- Email: [email protected]
Serum neuropeptide Y (NPY) and leptin concentrations in pigs selected for components of efficient lean growth
Domestic Animal Endocrinology 24 (2003) 15–29
Serum neuropeptide Y (NPY) and leptin concentrations in pigs selected for components of efficient lean growth N.D. Cameron∗ , E. McCullough, K. Troup, J.C. Penman Roslin Institute, Roslin, Midlothian, EH25 9PS, Scotland, UK Received 15 April 2002; accepted 6 June 2002
Abstract Responses in log transformed serum neuropeptide Y (NPY) concentration and leptin concentration after six generations of divergent selection on components of efficient lean growth in pigs were measured. From an animal breeding perspective, serum NPY and/or leptin concentrations could be used as physiological predictors of genetic merit if there were significant responses to selection. At 90 kg liveweight, log transformed serum NPY concentrations were increased with divergent selection for low food conversion ratio (LFC) (6.31 versus 5.72, SED 0.09 log(pmol/L)) or for high lean growth rate (LGA) (5.80 versus 5.37 log(pmol/L)) but not with selection on daily food intake (DFI) (6.26 versus 6.14 log(pmol/L)). Selection for high DFI was associated with increased serum leptin concentration (3.06 versus 2.45, SED 0.21 ng/mL human equivalent (HE)) as was selection for low LFC (3.04 versus 2.46 ng/mL HE). Correlations between leptin and predicted lipid weight increased with stage of test (0.13, 0.34 and 0.43, SE 0.08 at 30, 50 and 75 kg). The high correlations between successive serum NPY concentrations (0.80, SE 0.11) suggest that changes in body composition with time would not be reflected in serum NPY concentrations. Serum NPY and, to a lesser extent, serum leptin concentrations were insensitive to dietary differences in total lysine: energy and indicated that studies using a genetic resource population of animals may be more powerful than nutritional studies using isoenergetic diets differing in lysine content to examine aspects of function of serum concentrations of NPY and leptin in pigs. © 2002 Elsevier Science Inc. All rights reserved.
1. Introduction Neuropeptide Y (NPY) and leptin have been respectively associated with stimulating and inhibiting appetite [1]. NPY is a neurotransmitter expressed in the hypothalamus with a ∗
Corresponding author. Tel.: +44-131-527-4258; fax: +44-131-440-0434. E-mail address: [email protected] (N.D. Cameron). 0739-7240/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 9 - 7 2 4 0 ( 0 2 ) 0 0 1 8 4 - 4
16
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
role in energy homeostasis, such that NPY administration in rats can increase energy intake, decrease energy expenditure and increase lipogenesis [2]. In contrast, leptin is synthesised and secreted by the adipocytes into the blood stream [3] and transported to the brain where it stimulates gene expression of catabolic neuropeptides and inhibits that of anabolic neuropeptides, such as NPY [2,4]. The interaction between NPY and leptin is demonstrated in leptin-deficient ob/ob mice and leptin-resistant (lack of sensitivity to leptin) db/db mice as the over-expression of NPY in ob/ob mice can be controlled by leptin administration but not in db/db mice [5]. An interaction between leptin and NPY affecting lipid mobilization has also been demonstrated at the adipocyte level [6]. High correlations between leptin concentrations in human cerebrospinal fluid (CSF) and plasma have been reported [7–10]. Obese women had higher plasma leptin concentrations than controls [10,11] that was consistent with the proposal that human obesity may be associated with leptin-resistance [12]. In contrast, no correlation between human CSF and plasma NPY concentrations was detected and there was no difference between obese and control women in CSF and plasma NPY concentrations [10]. However, higher plasma NPY concentrations in bulimia nervosa women were detected than in anorexia nervosa women [11], that may have resulted from the increased experimental power due to the use of “extreme” phenotypes. “Extreme” pig genotypes have been generated in the Edinburgh lean growth experiment by several generations of divergent selection for components of efficient lean growth rate [13]. From an animal breeding perspective, it would be useful if plasma leptin and/or NPY concentrations could be used as physiological predictors of genetic merit and informative if they provided a physiological explanation of the observed responses in growth and carcass traits [14]. For example, were the responses in carcass fat content consistent with increased leptin-resistance, as in humans and db/db mice, or with leptin-deficiency, as in ob/ob mice ? Higher leptin mRNA and serum concentrations have been reported in pigs with greater backfat depths [15,16], but there is an absence of information concerning the relationship between nutrient intake and leptin concentrations [17] and, secondly, serum NPY concentrations in pigs have not been previously reported. Therefore, the current study was superimposed on a genotype-nutrition interaction experiment [18] to measure the responses in serum leptin and NPY concentrations to selection for components of efficient lean growth and determine the sensitivity of each genotype, in terms of serum leptin and NPY concentrations, to nutritional inputs.
2. Materials and methods 2.1. Animals and performance test Details on establishment of the Large White pig population and seven generations of divergent selection for daily food intake (DFI), lean food conversion (LFC) and lean growth rate (LGA) are available [13]. Each selection line consisted of progeny from 10 sires with two dams mated to each sire and the selection lines were derived from the same base population. In the current study, there were five litters of five full-sibs from each of the six
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
17
Table 1 Composition and specification of diets (g/kg) Diet
P
Q
R
S
T
U
X
Barley Wheat Wheatfeed Hi-pro soya Full fat soya Pellet binder Fat (mixer and spray) Chilean fishmeal Vitaminsa Protein Total lysine Lysine: DE
293.5 429 114.5 0 89 12.5 31 0 30.5 135 8.14 0.58
199 434 150 86 61 12.5 28 0 29.5 161 9.66 0.69
192 385.7 150 203.5 0 12.5 28 0 28.3 187 11.30 0.81
175.5 355.5 149 256.5 0 12.5 22.5 0 28.5 207 12.73 0.91
178 347 150.5 257.5 0 12.5 22 0 32.5 210 14.16 1.01
200 292.5 161 257.5 0 12.5 23 18.5 35 222 15.75 1.12
200 277 160 260 0 12.5 15 47 28.5 240 17.20 1.23
a
Vitamins and minerals, etc.
selection lines with one litter from five sires chosen at random. Pigs were performance tested on a phase-feeding regime [19] with each full-sib ad libitum fed one of five isoenergetic (14.0 MJ DE/kg) diets differing in total lysine: energy (Table 1) with diets R, S, T, U or X, then Q, R, S, T or U and finally P, Q, R, S or T offered in three test-periods. For example, the first sequence of diets fed during the three test-periods would be R then Q and then P. At the start of the performance test, a full-sib within each litter was allocated to each of five diets, such that within a selection line, five animals were performance tested on each of the five diet combinations. Each litter consisted of two boars and three gilts or three boars and two gilts with consideration of sex for allocation of animals to diets. The intended start weights for each test-period, lasting 14 days, were 30 ± 3, 50 ± 4 and 75 ± 5 kg (i.e., ranges of 27–33, 46–54 and 70–80 kg) with realized mean (SD) weights of 29.3 (1.6), 49.5 (2.1) and 73.5 (2.1) kg. Pigs were penned individually, and prior to the start of test and between test-periods, pigs were fed diet S. There were no diet effects or between-line within-selection group effects on the time taken to reach the required start weight for second and third test-periods, which averaged 13.9 (SD 4.1) and 15.5 (SD 4.6) days. Although, the interval between test-periods 2 and 3 was longer for the low LFC line than for other selection lines (18.2 versus 15.0, SED 1.7 days). The different patterns of NPY and leptin concentrations over time due to the selection lines and diets were interpretable on a consistent basis, given the relatively constant time-interval between blood sampling occasions. An additional 20 pigs from each of the high and low DFI and LGA lines were performance tested with the diet-choice procedure [19]. Animals had ad libitum access to two diets (Q and U) in separate food hoppers with start and end of test of 30 ± 3 and 90 ± 5 kg (i.e., ranges of 27–33 and 85–95 kg). Food intake of diets Q and U were measured mid-test at 50 ±4 kg and at 75±5 kg and at the end of test. In practice, the mean (SD) liveweights for measurements were 29.6 (1.9), 49.3 (2.1), 73.1 (2.4) and 88.4 (2.6) kg. The additional 20 animals formed 10 pairs of full-sibs, a boar and a gilt, of which five pairs were full-sibs of animals performance tested with phase-feeding and five pairs were half-sibs of those tested with phase-feeding. Further details of the performance test procedure with phase-feeding and diet-choice are available [18].
