Introduction
The gastrointestinal tract is responsible for the digestion and absorption of nutrients, and acts as a barrier against luminal pathogens. Moreover, the gastrointestinal tract cooperates in controlling the metabolism through hormones secreted from enteroendocrine cells, which are the body’s largest endocrine organ(Reference Rehfeld1). Enteroendocrine cells are capable of responding to luminal content because their apical side has chemosensing machinery such as taste receptors (TASR), G protein-coupled receptors (GPCR), specific transporters and channels. Their secretory products are stored in characterised secretory vesicles, before being secreted through the basolateral membrane by exocytosis(Reference Gunawardene, Corfe and Staton2,Reference Sternini, Anselmi and Rozengurt3) . When luminal content moves through the gastrointestinal tract, specific macronutrients stimulate the chemosensing machinery, which leads to the modulation of gut hormone release. Gut hormones exert their effect via vagal nerve or endocrine/paracrine signalling, through the interaction of specific receptors expressed in different tissues of the body. These hormones, which are mainly glucagon-like peptide-1 (GLP-1), cholecystokinin (CCK), peptide YY (PYY), gastric inhibitory polypeptide (GIP) and ghrelin, influence the functioning of the digestive tract, but also modulate insulin secretion from the pancreas, the energy storage of adipose tissue and neuronal signalling in appetite centres in the brain to mediate the regulation of food intake by terminating hunger and inducing satiety.
Since dietary compounds modulate enterohormone secretion, and given the central role of enterohormones in body homeostasis, such an interaction could have beneficial health implications(Reference Pinent, Blay and Serrano4). In this context, protein and protein hydrolysates are currently being studied to determine their effects on GLP-1 modulation, either through secretion or clearance, which may influence the processes regulated by this hormone such as regulation of glycaemia homeostasis and food intake control. The nutrient-sensing machinery of carbohydrates and lipids is better understood than the detection and pathways followed by protein digestion. The main reasons for this gap in knowledge is the redundant signalling in the gut for the different protein digestion products and the complexity of protein digests(Reference Santos-Hernández, Miralles and Amigo5). Here we review the literature on this subject in order to determine if the evidence supports differential effects of food proteins on GLP-1 profile. We will introduce the relevance of GLP-1 signalling on health. Then we will focus on the effects on GLP-1 secretion of proteins and its hydrolysates, and the suggested mechanisms. Finally, we will briefly review the use of protein hydrolysates as dipeptidyl peptidase-4 (DPP4) inhibitors. We compile a significant number of scientific studies to highlight the importance of the different protein sources, the hydrolysis conditions applied to them, and the resulting digestion products.
Relevance of glucagon-like peptide-1 signalling in health
There is evidence to suggest that specific enterohormones administered at physiological concentrations can influence the appetite of rodents and human subjects (for a review, see Murphy & Bloom(Reference Murphy and Bloom6)). Likewise, the effects of gut hormones on food intake and body weight have been observed in bariatric surgery (such as Roux-en-Y gastric bypass), which induces a huge increase in GLP-1 and peptide YY (PYY) secretion and is used to treat obesity(Reference Le Roux, Aylwin and Batterham7). Therefore, the modulation of enterohormone signalling may be an important target in the prevention of obesity and related/associated pathologies. Moreover, endogenous gut hormones regulate appetite physiologically, unlike the drugs that are currently available, which mainly influence the central neurotransmitter systems. Therefore, gut hormone-based therapies might lead to fewer side effects(Reference Murphy and Bloom6).
