Most of the current feed evaluation systems used to estimate dairy cow requirements and formulate diets cannot predict the effects of dietary changes on milk composition, in particular on milk fat content (MFC) and milk fat yield (MFY) responses. Dietary changes alter dairy cow MFC and MFY through modifications in the supply of nutrients, which act as precursors or inhibitors of mammary fat synthesis. Taking into account the supply of nutrients could allow feed evaluation systems to predict milk composition responses to dietary changes. We described an empirical model based on nutrient flows to predict MFC and MFY responses to dietary changes. This model was built by coupling a set of published empirical equations estimating nutrient flows (acetate, propionate, butyrate, glucose and digestible protein) from dietary characteristics and the measured ruminal VFA proportions to another set of equations estimating the response of milk fat to the supply of these nutrients. The model was evaluated by comparing the predicted and observed MFC and MFY responses on three databases derived from published feeding studies in dairy cows. The databases compiled published trials involving a change in the forage-to-concentrate ratio (‘F:C ratio’ sub-database), a change in the carbohydrate source (‘CHO source’ sub-database) or a change in both F:C ratio and CHO source (‘ratio and source’ sub-database). Overall, the current version of the model did not accurately predict milk fat responses: the model did not enable to explain the MFC and MFY responses in the ‘CHO source’ sub-database (P>0.10 and R2=0.02); in the ‘F:C ratio’ sub-database, agreement between observed and predicted responses was better (the slopes were not different from unity and were respectively 0.84 and 0.94 for MFC and MFY), but R2 were low (R2=0.36 for MFC and R2=0.43 for MFY). In the ‘ratio and source’ database, MFY was quite correctly predicted (the slope was 0.93 and R2=0.42), but not the MFC responses (P>0.10 and R2=0.01). The mean errors of prediction for MFC were 2.61, 3.79 and 3.55g/kg, respectively in the ‘F:C ratio’, ‘CHO source’ and the ‘ratio and source’ sub-databases. The effects of several interfering factors related to cows, diets or experimental conditions were tested to explain differences between observed and predicted responses in each sub-database. They enabled to identify several potential sources of bias in the model: especially, bias in estimation of OM digestibility, bias in estimation of glucose effects, lack of the effects of fatty acids in the model. These constitute interesting ways to improve the model.

An ideal profile of essential AA (EAA) can improve the efficiency of metabolizable protein (or PDIE, the equivalent in the INRA feeding system) utilization in dairy cows. Compared with other EAA, existing recommendations for the requirements of Arg, Ile, and Val are few and inconsistent. Four multiparous Holstein dairy cows at 22±6 wk of lactation received 4 treatments (duodenal infusions of 445±22.4g/d of an EAA mixture complementing a low-protein diet in a 4×4 Latin square design with a period length of 1 wk). The control treatment provided a balanced supply (in % of PDIE) of 5.1% Arg, 5.2% Ile, and 5.9% Val, whereas in the 3 subsequent treatments of −Arg, −Ile, and −Val, the concentrations of these 3 EAA were reduced to 3.5, 4.1, and 4.5%, respectively. All treatments were made isonitrogenous and were balanced to provide 7 other EAA (Lys, Met, His, Leu, Phe, Thr, and Trp), according to the recommendations described in the literature. Combined, the diet and the infusions provided 14.3±0.1% crude protein on a dry matter basis, and 66.0±1.2g of PDIE/Mcal of net energy for lactation. Neither dry matter intake (19.2kg/d) nor milk yield (30.4±0.4kg/d) was affected by treatments. The −Arg and −Ile treatments did not modify milk protein synthesis or the efficiency of N utilization. However, the −Val treatment decreased milk protein content by 4.9% and milk crude protein content by 4.3%, and tended to decrease the efficiency of N use for milk protein yield by 3.7% (compared with the control). These effects of Val were related to a decrease in the plasma concentration of Val as well as a trend toward decreasing plasma concentrations of Met, His, and the sum of all EAA and nonessential AA in the −Val treatment, which indicates a different utilization of all AA in response to the Val deficit. The deletion of Ile, compared with the deletion of Val, tended to decrease the milk protein-to-fat ratio by 3.8%. In conclusion, the supply of Arg at 3.5% of PDIE was not limiting for milk protein synthesis. The slight effect on the milk protein-to-fat ratio caused by decreasing the supply of Ile suggests a need to reevaluate the Ile requirement more precisely. A low Val supply could be limiting for milk protein synthesis, provided that the requirements of Lys, Met, and His are met.