18
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Protein and lipid weights at different stages of the performance test were predicted using equations derived from chemical composition studies of carcasses in an earlier generation [20]. Briefly, six litters of five full-sibs in each selection line were slaughtered at 30, 45, 60, 75 or 90 ± 3 kg, with a full-sib in each litter allocated at random to each slaughter weight, for subsequent chemical analysis of carcass and non-carcass components. For the purpose of inputs to the nutrient utilisation equations in the current study, predicted protein and lipid weights were based on liveweight and ultrasonic backfat depth. Prediction equations for lipid and protein weights were estimated as 0.165 W + 0.544 BF and 0.176 W − 0.117 BF, where W and BF were the appropriate liveweight and ultrasonic backfat depth, with the prediction equations proportionately accounting for 0.91 and 0.98 of the variation in lipid and protein weights. 2.2. Serum NPY and leptin assays Prior to each performance test measurement of phase-fed and diet-choice animals, 10 mL blood samples were withdrawn by jugular venepuncture from the vena cava, kept at −4◦ C for 24 h, then centrifuged for 30 min at 4◦ C. The red blood cell-free serum was decanted and frozen at −20◦ C until required for assay. Blood sampling was performed from 9:00 h each day and food was available to animals at all times due to the ad libitum feeding procedure. Animals were sampled under the same feeding procedure, although between-selection line variation on feeding patterns may have influenced serum NPY and leptin concentrations. Serum NPY concentrations, expressed as picomole per liter, were determined in duplicate with a commercially available competitive radio-immunoassay procedure (EuroDiagnostica, The Netherlands) using an antiserum raised against synthetic NPY. NPY consists of 36 amino acids with the human and porcine NPY amino acid sequences only differing by replacement of leucine with methione at position 17 [21]. The detection limit of the assay was 6 pmol/L (Euro-Diagnostica, The Netherlands) and the repeatability of the NPY assay in the current study, calculated from the between- and within-animal variance components from data on all 230 animals, averaged 0.92 (SE 0.04) for the six sampling times. The phenotypic SD of the six log transformed NPY sample times were 0.30, 0.30, 0.28, 0.27, 0.32, and 0.44 with an average coefficient of variation (s/x) ¯ of 0.050. Quality controls were performed in duplicate at the start and end of each assay, but with only five assays estimation of inter-assay coefficient of variation was imprecise. Serum leptin concentrations, expressed as nanogram per milliliter human equivalent (HE), were determined in duplicate with a commercially available radio-immunoassay procedure (Linco Research, Missouri) using an antibody raised against human leptin which displayed 67% cross-reactivity to porcine leptin (Linco Research, Missouri). The same assay was used to study serum leptin concentrations in gilts [22]. The human and porcine cDNA and amino acid sequences have 88 and 83% homology, respectively [16]. The detection limit of the serum leptin assay was 1 ng/mL HE (Linco Research, Missouri) and the repeatability of the assay averaged 0.95 (SE 0.03) for the six sampling times. The phenotypic SD of the six leptin sample times were 0.38, 0.37, 0.49, 0.52, 0.48 and 0.54 with coefficients of variation (s/x) ¯ of 0.18, 0.18, 0.24, 0.24, 0.21 and 0.20, respectively. The NPY and leptin assays were both performed at Roslin Institute.
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
19
2.3. Statistical analysis All traits were analysed using the residual maximum likelihood (REML) algorithm of Genstat Committee [23]. The model for phase-fed animals included selection line, diet, sex and the selection line with diet interaction as fixed effects with litter and laboratory assay fitted as random effects. There were five laboratory assays of 240 samples measured in duplicate, given a total of 1220 blood samples, and to remove any experimental noise attributable to inter-laboratory assay variation, laboratory assay was included in the model as a random block effect. The selection line with diet interaction was examined with selection line and diet fitted as fixed effects and with diet fitted as a covariate. Covariate analyses determined if there were between-selection line differences in the mean performance, averaged over diets, and in the rate of change in a trait from changing diet. For diet-choice fed pigs, the model included selection line and sex as fixed effects with litter and laboratory assay as random effects. A log transformation was applied to serum NPY concentrations as the distributions of observed values were skewed. Throughout the text, all effects which are different at the 0.05 significance level are referred to as being significant. The diets differed in several aspects other than in the total lysine: energy ratio, but the diets were formulated to ensure that lysine was always the first-limiting amino acid. The data was analysed from the perspective of dietary differences in total lysine: energy ratio; an approach which has been used in previous studies [24,25]. 2.4. Correlations between traits REML estimates of the between- and within-litter variance components were used to determine upper limits of the heritability, as the between-litter variance component included the common-environmental variance. The experimental design maximized the number of full-sib litters to minimize the between-selection line variance and increase the precision of the study. The estimated phenotypic correlation matrices were examined using principal component analysis to determine linear functions of traits that describe the overall pattern of correlations between traits [26]. In a plot of coefficients of the two eigenvectors with largest eigenvalues, weighted by the square root of their eigenvalue, the cosine of the angle between two points represents the estimated correlation using only information from the two eigenvectors and eigenvalues. For example, points clustering together are highly correlated, points at right angles to each other are uncorrelated and points diametrically opposite are highly negatively correlated. The distance of a point from the origin is a measure of the information of a trait provided by the first and second eigenvectors and eigenvalues.
3. Results 3.1. Phase-fed animals Significant selection line differences were detected for log transformed serum NPY concentrations on all sampling occasions (Fig. 1). The high LGA line had higher NPY concentrations than the low LGA line and conversely the high LFC line had lower concentrations
20
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Fig. 1. Log transformed serum NPY concentrations for each selection line, averaged over diets, at the start and end of the three test-periods (SE of between-selection line differences were 0.11). There were 25 animals in each selection line. The high–low LFC and high–low LGA selection line differences were statistically significant (P < 0.05) at all sampling times.
than the low LFC line. In contrast, there was no difference in NPY concentrations between the DFI lines although the DFI lines and the low LFC line had higher concentrations than the low LGA line. Over the whole test-period, the between-selection line differences in log transformed NPY concentrations were relatively constant in magnitude (DFI: 0.2, LFC: −0.6 and LGA: 0.5 log(pmol/L)). The only significant change in NPY concentrations during a test-period was the greater reduction in log transformed NPY concentration in the high LFC line than in the low line (−0.30 versus 0.05, SED 0.12 log(pmol/L)) during the first test-period. The effect of diet on NPY concentrations was negligible at each sampling occasion (Fig. 2) with the range of estimated dietary means being less than the SED. There was no evidence of a selection line with diet interaction for log transformed NPY concentration when diet was included in the model either as a fixed effect or as a covariate. Leptin concentrations were significantly higher in the high DFI line than in the low DFI line on all sampling occasions (Fig. 3). There was no significant difference between the high and low LGA lines on any of the sampling occasions, but the low LFC line had higher leptin concentrations than the high line during the third test-period. When averaged over selection lines, leptin concentrations were similar in the first and second test-periods with diet having no significant effect (Fig. 4), but during the third test-period leptin concentration increased with the magnitude of the change decreasing as the dietary lysine: energy increased (0.63,
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
21
Fig. 2. Log transformed serum NPY concentrations by diet group (e.g., diets R, Q and then P in periods 1, 2 and 3), averaged over selection lines, at the start and end of the three test-periods (SE of between-diet differences were 0.08). There were 30 animals performance tested on each diet combination. Within each sampling time, there were no statistically significant (P < 0.05) between-diet differences.
0.57, 0.50, 0.17, 0.24; SED 0.15 ng/mL HE). There was no evidence of selection line with diet interaction for leptin concentration when diet was included in the model either as a fixed effect or as a covariate. When selection line-diet subclass was included in the model as a random effect, there was no increase in the selection line-diet subclass variance components
Fig. 3. Serum leptin concentrations for each selection line, averaged over diets, at the start and end of the three test-periods (SE of between-selection line differences were 0.17). The high–low DFI selection line differences were statistically significant (P < 0.05) at all sampling times, as was the high–low LFC difference at the end of the third test-period (E3).
