![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
UMR INRA-INAPG de Physiologie de la Nutrition et du Comportement Alimentaire, Institut National Agronomique Paris-Grignon, F-75231 Paris Cedex 05, France
2To whom correspondence should be addressed. E-mail: fouillet{at}inapg.inra.fr.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
35% of ingested N at 8 h after both meals. However, dietary N intestinal absorption and its appearance in splanchnic free amino acids were predicted to be more rapid from soy protein and were associated with a higher deamination, concomitant with a higher efficiency of incorporation of dietary N into proteins in the splanchnic bed. In contrast, soy protein was predicted to cause a reduction in peripheral dietary N uptake, as a consequence of both similar splanchnic retention and increased oxidation compared with milk protein. In addition, protein synthesis efficiency was reduced in the peripheral area after soy protein intake, leading to dietary N incorporation in peripheral proteins that fell from 26 to 19% of ingested N 8 h after milk and soy protein ingestion, respectively. Such a model thus enables a description of the processes involved in the differential metabolic utilization of dietary proteins and constitutes a valuable tool for further definition of the notion of protein quality during the period of protein gain.
KEY WORDS: • interorgan amino acid metabolism • milk protein • soy protein • postprandial period • parameter estimation H
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Although a classical approach measures the whole-body N balance over several days, the acute utilization of N from dietary protein may be determined more precisely by its net deposition during the postprandial phase, which is the critical step for the metabolic orientation of dietary AA (2
–4
). In addition, a growing body of evidence suggests that the ability of dietary N to promote postprandial protein anabolism in different tissues (splanchnic and/or peripheral) may be relevant to the nutritional efficiency of dietary proteins (4
,5
). However, little is yet known, particularly in humans, about the kinetics of dietary N distribution and orientation in the anabolic and catabolic pathways of different organs after protein ingestion.
Indeed, among the studies focusing on region-specific N metabolism in the fed state, organ-balance studies using the arteriovenous catheterization technique (5
–7
) do not follow the specific metabolic fate of dietary N, whereas the multiple tracer approach (8
–10
) does not determine that part of the splanchnic (i.e., intestinal plus hepatic) uptake that is used specifically for protein synthesis (8
). In contrast, we previously developed a compartmental model to simulate specifically the metabolic fate of dietary N in both splanchnic and peripheral tissues, and thus to determine its regional retention and utilization for protein synthesis in the fed state in humans (11
,12
). This is a linear, 11-compartment model, which enables simulation of the kinetics of dietary N absorption, elimination and distribution between free AA and proteins in both the splanchnic and peripheral areas, based on experimental measurements of 15N kinetics in ileum, blood and urine after the ingestion of a single 15N-labeled protein meal in humans.
In the present study, this compartmental model is applied to estimating the respective contributions of splanchnic and peripheral tissues to the postprandial retention and metabolism of dietary N from milk and soy proteins in humans. Recently, animal (5
) and human (12
,13
) studies reported a higher whole-body retention of milk than soy protein in the postprandial phase, because of a lower oroileal absorption and a higher degradation to urea of dietary N from soy protein. The reasons for these differences are still unclear but may involve variations in both the pattern and kinetics of dietary AA reaching the tissues, which are potent modulators of protein synthesis and oxidation (14
–17
). We resorted to modeling to answer the question whether the lower level of whole-body retention observed experimentally for soy protein involved less splanchnic and/or peripheral dietary N retention and incorporation into protein than that induced by the better-retained milk protein. Such a model provides insight into the dynamics of the system in the nonsteady postprandial state, and enables a description of the processes involved in the differential metabolic utilization of different proteins.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Meal preparation, experimental data collection, analytic methods and calculations were described in full detail previously (13
,18
). The protocol was approved by the Institutional Review Board for Saint-Germain-en-Laye Hospital, and informed consent was obtained from each subject. Briefly, after an overnight fast, healthy humans ingested a liquid meal composed of 100 g sucrose with 30 g 15N-labeled milk protein (SMP meal, n = 9) or 30 g 15N-labeled soy protein isolate (SSP meal, n = 10). Both meals were mixed with water to 500 mL. The meals contained the following amounts of indispensable (or conditionally indispensable) AA, as determined by ion-exchange chromatography (mg/g protein, N · 5.95): His 27.1 vs. 27.7, Ile 60.6 vs. 52.5, Leu 101.4 vs 89.4, Lys 86.6 vs 63.9, Met 25.1 vs 12.2, Cys 8.9 vs. 14.5, Phe 51.0 vs. 58.7, Tyr 42.2 vs. 43.1, Thr 50.9 vs. 43.9, Trp 14.3 vs. 13.7, Val 67.4 vs. 52.5, for SMP vs. SSP, respectively. The energy and N contents of the experimental meals were 2150 kJ and 295 mmol N for SMP and 2174 kJ and 316 mmol N for SSP. This small difference in ingested N (7%) was <3% in terms of the dietary N absorbed, because of differences in real ileal digestibility. Thus, the meals were considered to be isonitrogenous and isoenergetic. Over a period of 8 h after the meal, samples of intestinal effluents were collected on a continuous basis using ileal tubing, with blood samples being drawn hourly using a catheter inserted in a forearm vein and urine completely collected under mineral oil every 2 h. Ileal flow rates were assessed using a slow-marker technique (19
). Dietary N levels in these intestinal, blood and urine samples were determined by measuring 15N enrichment by isotope ratio mass spectrometry in ileal effluents, plasma free AA and urea, and urinary urea and ammonia. Body urea was calculated from plasma urea by considering that urea mixes uniformly in total body water; the latter value was determined using the equations developed by Watson et al. (20
). The amount of dietary N present in plasma free AA was calculated by assuming that the plasma AA concentration was 100 mg/L and the mean plasma volume was 5% of the body mass (11
). All data (dietary N recovered from cumulated ileal effluents, plasma free AA, body urea, cumulated urinary urea and cumulated urinary ammonia) were expressed as a percentage of ingested N. Urinary data were interpolated and ileal effluent data were pooled in such a way as to obtain the same 1-h data step size.
