I took a good revision break and did some searching myself!! This is a really good article - I’ve cut and pasted the most interesting bits (which actually ended up being just about all of it!)
Glycemia and insulinemia in healthy subjects after lactose-equivalent meals of milk and other food proteins: the role of plasma amino acids and incretins1,2,3
Mikael Nilsson, Marianne Stenberg, Anders H Frid, Jens J Holst and Inger ME Bj?rck
Milk products deviate from other carbohydrate-containing foods in that they produce high insulin responses, despite their low GI. The insulinotropic mechanism of milk has not been elucidated.
[I am still surprised by this!!]
Interestingly, there is epidemiologic evidence suggesting that overweight subjects with a high intake of milk and dairy products are at a lower risk of developing diseases related to the insulin resistance syndrome (13). However, the insulinotropic effect of milk has not been sufficiently acknowledged and the mechanism, as well as the potential health implications remain unclear.
The insulin response to milk products does not relate solely on the lactose component. Consequently, when testing pure lactose in healthy subjects, the II paralleled the GI, suggesting that some other noncarbohydrate component is responsible for the insulinotropic effect of milk, eg, the milk proteins. It is well known that different food proteins differ in their effect on glucose metabolism in humans (14-18), and several amino acids are potent in stimulating insulin secretion (19-23).
Glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) are released in response to a meal and enhance insulin secretion (24), and protein has been shown to increase the GIP response in both type 2 diabetic patients and healthy subjects (25).
Objective: The objective was to evaluate the effect of common dietary sources of animal or vegetable proteins on concentrations of postprandial blood glucose, insulin, amino acids, and incretin hormones [glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1] in healthy subjects.
Results: A correlation was found between postprandial insulin responses and early increments in plasma amino acids; the strongest correlations were seen for leucine, valine, lysine, and isoleucine. A correlation was also obtained between responses of insulin and GIP concentrations. Reconstituted milk powder and whey had substantially lower postprandial glucose areas under the curve (AUCs) than did the bread reference (?62% and ?57%, respectively). Whey meal was accompanied by higher AUCs for insulin (90%) and GIP (54%).
Amino acid content in the test meals
Concentrations of the amino acids in the test meals are presented in Table 2. The concentrations of branched-chain amino acids in the milk-based products and the cod meal were in the same range. However, the content of leucine was somewhat lower in cod than in milk, whey, and cheese. The cod meal contained almost the same amount of valine as milk and cheese, whereas the whey showed a considerably higher amount. Lysine was slightly more represented in cod than in the dairy products. GH contained somewhat lower amounts of lysine and the branched-chain amino acids compared with the other test meals.
Although the blood glucose responses after the whey meal were considerably lower than those after the reference meal (?57%), the serum insulin AUC (Table 3) was significantly higher (90%) (P < 0.05). The insulin response registered after whey deviated from all other test meals by being significantly higher. The milk and cheese meals showed significantly higher insulin AUCs than did the GL.
Compared with the reference, whey resulted in increased insulin concentrations at 15, 30, 45, and 75 min (Figure 2). Also at 30 min, the insulin responses after the milk and the cheese meals were significantly higher than after the reference (P < 0.05). Serum insulin concentrations increased 15, 45, 60, and 75 min after whey ingestion compared with all other test meals. At 30 min, insulin concentrations were higher after the whey meal than after the other test meals, excluding milk.
Although the postprandial blood glucose response after the test meal with reconstituted skim milk powder was low, the insulin response after milk was not significantly distinguishable from that after the WWB reference. Thus, the present results confirm those from a previous study in which the ingestion of pasteurized milk resulted in a discrepancy between blood glucose (GI = 30) and the insulin response (II = 90), which was not present after a carbohydrate equivalent load of pure lactose (GI = 68; II = 50) (9). In that study, it was hypothesized that some milk component in addition to lactose appears to stimulate insulin secretion. As judged from similar and high IIs for reconstituted skim milk (<0.1% fat, present study) and pasteurized 3%-fat milk (9, 10), neither the fat content per se nor the drying process appears to be involved in the insulinotropic mechanism. Instead, we supposed an involvement of milk proteins. About 80% of milk proteins are casein and 20% are whey. When rennet (used in cheese making) is added to milk, casein proteins aggregate and form a gel but whey proteins remain soluble. Also, when the pH is decreased, casein proteins clot; hence, the acidity in the stomach makes casein, but not whey, to aggregate into a gel.
