment than before because of greater environmental variation resulting from such a stress. Dunnington et al. (1977) reported a heritability of .31 ± .07 for serum cholesterol by sire-son regression in an unselected randombred line of mice. Realized heritabilities for plasma cholesterol in mice after five generations of selection were .51 ± .05 for males and .50 ± .03 for females (Weibust, 1973); however, actual data were not analyzed in that report. Because the variance was not independent of the mean, data were transformed to logarithms before analysis and calculation of heritabilities. For both sexes, realized h 2 estimates were greater in the low-line mice.
TABLE 5. Mid-parent heritability estimates for pre-ACTH plasma cholesterol in high and low lines of quail Generation
High line
Low line
1 2 3 4 5 6 7 8
.24 -.05 .44 .18 .19 .14 .54 .29
.18 .19 .42 .16 .12 .04 -.06 .18
X
.25
± ± ± ± ± ± ± ±
.05 .15 .14 .07 .09 .07 .12 .11
.16
± ± ± ± ± ± ± ±
.07 .13 .14 .13 .16 .18 .07 .18
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
Intensity of selection (i) values obtained by dividing the selection differential by the phenotypic standard deviation o p of the population (Falconer, 1960) also illustrates the importance of scaling on the magnitude of the selection differentials of the two lines (Table 3). Although i values were higher in the high line than in the low line, differences were smaller than those between actual selection differentials (Table 3). The heritability of plasma cholesterol after ACTH injection in quail appears to be of relatively low magnitude. The mid-parent heritability estimates showed considerable variation across generations in both the high and the low lines. The mean of generation estimates was .16 in the high line and .13 in the low (Table 4). These values are in good agreement with realized estimates from alternate methods of calculation (Table 4). Realized estimates obtained by dividing the cumulative gains by cumulative selection differentials (h 2 = R/S) yielded values of .18 in the high line and .14 in the low. Regression of generation means on cumulative selection differentials resulted in estimates of .13 ± .09 in the high and .11 ± .12 in the low. These data indicate little difference in the heritability of plasma post-ACTH cholesterol in the upward or downward direction. The similarity of heritability estimates between lines indicates that the greater progress in the upward direction than in the downward direction was due primarily to the greater magnitude of the selection differentials in the high line (Table 3).
DIVERGENT SELECTION IN JAPANESE QUAIL
ACKNOWLEDGMENTS We thank J. W. Latimer, Joyce Hinesely, Jennifer Windham, and Vivian Patten for their technical assistance. REFERENCES Bransome, E. D., Jr., 1968. Adrenal cortex. Ann. Rev. Physiol. 30:171-212. Brown, K. I., and K. E. Nestor, 1973. Some physiological responses of turkeys selected for high and low adrenal response to cold stress. Poultry Sci. 52:1948-1954. Cherms, F. L., Jr., F. H. Wilcox and C. S. Shaffner,
1960. Genetic studies of serum cholesterol levels in the chicken. Poultry Sci. 39:889-892. Dunnington, E. A., J. M. White, and W. E. Vinson, 1977. Genetic parameters of serum cholesterol levels, activity and growth in mice. Genetics 85:659-668. Edens, F. W., and H. S. Siegel, 1975. Adrenal responses in high and low ACTH response lines of chickens during acute heat stress. Gen. Comp. Endocrinol. 25:67-73. Falconer, D. S., 1960. Introduction to quantitative genetics. Ronald Press Co., NY. Frankel, A. J., 1970. Neurohumoral control of the avian adrenal: A review. Poultry Sci. 49: 869-921. Gale, C. C , 1973. Neuroendocrine aspects of thermoregulation. Ann. Rev. Physiol. 35:391-430. Hollands, K. G., A. A. Grunder, and C. J. Williams, 1980. Response to five generations of selection for blood cholesterol levels in White Leghorns. Poultry Sci. 59: 1316-1323. Huang, T. C , C. P. Chen, V. Wefler, and A. Raftery, 1961. A stable reagent for the LiebermannBurchand Reaction. Application to rapid serum cholesterol determination. Anal. Chem. 33: 1405-1407. Marks, H. L., and K. W. Washburn, 1977. Divergent selection for yolk cholesterol in laying hens. Brit. Poultry Sci. 18:179-188. Mulrow, P. J., 1973. The adrenal cortex. Ann. Rev. Physiol. 34:409-424. Seyle, H., and G. Heuser, 1956. The fifth annual report on stress. M. D. Publications, NY. Siegel, H. S., 1971. Adrenals, stress and the environment. World's Poultry Sci. J. 27:327-349. Weibust, R. S., 1973. Inheritance of plasma cholesterol levels in mice. Genetics 73:303 — 312.
Downloaded from http://ps.oxfordjournals.org/ by guest on May 21, 2015
However, when the actual data expressed as mg/100 ml were considered the percentage of deviation from control was 13 to 23% greater in the high line than in the low line. Also during the first three generations, the deviation was two to three times greater in the high line than in the low line. This selection response in mice plasma cholesterol agrees with our results (Figs. 1 and 2). The higher heritabilities observed in plasma cholesterol in mice than in quail may be due to the fact that estimates for quail were obtained before maturity, whereas estimates for mice were obtained after maturity (60 to 100 days of age). Possibly the fact that less environmental variation is present after maturity than during the developmental phases would result in higher heritabilities.
1705