To the best of our knowledge, this is the first study where human milk is evaluated with respect to glycemic and insulinemic effects. The study was performed in healthy adults with a GI setting, thus post prandial glycemic and hormonal responses to the human and bovine milk were compared to an equicarbohydrate amount of WWB. In the present study, we found that human milk displayed insulinogenic properties and resulted in a postprandial glycemic response (GI 57) that was in the same range as that to bovine milk (GI 43) or the whey and casein fractions, respectively (GI 61 and 44). The human milk showed the lowest insulin response (iAUC 0–120 min) in comparison to all the other meals, thus indicating that human milk display superior insulin economy. Worth emphasizing is that the protein content of the human milk meal was markedly lower than that of bovine milk, whey and casein meals. According to our analysis, the human milk meal contained 3.5g protein as calculated from nitrogen analysis using 6.38 as conversion factor. However, it has been has reported that the true protein content of human milk is overestimated from nitrogen analysis due to its high non protein nitrogen (NPN) content (20-25% of total nitrogen)
[16, 17]. Consequently, the protein amount in the human milk, used in this study, may have been overestimated. Although the whey:casein ratio of the human and bovine milk in this study was not analyzed, it has been reported to range from 50:50 – 80:20 in human milk depending on the lactation period
, compared to approximately 20:80 in commercial bovine milk
. Our results thus indicate that human milk may be more insulinogenic per unit protein compared to bovine milk, and the higher proportion of whey protein may be a contributing factor.
All the test products in our study yielded higher postprandial plasma amino acid responses compared to the WWB reference meal, although not reaching significance for all products. The bovine milk and the reconstituted whey and casein meals were especially prone to affect responses of the BCAA; val and leu, and also lys and thr. We also noticed a tendency to a rise in these amino acids following human milk in comparison to the WWB. We recently showed in vitro that the amino acids ile, leu, val, lys and thr induce strong insulin secreting properties in mouse pancreatic islets
. In particular, leu is recognized as a potent insulin secretagogue
[25, 26]. Moreover it has been reported that also taurine, the second most common “amino acid” and the largest part of the NPN fraction in human milk
[27, 28], may be involved in insulin secretion
[29, 30]. In the present study, we also observed positive correlations between individual plasma amino acids and serum insulin and plasma incretin secretion in the postprandial phase (iAUC 0–60 min) as well as negative correlations to the glycemic response (iAUC 0-60min). Altogether this suggests that the amino acids play an important role in the insulinogenic properties of the milk proteins and contribute to lowered postprandial glycemia. Thus, it can be suggested that amino acids appearing in plasma following milk ingestion, may affect insulin secretion in two ways; directly by acting on the pancreatic β-cells and indirectly by promoting incretin release. We thus found that human milk, bovine whey and casein respectively, affect release of GLP-1, with the human milk resulting in a significantly higher early response compared to the WWB meal. In relation to the human milk, the whey meal yielded a higher GLP-1 response. However, considering the fact that the protein content in the human milk meal constitutes only about 22% of that in the bovine whey meal, the early GLP-1 response (iAUC 0–30 min) still reached more than 50% of that seen following whey. Thus, it could be suggested that human milk is prone to stimulate GLP-1. It remains to be shown if the GLP-1 response is solely related to the proteins in the human milk, or if other bioactive components may also be involved.
Strong positive correlations were found between the early GLP-1 response (iAUC 0–15 min) and postprandial insulin release up to 30 min, as well as negative correlations to blood glucose response (iAUC 0–90 min), an indication that early GLP-1 release may be important for the modulation of glycemia following both human and bovine milk. In support of such an opinion, we recently observed that ingestion of whey protein and free amino acid caused an early GLP-1 secretion that reduced postprandial glycemia in absence of a hyper-insulinemic peak
. In addition to GLP-1, the human milk also resulted in a small elevation of early phase GIP that was significantly higher than after the WWB reference meal at 7.5 min. GIP may be present in human milk
, and the response we found might result from a higher intake of this incretin, and could possibly attenuate the postprandial insulin response and facilitate glycemic regulation. Although, little is known about to what extent ingested GIP can maintain activity following exposure to the acidic environment in the stomach, it could be hypothesized that ingestion of intrinsic GIP could possibly contribute to the incretin effect of human milk. In previous studies of incretin responses we have preferentially seen increase in GIP after whey ingestion
. Thus, it is interesting that the whey meal, as well as the human milk meal, in this study significantly increased both GLP-1 and GIP response compared to all the other meals. GLP-1 secretion following whey ingestion has previously been reported by Hall et al.
. However, in that study the test meals contained 38E% of fat, which may have influenced GLP-1 secretion. Both GLP-1 and GIP have been identified as strong insulinotropic agents
[13, 34], and a release of both these incretins after whey ingestion may add to its insulinogenic properties. Interestingly, it was recently reported that whey protein inhibits dipeptidyl peptidase IV (DPP-4), an enzyme that inactivates GLP-1
. Consequently, whey protein may both induce GLP-1 secretion as well as prolong the activity of this incretin.
In contrast to whey, the casein meal resulted in a small incremental insulin response and had the lowest glycemic response compared to the other test meals. This differences in glycemic and insulinemic responses when, comparing the casein and whey meal, is in contrast with observations by Hall et al.
, who found no differences on neither postprandial glucose nor insulin responses between these proteins. However, in accordance with the present study, they observed that whey protein, in comparison to casein, increased plasma incretins, with higher levels of both GLP-1 and GIP
There were some limitations of the present study. All the test meals resulted in lower blood glucose increments compared to the WWB reference meal. It should, however, be noted that lactose was the carbohydrate source in all the test meals, whereas the reference WWB meal contained starch. In previous studies, we have shown that lactose has a lower glycemic index (GI 68) and insulinemic index (II 50) than the starchy WWB (GI 100 and II 100)
. Consequently, the comparatively low glycemic properties seen with all the present test meals could partly be explained by the lactose per se. The number of test subject is slightly lower (n = 9) than recommended for standardized glycemic index determinations (n = 10)
. Also, although the test subjects were used to regular Swedish food habits where milk consumption is frequent, differences in lactase levels might affect glycemia to milk.