Providing a well-balanced supply of essential AA (EAA) can serve as an opportunity to reduce the protein intake for dairy cows by increasing the efficiency of metabolizable protein (or PDIE, its equivalent in the INRA feeding system) utilization for milk protein yield. Our objectives were to compare the effect of supplying an "ideal" EAA profile (EAA+) with an imbalanced AA profile (control) at 2 levels of PDIE/NE(L) (net energy for lactation) supplies to study the interaction between PDIE and AA profiles, and to compare this ideal profile with a simple mixture of the 4 most deficient EAA (4 EAA) in the diets of dairy cows. Six lactating multiparous Holstein cows received 6 treatments with 2 different levels of PDIE supplied by diets and AA infusions in the duodenum according to a changeover design with 3-wk periods. Within each PDIE supply level, the cows received 3 different AA infusions in the duodenum according to a 3×3 Latin square design with 1-wk subperiods, which corresponded to the following treatment groups: control (Glu), 4EAA (Lys, Met, His, Leu), and EAA+ (4 EAA plus Ile, Val, Phe, Trp, and Tyr). In the low and high PDIE treatments, diets and infusions provided 54.7 and 64.0 g/Mcal of PDIE/NE(L), respectively, which corresponded to crude protein levels of 13.6 and 15.2%, respectively. High-PDIE supplies increased the milk protein yield by 163 g/d, the milk protein content by 1.4 g/kg, the milk yield by 4.1 kg/d, and the lactose yield by 178 g/d and decreased the PDIE efficiency of utilization by 12.4%, whereas the N efficiency of utilization remained unaffected. Supplying the 2 EAA profiles (4EAA and EAA+) increased the milk protein yield by 67 g/d, the milk protein content by 1.3g/kg, and the milk yield by 0.9 kg/d, whereas the milk fat and milk lactose contents were decreased by 2.4 and 1.6g/kg, respectively. The responses regarding milk yield and its composition were similar whether the cows received the 4 EAA or the EAA+ treatment. The responses were similar for the milk yield and composition whether the EAA were supplied by low- or high-PDIE supplies. In conclusion, the efficiency of PDIE utilization was improved by 6.6% and the N efficiency was improved by 7.0% by correcting the EAA profiles, independent of the level of PDIE supplied. In addition, the increased efficiency observed, associated with provision of the 4 EAA, was similar to the provision of all EAA (EAA+) in this experiment.

Although in dairy cows the mammary gland (MG) is the major net user of essential AA (EAA) supply, milk protein synthesis from absorbed EAA is not a straightforward process. Early studies identified 2 groups of EAA based on different pattern of mammary utilization: group 1 [Met, Phe (+Tyr), Trp], where MG uptake was similar to secretion in milk protein, and group 2 (Arg, Ile, Leu, Lys, Thr, and Val), where uptake exceeded milk protein output. This review examines the validity of this classification under variable protein supply through a meta-analysis, with the outcomes then explained with studies in which the fates of individual EAA were monitored using isotope approaches. For the meta-analysis, the Fick principle, based on stoichiometric transfer of Phe+Tyr uptake to milk protein, was used to estimate mammary plasma flow across all studies. This approach was judged acceptable because doubling Phe supply did not result in mammary oxidation of Phe+Tyr and either limited or no contribution of peptides to Phe and Tyr mammary supply could be detected. The AA content of proteins synthesized by the MG was estimated from milk protein composition, and the uptake-to-output ratio (U:O) for individual AA was re-calculated based on these assumptions. Analysis of individual samples by isotopic dilution resulted in reduced variance compared with analysis on pooled samples performed with an AA analyzer. Globally, the U:O of His and Met is maintained close to unity under variable protein supply. The group 2 AA could be subdivided. First, the U:O for group "2v" AA (Ile, Leu, Val, and Lys) is greater than 1 and varied with protein supply. Accordingly, the increased U:O of Leu, induced by duodenal casein infusion, led to extra-mammary Leu oxidation. Decreasing Lys supply decreased Lys U:O and the associated transfer of N to non-EAA, mainly to Glx, Asx, Ser, and Ala. Second, the U:O of group "2nv" AA, Arg and Thr, does not vary with protein supply. The Arg U:O averages 2.5, whereas the Thr U:O, albeit averaging 1.2, does not differ from unity. Excess of both these AA is probably directed toward the synthesis of non-EAA rather than energy supply. Overall, the ability of the MG to use excess EAA-N supply offers alternative sources of N and C for energy provision, lactose synthesis and non-EAA synthesis. The latter function spares dietary non-EAA for other necessary processes, such as gluconeogenesis and energy supply, in other tissues to support lactation.