Linear compartmental model.
We used a linear compartmental model (Fig. 1
) developed previously using data obtained after the ingestion of a single milk protein meal in humans (11
). This model, which enables simulation of the absorption, elimination and regional distribution of dietary N during the postprandial phase, was developed using SIMUSOLV Software (21
). This model was validated at each stage of its development by testing successively its a priori (theoretical) and a posteriori (numerical) identifiability (11
). It consists of 11 compartments representing distinct amounts of dietary N, and 15 different pathways of exchange between these compartments, each being characterized by a constant diffusion coefficient ki,j. These exchange rate constants ki,j represent the fraction of dietary N in compartment j transferred to compartment i per unit time (min). The model structure includes three subsystems, i.e., the gastrointestinal tract subsystem built using ileal effluent data, the deamination subsystem covering body urea, urinary urea and urinary ammonia data, and the retention subsystem built using plasma free AA data. The gastrointestinal tract subsystem consists of an unidirectional chain of three compartments representing dietary N in the stomach (G), dietary N in the intestinal lumen (Il) and dietary N in the ileal effluents (E). The deamination subsystem consists of three compartments representing dietary N in body urea (BU), dietary N in urinary urea (UU) and dietary N in urinary ammonia (UA). The retention subsystem was structured with the aim of clarifying N distribution between splanchnic and peripheral tissues, and its selected structure distinguishes between free and protein bound AA in both areas. The retention subsystem consists of five compartments representing dietary N in splanchnic free AA (SA), dietary N in splanchnic proteins (SP), dietary N in plasma free AA (PL), dietary N in peripheral free AA (PA) and dietary N in peripheral proteins (PP). Parsimonious modeling was applied to the choice of each subsystem structure so that it would be the minimum necessary to fit the sampled compartments. We tested the goodness-of-fit of different models of increasing order, and retained the simplest model structure that adequately fit the data vs. higher order models that do not significantly improve the fit (11
). The development of the present model is fully detailed elsewhere (11
). Furthermore, because this model is intended for use in a predictive setting, its validation with experimental data other than those used for its development (i.e., cross-validation) has been performed. By confronting the model with data obtained after the ingestion of milk protein with the addition of either sucrose or fat, it has been proven that this model is capable of discriminating between different nutritional conditions (12
).
|
,22
). Similarly, we defined the regional PSE of dietary N as its specific ability to be used for anabolic purposes in each area. The regional PSE of dietary N was thus calculated as the flux of its incorporation into protein divided by the flux of its appearance in the free AA pool, i.e., (k7,4 · SA)/[(k4,2 · IL) + (k4,5 · PL) + (k4,7 · SP)] in the splanchnic bed and (k8,6 · PA)/[(k6,5 · PL) + (k6,8 · PP)] in the peripheral area. Parameter estimation.
This compartmental model was confronted with the experimental data obtained after ingestion of the SMP and SSP meals. SIMUSOLV Software was used to estimate those parameter values that would produce the closest model predictions for each meal, by adjusting the rate constant values until the model predictions fit the data for all sampled compartments simultaneously. The objective function iteratively maximized during the parameter estimation process under SIMUSOLV was the log of the Likelihood Function, which represents the joint probability of obtaining our experimental data for each sampled pool in the context of a given set of fitted parameters, taking account of the fact that an experimental error is always associated with experimental measurements (11
,21
). The heteroscedasticity parameter (
), representing the heterogeneous error of each experimental data set, was also adjusted during the optimization process (11
). The optimization process started by exploring a broad range of variations for all parameters, allowing values to change individually or in various combinations to fit the data. We then focused on variations in parameters that had the strongest influence on the system during the fitting process by widely exploring the effect of their combined variations on model fitting. This first step allowed us to circumscribe the domain of optimal values for parameters, before optimizing all parameters simultaneously. Furthermore, different values for initial parameter estimates were tested to reduce the probability of falling into a local optimum if the starting point was not in the neighborhood of the global optimum. We applied this optimization process to the SMP meal and then used the parameter estimate values obtained with SMP as the initial values for parameter estimation for the SSP meal. The correlation matrix provided from the statistical output showed that the two parameters of the bidirectional pathways were always strongly correlated. We therefore decided to keep one of these parameters for each pair (k2,4, k4,5, k4,7, k5,6 and k6,8) constant and equal to the estimate value obtained with SMP; changes to other parameter values were both necessary and sufficient to explain the observed differences in kinetics. In this way, by making minor adjustments to the smallest set of parameters, we took advantage of Berman’s minimal change postulate (23
). In the context of our study, this implies that we could characterize the changes ensuing from our experimental perturbation (type of protein in the meal), by exploring the minimal change to the "reference" model (SMP meal), thus bringing the new predictions into line with the new data. These parameters, in which changes were identified as making the greatest contribution to the separation of meals, were considered to be regulatory steps involved in the metabolic response to variations in meal composition. Final parameter estimates were verified as providing the best possible fit, and not a local optimum. Thus, for each meal, the model was quantified for each subject and for the mean of individual data using parameter estimation. Results concerning dietary N postprandial distribution after each meal were obtained by optimization using the mean of individual data for each meal, whereas individual fits were used to assess the discriminatory capacity of the model and statistical differences between meals.
Sensitivity analysis.