Furthermore, modulation of endogenous incretin hormones (GLP-1 and GIP) could be an interesting strategy for preventing and/or managing type 2 diabetes mellitus (T2DM)(Reference Kreymann, Williams and Ghatei8). T2DM is the most common endocrine disorder, characterised by insulin resistance and impaired insulin secretion, and it is one of the fastest growing non-communicable diseases in the world(Reference Shaw, Sicree and Zimmet9). The main goal in the treatment of T2DM is to keep blood glucose levels within the normal physiological range. In this regard, GLP-1 and GIP are therapeutically interesting peptides because they are important mediators of glycaemic homeostasis, as they are responsible for approximately 50–70 % of the total insulin secreted after glucose intake(Reference Baggio and Drucker10). GLP-1, together with GIP, is responsible for the incretin effect, since it binds to GLP-1 receptor in β-cells in the pancreas leading to an increase in intra-cellular Ca and a subsequent insulin secretion in response to glucose(Reference Vilsbøll and Holst11). It has also been shown that GLP-1 enhances markers of proliferation and differentiation, and decreases markers of apoptosis in the pancreas of Zucker diabetic rats(Reference Pick, Clark and Kubstrup12,Reference Farilla, Hui and Bertolotto13) . Furthermore, GLP-1 improves the glycaemic profile by inhibiting glucagon secretion and improves glucose disposal in peripheral tissues(Reference Baggio and Drucker10). In that way, for patients with T2DM, a non-pharmacological therapeutic approach could be achieved by targeting these incretins (GLP-1 and GIP) through protein- and protein hydrolysate-based strategies. This approach would be mainly focused on increasing GLP-1 levels rather than stimulating GIP because in these patients the responsiveness of their β-cells to GIP action is decreased(Reference Nauck14). Furthermore, only GLP-1 exerts an appetite-suppressing effect, while GIP does not seem to do the same(Reference Baggio and Drucker10). Accordingly, many incretin-based therapies focus on using GLP-1 analogues, promoting endogenous GLP-1 secretion or using DPP4 inhibitors.
DPP4 is a ubiquitous aminodipeptidase that exists essentially as a membrane-anchored cell-surface enzyme(Reference Filippatos, Athyros and Elisaf15). It is expressed throughout the body tissues, such as kidneys, the gastrointestinal tract, liver, pancreas, and the endothelial and epithelial cells on the vascular bed. Its soluble form is found in plasma and therefore it is in close proximity with hormones circulating in the blood(Reference Yu, Yao and Chowdhury16,Reference Mentlein17) . The main activity of DPP4 is to remove N-terminal dipeptides from polypeptides(Reference Thoma, Löffler and Stihle18), which preferably have a proline or alanine in the second position from the N-terminal. Some of the main DPP4 substrates are GLP-1 and the other incretin hormone GIP, which are peptides with N-terminal Tyr–Ala and His–Ala, respectively(Reference Havale and Pal19). The intact GLP-1 is rapidly hydrolysed by DPP4 into a shorter, inactive form, once it reaches the plasma. GLP-1 has a half-life of 1–2 min(Reference Thoma, Löffler and Stihle18). Only 25 % of the active GLP-1 reaches the portal circulation and subsequently the liver, where a further 40–50 % is digested by the DPP4 in hepatocytes. This means that only 15 % of the secreted GLP-1 enters the systemic circulation and may reach other tissues, such as the pancreas or the brain(Reference Holst20). Therefore, DPP4 is responsible for inactivating more than 80 % of the secreted GLP-1(Reference Thoma, Löffler and Stihle18). Studies focus not only in the development of DPP4-inhibitory drugs, but also on peptides derived from food sources with DPP4-inhibitory capacity.
Although pharmacological compounds are being studied(Reference Khera, Murad and Chandar21), natural compounds might be used to prevent the development of overweight- and obesity-related problems from early preclinical stages through interaction with the enteroendocrine system(Reference Serrano, Casanova-Martí and Blay22).
Dietary regulation of glucagon-like peptide-1 secretion
Nutrient ingestion is the primary physiological stimulus for inducing GLP-1 secretion by L cells, located in the ileum and colon in the human gastrointestinal tract. GLP-1 secretion occurs in a biphasic pattern, which consists of a rapid release in 15–30 min after a meal, followed by a second minor peak that occurs in 60–120 min. Enteroendocrine cells have been shown to respond to carbohydrates, lipids and proteins.
Glucose and fat have been reported to be strong GLP-1-secretagogues after they have been ingested(Reference Elliott, Morgan and Tredger23), or directly administered into the intestine(Reference Rocca and Brubaker24,Reference Roberge and Brubaker25) or into perfused ileal segments(Reference Cordier-Bussat, Bernard and Levenez26). In the murine model, glucose-stimulated GLP-1 release is blocked using Na-dependent GLUT-1 (SGLT-1) knockout mice and SGLT-1 inhibitors(Reference Gorboulev, Schürmann and Vallon27,Reference Kuhre, Frost and Svendsen28) , which suggests that glucose metabolism uses glucose transport via SGLT-1 to induce GLP-1 secretion. It has also been proposed that sweet taste receptors (T1R2, T1R3) are involved in the glucose-sensing mechanism, but there is still some controversy about whether this is so(Reference Steinert, Gerspach and Gutmann29,Reference Kokrashvili, Mosinger and Margolskee30) . On the other hand, it has been reported that G-protein-coupled receptors (GPCR) are activated by dietary fat to stimulate GLP-1 release, including GPR40 and GPR120 by medium-chain fatty acids, long-chain fatty acids and long-chain unsaturated FA; and GPR41 and GPR43 by SCFA (for reviews, see Hirasawa et al. (Reference Hirasawa, Hara and Katsuma31) and Reimann(Reference Reimann32)).
