The ability of animals to select diets best suited for their physiologic status has been examined in numerous species. For example, as rats age, their dietary protein requirement decreases. This is presumably associated with a decease in the rate of protein accretion and an overall decrease in growth rate as the animals reach maturity (Hartsook and Mitchell 1956). Similarly, growing pigs are able to select between high and low protein diets, and the dietary selection pattern changes with the changes in the relative amounts of protein and lipid accretion that occur as the animals age (Bradford and Gous 1992, Kyriazakis et al. 1990 ).

In addition to these age-associated changes in composition of gain and macronutrient selection, other factors, such as pregnancy, lactation and endocrine manipulations can alter macronutrient requirements. This work is focused on the effects of exogenous somatotropin on these variables. Pigs treated with somatotropin (STH)⁵ show an increase in protein accretion and a decrease in lipid accretion, with an overall decrease in feed intake (Boyd et al. 1991, Etherton et al. 1987, Lee et al. 1994). Pituitary intact rats treated with STH also show an increase in protein accretion, a decrease or no change in lipid accretion, but an increase in feed intake (Azain et al. 1995b, Grosbeck et al. 1987). We have recently demonstrated that the increase in intake by STH-treated rats can be accounted for by an increase in protein intake. Rats treated with STH and allowed to select between high and low protein diets select a greater amount of the high protein diet compared with non-STH treated rats (Roberts et al. 1995). If rats are offered a single diet, the net result of this signal is an increase in intake (Azain et al. 1995b). When rats are allowed to select, the result is specific for the high protein diet (Roberts et al. 1995).

Using a design similar to the previous rat study (Roberts et al. 1995), we wished to determine if the change in intake associated with STH treatment in pigs would alter diet selection. The present experiment examined the effect of STH treatment on feed intake and diet selection in pigs allowed to select between high and low protein diets. Because STH treatment of pigs reduces feed intake and lipid accretion, but increases protein accretion, we hypothesized that the decrease in intake would be accounted for by a reduced intake of the low protein diet.

MATERIALS AND METHODS

Diet composition is shown in Table 1. The most practical way to alter crude protein is to blend ingredients with different protein contents to achieve the desired dietary concentration. However, because the amino acid pattern of ingredients such as corn and soybean meal differs, changing the ratio of corn to soybean meal not only alters crude protein but also disrupts the pattern or relative proportion of amino acids to each other. In this study, cornstarch was used to alter dietary protein concentration, and the ratio of corn to soybean meal was the same in both diets. In this way, the protein content was altered (by the use of starch) without affecting the relative amino acid profile.

Table 1. Diet composition [View Table]

Castrated male pigs (n = 16, initial weight 69 kg) used in this study were the progeny of Yorkshire × Chester dams and Hampshire or Hampshire × Duroc sires and were from the University of Georgia herd. All animal procedures were approved by the University of Georgia Institutional Animal Care and Use Committee. Pigs were individually penned (1.2 × 1.8 m) in an environmentally controlled building with a 12-h light:dark cycle (on at 0700 h) and unlimited access to water and to two feeders containing either the 12 or 24% crude protein (CP) diets. The placement of diets was alternated between feeders daily. Feed was vacuumed from the feeders and weighed daily between 0900 and 1000 h. Intake of the 12 and 24% CP diets was measured. Total feed, protein and lysine intakes were calculated from the sum of the two diets and from chemical analysis of the diets (Table 1). There was a 7-d diet adjustment period (pretreatment) during which time the pigs were allowed to select between the diets. Beginning on d 3, all pigs received saline injections to acclimate them to the injection protocol. During the pretreatment period, a range of individual preferences for the two diets was noted. Despite the daily repositioning of the two diets in the feeders, some pigs consistently preferred one of the diets, whereas others selected a more even balance. Because of this, pigs were blocked based on the amount of 24% CP diet consumed (d 0-4) and body weight (d 4) and assigned to either the control or STH treatment groups. On d 7 and continuing through d 20, somatotropin (0 or 4 mg/d) was injected subcutaneously in the neck between 1000 and 1100 h daily. STH treatment was withdrawn on d 21 and all pigs continued to receive saline injections for the remainder of the study. Final body weight, back fat thickness and feed intake measurements were recorded on d 26.