22
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Fig. 4. Serum leptin concentrations by diet group (e.g., diets R, Q and then P in periods 1, 2 and 3), averaged over selection lines, at the start and end of the three test-periods (SE of between-diet differences were 0.11). There were statistically significant (P < 0.05) between-diet differences at the end of the third test-period (E3) and the diet combination U–T–S resulted in a lower leptin concentration at the start of the second test-period (S2).
between the start and end of the each test-period which confirmed the lack of dietary effects on serum NPY and leptin concentrations. For example, the subclass variances for leptin concentration at the start and end of each test-period were 0.22 and 0.17, 0.15 and 0.20, 0.10 and 0.27 (SE 0.16 (ng/mL HE)). 3.2. Diet-choice animals Log transformed serum NPY concentration and serum leptin concentration of DFI and LGA selection line animals tested on the diet-choice procedure are presented in Fig. 5. As with phase-fed animals, there was no difference in NPY concentration between the high and low DFI lines during the performance test, while the high LGA line had higher NPY concentrations than the low line at start (30 kg) and end (90 kg) of test. The higher leptin concentrations of the high DFI line relative to the low DFI line and the similar high and low LGA line concentrations were consistent with the selection line comparisons of phase-fed animals. 3.3. Heritability estimates and phenotypic correlations between traits Heritability estimates derived from within-selection line between full-sib litter variation for log transformed serum NPY concentration and for serum leptin concentration are presented in Table 2. There was no pattern in heritability or phenotypic SD estimates due to test-period or a difference in parameters at the start and end of each test-period. The phenotypic correlation matrix for log transformed serum NPY concentration, leptin concentration, predicted protein and lipid weights, and lysine and energy intakes is
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
23
24
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Table 2 Estimated heritabilities (h2 ) and phenotypic SD (σ P ) for log transformed serum NPY concentration and serum leptin concentration for animals tested with phase-feeding Sample taken at
Start test-period 1 End test-period 1 Start test-period 2 End test-period 2 Start test-period 2 End test-period 3
Log(NPY) (pmol/L)
Leptin (ng/mL HE)
h2
SE
σP
h2
SE
σP
0.19 0.22 0.13 0.51 0.39 0.26
0.18 0.18 0.16 0.20 0.19 0.18
0.30 0.30 0.28 0.27 0.32 0.34
0.05 0.34 0.46 0.10 0.30 0.42
0.15 0.19 0.20 0.16 0.18 0.19
0.38 0.37 0.49 0.52 0.48 0.54
represented graphically in Fig. 6. The first two eigenvectors proportionally accounted for 0.27 and 0.15 of the total information on the relationships between traits. The linear combinations of traits contrasted predicted protein weight with predicted lipid weight, with log transformed NPY concentrations forming a perpendicular group to the axis defined by protein and lipid weights. The three groups of traits: serum NPY, predicted protein and lipid weights illustrate that traits clustering together are highly correlated, traits at right angles to each other are uncorrelated and traits diametrically opposite are highly negatively correlated. Successive NPY concentrations were more highly correlated than successive leptin
Fig. 6. The eigenvector with the largest eigenvalue (X-axis) plotted against the eigenvector with the second largest eigenvalue (Y-axis), with the eigenvectors weighted by the square root of their eigenvalues, of the correlation matrix for phase-fed animals (npyi and lepi: log transformed NPY concentrations and leptin concentrations, lwti and pwti predicted lipid and protein weights at the start (i = 1, 3, 5) and end (i = 2, 4 ,6) of each test-period; eini and lini: energy and lysine intake in test-period i).
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
25
concentrations (0.80 versus 0.31, SED 0.11) as reflected by the tighter clustering of NPY compared to leptin concentrations in Fig. 6. Correlations between NPY and leptin concentrations were less than 0.20, in magnitude, with an average of 0.06 (SE 0.08). Leptin was positively correlated with predicted lipid weight with the correlation increasing with stage of test (0.13, 0.13, 0.34, 0.24, 0.43 and 0.47). In contrast, predicted protein weight was not significantly correlated with NPY (0.02) but was negatively correlated with leptin (−0.20) concentration. Energy and lysine intakes were not significantly correlated with either NPY or leptin concentrations during each test-period. The third eigenvector proportionally accounted for 0.11 of the total information on the relationships between traits and consisted of predicted protein weight, energy and lysine intakes. Therefore a three-dimensional representation of the correlation matrix would have predicted protein and lipid weight forming one axis, NPY concentration the second axis with lysine and energy intakes forming the third axis. The “clear” symbol for lysine and energy intakes in Fig. 6 indicated that the points lay in a different plane. The position of leptin concentrations and energy intake in Fig. 6 was a function of collapsing a three-dimensional representation into two dimensions.
4. Discussion The study detected genetic but not nutritional effects on log transformed serum NPY and serum leptin concentrations. Interpretation of the physiological responses to selection for components of efficient lean growth should be considered in conjunction with the selection objectives and correlated responses in growth and carcass traits. The DFI lines were selected for high and low DFI, while the selection objectives in the LGA and LFC lines were to obtain equal correlated responses, measured in phenotypic SD units, in carcass lean content with growth rate and carcass lean content with food conversion ratio, respectively. Estimated responses to selection in growth and predicted carcass traits during the third test-period are given in Table 3. Selection on DFI altered food intake and fat deposition, while selection on LFC changed protein and lipid deposition in equal but opposite directions, but there Table 3 Estimated means of high and low DFI, LFC and LGA selection lines for performance traits during the third test-period, predicted protein and lipid weights, log transformed serum NPY and serum leptin concentrations at the end of the third test-period DFIa
Food intake (g per day) Growth rate (g per day) Food conversion ratio Predicted protein weight (kg) Predicted lipid weight (kg) Log(serum NPY) Serum leptin a b
LFCa
LGAa
SED
High
Low
High
Low
High
Low
2826 889 3.24 13.6 23.8 6.26 3.06
2337b
2225 708 3.20 13.7 18.9 5.72 2.46
2207 619 3.57b 12.7b 21.5b 6.31b 3.04b
2673 896 3.09 14.3 20.3 5.80 2.70
2426b 715b 3.49 13.0b 21.9b 5.37b 2.74
861 2.75b 13.7 20.5b 6.14 2.45b
There were 25 animals within each selection line. High–low selection line differences were statistically significant (P < 0.05).
104 56 0.16 0.2 0.5 0.09 0.21
26
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
were responses in food intake, growth rate, protein and lipid deposition with selection on LGA. The three selection strategies with different emphases on components of efficient lean growth resulted in different combinations of nutrient intake, protein and lipid deposition rates [14,20]. The response in serum leptin concentration to selection on DFI was consistent with the observations in ob/ob mice [5] and human [11] studies that high fat deposition was associated with increased leptin-resistance. The higher food intake of the high DFI line was possibly due to decreased sensitivity to leptin, as secreted by the adipocyte. The DFI lines had not only similar serum NPY concentrations but had among the highest NPY concentrations of the selection lines. Plasma leptin concentrations differed between obese and normal women, that differed in body fat weight (45 versus 16 kg), but not in lean weight (42 versus 40 kg) but not plasma NPY nor CSF NPY concentrations [10]. It was proposed that “prior sustained stable adiposity may uncouple (the leptin–NPY) mechanism of homeostatic regulation of body weight” [10]. Ad libitum feeding of pigs in the current study may have enabled attainment and sustentation of genetically determined body fat such that the hypothesized regulation of NPY by leptin was not observed at the serum level. Responses in serum leptin concentration with selection on DFI and LFC were consistent with the greater lipid deposition of the “fat” genotypes [3]. However, the smaller difference in predicted lipid weight between the LGA lines than in the DFI and LFC lines (1.6 versus 3.3 and 2.6, SED 0.5 kg) may not have been sufficient to detect a statistically significant response in serum leptin concentration. The responses in serum NPY concentration of similar magnitude but of opposite sign with selection on LFC and LGA were contrary to expectation, given the different responses in food intake but responses in predicted protein and lipid weights of the same direction. The high correlations between successive serum NPY concentrations suggest that changes in body composition with time are unrelated to serum NPY concentration. Similarly, the low correlations of log transformed serum NPY concentration with energy or lysine intake implied that the value of serum NPY concentration to provide information on the responses to selection at a physiological level or as a physiological predictor of genetic merit, in an animal breeding framework, were limited. The response in leptin concentrations to the three selection strategies indicate that knowledge of the selection criterion or history is required to interpret the correlated responses in physiological traits. For example, the DFI and LGA lines represent two sets of lean and fat pigs derived from the one base population, but in one set the lean line had higher serum NPY concentrations than the fat line but no difference in serum leptin concentrations and vice versa for the other set of lines. Higher leptin mRNA levels in fat than in lean pigs were detected [16,27] but, as the genotypes were derived from different genetic populations, the difference in leptin concentrations may not have been solely due to differences in fat deposition. The significant responses in serum NPY and leptin concentrations to appropriate selection strategies and the moderate heritabilities indicate that there is substantial genetic covariance between serum NPY and leptin with components of efficient lean growth in pigs. Therefore, further study is merited for interpretation on serum NPY concentration and for inclusion in selection criteria of serum leptin concentration as a physiological predictor of genetic merit for fat deposition.