Other food components could also modulate GLP-1 secretion. Flavonoid structures, present in several vegetables, also stimulate GLP-1 secretion(Reference Domínguez Avila, Rodrigo García and González Aguilar33). In both ex vivo (Reference Casanova-Martí, Serrano and Blay34) and rat models(Reference González-Abuín, Martínez-Micaelo and Margalef35), these compounds have been shown to improve the metabolic status altered by a cafeteria diet treatment(Reference Gonzalez-Abuin, Martinez-Micaelo and Blay36).
Effects of proteins on glucagon-like peptide-1 secretion
Dietary proteins undergo digestion by gastric (pepsin) and pancreatic (chymotrypsin and trypsin) proteases and membrane digestion by peptidases associated with the brush-border membrane of enterocytes. The different digestive proteases cleave the peptide bonds at preferential positions. The primary endproducts are dipeptides and tripeptides, which will enter the cell through peptide transporters. Free amino acids are also released after luminal protein digestion and after peptide hydrolysis within the intestinal cells, and then exit across the basolateral membrane via specific amino acid transporters.
GLP-1 release is activated by luminal intestinal chemosensors, which could be reached by peptides of different sizes, mixed with free amino acids.
Studies in human, animal and enteroendocrine cells have shown increased GLP-1 secretion by free amino acids such as l-phenylalanine, l-alanine and l-glutamine(Reference Pais, Gribble and Reimann37,Reference Reimann, Williams and Da Silva Xavier38) and l-asparagine(Reference Mace, Schindler and Patel39). The effect of glutamine has been confirmed in healthy, obese and diabetic human subjects(Reference Greenfield, Farooqi and Keogh40,Reference Samocha-Bonet, Wong and Synnott41) . Tolhurst et al. (Reference Tolhurst, Zheng and Parker42) demonstrated this effect in isolated mouse L cells and reported that the mechanisms were associated with an increase in cyclic AMP (cAMP) and cytosolic Ca2+ levels. They also found evidence to suggest that electrogenic Na-coupled amino acid uptake is responsible for initiating membrane depolarisation and voltage gated Ca2+, while a second pathway increases intracellular cAMP levels. Young et al. (Reference Young, Rey and Sternini43) also reported similar results with l-proline, l-serine, l-alanine, l-glycine, l-histidine, l-cysteine and l-methionine in the STC-1 cell line.
When analysing the effects of protein on GLP-1 release, many studies focus on the effects of protein hydrolysates, produced by the hydrolysis of food protein with commercial enzymes (summarised in Tables 1–3). Sometimes, especially in in vitro studies, these are digestive enzymes that simulate intestinal digestion. However, many different hydrolysates are obtained through treatment with enzymes other than pepsin, chymotrypsin or trypsin. Protein hydrolysis can have two main benefits: (1) protein will be more quickly digested after intake; and (2) bioactive peptides(Reference Mine, Li-Chan, Jiang, Mine, Li-Chan and Jiang44–Reference Udenigwe and Aluko52) might be released. Thus, the degree of protein digestion may impact the capability of protein to stimulate GLP-1 release, as discussed below.
Table 1. Stimulation of glucagon-like peptide-1 (GLP-1) secretion by protein and protein hydrolysates in humans
↑, GLP-1 secretion is incremented v. the control, specified in each row; N.D., hydrolysis conditions not described; T2DM, type 2 diabetes mellitus; CGMP, casein glycomacropeptide.
* Number of subjects per group.
† Time not known.
Table 2. Stimulation of glucagon-like peptide-1 (GLP-1) secretion by protein and protein hydrolysates in vitro*
↑ GLP-1 secretion is incremented v. the control, specified in each row; –, GLP-1 secretion is not altered v. the control, specified in each row; ↓ GLP-1 secretion is reduced v. the control, specified in each row; BSA, bovine serum albumin; Corolase PP, a porcine pancreatic enzyme preparation; DH32, 32 % degree of hydrolysis; DH45, 45 % degree of hydrolysis; DPS, Dutch Protein Services; H, hydrolysis; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; IVD, in vitro digestion with pepsin and pancreatin, always indicates the same hydrolysis conditions as the protein that is compared with; KRB, Krebs–Ringer modified buffer; N.D., hydrolysis conditions not described; PSE, proline-specific endoprotease; UCVP, undigested cuttlefish viscera protein.