Individual body weights and 10th-rib backfat thickness (Lean-Meater, Renco, Minneapolis, MN) were recorded on d 0, 4, 7, 11, 14, 18, 21 and 26. Blood samples were obtained from all pigs by jugular puncture on d 5 and 19, representing pretreatment and d 12 of treatment, respectively. The blood samples were obtained 3 h after somatotropin treatment and were immediately transferred to tubes containing EDTA. Plasma was obtained by centrifugation at 1200 × g for 15 min at 4°C and stored at 20°C. Plasma samples were assayed for glucose, urea nitrogen and triglycerides with the use of commercially available kits (ICN Biochemicals, Costa Mesa, CA; Sigma Chemical, St. Louis, MO and Wako Chemical, Dallas, TX). Glucose was determined colorimetrically using the glucose oxidase method. Triglycerides were measured colorimetrically via glycerol production. Urea nitrogen was assayed using urease and colorimetric procedures. Insulin (ICN Biochemicals) and insulin-like growth factor-I (IGF-I, Azain et al. 1991) were measured by RIA. Amino acid concentrations of deproteinized plasma were determined by using a Beckman System 6300 high performance amino acid analyzer (Beckman Instruments, Fullerton, CA). Samples were deproteinized by addition of 5 g/L sulfosalicylic acid.

Data were analyzed using the general linear models procedure in SAS (1985), with the main effect of STH in the model. For determination of the effect of treatment on intake of the 12 and 24% CP diets, as well as total intake, protein and lysine intake, pretreatment intake (d 0-7) for each diet or corresponding calculated variable was used as a covariate. Results are reported as least-squares means ± pooled SEM.

RESULTS

Table 2. Effect of somatotropin (STH) on body weight and rate of gain in pigs1,2 [View Table]

There was considerable variation among pigs in diet selection patterns. The coefficient of variation (CV) for intake of the 12 and 24% CP diets was 48.2 and 51.3%, respectively. Total intake (the sum of the 12 and 24%CP diets ) was much less variable (CV = 12.7%). The assignment of treatments on the basis of pretreatment selection pattern was essential and resulted in a more even distribution of pigs preferring one diet over the other in each STH treatment group.

Total feed intake during the pretreatment period for pigs that would subsequently be assigned to the control or STH groups was not different during d 0-7, as shown in Table 3 and Figure 1. Pooled across treatment groups, pigs consumed ~59 and 41% of their feed from the low (1.57 ± 0.27 kg/d) and high (1.09 ± 0.27 kg/d) CP diets during the week of pretreatment, respectively. This resulted in pigs selecting a 16% protein diet, which is greater than the recommended 13% (NRC 1988), but is consistent with current commercial standards for pigs of this age or weight. Total protein or lysine intake (g/d) was not different for the groups. Approximately 42 and 58% of total protein intake stemmed from the low (184 ± 32 g/d) and high (256 ± 64 g/d) protein diets, respectively.

Table 3. Pretreatment feed intake and diet selection in pigs on d 0-71 [View Table]

STH altered the pattern of diet selection. STH treatment decreased total feed intake during d 7-14 (2.51 vs. 3.07 ± 0.14 kg/d, P <0.01) and d 14-21 (2.44 vs. 3.29 ± 0.15 kg/d, P <0.01), as shown in Table 4 and Figure 1. Total energy intake was reduced with STH treatment (38.7 vs. 50.2 ± 2.1 MJ/d, P < 0.01). The pattern of selection of the 12 and 24% CP diets is shown in Figure 2. The reduced intake was accounted for by a decrease in the amount of 12% CP diet selected (average: 1.00 vs. 2.00 ± 0.22 kg/d, P <0.01). The decrease in the amount of the 12% CP diet selected was significant by the fifth day of STH treatment (d 12 of the study: 2.15 vs. 0.69 ± 0.33 kg, for control and STH, respectively, P < 0.01) and persisted throughout the STH treatment period.

Table 4. Effect of somatotropin (STH) on feed intake and diet selection in pigs on d 7-211 [View Table]

There was a trend for the amount of the 24% CP diet selected to increase during the STH treatment phase (d 7-21, 1.51 vs. 1.15 ± 0.19 kg/d, P < 0.10). Other statistical manipulations of the intake data yielded similar results. Using a repeated-measures analysis, it was found that for the 12% CP diet, there was a significant day × treatment interaction, indicating a change in intake with the initiation of treatment. However, for the 24% CP diet, there was no significant interaction of day and treatment (P > 0.20) and the treatment effect was not significant (P > 0.20). However, there was a significant time effect (P < 0.01). We decided that intake was best compared using pretreatment intake of each diet as a covariate in the analysis of the effects of treatment.

Although the intake of the 24% CP diet was not different on an absolute basis, it was of greater relative importance in the STH-treated pigs. During d 7-21, 61% of the total feed intake in STH-treated pigs was accounted for by consumption of the 24% CP diets compared with 37% in non-STH treated pigs (SEM = 6.4%, P < 0.05). Because of the numerically greater intake of the 24% CP diet by the STH-treated pigs, total protein intake was not different (449 vs. 487 ± 25 g/d). Similarly, lysine intake was not different during the STH treatment period (28.4 vs. 26.5 ± 1.6 g/d).