Sensitivity analysis of the model was performed by evaluating the effect of a 1% change in parameter value on the prediction of a variable response, i.e., by calculating a sensitivity coefficient for each pair: 
(model response)/(model parameters). However, to eliminate the bias caused by the magnitude of parameter values, sensitivity coefficients were log-normalized and calculated using the direct decoupled method under SIMUSOLV (21
). Sensitivity analysis was performed on the parameter estimate values obtained for each meal, by evaluating their influence on the model responses for each compartment and also for the retention subsystem, following definition of a new variable calculated as being the sum of SP, SA, PL, PA, and PP.
Statistical analysis.
The discriminatory capacity of the model was tested by discriminant analysis using the SYSTAT statistical package (SYSTAT, Evanston, IL). The overall effects of meal, time and interactive effects on model predictions were assessed by ANOVA with time as a repeated factor (General Linear Models procedure, SAS/STAT Version 6.03, SAS Institute Cary, NC). For particular criteria such as gastric emptying half-times, intestinal absorption half-times, intestinal transit half-times and regional PSE, differences between meals were assessed using two-tailed unpaired t tests. Differences between parameter estimates obtained using the mean of individually fitted parameters or when directly fitting the mean of individual data were assessed using nonparametric Wilcoxon’s matched-pairs signed rank test (two-tailed). A probability < 0.05 was considered to be significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
For each meal, the model was quantified for each subject and for the mean of individual data using parameter estimation. The model fit all subjects satisfactorily, but a better fit was obviously obtained for each compartment using the mean of individual data (Fig. 2A
–E). The distribution of parameter estimates did not differ significantly when obtained using the mean of individually fitted parameters or when directly fitting the mean of individual data, as assessed by Wilcoxon matched-pairs signed-rank tests. The optimization criteria and parameter estimates obtained after optimization using the mean of individual data for each meal are given in Tables 1
and 2. As shown in Table 1
, the fit appeared adequate for both meals because the percentage of variations explained was generally higher than 95% for all compartments, except for E after SMP (Table 1)
. To further ascertain the adequacy of the model, its numerical identifiability was tested successively by checking the goodness-of-fit and the reliability of parameter estimates for each meal (24
).
|
|
). Goodness-of-fit, which appeared to be highly acceptable from visual inspection of a plot of model predictions vs. experimental data (Fig. 2)
, was further assessed by an analysis of residuals to check the underlying assumptions of both the normality and randomness of the data error distribution involved in optimization (24
,25
). The standardized residuals of sampled compartments were generally consistent with the hypothesis of normality because 95% of them were within the range [-1.96; +1.96] (25
) for both meals (98% for SMP and 93% for SSP). As formally tested using the runs test (24
), the residuals of each sampled compartment were generally consistent with the hypothesis of randomness, except E in both the SMP and SSP groups (P < 0.05). These systematic but slight deviations between experimental data and model predictions for ileal effluents suggested, as previously reported (12
), that delayed intestinal transit in the presence of sucrose may require at least one more compartment in the gastrointestinal tract subsystem to better fit the ileal effluent data. However, we were satisfied with the model response, even in these two less adequately fitted compartments because these two sampled compartments constitute traps from which there is no return to another compartment; this lack of structural influence on the remainder of the system limits the consequences of any error committed. Moreover, the reliability of parameter estimates was judged to be highly acceptable. The precision of fitted parameters is commonly expressed by their CV, and parameter values with a CV of <50% are usually considered to be adequately estimated (26
). As shown in Table 2
, the CV for fitted parameters did not exceed 14 and 20% for the SMP and SSP meals, respectively, indicating that the parameters were estimated with high precision.
|
The discriminatory capacity of the model was tested using a discriminant analysis that optimally separates all individual sets of ki,j into groups of closer characteristics, on the basis of linear combinations of the parameters. The discriminatory capacity of the model was assessed by determining the extent to which subjects who received the same meal were a posteriori correctly grouped together; 100% of subjects were correctly replaced in their original group, and discriminant analysis further indicated that k3,2, the intestinal transit rate, k6,5, the peripheral tissue transfer rate, and k4,2, the intestinal absorption rate, accounted for most of the difference between meals.
Sensitivity analysis, which enabled identification of those parameters with the greatest influence on the system, was performed on the parameter estimate values obtained for each meal and produced similar results for both meals. After both meals, and whatever the compartment and subsystem, k2,1 and k4,2 exhibited considerable initial influence, which then declined slowly. The retention subsystem was rapidly positively sensitive to k5,4 and k7,4, whereas it was negatively influenced by variations in k9,4 and k3,2, in descending order, with these trends increasing over time. SP was most positively sensitive to variations in k7,4 and negatively sensitive to variations in k5,4 and k9,4, in descending order. Moreover, PP was positively sensitive to variations in k5,4 and k8,6, in descending order, and negatively to those in k7,4 and k9,4, with these latter trends increasing over time. To summarize, k2,1 and k4,2 on the one hand, and k5,4, k7,4 and k9,4 on the other hand were systematically identified as important governing parameters, as previously reported (12
). These results agreed with the model structure and our knowledge of system behavior.
Postprandial retention, distribution and metabolism of dietary N from milk and soy proteins.
Experimental data indicated a lower level of whole-body retention of dietary N from soy protein, because of higher ileal losses and splanchnic oxidation compared with milk protein (Fig. 2
F). Hence, the whole-body retention of dietary N was significantly reduced from 80 to 72% 8 h after the milk and soy protein meals, respectively. Unlike experimental data, the compartmental model enabled simulation of the successive transfers of dietary N between different compartments after each meal, thus allowing evaluation of the kinetics of dietary N absorption and transfer to the splanchnic and peripheral organs.