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
27
Incorporation of animals tested with the diet-choice procedure confirmed the result from phase-fed animals that serum NPY and, to a lesser extent, serum leptin concentrations were insensitive to dietary differences in total lysine: energy. In particular, the substantial between-selection line variation in log transformed serum NPY concentrations relative to the between-diet variation indicated that studies using a genetic resource population of animals may be more powerful than nutritional studies using isoenergetic diets differing in lysine content to examine aspects of function and expression of NPY and leptin. For example, no difference in plasma leptin concentration between sheep given high and medium supplementation was detected [28], despite significant differences in ultrasonic backfat depth. The lack of a significance difference in serum leptin concentration between the high and low LGA lines, despite differences in food intake, growth rate and predicted protein and lipid weights indicates that the choice of selection strategy is important when using selected lines of animals for physiological studies. The heritability and mean serum NPY concentration did not change substantially over the three test-periods suggesting that if serum NPY concentration was used as a genetic predictor of some aspect of food intake, then an early measurement would be sufficient, assuming that the genetic correlation with the trait of interest did not diminish with time. Serum leptin concentrations remained relatively constant during the first and second test-period, but the increase during the third test-period was associated with a larger phenotypic SD and marginally higher heritability. The change in variance, at both phenotypic and genetic levels, in serum leptin concentration was associated with an increase in the correlation between serum leptin concentration and predicted lipid weight over time, from 0.13 at the start of test to 0.47 at the end of test. Therefore, if serum leptin concentration was incorporated in a selection criterion as an indicator of genetic merit for fat deposition or food intake, then measurement at a later weight, such at 90 kg, would be required. In summary, the study detected genetic but not nutritional effects on log transformed serum NPY and serum leptin concentrations. Serum NPY concentrations were increased with selection for low food conversion ratio (LFC) or for high lean growth rate (LGA) but not with divergent selection on DFI. However, selection for high DFI was associated with increased serum leptin concentration as was selection for low LFC, but there was no response with divergent selection for LGA. The response in serum leptin concentration to selection on DFI was consistent with the observations in ob/ob mice and human studies that high fat deposition was associated with increased leptin-resistance and the correlation between leptin and predicted lipid weight increased with stage of test. The high correlations between successive serum NPY concentrations between 30 and 90 kg suggest that changes in body composition would not be reflected by serum NPY concentrations. The low correlations of serum NPY concentration with energy or lysine intake implied that the value of serum NPY concentration to provide information on the responses to selection at a physiological level or as a physiological predictor of genetic merit, in an animal breeding framework, were limited. Serum NPY and, to a lesser extent, serum leptin concentrations were insensitive to dietary differences in total lysine: energy and indicated that studies using a genetic resource population of animals may be more powerful than nutritional studies using isoenergetic diets differing in lysine content to examine aspects of function and expression of NPY and leptin in pigs.
28
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Acknowledgments The project was funded by the Ministry of Agriculture, Fisheries and Food, now Department for Environment, Food and Rural Affairs. The diets were designed by Philip Boyd and manufactured by Cranswick Mill Ltd.
References [1] Kalra SP, Dube MG, Pu S, Xu B, Horvath TL, Kalra PS. Interacting appetite-regulating pathways in the hypothalamic regulation of body weight. Endocr Rev 1999;20:68–100. [2] Woods SC, Seeley RJ, Porte D, Schwartz MW. Signals that regulate food intake and energy homeostasis. Science 1998;280:1378–83. [3] Houseknecht KL, Baile CA, Matteri RL, Spurlock ME. The biology of leptin: a review. J Anim Sci 1998;76:1405–20. [4] Houseknecht KL, Portocarrero CP. Leptin and its receptors: regulators of whole-body energy homeostasis. Domest Anim Endocrinol 1998;15:457–75. [5] Schwartz MW, Baskin DG, Bukowski TR, Kujiper JL, Foster D, Lasser G, Prunkard DE, Porte D, Woods SC, Seeley RJ, Weigle DS. Specificity of leptin action on elevated blood glucose levels and hypothalamic neuropeptide Y gene expression in ob/ob mice. Diabetes 1996;45:531–5. [6] Martinez JA, Aguado M, Fruhbeck G. Interactions between leptin and NPY affecting lipid mobilization in adipose tissue. J Physiol Biochem 2000;56:1–7. [7] Schwartz MW, Peskind E, Raskind M, Boyko EJ, Porte D. Cerebrospinal fluid leptin levels: relationship to plasma levels and adiposity in humans. Nat Med 1996;2:589–93. [8] Dötsch J, Adelmann M, Englaro P, Dötsch A, Hänze J, Blum WF, Kiess W, Rascher W. Relation of leptin and neuropeptide Y in human blood and cerebrospinal fluid. J Neurol Sci 1997;151:185–8. [9] Mantzoros C, Flier JS, Lesem MD, Brewerton TD, Jimerson DC. Cerebrospinal fluid leptin in anorexia nervosa: correlation with nutritional status and potential role in resistance to weight gain. J Clin Endocrinol Metab 1997;82:1845–51. [10] Nam S, Kratzsch J, Kim KW, Kim KR, Lim S, Marcus C. Cerebrospinal fluid and plasma concentrations of leptin, NPY, and ␣-MSH in obese women and their relationship to negative energy balance. J Clin Endocrinol Metab 2001;86:4849–53. [11] Baranowska B, Wolinska-Witort E, Wasilewska-Dziubinska E, Roguski K, Chmielowska M. Plasma leptin, neuropeptide Y (NPY) and galanin concentrations in bulimia nervosa and anorexia nervosa. Neuroendocrinol Lett 2001;22:356–8. [12] Maffei M, Halaas J, Ravussin E, Pratley RE, Lee GH, Zhang Y, Fei H, Kim S, Lallone R, Ranganathan S, Kern PA, Friedman JM. Leptin levels in human and rodent: measurement of serum leptin and ob RNA in obese and weight-reduced subjects. Nat Med 1995;1:1155–61. [13] Cameron ND. Selection for components of efficient lean growth rate in pigs. 1. Selection pressure applied and direct responses in a Large White herd. Anim Prod 1994;59:251–62. [14] Cameron ND, Curran MK. Responses in carcass composition to divergent selection for components of efficient lean growth rate in pigs. Anim Sci 1995;61:347–59. [15] Spurlock ME, Frank GR, Cornelius SG, Ji S, Willis GM, Bidwell CA. Obese gene expression in porcine adipose tissue is reduced by food deprivation but not by maintenance or submaintenance food intake. J Nutr 1998;128:677–82. [16] Ramsay TG, Yan X, Morrison C. The obesity gene in swine: sequence and expression of porcine leptin. J Anim Sci 1998;76:484–90. [17] Barb CR, Hausman GJ, Houseknecht KL. Biology of leptin in the pig. Domest Anim Endocrinol 2001;21:297– 317. [18] Cameron ND, Garth GB, Penman JC, Fiskin A. Genotype sensitivity to dietary lysine content in growing pigs. Anim Sci 2002, submitted for publication. [19] Bradford MMV, Gous RM. A comparison of phase-feeding and choice feeding as methods of meeting the amino acid requirements of growing pigs. Anim Prod 1991;52:323–30.
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
29
[20] Cameron ND. Genotype with nutrition interaction for protein and lipid deposition in pigs. In: Proceedings of the BSAS Annual Meeting, Scarborough, Paper no. 20, 2000. [21] Corder R, Emson PC, Lowry PJ. Purification and characterization of human neuropeptide Y from adrenal-medullary phaeochromocytoma tissue. Biochem J 1984;219:699–706. [22] Barb CR, Barrett JB, Kraeling RR, Rampacek GB. Serum leptin concentrations, luteinizing hormone and growth hormone secretion during feed and metabolic fuel restriction in the prepubertal gilt. Domest Anim Endocrinol 2001;20:47–63. [23] Genstat 5.3 Committee. Genstat 5.3 reference manual. Oxford: Clarendon Press, 1993. [24] Rao DS, McCracken KJ. Protein requirements of boars of high genetic potential for lean growth. Anim Prod 1990;51:179–87. [25] Van Lunen TA, Cole DJA. The effect of lysine/digestible energy ratio on growth performance and nitrogen deposition of hybrid boars, gilts and castrated male pigs. Anim Sci 1996;63:465–75. [26] Rao CR. Linear statistical inference and its applications. 2nd ed. New York: Wiley, 1973. [27] Robert C, Palin M-F, Coulombe N, Roberge C, Silversides FE, Benkel BF, McKay RM, Pelletier G. Backfat thickness in pigs is positively associated with leptin mRNA levels. Can J Anim Sci 1998;78:473–82. [28] Blanche D, Tellam RL, Chagas LM, Blackberry MA, Vercoe PE, Martin GB. Level of nutrition affects leptin concentrations in plasma and cerebrospinal fluid in sheep. J Endocrinol 2000;165:625–37.