The 24% CP diet accounted for a significantly greater proportion of the daily protein intake in the STH-treated pigs (73%) than in the control group (50%, SEM = 6.8%, P < 0.05). Control pigs selected a 15.4% protein diet during d 7-21, which is similar to the pretreatment level. In contrast, STH-treated pigs selected an 18.0% protein diet (SEM = 0.7, P <0.02). STH treatment improved feed efficiency during d 7-21 ( 0.51 vs. 0.33 ± 0.03 kg gain/kg feed, P < 0.005).

Although total feed intake for pigs previously treated with STH was not different than control pigs during the withdrawal period (Table 5), the selection pattern remained similar to that during the treatment phase. Total feed intake recovered by the d 3 of withdrawal (Fig. 1; for d 24: control 3.36, STH 3.13 kg/d, SEM = 0.15), but the recovery of total feed intake was accounted for by increased intake of both the 12 and 24% CP diets (Fig. 2). During the withdrawal period, STH-treated pigs continued to select a higher percentage protein diet. The percentages of intake accounted for by each diet for d 21-26 (Table 5) were similar to those during d 7-21 (Table 4). The amount of the low protein diet selected by the pigs that had previously been treated with STH remained numerically lower than that in the control group throughout the withdrawal period, but was not significantly different (P > 0.20) during the last 3 d of the study. Similarly, the amount of high protein diet selected by the pigs that had previously been treated with STH remained numerically greater than the control group. For example, on d 25, the amount of high protein diet selected was 49% greater (1.68 vs. 1.13 ± 0.30 kg/d, P = 0.16) in the pigs that had been treated with STH compared with control pigs. Total protein or lysine intake was not different during the recovery phase.

Table 5. Effect of previous treatment with somatotropin (STH) on feed intake and diet selection in pigs on d 21-261 [View Table]

There were no differences in circulating plasma levels of glucose, urea nitrogen, triglyceride, insulin and IGF-I levels in the pretreatment (d 5) blood samples (Table 6). On d 19, urea nitrogen levels were 46% lower, and IGF-I and insulin concentrations were 42 and 135% greater in STH-treated pigs, respectively. Treatment was without effect on glucose levels; however, at d 19, glucose concentrations across treatment groups were higher than d 5 values (6.5 vs. 5.2 ± 0.2 mmol/L, P < 0.05). Additionally, insulin concentrations at d 19 for the non-STH treated pigs were also significantly higher than those at d 5 (459 vs. 142 pmol/L ± 12, P < 0.001).

Table 6. Effect of somatotropin (STH) administration in pigs on plasma metabolites during the pretreatment (d 5) and treatment phase (d 19)1 [View Table]

STH treatment altered plasma amino acid profile. The levels of tyrosine (P < 0.05), glutamic acid (P < 0.02), glutamine (P < 0.01), ammonia (P < 0.05) and alanine (P < 0.10) were greater in STH-treated pigs (Table 7). In addition, threonine and serine tended to be lower (P < 0.10) in STH-treated pigs.

Table 7. Effect of somatotropin (STH) treatment on circulating free amino acids in pigs1 [View Table]

DISCUSSION

The objective of this experiment was to examine whether pigs offered a choice of dietary protein levels and treated with STH could select a dietary combination that matched their energy and protein requirements. The results of this experiment show that pigs treated with STH and offered a choice between high and low protein diets decrease their overall feed intake, primarily by selecting a lower amount of the low protein diet when compared with control pigs. The decreased energy intake coupled with a maintenance of protein (and lysine) intake is the expected response to STH treatment in pigs. We and others have demonstrated that STH alters composition of gain in pigs (Lee et al. 1994, NRC 1994); lipid accretion is decreased and protein accretion is increased. However, the decrease in lipid accretion is of greater relative magnitude than the increase in protein accretion, and the total carcass energy retained is less in somatotropin-treated pigs (Campbell et al. 1991, NRC 1994). Because of this, the predicted digestible energy intake (NRC 1988), which accounts for lipid and protein accretion, maintenance and a correction for thermogenesis, is reduced in STH-treated pigs relative to controls. Because of changes in the efficiency of utilization of dietary essential amino acids (Caperna et al. 1995, Krick et al. 1993), STH-treated pigs are able to increase protein deposition without a significant change in protein intake. Thus, we observe that STH treatment results in a reduction in energy intake as a means to compensate for reduced lipid accretion, but a maintenance of protein intake to support protein accretion. This separate regulation of energy and protein intakes is in agreement with an earlier proposal (Webster 1993). STH-treated pigs select a higher percentage protein diet similar to that used in previous single diet studies (Etherton et al. 1987, Krick et al. 1992). Failure to provide STH-treated pigs with a sufficiently high concentration of dietary protein results in loss of the protein accretion benefit (Campbell et al. 1991, Goodband et al. 1990, Smith and Kasson 1991). This implies that the signals indicating the need to reduce energy intake override the signals for protein when the pigs are offered a single diet.