ABSTRACT: Although in dairy cows the mammary gland (MG) is the major net user of essential AA (EAA) supply, milk protein synthesis from absorbed EAA is not a straightforward process. Early studies identifi ed 2 groups of EAA based on different pattern of mammary utilization: group 1 [Met, Phe (+Tyr), Trp], where MG uptake was similar to secretion in milk protein, and group 2 (Arg, Ile, Leu, Lys, Thr, and Val), where uptake exceeded milk protein output. This review examines the validity of this classifi cation under variable protein supply through a meta-analysis, with the outcomes then explained with studies in which the fates of individual EAA were monitored using isotope approaches. For the meta-analysis, the Fick principle, based on stoichiometric transfer of Phe+Tyr uptake to milk protein, was used to estimate mammary plasma fl ow across all studies. This approach was judged acceptable because doubling Phe supply did not result in mammary oxidation of Phe+Tyr and either limited or no contribution of peptides to Phe and Tyr mammary supply could be detected. The AA content of proteins synthesized by the MG was estimated from milk protein composition, and the uptake-to-output ratio (U:O) for individual AA was re-calculated based on these assumptions. Analysis of individual samples by isotopic dilution resulted in reduced variance compared with analysis on pooled samples performed with an AA analyzer. Globally, the U:O of His and Met is maintained close to unity under variable protein supply. The group 2 AA could be subdivided. First, the U:O for group “2v” AA (Ile, Leu, Val, and Lys) is greater than 1 and varied with protein supply. Accordingly, the increased U:O of Leu, induced by duodenal casein infusion, led to extra-mammary Leu oxidation. Decreasing Lys supply decreased Lys U:O and the associated transfer of N to non-EAA, mainly to Glx, Asx, Ser, and Ala. Second, the U:O of group “2nv” AA, Arg and Thr, does not vary with protein supply. The Arg U:O averages 2.5, whereas the Thr U:O, albeit averaging 1.2, does not differ from unity. Excess of both these AA is probably directed toward the synthesis of non-EAA rather than energy supply. Overall, the ability of the MG to use excess EAA-N supply offers alternative sources of N and C for energy provision, lactose synthesis and non-EAA synthesis. The latter function spares dietary non-EAA for other necessary processes, such as gluconeogenesis and energy supply, in other tissues to support lactation.

Dietary changes alter dairy cow milk fat concentration (MFC) and yield (MFY) through modifications in the supply of nutrients, which act as precursors or inhibitors of mammary fat synthesis. The current models used to formulate dairy cow diets cannot predict changes in milk fat. The knowledge of the effects of the nutrients on milk fat would help to progress toward this prediction. To this end, we quantified and compared the milk fat responses to variations in the supply of seven nutrients derived from digestion: volatile fatty acids, glucose, proteins, long-chain fatty acids (LCFA) and t10,c12-conjugated linoleic acid (CLA). A database was compiled from studies involving digestive infusions of these nutrients in dairy cows. It included 147 comparisons between a nutrient infusion and a control treatment. The nutrient infusions were limited to the range of physiological variations to mimic nutrient changes after dietary modifications. We established models for the response of MFC, MFY and milk fatty acid (FA) composition to the supply of each nutrient. MFC and MFY responses to the nutrients were significant and linear, except for the MFC response to glucose that was curvilinear. The nutrients differed in their effects on MFC and MFY: acetate, butyrate and LCFA increased MFC and MFY, whereas propionate, glucose and t10,c12-CLA decreased them. Protein infusions increased MFY and decreased MFC because of an increase in milk yield. The effects of numerous interfering factors related to animals, diets or experimental conditions were tested on the residuals of the response models. The responses of milk FA percentages are also provided. When adjusted to the in vivo variations in the nutrients observed after dietary changes, the effects of the different nutrients were moderate. Finally, this study showed that several of these nutrients could contribute to the changes in milk fat production and composition observed after dietary changes. This is a first step toward predicting milk fat response to changes in nutrient supply.