Serum neuropeptide Y (NPY) and leptin concentrations in pigs selected for components of efficient lean growth N.D. Cameron∗ , E. McCullough, K. Troup, J.C. Penman Roslin Institute, Roslin, Midlothian, EH25 9PS, Scotland, UK Received 15 April 2002; accepted 6 June 2002
Abstract Responses in log transformed serum neuropeptide Y (NPY) concentration and leptin concentration after six generations of divergent selection on components of efficient lean growth in pigs were measured. From an animal breeding perspective, serum NPY and/or leptin concentrations could be used as physiological predictors of genetic merit if there were significant responses to selection. At 90 kg liveweight, log transformed serum NPY concentrations were increased with divergent selection for low food conversion ratio (LFC) (6.31 versus 5.72, SED 0.09 log(pmol/L)) or for high lean growth rate (LGA) (5.80 versus 5.37 log(pmol/L)) but not with selection on daily food intake (DFI) (6.26 versus 6.14 log(pmol/L)). Selection for high DFI was associated with increased serum leptin concentration (3.06 versus 2.45, SED 0.21 ng/mL human equivalent (HE)) as was selection for low LFC (3.04 versus 2.46 ng/mL HE). Correlations between leptin and predicted lipid weight increased with stage of test (0.13, 0.34 and 0.43, SE 0.08 at 30, 50 and 75 kg). The high correlations between successive serum NPY concentrations (0.80, SE 0.11) suggest that changes in body composition with time would not be reflected in serum NPY concentrations. Serum NPY and, to a lesser extent, serum leptin concentrations were insensitive to dietary differences in total lysine: energy and indicated that studies using a genetic resource population of animals may be more powerful than nutritional studies using isoenergetic diets differing in lysine content to examine aspects of function of serum concentrations of NPY and leptin in pigs. © 2002 Elsevier Science Inc. All rights reserved.
1. Introduction Neuropeptide Y (NPY) and leptin have been respectively associated with stimulating and inhibiting appetite [1]. NPY is a neurotransmitter expressed in the hypothalamus with a ∗
Corresponding author. Tel.: +44-131-527-4258; fax: +44-131-440-0434. E-mail address: [email protected] (N.D. Cameron). 0739-7240/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 9 - 7 2 4 0 ( 0 2 ) 0 0 1 8 4 - 4
16
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
role in energy homeostasis, such that NPY administration in rats can increase energy intake, decrease energy expenditure and increase lipogenesis [2]. In contrast, leptin is synthesised and secreted by the adipocytes into the blood stream [3] and transported to the brain where it stimulates gene expression of catabolic neuropeptides and inhibits that of anabolic neuropeptides, such as NPY [2,4]. The interaction between NPY and leptin is demonstrated in leptin-deficient ob/ob mice and leptin-resistant (lack of sensitivity to leptin) db/db mice as the over-expression of NPY in ob/ob mice can be controlled by leptin administration but not in db/db mice [5]. An interaction between leptin and NPY affecting lipid mobilization has also been demonstrated at the adipocyte level [6]. High correlations between leptin concentrations in human cerebrospinal fluid (CSF) and plasma have been reported [7–10]. Obese women had higher plasma leptin concentrations than controls [10,11] that was consistent with the proposal that human obesity may be associated with leptin-resistance [12]. In contrast, no correlation between human CSF and plasma NPY concentrations was detected and there was no difference between obese and control women in CSF and plasma NPY concentrations [10]. However, higher plasma NPY concentrations in bulimia nervosa women were detected than in anorexia nervosa women [11], that may have resulted from the increased experimental power due to the use of “extreme” phenotypes. “Extreme” pig genotypes have been generated in the Edinburgh lean growth experiment by several generations of divergent selection for components of efficient lean growth rate [13]. From an animal breeding perspective, it would be useful if plasma leptin and/or NPY concentrations could be used as physiological predictors of genetic merit and informative if they provided a physiological explanation of the observed responses in growth and carcass traits [14]. For example, were the responses in carcass fat content consistent with increased leptin-resistance, as in humans and db/db mice, or with leptin-deficiency, as in ob/ob mice ? Higher leptin mRNA and serum concentrations have been reported in pigs with greater backfat depths [15,16], but there is an absence of information concerning the relationship between nutrient intake and leptin concentrations [17] and, secondly, serum NPY concentrations in pigs have not been previously reported. Therefore, the current study was superimposed on a genotype-nutrition interaction experiment [18] to measure the responses in serum leptin and NPY concentrations to selection for components of efficient lean growth and determine the sensitivity of each genotype, in terms of serum leptin and NPY concentrations, to nutritional inputs.
2. Materials and methods 2.1. Animals and performance test Details on establishment of the Large White pig population and seven generations of divergent selection for daily food intake (DFI), lean food conversion (LFC) and lean growth rate (LGA) are available [13]. Each selection line consisted of progeny from 10 sires with two dams mated to each sire and the selection lines were derived from the same base population. In the current study, there were five litters of five full-sibs from each of the six
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
17
Table 1 Composition and specification of diets (g/kg) Diet
P
Q
R
S
T
U
X
Barley Wheat Wheatfeed Hi-pro soya Full fat soya Pellet binder Fat (mixer and spray) Chilean fishmeal Vitaminsa Protein Total lysine Lysine: DE
293.5 429 114.5 0 89 12.5 31 0 30.5 135 8.14 0.58
199 434 150 86 61 12.5 28 0 29.5 161 9.66 0.69
192 385.7 150 203.5 0 12.5 28 0 28.3 187 11.30 0.81
175.5 355.5 149 256.5 0 12.5 22.5 0 28.5 207 12.73 0.91
178 347 150.5 257.5 0 12.5 22 0 32.5 210 14.16 1.01
200 292.5 161 257.5 0 12.5 23 18.5 35 222 15.75 1.12
200 277 160 260 0 12.5 15 47 28.5 240 17.20 1.23
a
Vitamins and minerals, etc.
selection lines with one litter from five sires chosen at random. Pigs were performance tested on a phase-feeding regime [19] with each full-sib ad libitum fed one of five isoenergetic (14.0 MJ DE/kg) diets differing in total lysine: energy (Table 1) with diets R, S, T, U or X, then Q, R, S, T or U and finally P, Q, R, S or T offered in three test-periods. For example, the first sequence of diets fed during the three test-periods would be R then Q and then P. At the start of the performance test, a full-sib within each litter was allocated to each of five diets, such that within a selection line, five animals were performance tested on each of the five diet combinations. Each litter consisted of two boars and three gilts or three boars and two gilts with consideration of sex for allocation of animals to diets. The intended start weights for each test-period, lasting 14 days, were 30 ± 3, 50 ± 4 and 75 ± 5 kg (i.e., ranges of 27–33, 46–54 and 70–80 kg) with realized mean (SD) weights of 29.3 (1.6), 49.5 (2.1) and 73.5 (2.1) kg. Pigs were penned individually, and prior to the start of test and between test-periods, pigs were fed diet S. There were no diet effects or between-line within-selection group effects on the time taken to reach the required start weight for second and third test-periods, which averaged 13.9 (SD 4.1) and 15.5 (SD 4.6) days. Although, the interval between test-periods 2 and 3 was longer for the low LFC line than for other selection lines (18.2 versus 15.0, SED 1.7 days). The different patterns of NPY and leptin concentrations over time due to the selection lines and diets were interpretable on a consistent basis, given the relatively constant time-interval between blood sampling occasions. An additional 20 pigs from each of the high and low DFI and LGA lines were performance tested with the diet-choice procedure [19]. Animals had ad libitum access to two diets (Q and U) in separate food hoppers with start and end of test of 30 ± 3 and 90 ± 5 kg (i.e., ranges of 27–33 and 85–95 kg). Food intake of diets Q and U were measured mid-test at 50 ±4 kg and at 75±5 kg and at the end of test. In practice, the mean (SD) liveweights for measurements were 29.6 (1.9), 49.3 (2.1), 73.1 (2.4) and 88.4 (2.6) kg. The additional 20 animals formed 10 pairs of full-sibs, a boar and a gilt, of which five pairs were full-sibs of animals performance tested with phase-feeding and five pairs were half-sibs of those tested with phase-feeding. Further details of the performance test procedure with phase-feeding and diet-choice are available [18].