* The salivary fluid does not contain enzymes.
† Pea protein origin: DPS, from Dutch Protein Services; Pisane, from Cosucra; SM, from Nutralys; HP90, from Triballat.
‡ Temperature or time not known.
§ This pea hydrolysate did not stimulate GLP-1 secretion; nor did the 10 kDa permeate. Nevertheless, the supernatant fraction obtained after centrifugation increased GLP-1 secretion compared with the control.
‖ Hydrolysis with cuttlefish hepatopancreas digestive proteases.
¶ Hydrolysis with cuttlefish smooth hound intestine digestive proteases.
Table 3. Stimulation of glucagon-like peptide-1 (GLP-1) secretion by protein and protein hydrolysates in animals
N.D., hydrolysis conditions not described; ↑, GLP-1 secretion is incremented v. the control, specified in each row; –, GLP-1 secretion is not altered v. the control, specified in each row; SPF, specific pathogen-free; BW, body weight; ZDF, Zucker diabetic fatty.
* Number of animals per group.
† Sprague–Dawley streptozotocin-induced diabetic rats.
‡ Changes in plasma GLP-1 after oral administration of the protein under the oral glucose tolerance test.
§ Changes in plasma GLP-1 after oral administration of the protein under the intraperitoneal glucose tolerance test.
In vitro studies on the STC-1 cell line showed a clear stimulation by whole dairy proteins (whey, casein, α-lactalbumin, β-lactoglobulin)(Reference Geraedts, Troost and Fischer53–Reference Gillespie and Green55). Moreover, the stimulation of GLP-1 by whey protein β-lactoglobulin in STC-1 cells was partially lost when treated with trypsin (β-lactoglobulin 7·3-fold increase and hydrolysates 2–5·8-fold increase, all v. vehicle control), and totally lost when digested with chymotrypsin for 60 min or more(Reference Gillespie, Calderwood and Hobson54). In the same cell line, the stimulatory effects of whey protein on GLP-1 were lost after extensive hydrolysis with microbial (not described) enzymes, or after a simulated gastrointestinal digestion that included a 90-min treatment with pepsin and a 150-min treatment with Corolase PP(Reference Power-Grant, Bruen and Brennan56). Another study showed that treating whey or casein with trypsin or DPP4 for 30 min did not lead to any loss of GLP-1-stimulatory properties(Reference Geraedts, Troost and Fischer53).
In humans, dairy protein is one of the most studied protein sources involving GLP-1 secretion. Intraduodenal infusion of whey protein hydrolysate has stimulated plasma GLP-1 in lean and obese subjects(Reference Hutchison, Feinle-Bisset and Fitzgerald57), reduced glucose concentration and suppressed energy intake(Reference Ryan, Feinle-Bisset and Kallas58) compared with saline. In these studies, hydrolysed, rather than intact, whey protein was selected because it more closely resembles partially digested protein. Also in patients with T2DM, a whey preload increased GLP-1 secretion, lowered plasma glucose levels and increased the insulin response(Reference Watson, Phillips and Wu59,Reference Jakubowicz, Froy and Ahrén60) compared with water and sucralose, respectively. It has been shown that whey, casein and casein hydrolysates increase GLP-1 secretion(Reference Bendtsen, Lorenzen and Gomes61-Reference Calbet and Holst63). However, there is no agreement about whether there are any differences between their effect on GLP-1 secretion. Hall et al. (Reference Hall, Millward and Long62) showed that 120 min after being ingested, whey protein induced a 2-fold increase in postprandial GLP-1 levels compared with casein protein. On the other hand, when comparing whey, casein and their hydrolysates, Calbet & Holst(Reference Calbet and Holst63) showed that the release of GLP-1 was not influenced by the source or hydrolysis process. Also, a commercially available whey protein hydrolysate showed a higher GLP-1 release 30 min after an oral glucose tolerance test (OGTT) than did casein glycomacropeptide (CGMP), but not compared with whey isolate or α-lactalbumin-enriched whey(Reference Mortensen, Holmer-Jensen and Hartvigsen64) (incremental AUC30min median; 593 (hydrolysate), 270 (CGMP); P=0·045). Thus, the studies performed with whey and whey hydrolysates do not show any differences in the effects of the two sources in terms of GLP-1 secretion. Calbet & Holst(Reference Calbet and Holst63) suggested that this is because the dairy protein hydrolyses rapidly in the intestine and there is a subsequent rise in peripheral amino acids independent of the fractionation.