Nam et al. (1995) used choice feeding in a study similar to the present experiment to test the hypothesis that somatotropin-treated pigs would increase dietary protein intake. It was concluded that somatotropin "did not stimulate an increase in protein consumption" and that pigs were unable to make appropriate adjustments in selection pattern to match changes in composition of gain. On the basis of previous work (Boyd et al. 1991, Krick et al. 1993), somatotropin-treated pigs have a protein requirement similar to that of nontreated pigs. The "requirement" for a higher percentage protein in other studies is accounted for by a reduction in energy needs and intake. Thus, an increase in protein intake should not have been expected. In Nam's study, there was a decrease in total feed intake and a numerical increase in the percentage protein selected by the somatotropin-treated pigs. This difference was less than in the present study and was not significant. In this work, the hypothesis tested was that the reduction in intake would be accounted for by a reduction in the consumption of the low protein diet. Differences in the response to somatotropin in the present study and that of Nam et al. (1995) may arise from differences in design. In contrast to our use of individually penned, castrated male pigs, Nam used group-penned gilts and an electronic feeding system.

Dietary lysine may have also played a role in the differences between studies. Pigs in the present experiment were consuming 24-28 g lysine/d and obtaining the level required to exhibit a protein accretion response to somatotropin (Goodband et al. 1990, Krick et al. 1993). In the Nam study (1995), lysine intake was ~20 g/d. Furthermore, our diets were formulated to maintain the amino acid profile across diets, by maintaining a constant corn:soybean meal ratio. In the Nam study, the ratio of barley to soybean meal was adjusted as a means to change percentage protein in the diet and thus, dietary amino acid profile was not maintained.

The difference in intake responses to somatotropin in pigs and rats can be accounted for by differences in the relative magnitude of somatotropin's effects on protein and lipid accretion in each species. In female rats that have reached the growth plateau and are treated with somatotropin, a dramatic increase in protein accretion with only a small decrease or no change in lipid accretion is observed (Azain et al. 1995b, Roberts et al. 1994). The net result is an overall increase in both energy retention (carcass lipid + carcass protein) and feed intake. We have shown that, when allowed to select between high and low protein diets, the increase in intake in rats treated with somatotropin is accounted for by an increased consumption of the high protein diet (Roberts et al. 1995).

Although the mechanism whereby rats are able to perceive the need for additional dietary protein is not known, the somatotropin-induced increase in carcass protein accretion may alter circulating amino acid profiles and initiate a signal that results in a greater protein intake (Roberts et al. 1995).

Despite an increase in protein accretion in somatotropin-treated pigs, the large decrease in lipid accretion results in an overall decrease in carcass energy retention (Campbell et al. 1991, NRC 1994) and thus a decrease in energy requirement. In support of this explanation for the species differences in the intake response, we have shown that the intake response to somatotropin in insulin-treated (Roberts et al. 1994) and genetically obese rats (Azain et al. 1995a) is similar to that in pigs. The composition of gain in both models (insulin treated and genetically obese) is largely lipid and thus, as in pigs, the primary effect of exogenous somatotropin is to decrease lipid accretion, reduce total energy retention and decrease intake. Diet selection studies in these rat models have not been conducted.

Failure to maintain the altered body composition upon withdrawal of somatotropin treatment has been demonstrated previously in both pigs (Azain et al. 1991, Lee et al. 1994) and rats (Groesbeck et al. 1987, Roberts et al. 1995). During the withdrawal period in the present study, total feed intake normalized by the third day. However, the diet selection pattern remained similar to that observed during the treatment phase, with consumption of the low protein diet by pigs previously treated with somatotropin remaining lower than that of the nontreated pigs. Consumption of the high protein diet tended to be greater than that in the control pigs. The basis for the continued alteration in selection pattern is not known.

In summary, the decrease in feed intake in STH-treated pigs was accounted for by a reduced intake of the low protein diet. We suggest that this effect is related to the STH-induced decrease in lipid accretion and resulting decrease in carcass energy retention. Pigs appear to select a diet that matches their energy and protein requirements. The exact mechanism involved in monitoring lipid and protein accretion, as well as overall energy retention, is not known and warrants further investigation.

ACKNOWLEDGMENTS