18
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Protein and lipid weights at different stages of the performance test were predicted using equations derived from chemical composition studies of carcasses in an earlier generation [20]. Briefly, six litters of five full-sibs in each selection line were slaughtered at 30, 45, 60, 75 or 90 ± 3 kg, with a full-sib in each litter allocated at random to each slaughter weight, for subsequent chemical analysis of carcass and non-carcass components. For the purpose of inputs to the nutrient utilisation equations in the current study, predicted protein and lipid weights were based on liveweight and ultrasonic backfat depth. Prediction equations for lipid and protein weights were estimated as 0.165 W + 0.544 BF and 0.176 W − 0.117 BF, where W and BF were the appropriate liveweight and ultrasonic backfat depth, with the prediction equations proportionately accounting for 0.91 and 0.98 of the variation in lipid and protein weights. 2.2. Serum NPY and leptin assays Prior to each performance test measurement of phase-fed and diet-choice animals, 10 mL blood samples were withdrawn by jugular venepuncture from the vena cava, kept at −4◦ C for 24 h, then centrifuged for 30 min at 4◦ C. The red blood cell-free serum was decanted and frozen at −20◦ C until required for assay. Blood sampling was performed from 9:00 h each day and food was available to animals at all times due to the ad libitum feeding procedure. Animals were sampled under the same feeding procedure, although between-selection line variation on feeding patterns may have influenced serum NPY and leptin concentrations. Serum NPY concentrations, expressed as picomole per liter, were determined in duplicate with a commercially available competitive radio-immunoassay procedure (EuroDiagnostica, The Netherlands) using an antiserum raised against synthetic NPY. NPY consists of 36 amino acids with the human and porcine NPY amino acid sequences only differing by replacement of leucine with methione at position 17 [21]. The detection limit of the assay was 6 pmol/L (Euro-Diagnostica, The Netherlands) and the repeatability of the NPY assay in the current study, calculated from the between- and within-animal variance components from data on all 230 animals, averaged 0.92 (SE 0.04) for the six sampling times. The phenotypic SD of the six log transformed NPY sample times were 0.30, 0.30, 0.28, 0.27, 0.32, and 0.44 with an average coefficient of variation (s/x) ¯ of 0.050. Quality controls were performed in duplicate at the start and end of each assay, but with only five assays estimation of inter-assay coefficient of variation was imprecise. Serum leptin concentrations, expressed as nanogram per milliliter human equivalent (HE), were determined in duplicate with a commercially available radio-immunoassay procedure (Linco Research, Missouri) using an antibody raised against human leptin which displayed 67% cross-reactivity to porcine leptin (Linco Research, Missouri). The same assay was used to study serum leptin concentrations in gilts [22]. The human and porcine cDNA and amino acid sequences have 88 and 83% homology, respectively [16]. The detection limit of the serum leptin assay was 1 ng/mL HE (Linco Research, Missouri) and the repeatability of the assay averaged 0.95 (SE 0.03) for the six sampling times. The phenotypic SD of the six leptin sample times were 0.38, 0.37, 0.49, 0.52, 0.48 and 0.54 with coefficients of variation (s/x) ¯ of 0.18, 0.18, 0.24, 0.24, 0.21 and 0.20, respectively. The NPY and leptin assays were both performed at Roslin Institute.
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
19
2.3. Statistical analysis All traits were analysed using the residual maximum likelihood (REML) algorithm of Genstat Committee [23]. The model for phase-fed animals included selection line, diet, sex and the selection line with diet interaction as fixed effects with litter and laboratory assay fitted as random effects. There were five laboratory assays of 240 samples measured in duplicate, given a total of 1220 blood samples, and to remove any experimental noise attributable to inter-laboratory assay variation, laboratory assay was included in the model as a random block effect. The selection line with diet interaction was examined with selection line and diet fitted as fixed effects and with diet fitted as a covariate. Covariate analyses determined if there were between-selection line differences in the mean performance, averaged over diets, and in the rate of change in a trait from changing diet. For diet-choice fed pigs, the model included selection line and sex as fixed effects with litter and laboratory assay as random effects. A log transformation was applied to serum NPY concentrations as the distributions of observed values were skewed. Throughout the text, all effects which are different at the 0.05 significance level are referred to as being significant. The diets differed in several aspects other than in the total lysine: energy ratio, but the diets were formulated to ensure that lysine was always the first-limiting amino acid. The data was analysed from the perspective of dietary differences in total lysine: energy ratio; an approach which has been used in previous studies [24,25]. 2.4. Correlations between traits REML estimates of the between- and within-litter variance components were used to determine upper limits of the heritability, as the between-litter variance component included the common-environmental variance. The experimental design maximized the number of full-sib litters to minimize the between-selection line variance and increase the precision of the study. The estimated phenotypic correlation matrices were examined using principal component analysis to determine linear functions of traits that describe the overall pattern of correlations between traits [26]. In a plot of coefficients of the two eigenvectors with largest eigenvalues, weighted by the square root of their eigenvalue, the cosine of the angle between two points represents the estimated correlation using only information from the two eigenvectors and eigenvalues. For example, points clustering together are highly correlated, points at right angles to each other are uncorrelated and points diametrically opposite are highly negatively correlated. The distance of a point from the origin is a measure of the information of a trait provided by the first and second eigenvectors and eigenvalues.
3. Results 3.1. Phase-fed animals Significant selection line differences were detected for log transformed serum NPY concentrations on all sampling occasions (Fig. 1). The high LGA line had higher NPY concentrations than the low LGA line and conversely the high LFC line had lower concentrations
20
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Fig. 1. Log transformed serum NPY concentrations for each selection line, averaged over diets, at the start and end of the three test-periods (SE of between-selection line differences were 0.11). There were 25 animals in each selection line. The high–low LFC and high–low LGA selection line differences were statistically significant (P < 0.05) at all sampling times.
than the low LFC line. In contrast, there was no difference in NPY concentrations between the DFI lines although the DFI lines and the low LFC line had higher concentrations than the low LGA line. Over the whole test-period, the between-selection line differences in log transformed NPY concentrations were relatively constant in magnitude (DFI: 0.2, LFC: −0.6 and LGA: 0.5 log(pmol/L)). The only significant change in NPY concentrations during a test-period was the greater reduction in log transformed NPY concentration in the high LFC line than in the low line (−0.30 versus 0.05, SED 0.12 log(pmol/L)) during the first test-period. The effect of diet on NPY concentrations was negligible at each sampling occasion (Fig. 2) with the range of estimated dietary means being less than the SED. There was no evidence of a selection line with diet interaction for log transformed NPY concentration when diet was included in the model either as a fixed effect or as a covariate. Leptin concentrations were significantly higher in the high DFI line than in the low DFI line on all sampling occasions (Fig. 3). There was no significant difference between the high and low LGA lines on any of the sampling occasions, but the low LFC line had higher leptin concentrations than the high line during the third test-period. When averaged over selection lines, leptin concentrations were similar in the first and second test-periods with diet having no significant effect (Fig. 4), but during the third test-period leptin concentration increased with the magnitude of the change decreasing as the dietary lysine: energy increased (0.63,
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
21
Fig. 2. Log transformed serum NPY concentrations by diet group (e.g., diets R, Q and then P in periods 1, 2 and 3), averaged over selection lines, at the start and end of the three test-periods (SE of between-diet differences were 0.08). There were 30 animals performance tested on each diet combination. Within each sampling time, there were no statistically significant (P < 0.05) between-diet differences.
0.57, 0.50, 0.17, 0.24; SED 0.15 ng/mL HE). There was no evidence of selection line with diet interaction for leptin concentration when diet was included in the model either as a fixed effect or as a covariate. When selection line-diet subclass was included in the model as a random effect, there was no increase in the selection line-diet subclass variance components
Fig. 3. Serum leptin concentrations for each selection line, averaged over diets, at the start and end of the three test-periods (SE of between-selection line differences were 0.17). The high–low DFI selection line differences were statistically significant (P < 0.05) at all sampling times, as was the high–low LFC difference at the end of the third test-period (E3).