Other protein sources have also been shown to stimulate GLP-1 release in vivo. A similar rise in rat plasma GLP-1 levels, comparable with that caused by dairy protein, has been observed after pea-protein meals(Reference Overduin, Guérin-Deremaux and Wils65). Furthermore, also in rats, pea protein and pea-protein hydrolysate have been shown to similarly stimulate GLP-1 release, although the hydrolysate showed stronger eating-inhibitory properties(Reference Häberer, Tasker and Foltz66) (total energy intake: 63 (SEM 6) kJ, 46 (SEM 3) kJ, 67 (SEM 5) kJ after pea protein, the hydrolysate and the control, respectively). In vitro studies with STC-1 cells showed that intact pea protein increases GLP-1 release. On the other hand, various pea-protein hydrolysates obtained by enzymic hydrolysis with subtilisin were tested, and only one of them maintained its GLP-1-secretory capacity(Reference Geraedts, Troost and Fischer53).
Cereal protein has also been shown to stimulate GLP-1. Maize protein zein (a major maize protein) hydrolysate attenuated glycaemia in rats under the intraperitoneal glucose tolerance test, associated with enhanced secretions of GLP-1 and GIP(Reference Higuchi, Hira and Yamada67) compared with water. In vitro (GLUTag cells), zein hydrolysate was shown to stimulate GLP-1 release more than egg albumin, country bean and meat hydrolysates(Reference Hira, Mochida and Miyashita68). However, the type of hydrolysis was different in the various sources, so the effect of the protein source per se cannot be concluded from this paper. The stimulation of GLP-1 secretion by maize zein hydrolysate in GLUTag cells is not affected by treatment with pepsin/pancreatin for 60 min, although it is reduced after pronase treatment(Reference Higuchi, Hira and Yamada67) compared with the positive control, KCl 70 mm. The authors suggested that the hydrolysate is not further cleaved by pepsin treatment (the degree of hydrolysis was only 8·6 %).
Oral administration of rice protein hydrolysates also increased total GLP-1 in plasma, and improved glycaemic response in rats(Reference Ishikawa, Hira and Inoue69) (the control used was 2 g/kg of glucose solution). In the same study, rice protein hydrolysates (degree of hydrolysis 5–10 %) stimulated GLP-1 in GLUTag cells, with the potency depending on the enzyme and the time of digestion(Reference Ishikawa, Hira and Inoue69) compared with the blank treatment. The effect of the whole rice protein was not assessed. The authors found that GLP-1 secretion was weaker after 60 min digests with pepsin in rice endosperm protein hydrolysates than after 30 min digests, which suggests that oligo- or larger peptides, rather than small peptides or free amino acids, might be responsible for this stimulation. The results for wheat protein were just the opposite. In GLUTag cells, a low-molecular fraction of wheat protein hydrolysate enhanced GLP-1 secretion while a high-molecular fraction did not(Reference Kato, Nakanishi and Tani70). The low-molecular fraction of wheat protein hydrolysate had a glucose-lowering effect mediated by GLP-1 in rats(Reference Kato, Nakanishi and Tani70) after an oral administration compared with 0·9 % NaCl. Also, in another study in a distal enteroendocrine cell model (GLUTag cells), the effect of wheat hydrolysate on the stimulation of GLP-1 secretion was largely enhanced by pepsin/pancreatin digestion relative to the blank(Reference Chen, Hira and Nakajima71).
For other protein sources, in vitro studies also showed that GLP-1-secreting activity of digested protein was greater than that of the original source. In a study performed with cuttlefish (Sepia officinalis) viscera, a hydrolysate (obtained from digestion with cuttlefish hepato-pancreatic enzymes) was found to exert GLP-1-secreting action while the undigested protein did not(Reference Cudennec, Balti and Ravallec72). These results were found with the samples solubilised in saliva, but they were subjected to further in vitro simulated gastrointestinal digestion (including treatment with pepsin and pancreatin). Results showed that gastrointestinal digestion increased the GLP-1-secretory effects of both the hydrolysate and the initially undigested protein, leading to no differences between the hydrolysate and the non-hydrolysate gastrointestinally digested samples. Also, intestinal digested bovine Hb protein had a greater effect on GLP-1 release than partially digested protein (saliva and gastric digest) in STC-1 cells(Reference Caron, Domenger and Belguesmia73).