It was previously observed that the ingestion of milk and other food proteins may stimulate insulin secretion (11, 12, 35). In the current study, the insulin response to the whey meal was even more pronounced than that to milk, which indicated that the insulinotropic component may be connected to the soluble milk proteins. Assuming that the protein fraction of milk contains an insulin secretagogue, the stimulating effect might be mediated through bioactive peptides or by specific amino acids released during digestion. Several amino acids are potent stimulators of insulin release, either when taken as a protein orally or when infused intravenously (21), and certain amino acids (eg, the branched-chain amino acids) are more insulinogenic than are others. van Loon et al (36) showed that the insulin response in healthy subjects was positively correlated with plasma leucine, phenylalanine, and tyrosine when ingested orally in the form of drinks in combination with glucose. Furthermore, it was concluded that protein hydrolysates stimulate insulin secretion to a higher extent than do intact protein because of a more rapid increase in postprandial plasma amino acid concentrations. In addition, Calbet and MacLean (37) described a close relation between the insulin response and the increase in plasma amino acid response, especially for leucine, isoleucine, valine, phenylalanine, and arginine. These findings indicate that the postprandial pattern of plasma amino acids may be an important entity for the insulinogenic properties of food proteins.
Of the 7 amino acids that reached the highest increments after the whey meal in the current study, the branched-chain amino acids (leucine, valine, and isoleucine), lysine, and threonine are all known to stimulate insulin secretion (20, 36, 38). Alanine might also have insulinotropic effects under select experimental conditions (39).
Whereas whey, milk, and to some extent cheese ingestion resulted in obvious amino acid responses, the remaining meals (GH, GL, cod, and WWB) resulted in only small increases in plasma amino acids. Generally, the amino acid responses to the cod meal occurred 60 min after ingestion. In contrast, peak amino acid responses to milk, whey, and cheese occurred more rapidly?within 30?45 min after ingestion?which indicated that milk proteins are highly digestible and result in a rapid release of amino acids into the circulation.
Instead of being related to amino acids per se, the insulinotropic effect of milk proteins might be related to bioactive peptides either present in the milk or formed during digestion in the small intestine. A possible pathway in the case of peptides may include the activation of the incretin system (24). Previous studies showed a protein-stimulated insulin response in type 2 diabetic patients (40) and healthy subjects (41) that did not parallel the rise in amino acids in the circulation, which suggests the involvement of the incretin hormones in protein-stimulated insulin release.
Conversely, Schmid et al (22) concluded that gut factors are only of minor importance and that amino acids are the major insulin secretagogue in the absence of carbohydrates. Whereas the GLP-1 responses to all of the test meals were similar in the current study, whey induced a particularly elevated GIP response. Thus, the higher GIP response after whey may have been one contributing factor to the observed elevated postprandial insulin response. The degree to which the GIP response explains the insulinotropic effect of whey proteins can, however, not be elucidated from the present data. Surprisingly, the GIP response to the milk meal was not elevated compared with the response to the reference meal. Similarly to whey, milk also showed an insulinogenic effect, although it was of a lower magnitude. This finding indicates that the stimulation of the incretin system may not solely explain the insulinotropic effects of whey.
In contrast with milk and whey, the postprandial blood glucose response after the meal consisting of cheese and lactose was not significantly different from that obtained after the WWB meal. However, serum insulin concentrations after the cheese meal were not significantly different from those after milk, although they were lower than those after whey. It is likely that cheese contains not only casein but also the remnants of whey proteins, and either this small amount of whey in the cheese curd is capable of enhancing insulin concentrations or the casein fraction itself may contain an insulin secretagogue. However, it is known that casein is more slowly digested than is whey (42, 43), and the different digestion rates of the proteins may effect the insulin response.
Wheat gluten in high and low amounts (the GH and GL meals, respectively) and cod affected glycemia and insulin response similarly to the reference meal, which suggests that both wheat gluten and cod have a poor capacity to stimulate insulin secretion. The lack of effect of wheat protein on the insulin response agrees with the consistency reported in GIs and IIs for a range of wheat products (6).
A synergistic effect of carbohydrates and proteins in stimulating insulin has been reported in diabetic subjects (44), whereas this effect was not seen in healthy persons (41). Although an additive effect of protein and carbohydrates (45) after the cod meal would be possible, the rise in plasma amino acids after the cod meal was modest compared with that after the milk and whey meals and presumably was too low to evoke an amino acid?induced insulin response.
Although whey and cod proteins are similar with respect to the content and distribution of amino acids, the postprandial plasma pattern of amino acids differed substantially after the test meals containing these proteins, most probably because of the different digestion and absorption rates of these proteins. It is especially interesting that several of the known insulinotropic amino acids (leucine, valine, isoleucine, lysine, and threonine) were among those amino acids that were observed to increase after the whey meal.
It can be concluded that food proteins differ in their capacity to stimulate insulin release, possibly by affecting the early postprandial concentrations of insulinotropic amino acids and incretin hormones differently. It cannot be excluded that an elevated plasma amino acid response is merely an indicator of the rapid digestion and absorption of whey proteins.
The results of the current study show that milk proteins have insulinotropic properties, with the whey fraction being a more efficient insulin secretagogue than casein. It remains to be shown whether the insulinotropic effect of whey and milk depends on an optimal and rapid postprandial release of certain amino acids to the blood, the release of a bioactive peptide, or an activation of the incretin system, particularly by enhancing GIP secretion. Also, the potential long-term effects of a noncarbohydrate?mediated insulin stimulus on metabolic variables should be evaluated in healthy persons and in persons with a diminished capacity for insulin secretion.