22
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Fig. 4. Serum leptin concentrations by diet group (e.g., diets R, Q and then P in periods 1, 2 and 3), averaged over selection lines, at the start and end of the three test-periods (SE of between-diet differences were 0.11). There were statistically significant (P < 0.05) between-diet differences at the end of the third test-period (E3) and the diet combination U–T–S resulted in a lower leptin concentration at the start of the second test-period (S2).
between the start and end of the each test-period which confirmed the lack of dietary effects on serum NPY and leptin concentrations. For example, the subclass variances for leptin concentration at the start and end of each test-period were 0.22 and 0.17, 0.15 and 0.20, 0.10 and 0.27 (SE 0.16 (ng/mL HE)). 3.2. Diet-choice animals Log transformed serum NPY concentration and serum leptin concentration of DFI and LGA selection line animals tested on the diet-choice procedure are presented in Fig. 5. As with phase-fed animals, there was no difference in NPY concentration between the high and low DFI lines during the performance test, while the high LGA line had higher NPY concentrations than the low line at start (30 kg) and end (90 kg) of test. The higher leptin concentrations of the high DFI line relative to the low DFI line and the similar high and low LGA line concentrations were consistent with the selection line comparisons of phase-fed animals. 3.3. Heritability estimates and phenotypic correlations between traits Heritability estimates derived from within-selection line between full-sib litter variation for log transformed serum NPY concentration and for serum leptin concentration are presented in Table 2. There was no pattern in heritability or phenotypic SD estimates due to test-period or a difference in parameters at the start and end of each test-period. The phenotypic correlation matrix for log transformed serum NPY concentration, leptin concentration, predicted protein and lipid weights, and lysine and energy intakes is
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
23
24
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Table 2 Estimated heritabilities (h2 ) and phenotypic SD (σ P ) for log transformed serum NPY concentration and serum leptin concentration for animals tested with phase-feeding Sample taken at
Start test-period 1 End test-period 1 Start test-period 2 End test-period 2 Start test-period 2 End test-period 3
Log(NPY) (pmol/L)
Leptin (ng/mL HE)
h2
SE
σP
h2
SE
σP
0.19 0.22 0.13 0.51 0.39 0.26
0.18 0.18 0.16 0.20 0.19 0.18
0.30 0.30 0.28 0.27 0.32 0.34
0.05 0.34 0.46 0.10 0.30 0.42
0.15 0.19 0.20 0.16 0.18 0.19
0.38 0.37 0.49 0.52 0.48 0.54
represented graphically in Fig. 6. The first two eigenvectors proportionally accounted for 0.27 and 0.15 of the total information on the relationships between traits. The linear combinations of traits contrasted predicted protein weight with predicted lipid weight, with log transformed NPY concentrations forming a perpendicular group to the axis defined by protein and lipid weights. The three groups of traits: serum NPY, predicted protein and lipid weights illustrate that traits clustering together are highly correlated, traits at right angles to each other are uncorrelated and traits diametrically opposite are highly negatively correlated. Successive NPY concentrations were more highly correlated than successive leptin
Fig. 6. The eigenvector with the largest eigenvalue (X-axis) plotted against the eigenvector with the second largest eigenvalue (Y-axis), with the eigenvectors weighted by the square root of their eigenvalues, of the correlation matrix for phase-fed animals (npyi and lepi: log transformed NPY concentrations and leptin concentrations, lwti and pwti predicted lipid and protein weights at the start (i = 1, 3, 5) and end (i = 2, 4 ,6) of each test-period; eini and lini: energy and lysine intake in test-period i).
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
25
concentrations (0.80 versus 0.31, SED 0.11) as reflected by the tighter clustering of NPY compared to leptin concentrations in Fig. 6. Correlations between NPY and leptin concentrations were less than 0.20, in magnitude, with an average of 0.06 (SE 0.08). Leptin was positively correlated with predicted lipid weight with the correlation increasing with stage of test (0.13, 0.13, 0.34, 0.24, 0.43 and 0.47). In contrast, predicted protein weight was not significantly correlated with NPY (0.02) but was negatively correlated with leptin (−0.20) concentration. Energy and lysine intakes were not significantly correlated with either NPY or leptin concentrations during each test-period. The third eigenvector proportionally accounted for 0.11 of the total information on the relationships between traits and consisted of predicted protein weight, energy and lysine intakes. Therefore a three-dimensional representation of the correlation matrix would have predicted protein and lipid weight forming one axis, NPY concentration the second axis with lysine and energy intakes forming the third axis. The “clear” symbol for lysine and energy intakes in Fig. 6 indicated that the points lay in a different plane. The position of leptin concentrations and energy intake in Fig. 6 was a function of collapsing a three-dimensional representation into two dimensions.
4. Discussion The study detected genetic but not nutritional effects on log transformed serum NPY and serum leptin concentrations. Interpretation of the physiological responses to selection for components of efficient lean growth should be considered in conjunction with the selection objectives and correlated responses in growth and carcass traits. The DFI lines were selected for high and low DFI, while the selection objectives in the LGA and LFC lines were to obtain equal correlated responses, measured in phenotypic SD units, in carcass lean content with growth rate and carcass lean content with food conversion ratio, respectively. Estimated responses to selection in growth and predicted carcass traits during the third test-period are given in Table 3. Selection on DFI altered food intake and fat deposition, while selection on LFC changed protein and lipid deposition in equal but opposite directions, but there Table 3 Estimated means of high and low DFI, LFC and LGA selection lines for performance traits during the third test-period, predicted protein and lipid weights, log transformed serum NPY and serum leptin concentrations at the end of the third test-period DFIa
Food intake (g per day) Growth rate (g per day) Food conversion ratio Predicted protein weight (kg) Predicted lipid weight (kg) Log(serum NPY) Serum leptin a b
LFCa
LGAa
SED
High
Low
High
Low
High
Low
2826 889 3.24 13.6 23.8 6.26 3.06
2337b
2225 708 3.20 13.7 18.9 5.72 2.46
2207 619 3.57b 12.7b 21.5b 6.31b 3.04b
2673 896 3.09 14.3 20.3 5.80 2.70
2426b 715b 3.49 13.0b 21.9b 5.37b 2.74
861 2.75b 13.7 20.5b 6.14 2.45b
There were 25 animals within each selection line. High–low selection line differences were statistically significant (P < 0.05).
104 56 0.16 0.2 0.5 0.09 0.21
26
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
were responses in food intake, growth rate, protein and lipid deposition with selection on LGA. The three selection strategies with different emphases on components of efficient lean growth resulted in different combinations of nutrient intake, protein and lipid deposition rates [14,20]. The response in serum leptin concentration to selection on DFI was consistent with the observations in ob/ob mice [5] and human [11] studies that high fat deposition was associated with increased leptin-resistance. The higher food intake of the high DFI line was possibly due to decreased sensitivity to leptin, as secreted by the adipocyte. The DFI lines had not only similar serum NPY concentrations but had among the highest NPY concentrations of the selection lines. Plasma leptin concentrations differed between obese and normal women, that differed in body fat weight (45 versus 16 kg), but not in lean weight (42 versus 40 kg) but not plasma NPY nor CSF NPY concentrations [10]. It was proposed that “prior sustained stable adiposity may uncouple (the leptin–NPY) mechanism of homeostatic regulation of body weight” [10]. Ad libitum feeding of pigs in the current study may have enabled attainment and sustentation of genetically determined body fat such that the hypothesized regulation of NPY by leptin was not observed at the serum level. Responses in serum leptin concentration with selection on DFI and LFC were consistent with the greater lipid deposition of the “fat” genotypes [3]. However, the smaller difference in predicted lipid weight between the LGA lines than in the DFI and LFC lines (1.6 versus 3.3 and 2.6, SED 0.5 kg) may not have been sufficient to detect a statistically significant response in serum leptin concentration. The responses in serum NPY concentration of similar magnitude but of opposite sign with selection on LFC and LGA were contrary to expectation, given the different responses in food intake but responses in predicted protein and lipid weights of the same direction. The high correlations between successive serum NPY concentrations suggest that changes in body composition with time are unrelated to serum NPY concentration. Similarly, the low correlations of log transformed serum NPY concentration with energy or lysine intake implied that the value of serum NPY concentration to provide information on the responses to selection at a physiological level or as a physiological predictor of genetic merit, in an animal breeding framework, were limited. The response in leptin concentrations to the three selection strategies indicate that knowledge of the selection criterion or history is required to interpret the correlated responses in physiological traits. For example, the DFI and LGA lines represent two sets of lean and fat pigs derived from the one base population, but in one set the lean line had higher serum NPY concentrations than the fat line but no difference in serum leptin concentrations and vice versa for the other set of lines. Higher leptin mRNA levels in fat than in lean pigs were detected [16,27] but, as the genotypes were derived from different genetic populations, the difference in leptin concentrations may not have been solely due to differences in fat deposition. The significant responses in serum NPY and leptin concentrations to appropriate selection strategies and the moderate heritabilities indicate that there is substantial genetic covariance between serum NPY and leptin with components of efficient lean growth in pigs. Therefore, further study is merited for interpretation on serum NPY concentration and for inclusion in selection criteria of serum leptin concentration as a physiological predictor of genetic merit for fat deposition.