The model simulated an N gastric emptying half-time of 100 min that was similar after both meals, but it predicted delayed intestinal transit and absorption kinetics of dietary N from milk protein compared with soy protein. Indeed, the half-times of intestinal transit and intestinal absorption of dietary N from milk protein were delayed by 50 min compared with soy protein (Table 3
). Consequently, dietary N incorporation into the splanchnic free AA (SA) occurred rapidly and transiently after SSP, with an acute peak reaching 8% of ingested N at 1 h 20 min, whereas SMP induced a delayed and higher peak reaching 11% of ingested N at 2 h 10 min after the meal (Fig. 3
B). Dietary N incorporation into splanchnic protein (SP), which occurred consecutively, was still rising 12 h after both meals, reaching a maximum value of 37% of ingested N at that time point (Fig. 3
B). After SSP (compared with SMP), SA was thus predicted to be significantly lowered, whereas SP tended to be increased; this last-mentioned effect was significant only during the first two postprandial hours (Fig. 3
B). As a consequence, PSE in the splanchnic bed was significantly affected by the protein source in the meal, and reached 23 and 30% 8 h after SMP and SSP, respectively (Fig. 4
). However, the total retention of dietary N in the splanchnic bed (SA + SP), which was still rising 12 h after both meals (38% of ingested N at 12 h), was not significantly affected by the protein source in the meal (Fig. 3
A).
|
|
|
D) due to a similar rate of dietary N delivery to PA (k6,5) (Table 2)
. Thus, PA peaked at 13% of ingested N 4.5 h after both meals (Fig. 3
D). In contrast, dietary N incorporated into peripheral protein (PP), which was still increasing 12 h after both meals, was predicted to be significantly lowered by soy protein as early as 7 h after meal ingestion, reaching maximum values of 34 and 26% of ingested N 12 h after SMP and SSP intake, respectively (Fig. 3
D). PSE in peripheral tissues was thus significantly affected by the protein source in the meal, reaching 32 and 24% 8 h after SMP and SSP intake, respectively (Fig. 4)
. Finally, the peripheral retention of dietary N (PA + PP) was still increasing 12 h after both meals, and reached maximum values of 39 and 31% of ingested N 12 h after SMP and SSP intake, respectively. The peripheral uptake was significantly affected by the protein source in the meal; the lowering effect of soy protein was significant as from 7 h after ingestion (Fig. 3
C). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Hence, it allowed us to predict both dietary N retention and incorporation into protein in each area after each meal. Indeed, the model made it possible to determine that the lower whole-body retention of dietary N observed experimentally with soy compared with milk protein was associated with the following: 1) the more rapid intestinal transit and absorption of dietary N from soy protein; 2) its increased transfer to urea concurrent with its similar sequestration in the splanchnic bed; and 3) its subsequent reduced uptake by the peripheral area. Interestingly, even though there was a similar splanchnic retention of dietary N after both meals, the AA part of this retention (SA) was significantly lower, whereas its protein part (SP) tended to be higher after soy protein ingestion. In contrast, even though there was a greater peripheral uptake of dietary N after milk protein, the AA part of this uptake (PA) was similar after both meals, whereas its protein part (PP) was significantly higher after milk protein ingestion. Consequently, the predicted PSE of dietary N was significantly higher for soy protein in the splanchnic bed (23 and 30% after milk and soy protein meals, respectively), whereas it was significantly higher after milk protein in the peripheral area (32 and 24% after milk and soy protein, respectively). The results show that the protein source affected protein synthesis differently in the splanchnic bed and peripheral area. This suggests that the region-specific anabolic response may be sensitive to the kinetics of absorption and appearance of dietary N in free AA pools and/or the AA composition of dietary protein.
|
Whatever the meal, the model predicted a predominant splanchnic uptake of dietary N during the early postprandial phase, which modulated the delivery of dietary AA to peripheral tissues. This is consistent with our current knowledge of the system (10
,27
–31
). These results are in agreement with the idea that the acute anabolic effect of a mixed meal occurs primarily in the splanchnic area (32
–36
), whereas muscular protein synthesis makes only a minority contribution to the whole-body anabolic response, despite the large mass of muscle (37
–39
). Splanchnic tissues are known to be involved mainly in the increase in protein synthesis associated with the fed state because of their high protein turnover (33
,37
,40
–42
). For instance, albumin synthesis has been reported to be under the regulation of insulin and enteral AA delivery (33
,35
,41
–43
). Consistent with these findings in the literature, the model predicted that the ingestion of SMP and SSP mixed meals, both inducing similar acute hyperinsulinemia (13
,18
), would give rise to a strong splanchnic anabolic response, as had previously been reported for the mixed sucrose and milk protein meal (12
). The range of variation covered by the model predictions for both meals in terms of splanchnic extraction, splanchnic anabolism, peripheral uptake and peripheral PSE are presented in Table 4
, together with the corresponding previous findings in the literature. The paucity of such findings in the literature highlights the need for additional data to further validate the model predictions. However, the plausibility and physiologic relevance of the model are supported by the consistency of its predictions with respect to our current knowledge of the system (44
).