Diets inducing milk fat depression (MFD) are known to alter ruminal lipid metabolism, leading to the formation of specific isomers [such as trans-10,cis-12 conjugated linoleic acid (CLA)] that inhibit milk fat synthesis in lactating dairy cows. However, ruminal outflow of these isomers does not fully account for the decreases in milk fat synthesis observed during diet-induced MFD. The high-concentrate diets inducing MFD also induce a greater production of propionate, suggesting a possible inhibition of milk fat by propionate associated with trans-10,cis-12-CLA during MFD. The present experiment aimed to study the combined effects of propionate and trans-10,cis-12-CLA (both inhibitors of milk fat synthesis) on milk fat secretion and the effects of the combination of 2 nutrients with opposite effects (acetate and propionate). Six Holstein cows were used in a 6×6 Latin square design with 21-d periods (14 d of nutrient infusion). The treatments were control; ruminal infusion of 1,500 g/d of acetate (A); ruminal infusion of 800 g/d of propionate (P); duodenal infusion of 1.60 g/d of trans-10,cis-12-CLA (CLA); ruminal infusion of 750 g/d of acetate+400 g/d of propionate (A+P); and duodenal infusion of 1.60 g/d of trans-10,cis-12-CLA+ruminal infusion of 800 g/d of propionate (CLA+P). The amounts of nutrients infused were chosen to induce a similar variation in milk fat content. Treatments A and P decreased dry matter intake. Compared with the control, P and CLA treatments decreased milk fat content and yield by 9% and 15% on average. Treatment A increased milk fat content by 6.5% but did not modify milk fat yield (because of a decrease in milk yield). The effects of A and P, and CLA and P on milk fat and fatty acid percentages and yield were additive (A+P and CLA+P treatments). With a same dose of trans-10,cis-12-CLA, the additional supply of propionate induced a decrease in milk fat 40% higher than that induced by trans-10,cis-12-CLA alone. The milk fatty acid profile obtained with CLA+P was similar to those observed with high-concentrate diets inducing MFD. In conclusion, under our experimental conditions, the effects of the 3 nutrients were additive on mammary lipogenesis, regardless of their separate effects. We also show that propionate could contribute to the milk fat reductions unaccounted for by trans-10,cis-12-CLA during MFD induced by high-concentrate diets.

Diet is a relatively simple way to modify milk fat yield and composition in dairy cows as the end-products of digestion are precursors or inhibitors of milk fat synthesis. The individual effects of these end-products are well-known, but it is still not known whether these nutrients have an additive effect or an interaction effect on milk fat secretion. Thus our objective was to investigate the effects of two of these nutrients on milk fat secretion,trans-10,cis-12 conjugated linoleic acid (CLA) and propionic acid (C3) supplied alone or together, under the same experimental conditions. Four Holstein dairy cows were used in a 4×4 Latin square design with 14-d periods. Treatments were control, CLA (duodenal infusion of 1·85 g/d oftrans-10,cis-12 CLA), C3 (ruminal infusion of 500 g/d of C3) and CLA+C3 (duodenal infusion of 1·85 g/d oftrans-10,cis-12 CLA plus ruminal infusion 500 g/d of C3). Infusions oftrans-10,cis-12 CLA reduced milk fat content and yield by 18% whereas C3 infusions had no significant effect on milk fat secretion.Trans-10,cis-12 CLA decreased the yields of all milk fatty acids (FA). This reduction was proportionally greater for FA synthesized de novo than for preformed long-chain FA. Infusions of C3 decreased the yields and percentages of 4:0 and 18:0 and increased the yields and percentages of all odd-chain FA. Interactions betweentrans-10,cis-12 CLA and C3 infusions on milk fat content, yield and FA composition were never significant. Overall, this study showed thattrans-10,cis-12 CLA has different and greater effects on milk fat secretion than C3. Moreover, under our experimental conditions, their effects on milk FA yields, which reflect their effects on mammary lipogenesis, were additive, whatever their individual effect.