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
27
Incorporation of animals tested with the diet-choice procedure confirmed the result from phase-fed animals that serum NPY and, to a lesser extent, serum leptin concentrations were insensitive to dietary differences in total lysine: energy. In particular, the substantial between-selection line variation in log transformed serum NPY concentrations relative to the between-diet variation indicated that studies using a genetic resource population of animals may be more powerful than nutritional studies using isoenergetic diets differing in lysine content to examine aspects of function and expression of NPY and leptin. For example, no difference in plasma leptin concentration between sheep given high and medium supplementation was detected [28], despite significant differences in ultrasonic backfat depth. The lack of a significance difference in serum leptin concentration between the high and low LGA lines, despite differences in food intake, growth rate and predicted protein and lipid weights indicates that the choice of selection strategy is important when using selected lines of animals for physiological studies. The heritability and mean serum NPY concentration did not change substantially over the three test-periods suggesting that if serum NPY concentration was used as a genetic predictor of some aspect of food intake, then an early measurement would be sufficient, assuming that the genetic correlation with the trait of interest did not diminish with time. Serum leptin concentrations remained relatively constant during the first and second test-period, but the increase during the third test-period was associated with a larger phenotypic SD and marginally higher heritability. The change in variance, at both phenotypic and genetic levels, in serum leptin concentration was associated with an increase in the correlation between serum leptin concentration and predicted lipid weight over time, from 0.13 at the start of test to 0.47 at the end of test. Therefore, if serum leptin concentration was incorporated in a selection criterion as an indicator of genetic merit for fat deposition or food intake, then measurement at a later weight, such at 90 kg, would be required. In summary, the study detected genetic but not nutritional effects on log transformed serum NPY and serum leptin concentrations. Serum NPY concentrations were increased with selection for low food conversion ratio (LFC) or for high lean growth rate (LGA) but not with divergent selection on DFI. However, selection for high DFI was associated with increased serum leptin concentration as was selection for low LFC, but there was no response with divergent selection for LGA. The response in serum leptin concentration to selection on DFI was consistent with the observations in ob/ob mice and human studies that high fat deposition was associated with increased leptin-resistance and the correlation between leptin and predicted lipid weight increased with stage of test. The high correlations between successive serum NPY concentrations between 30 and 90 kg suggest that changes in body composition would not be reflected by serum NPY concentrations. The low correlations of serum NPY concentration with energy or lysine intake implied that the value of serum NPY concentration to provide information on the responses to selection at a physiological level or as a physiological predictor of genetic merit, in an animal breeding framework, were limited. Serum NPY and, to a lesser extent, serum leptin concentrations were insensitive to dietary differences in total lysine: energy and indicated that studies using a genetic resource population of animals may be more powerful than nutritional studies using isoenergetic diets differing in lysine content to examine aspects of function and expression of NPY and leptin in pigs.
28
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
Acknowledgments The project was funded by the Ministry of Agriculture, Fisheries and Food, now Department for Environment, Food and Rural Affairs. The diets were designed by Philip Boyd and manufactured by Cranswick Mill Ltd.
References [1] Kalra SP, Dube MG, Pu S, Xu B, Horvath TL, Kalra PS. Interacting appetite-regulating pathways in the hypothalamic regulation of body weight. Endocr Rev 1999;20:68–100. [2] Woods SC, Seeley RJ, Porte D, Schwartz MW. Signals that regulate food intake and energy homeostasis. Science 1998;280:1378–83. [3] Houseknecht KL, Baile CA, Matteri RL, Spurlock ME. The biology of leptin: a review. J Anim Sci 1998;76:1405–20. [4] Houseknecht KL, Portocarrero CP. Leptin and its receptors: regulators of whole-body energy homeostasis. Domest Anim Endocrinol 1998;15:457–75. [5] Schwartz MW, Baskin DG, Bukowski TR, Kujiper JL, Foster D, Lasser G, Prunkard DE, Porte D, Woods SC, Seeley RJ, Weigle DS. Specificity of leptin action on elevated blood glucose levels and hypothalamic neuropeptide Y gene expression in ob/ob mice. Diabetes 1996;45:531–5. [6] Martinez JA, Aguado M, Fruhbeck G. Interactions between leptin and NPY affecting lipid mobilization in adipose tissue. J Physiol Biochem 2000;56:1–7. [7] Schwartz MW, Peskind E, Raskind M, Boyko EJ, Porte D. Cerebrospinal fluid leptin levels: relationship to plasma levels and adiposity in humans. Nat Med 1996;2:589–93. [8] Dötsch J, Adelmann M, Englaro P, Dötsch A, Hänze J, Blum WF, Kiess W, Rascher W. Relation of leptin and neuropeptide Y in human blood and cerebrospinal fluid. J Neurol Sci 1997;151:185–8. [9] Mantzoros C, Flier JS, Lesem MD, Brewerton TD, Jimerson DC. Cerebrospinal fluid leptin in anorexia nervosa: correlation with nutritional status and potential role in resistance to weight gain. J Clin Endocrinol Metab 1997;82:1845–51. [10] Nam S, Kratzsch J, Kim KW, Kim KR, Lim S, Marcus C. Cerebrospinal fluid and plasma concentrations of leptin, NPY, and ␣-MSH in obese women and their relationship to negative energy balance. J Clin Endocrinol Metab 2001;86:4849–53. [11] Baranowska B, Wolinska-Witort E, Wasilewska-Dziubinska E, Roguski K, Chmielowska M. Plasma leptin, neuropeptide Y (NPY) and galanin concentrations in bulimia nervosa and anorexia nervosa. Neuroendocrinol Lett 2001;22:356–8. [12] Maffei M, Halaas J, Ravussin E, Pratley RE, Lee GH, Zhang Y, Fei H, Kim S, Lallone R, Ranganathan S, Kern PA, Friedman JM. Leptin levels in human and rodent: measurement of serum leptin and ob RNA in obese and weight-reduced subjects. Nat Med 1995;1:1155–61. [13] Cameron ND. Selection for components of efficient lean growth rate in pigs. 1. Selection pressure applied and direct responses in a Large White herd. Anim Prod 1994;59:251–62. [14] Cameron ND, Curran MK. Responses in carcass composition to divergent selection for components of efficient lean growth rate in pigs. Anim Sci 1995;61:347–59. [15] Spurlock ME, Frank GR, Cornelius SG, Ji S, Willis GM, Bidwell CA. Obese gene expression in porcine adipose tissue is reduced by food deprivation but not by maintenance or submaintenance food intake. J Nutr 1998;128:677–82. [16] Ramsay TG, Yan X, Morrison C. The obesity gene in swine: sequence and expression of porcine leptin. J Anim Sci 1998;76:484–90. [17] Barb CR, Hausman GJ, Houseknecht KL. Biology of leptin in the pig. Domest Anim Endocrinol 2001;21:297– 317. [18] Cameron ND, Garth GB, Penman JC, Fiskin A. Genotype sensitivity to dietary lysine content in growing pigs. Anim Sci 2002, submitted for publication. [19] Bradford MMV, Gous RM. A comparison of phase-feeding and choice feeding as methods of meeting the amino acid requirements of growing pigs. Anim Prod 1991;52:323–30.
N.D. Cameron et al. / Domestic Animal Endocrinology 24 (2003) 15–29
29
[20] Cameron ND. Genotype with nutrition interaction for protein and lipid deposition in pigs. In: Proceedings of the BSAS Annual Meeting, Scarborough, Paper no. 20, 2000. [21] Corder R, Emson PC, Lowry PJ. Purification and characterization of human neuropeptide Y from adrenal-medullary phaeochromocytoma tissue. Biochem J 1984;219:699–706. [22] Barb CR, Barrett JB, Kraeling RR, Rampacek GB. Serum leptin concentrations, luteinizing hormone and growth hormone secretion during feed and metabolic fuel restriction in the prepubertal gilt. Domest Anim Endocrinol 2001;20:47–63. [23] Genstat 5.3 Committee. Genstat 5.3 reference manual. Oxford: Clarendon Press, 1993. [24] Rao DS, McCracken KJ. Protein requirements of boars of high genetic potential for lean growth. Anim Prod 1990;51:179–87. [25] Van Lunen TA, Cole DJA. The effect of lysine/digestible energy ratio on growth performance and nitrogen deposition of hybrid boars, gilts and castrated male pigs. Anim Sci 1996;63:465–75. [26] Rao CR. Linear statistical inference and its applications. 2nd ed. New York: Wiley, 1973. [27] Robert C, Palin M-F, Coulombe N, Roberge C, Silversides FE, Benkel BF, McKay RM, Pelletier G. Backfat thickness in pigs is positively associated with leptin mRNA levels. Can J Anim Sci 1998;78:473–82. [28] Blanche D, Tellam RL, Chagas LM, Blackberry MA, Vercoe PE, Martin GB. Level of nutrition affects leptin concentrations in plasma and cerebrospinal fluid in sheep. J Endocrinol 2000;165:625–37.