|
Experimental data showed the later appearance of dietary AA in the plasma after SMP than after SSP, with a peak achieved 4 h after SMP ingestion, i.e., 1 h later than after SSP (Fig. 2
C). The compartmental analysis accomplished in this study enabled a distinction between the role of the kinetics of gastric emptying, intestinal absorption and first-pass metabolism in the differences in dietary AA plasma appearance in the meals. The gastric emptying rates were predicted to be similar after both meals, but the dietary N kinetics of both intestinal transit and absorption were predicted to be delayed with milk protein (Table 3)
. These predictions were consistent with findings in rats showing the same gastric emptying half-time after soy-protein isolate or casein meals (both with sucrose), together with delayed small intestinal transit of the liquid phase of the chyme after the casein meal (45
). This effect may be due to the release during digestion of active opioid regulatory peptides from casein, which have been shown to slow gastrointestinal motility (46
). The predicted delay in dietary N absorption from the intestinal lumen (1 h) was sufficient to account for the delay observed in dietary AA plasma appearance after milk protein compared with soy protein. Moreover, the more rapid intestinal absorption of dietary N from soy protein was associated with its concurrent higher deamination and greater incorporation into protein in the splanchnic bed during the early postprandial phase. Sensitivity analysis further indicated that the rate of dietary N appearance in SA (k4,2, the rate of intestinal absorption) exerted a markedly positive initial influence on dietary N deamination (BU + UU + UA) and incorporation in SP, which then declined slowly. The predicted differences in dietary N absorption kinetics may therefore be involved in the differences between soy and milk proteins in terms of the splanchnic anabolism and catabolism of dietary N. Consistent with this idea, Boirie et al. (16
) had already developed the concept of "slow" and "fast" proteins, according to the rate at which dietary proteins are digested and absorbed from the gut. These authors concluded that a rapidly absorbed protein transiently induced a greater enhancement of both whole-body protein synthesis and dietary AA oxidation compared with a more slowly absorbed protein. Our findings both confirm and complete these data concerning the specific metabolic fate of dietary N.
Differences in the AA composition of dietary proteins: effects on splanchnic and peripheral metabolism.
The total amount of indispensable AA in soy is lower (85%) than that found in milk protein. Although soy protein can supply total sulfur AA in approximately adequate quantities compared with the reference protein (47
), the methionine content in soy (vs. milk protein) may be considered to be very low. Thus, the predicted differences in the splanchnic metabolism of dietary N (higher deamination and PSE of dietary N from soy compared with milk protein) could also be explained by certain differences in the AA composition of the dietary proteins, which affect the pattern of AA reaching the splanchnic tissues. This idea is consistent with the findings of various studies in rats (48
–52
). For instance, the ingestion of a diet deficient in an essential AA by rats has been reported to induce a simultaneous increase in catabolism and the protein synthesis rate in the liver, compared with supplemented control animals (51
). In contrast, it has recently been reported in pigs receiving a constant infusion in the stomach that splanchnic protein synthesis tended to be lower with soy than with casein (5
). However, during another study in humans, although albumin synthesis in the liver was reduced after 10 d of consuming a predominantly vegetarian diet, supplementation with soy protein reversed this effect. After supplementation, both fractional and absolute synthesis rates of albumin were indeed similar to those observed with an animal protein–rich diet (53
).
Furthermore, differences in the AA composition of dietary proteins may be involved in the differential effect of the protein source on regional dietary N metabolism. Indeed, after soy protein intake (compared with milk protein), dietary N incorporation into protein was transiently increased in the splanchnic bed and significantly decreased in the peripheral area. Similarly, the PSE of dietary N was concurrently higher in splanchnic tissues and lower in peripheral tissues after soy vs. milk protein. These predictions were consistent with various findings in the literature. The acute utilization of dietary AA for protein synthesis has been reported to rapidly be both slightly increased in liver and decreased in peripheral tissues under conditions of AA imbalance in rats (52
). When rats consuming a diet deficient in an essential AA are compared with supplemented control rats, the rate of protein synthesis was slower in peripheral tissues such as the muscle (49
,50
) or skin (51
), whereas it was more rapid in the splanchnic bed for liver and serum proteins (49
–51
). Moreover, in rats fed a legume protein diet vs. a casein diet, protein synthesis was reported to be slightly increased in the liver and markedly decreased in muscle (48
). Similarly, it has been shown in humans during insulin infusion that the acute, limited availability of a single essential AA, compared with complete hyperaminoacidemia, can adversely affect peripheral protein synthesis, while concomitantly upholding the fractional synthetic rates of the two hepatic proteins albumin and fibrinogen (15
). Moreover, milk protein contained higher levels (120%) of branched-chain AA (BCAA) than soy protein, AA that are known to be largely transferred to extrasplanchnic tissues (8
,54
). Thus, in the peripheral area, both relative and absolute amounts of BCAA were probably higher after milk protein. Because BCAA are stimulators of muscle protein anabolism (55
–57
), a higher proportion of BCAA could explain the higher peripheral PSE reported in the present study after milk protein ingestion. Last, because the AA compositions of milk and soy also differ greatly in their nonessential AA content, the latter may also be involved in the predicted differences in regional anabolism of dietary N, although few data are available concerning this issue.
In conclusion, the compartmental model allowed us to explain the lower whole-body retention of soy protein compared with milk protein observed experimentally during the postprandial state. This was not due to a lower retention of dietary N in the splanchnic bed, despite its higher splanchnic oxidation, but was associated with a reduction in peripheral dietary N uptake and use for anabolic purposes. The PSE of dietary N was thus concurrently increased in the splanchnic bed and decreased in the peripheral area by soy protein intake. Moreover, the model showed important differences in the gastrointestinal and absorption kinetics of dietary N between soy and milk proteins. This resulted in different patterns of dietary N appearance in splanchnic free AA after the meals. Thus, the higher deamination and PSE values in the splanchnic bed after soy protein could be explained both by differences in dietary N gastrointestinal kinetics and by differences in the AA composition of the dietary proteins. In contrast to the splanchnic bed, the PSE of dietary N in the peripheral area was significantly affected by the protein source in the meal, without any kinetic effect on dietary N availability at the time of its arrival in the peripheral tissues. Differences in the protein AA composition thus seemed to be implicated mainly in differences in the peripheral metabolic utilization of dietary N between milk and soy proteins. Finally, the model proved to be accurate and sensitive in discriminating between different nutritional conditions. It could be used as a predictive tool for the noninvasive determination of the organ-specific utilization of different protein sources in the postprandial nonsteady state in humans, also allowing further definition of the notion of protein quality in the phase of gain. Additional results from direct experiments must be compared with the predictions of the model to gain further insight in this research area. New experimental data (e.g., precise plasma concentrations or sampling of other pools) could also be used to confirm or improve the model predictions. Further development of the model would include obtaining experimental data to enable a distinction between the relative contributions of the gut and liver to the splanchnic metabolism of dietary N, with the aim of ascertaining its nutritional importance and modulation.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
3 Abbreviations used: AA, amino acid; BCAA, branched-chain amino acids; BU, dietary N in body urea; E, dietary N in ileal effluents; G, gastric dietary N content; Il, dietary N in intestinal lumen; PA, dietary N in peripheral free AA; PL, dietary N in plasma free AA; PP, dietary N in peripheral proteins; PSE, protein synthesis efficiency; SA, dietary N in splanchnic free AA; SMP, sucrose and milk protein meal; SP, dietary N in splanchnic proteins; SSP, sucrose and soy protein meal; UA, dietary N in urinary ammonia; UU, dietary N in urinary urea.