This study analyzed the effect of propionate (C3) and casein (CN) on whole-body and mammary metabolism of energetic nutrients. Three multiparous Holstein cows fitted with both duodenal and ruminal cannulas were used in 2 replicated Youden squares with 14-d periods. Effects of CN (743g/d in the duodenum) and C3 (1,042g/d in the rumen) infusions, either separately or in combination as supplements to a grass silage diet, were tested in a factorial arrangement. The control diet provided 97% of energy and protein requirements. Within each period, blood samples were taken (d 11) from the carotid artery and the right mammary vein to determine net uptake of energetic nutrients. Plasma blood flow was calculated using the Fick principle (based on Phe and Tyr). On d 13, [6,6-2H2]glucose was infused in the jugular vein to determine whole-body glucose rate of appearance (Ra) based on enrichments in arterial plasma. Both C3 and CN treatments increased whole-body Ra (17% and 13%, respectively) but only CN increased milk (18%) and lactose (14%) yields, suggesting no direct link between whole-body Ra and milk yield. When CN was infused alone, the apparent ratio of conversion of CN carbon into glucose carbon was 0.31 but, when allowance was made for the CN required to support the extra milk protein output, the ratio increased to 0.40, closer to the theoretical ratio (0.48). This may relate to the observed increases in arterial glucagon concentrations for CN alone. Conversely, the apparent conversion of infused C3 carbon alone to glucose was low (0.31). With C3, mammary plasma flow increased as did uptakes of lactate, Ala, and Glu whereas the uptake for β-hydroxybutyrate (BHBA) decreased. Mammary net carbon balance suggested an increase with C3 treatment in glucose, lactate, Ala, and Glu oxidation within the mammary gland. Mammary glucose uptake did not increase with CN treatment, despite an increase in glucose arteriovenous difference and extraction rate, because plasma flow decreased (−17%). Whereas CN, alone or in combination with C3, increased both lactose and protein yields, only mammary AA (and BHBA in CN alone) uptake increased because plasma flow decreased (−17%). These data suggest that the observed variations of milk lactose yield (and other milk components) are linked to metabolic interchanges between several energetic nutrients at both the whole-body and mammary levels and are not explained by increases in whole-body glucose availability.

A decreased fat content of milk seems to be a necessity for the future at least in European countries. This can be obtained by genetic or nutritional ways. This paper tries to quantify possibilities offered by nutrition. Fat is the milk solid that can be the most easily modified in a nutritional way. Propionic acid, glucose, trans10 C18:1 and trans(10), cis(12) CLA are the nutrients identified to reduce milk fat content and acetic acid and butyric acid are those that increase milk fat content. Seven meta-analyses of literature data of trials giving these nutrients as perfusions or as protected forms were performed. As a result, the efficiency to reduce milk fat content is - 0.006, - 0.003, - 0.09, and - 1.8 g/kg per gram of added nutrient for propionic, glucose, trans(10) C18:1, and trans(10), cis(12) CLA and - 0.007, - 0.002 g/kg per gram of subtracted nutrient for butyric and acetic acids respectively. The mechanisms involved in these responses are reviewed. In the decrease of milk fat content induced by a 30 % increase of concentrate, glucogenic precursors (propionic acid and glucose) explain 54 % of the drop of milk fat content, trans(10) containing fatty acids (trans(10) C18:1 and trans(10), cis(12) CLA) explain 34 %, and precursors of de novo fatty acid synthesis (acetic, and butyric acids) explain 21 %. In conclusion, a decrease of milk fat content with some diets is multi-factorial, and a pluri-nutrient approach is proposed in replacement of the usual mono-nutrient approach in order to understand the variations of milk fat content.