Manuscript received 19 June 2001. Initial review completed 2 August 2001. Revision accepted 26 September 2001.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
1. Munro, H. N. (1964) General aspects of the regulation of protein metabolism by diet and by hormones. Munro, H. N. Allison, J.B. eds. Mammalian Protein Metabolism 1964 Academic Press New York, NY. .
2.
Marchini, J. S., Cortiella, J., Hiramatsu, T., Chapman, T. E. & Young, V. R. (1993) Requirements for indispensable amino acids in adult humans: longer-term amino acid kinetic study with support for the adequacy of the Massachusetts Institute of Technology amino acid requirement pattern. Am. J. Clin. Nutr. 58:670-683.
3. Millward, D. J. & Pacy, P. J. (1995) Postprandial protein utilization and protein quality assessment in man. Clin. Sci. (Lond.) 88:597-606.[Medline]
4.
Tomé, D. & Bos, C. (2000) Dietary protein and nitrogen utilization. J. Nutr. 130:1868S-1873S.
5.
Deutz, N. E., Bruins, M. J. & Soeters, P. B. (1998) Infusion of soy and casein protein meals affects interorgan amino acid metabolism and urea kinetics differently in pigs. J. Nutr. 128:2435-2445.
6. Meek, S. E., Persson, M., Ford, G. C. & Nair, K. S. (1998) Differential regulation of amino acid exchange and protein dynamics across splanchnic and skeletal muscle beds by insulin in healthy human subjects. Diabetes 47:1824-1835.[Abstract]
7. Volpi, E., Ferrando, A. A., Yeckel, C. W., Tipton, K. D. & Wolfe, R. R. (1998) Exogenous amino acids stimulate net muscle protein synthesis in the elderly. J. Clin. Investig. 101:2000-2007.[Medline]
8.
Biolo, G., Tessari, P., Inchiostro, S., Bruttomesso, D., Fongher, C., Sabadin, L., Fratton, M. G., Valerio, A. & Tiengo, A. (1992) Leucine and phenylalanine kinetics during mixed meal ingestion: a multiple tracer approach. Am. J. Physiol. 262:E455-E463.
9.
Biolo, G. & Tessari, P. (1997) Splanchnic versus whole-body production of 
-ketoisocaproate from leucine in the fed state. Metabolism 46:164-167.[Medline]
10.
Krempf, M., Hoerr, R. A., Pelletier, E. A., Marks, L. M., Gleason, R. & Young, E. R. (1993) An isotopic study of the effect of dietary carbohydrate on the metabolic fate of dietary leucine and phenylalanine. Am. J. Clin. Nutr. 57:161-169.
11.
Fouillet, H., Gaudichon, C., Mariotti, F., Mahe, S., Lescoat, P., Huneau, J. F. & Tomé, D. (2000) Compartmental modeling of postprandial dietary nitrogen distribution in humans. Am. J. Physiol. 279:E161-E115.
12.
Fouillet, H., Gaudichon, C., Mariotti, F., Bos, C., Huneau, J. F. & Tomé, D. (2001) Energy nutrients modulate the splanchnic sequestration of dietary nitrogen in humans: a compartmental analysis. Am. J. Physiol. 281:E248-E260.
13.
Mariotti, F., Mahe, S., Luengo, C., Benamouzig, R. & Tomé, D. (2000) Postprandial modulation of dietary and whole-body nitrogen utilization by carbohydrates in humans. Am. J. Clin. Nutr. 72:954-962.
14. Buse, M. G. & Reid, S. S. (1975) Leucine. A possible regulator of protein turnover in muscle. J. Clin. Investig. 56:1250-1261.
15.
Lecavalier, L., De Feo, P. & Haymond, M. W. (1991) Isolated hypoisoleucinemia impairs whole body but not hepatic protein synthesis in humans. Am. J. Physiol. 261:E578-E586.
16.
Boirie, Y., Dangin, M., Gachon, P., Vasson, M. P., Maubois, J. L. & Beaufrere, B. (1997) Slow and fast dietary proteins differently modulate postprandial protein accretion. Proc. Natl. Acad. Sci. U.S.A. 94:14930-14935.
17.
Dangin, M., Boirie, Y., Garcia-Rodenas, C., Gachon, P., Fauquant, J., Callier, P., Ballevre, O. & Beaufrere, B. (2001) The digestion rate of protein is an independent regulating factor of postprandial protein retention. Am. J. Physiol. 280:E340-E348.
18.
Gaudichon, C., Mahe, S., Benamouzig, R., Luengo, C., Fouillet, H., Dare, S., Van Oycke, M., Ferriere, F., Rautureau, J. & Tomé, D. (1999) Net postprandial utilization of [15N]-labeled milk protein nitrogen is influenced by diet composition in humans. J. Nutr. 129:890-895.
19.
Mahe, S., Huneau, J. F., Marteau, P., Thuillier, F. & Tomé, D. (1992) Gastroileal nitrogen and electrolyte movements after bovine milk ingestion in humans. Am. J. Clin. Nutr. 56:410-416.
20.
Watson, P. E., Watson, I. D. & Batt, R. D. (1980) Total body water volumes for adult males and females estimated from simple anthropometric measurements. Am. J. Clin. Nutr. 33:27-39.
21. Dow Chemical Company (1990) Simusolv-Modeling and Simulation Software 1990 Dow Chemical Midland, MI. .
22.
Volpi, E., Mittendorfer, B., Wolf, S. E. & Wolfe, R. R. (1999) Oral amino acids stimulate muscle protein anabolism in the elderly despite higher first-pass splanchnic extraction. Am. J. Physiol. 277:E513-E520.
23. Berman, M. (1963) A postulate to aid in model building. J. Theor. Biol. 4:229-236.[Medline]
24. Cobelli, C. & Foster, D. M. (1998) Compartmental models: theory and practice using the SAAM II software system. Adv. Exp. Med. Biol. 445:79-101.[Medline]
25. Shyr, L. J., Griffith, W. C. & Boecker, B. B. (1991) An optimization strategy for a biokinetic model of inhaled radionuclides. Fund. Appl. Toxicol. 16:423-434.[Medline]
26. Miller, L. V., Krebs, N. F. & Hambidge, K. M. (1998) Human zinc metabolism: advances in the modeling of stable isotope data. Adv. Exp. Med. Biol. 445:253-269.[Medline]
27. Battezzati, A., Haisch, M., Brillon, D. J. & Matthews, D. E. (1999) Splanchnic utilization of enteral alanine in humans. Metabolism 48:915-921.[Medline]
28. Ferrannini, E., DeFronzo, R. A., Gusberg, R., Tepler, J., Jacob, R., Aaron, M., Smith, D. & Barrett, E. J. (1988) Splanchnic amino acid and glucose metabolism during amino acid infusion in dogs. Diabetes 37:237-245.[Abstract]
29. Mariotti, F., Huneau, J. F., Mahe, S. & Tomé, D. (2000) Protein metabolism and the gut. Curr. Opin. Clin. Nutr. Metab. Care 3:45-50.[Medline]
30. Soeters, P. B., de Blaauw, I., van Acker, B. A., von Meyenfeldt, M. F. & Deutz, N. E. (1997) In vivo inter-organ protein metabolism of the splanchnic region and muscle during trauma, cancer and enteral nutrition. Baillieres Clin. Endocrinol. Metab. 11:659-677.[Medline]
31.
Yu, Y. M., Wagner, D. A., Tredget, E. E., Walaszewski, J. A., Burke, J. F. & Young, E. R. (1990) Quantitative role of splanchnic region in leucine metabolism: L-[1-13C,15N]leucine and substrate balance studies. Am. J. Physiol. 259:E36-E51.
32. Cayol, M., Boirie, Y., Prugnaud, J., Gachon, P., Beaufrere, B. & Obled, C. (1996) Precursor pool for hepatic protein synthesis in humans: effects of tracer route infusion and dietary proteins. Am. J. Physiol. 270:E980-E987.
33.
De Feo, P., Horber, F. F. & Haymond, M. W. (1992) Meal stimulation of albumin synthesis: a significant contributor to whole body protein synthesis in humans. Am. J. Physiol. 263:E794-E799.
34.
Stoll, B., Burrin, D. G., Henry, J., Yu, H., Jahoor, F. & Reeds, P. J. (1998) Dietary amino acids are the preferential source of hepatic protein synthesis in piglets. J. Nutr. 128:1517-1524.
35. Volpi, E., Lucidi, P., Cruciani, G., Monacchia, F., Reboldi, G., Brunetti, P., Bolli, G. B. & De Feo, P. (1996) Contribution of amino acids and insulin to protein anabolism during meal absorption. Diabetes 45:1245-1252.[Abstract]
36.
Morens, C., Gaudichon, C., Fromentin, G., Marsset-Baglieri, A., Bensaïd, A., Larue-Achagiotis, C., Luengo, C. & Tomé, D. (2001) Daily delivery of dietary nitrogen to the periphery is stable in rats adapted to increased protein intake. Am. J. Physiol. 281:E826-E836.
37.
Scornik, O. A., Howell, S. K. & Botbol, E. (1997) Protein depletion and replenishment in mice: different roles of muscle and liver. Am. J. Physiol. 273:E1158-E1167.
38. Waterlow, J. C. (1995) Whole-body protein turnover in humans—past, present, and future. Annu. Rev. Nutr. 15:57-92.[Medline]
39. McNurlan, M. A. & Garlick, P. J. (1989) Influence of nutrient intake on protein turnover. Diabetes Metab. Rev 5:165-189.[Medline]
40. Deutz, N. E., Ten Have, G.A.M., Soeters, P. B. & Moughan, P. J. (1995) Increased intestinal amino-acid retention from addition of carbohydrates to a meal. Clin. Nutr. 14:354-364.
41. De Feo, P., Gaisano, M. G. & Haymond, M. W. (1991) Differential effects of insulin deficiency on albumin and fibrinogen synthesis in humans. J. Clin. Investig. 88:833-840.
42. Tessari, P. (1994) Effects of insulin on whole-body and regional amino acid metabolism. Diabetes Metab. Rev 10:253-285.[Medline]
43. Ballmer, P. E., McNurlan, M. A., Essen, P., Anderson, S. E. & Garlick, P. J. (1995) Albumin synthesis rates measured with [2H5ring]phenylalanine are not responsive to short-term intravenous nutrients in healthy humans. J. Nutr. 125:512-519.
44. Cobelli, C., Carson, E. R., Finkelstein, L. & Leaning, M. S. (1984) Validation of simple and complex models in physiology and medicine. Am. J. Physiol. 246:R259-R266.[Medline]
45. Hara, H., Nishikawa, H. & Kiriyama, S. (1992) Different effects of casein and soyabean protein on gastric emptying of protein and small intestinal transit after spontaneous feeding of diets in rats. Br. J. Nutr. 68:59-66.[Medline]
46. Daniel, H., Vohwinkel, M. & Rehner, G. (1990) Effect of casein and ß-casomorphins on gastrointestinal motility in rats. J. Nutr. 120:252-257.
47. FAO/WHO (1990) Joint FAO/WHO Expert Consultation 1990 Protein Quality Evaluation. Food and Agriculture Organization/World Health Organization Rome, Italy. .
48. Martinez, J. A., Goena, M., Santidrian, S. & Larralde, J. (1987) Response of muscle, liver and whole-body protein turnover to two different sources of protein in growing rats. Ann. Nutr. Metab. 31:146-153.[Medline]
49. Sidransky, H. & Verney, E. (1965) Chemical pathology of acute amino acids deficiency. VIII. Influence of amino acid intake on the morphologic and biochemical changes in young rats force-fed a threonine-devoid diet. J. Nutr. 86:73-80.
50. Sidransky, H. & Farber, E. (1958) Chemical pathology of acute amino acids deficiency. II. Biochemical changes in rats fed threonine- or methionine-devoid diets. Arch. Pathol. 66:135-141.
51. Nimni, M. E. & Bavetta, L. A. (1961) Dietary composition and tissue protein synthesis. I. Effect of tryptophan deficiency. Proc. Soc. Exp. Biol. Med. 108:38-45.
52. Yoshida, A., Leung, P. M., Rogers, Q. R. & Harper, A. E. (1966) Effect of amino acid imbalance on the fate of the limiting amino acid. J. Nutr. 89:80-90.
53.
Caso, G., Scalfi, L., Marra, M., Covino, A., Muscaritoli, M., McNurlan, M. A., Garlick, P. J. & Contaldo, F. (2000) Albumin synthesis is diminished in men consuming a predominantly vegetarian diet. J. Nutr. 130:528-533.
54. Wahren, J., Felig, P. & Hagenfeldt, L. (1976) Effect of protein ingestion on splanchnic and leg metabolism in normal man and in patients with diabetes mellitus. J. Clin. Investig. 57:987-999.
55. Kimball, S. R. & Jefferson, L. S. (2001) Regulation of protein synthesis by branched-chain amino acids. Curr. Opin. Clin. Nutr. Metab. Care 4:39-43.[Medline]
56. Zanetti, M., Barazzoni, R., Kiwanuka, E. & Tessari, P. (1999) Effects of branched-chain-enriched amino acids and insulin on forearm leucine kinetics. Clin. Sci. (Lond.) 97:437-448.[Medline]
57.
Wolfe, R. R. (2000) Protein supplements and exercise. Am. J. Clin. Nutr. 72:551S-557S.
58.
Hoerr, R. A., Matthews, D. E., Bier, D. M. & Young, V. R. (1993) Effects of protein restriction and acute refeeding on leucine and lysine kinetics in young men. Am. J. Physiol. 264:E567-E575.
59.
Stoll, B., Burrin, D. G., Henry, J., Jahoor, F. & Reeds, P. J. (1997) Phenylalanine utilization by the gut and liver measured with intravenous and intragastric tracers in pigs. Am. J. Physiol. 273:G1208-G1217.
60.
Stoll, B., Burrin, D. G., Henry, J. F., Jahoor, F. & Reeds, P. J. (1999) Dietary and systemic phenylalanine utilization for mucosal and hepatic constitutive protein synthesis in pigs. Am. J. Physiol. 276:G49-G57.
61.
Elia, M., Folmer, P., Schlatmann, A., Goren, A. & Austin, S. (1989) Amino acid metabolism in muscle and in the whole body of man before and after ingestion of a single mixed meal. Am. J. Clin. Nutr. 49:1203-1210.
62. Capaldo, B., Gastaldelli, A., Antoniello, S., Auletta, M., Pardo, F., Ciociaro, D., Guida, R., Ferrannini, E. & Sacca, L. (1999) Splanchnic and leg substrate exchange after ingestion of a natural mixed meal in humans. Diabetes 48:958-966.[Abstract]
63.
Morens, C., Gaudichon, C., Metges, C. C., Fromentin, G., Baglieri, A., Even, P. C., Huneau, J. F. & Tomé, D. (2000) A high-protein meal exceeds anabolic and catabolic capacities in rats adapted to a normal protein diet. J. Nutr. 130:2312-2321.
This article has been cited by other articles:
|
|
 |
|
  B. Juillet, H. Fouillet, C. Bos, F. Mariotti, N. Gausseres, R. Benamouzig, D. Tome, and C. Gaudichon Increasing habitual protein intake results in reduced postprandial efficiency of peripheral, anabolic wheat protein nitrogen use in humans Am. J. Clinical Nutrition, March 1, 2008; 87(3): 666 - 678. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W Hartman, J. E Tang, S. B Wilkinson, M. A Tarnopolsky, R. L Lawrence, A. V Fullerton, and S. M Phillips Consumption of fat-free fluid milk after resistance exercise promotes greater lean mass accretion than does consumption of soy or carbohydrate in young, novice, male weightlifters Am. J. Clinical Nutrition, August 1, 2007; 86(2): 373 - 381. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