Taken together, all these studies prove that several protein sources increase GLP-1 secretion, which is associated to benefits such as food intake or glucose homeostasis regulation. In vivo studies do not fully clarify whether previous hydrolysis of the protein sources with commercial enzymes leads to stronger GLP-1-secreting effects. In vitro data show that many protein sources, including purified proteins, activate GLP-1 release. However, digestion as it might physiologically happen upon protein intake might stimulate or reduce the effect of the undigested protein, depending on the original source. This suggests that some high-molecular-weight peptides might reach enteroendocrine cells and activate GLP-1 secretion, while in other cases the lower-molecular-weight peptides or the amino acids released after digestion are responsible for the secretion.
Mechanisms involved in the effects of protein as glucagon-like peptide-1 secretagogue
The mechanisms through which the proteins and peptides released after protein hydrolysis (either ‘synthetic’ or simulated digestion) act as secretagogues are still not fully understood, but several pathways have been shown to be involved. Studies on the mechanisms through which protein and protein hydrolysates stimulate GLP-1 secretion are carried out using in vitro (i.e. enteroendocrine cell lines such as STC-1 and GLUTag) and ex vivo (i.e. perfused intestine and intestinal explants) models, and also primary cultures.
Many of the studies that focus on the mechanisms that stimulate GLP-1 secretion use commercial meat peptones, that is meat hydrolysates produced by the digestion of meat with proteolytic enzymes which lead to a complex mixture of partially metabolised proteins.
With this protein source, it seems that one key player in the oligopeptide stimulation of GLP-1 release is peptide transporter 1 (PepT1) (Fig. 1). Meat peptone was shown to stimulate GLP-1 secretion in mouse colonic primary culture through PepT1-dependent uptake, followed by an increase in intracellular Ca, and activation of Ca-sensing receptor (CaSR)(Reference Diakogiannaki, Pais and Tolhurst74). Very recently Modvig et al. (Reference Modvig, Kuhre and Holst75) used isolated perfused rat small intestine to study GLP-1 secretion stimulated by meat peptone. The sensory mechanisms underlying the response depended on di-/tripeptide uptake through PepT1 and subsequent basolateral activation of the amino acid-sensing receptor (CaSR) (Fig. 2). CaSR might also be activated by free amino acids taken up from the intestinal lumen by different amino acid transporters(Reference Modvig, Kuhre and Holst75).
Fig. 1. The intestinal transporter form PEPT1 (SLC15A1) is located in apical membranes with a functional coupling to the apical Na+/H+ antiporter (NHE3) for pH recovery from the peptide-transport-induced intracellular acid load. Adapted from Daniel et al. (Reference Daniel, Spanier and Kottra103).
Fig. 2. Illustration of the endocrine L cell and the proposed mechanisms by which peptone stimulates glucagon-like peptide-1 (GLP-1) release. Di-/tripeptides are taken up by PepT1 and are degraded by cytosolic peptidases to their respective amino acids (AA). Intracellular amino acids are then transported to the interstitial side through basolateral amino acid transporters, wherefrom they stimulate the L cells by activating amino acid sensors, like calcium-sensing receptor (CaSR), situated on the basolateral membrane. IP3, inositol trisphosphate; PLC, phospholipase C. Adapted from Modvig et al. (Reference Modvig, Kuhre and Holst75).
It has been pointed out that it is difficult to determine the PepT1-dependent oligopeptide-sensing pathway in GLUTag and STC-1 cell lines, because the expression of endogenous PepT1 is lower than in native L cells(Reference Diakogiannaki, Pais and Tolhurst74). Therefore, the effects of peptones observed in both cell lines may be due to the free amino acids that some of these peptones contain, as has been suggested in an in vitro study on the effects of salmon hydrolysate(Reference Harnedy, Parthsarathy and McLaughlin76) carried out in GLUTag cells. However, other studies on these cell lines do not share this view. As mentioned above, GLP-1 secretion is activated by dairy proteins(Reference Geraedts, Troost and Fischer53–Reference Gillespie and Green55), low-molecular-weight wheat (with less than 1 % free amino acids)(Reference Kato, Nakanishi and Tani70), intact pea-protein(Reference Geraedts, Troost and Fischer53) or peptin-resistant zein hydrolysate(Reference Higuchi, Hira and Yamada67). Furthermore, three synthetic peptide sequences (ANVST, TKAVEH and KAAT) were reported to be able to enhance GLP-1 secretion in STC-1 cells(Reference Caron, Cudennec and Domenger77). The authors concluded that the incretin effect of proteins is associated with the amino acid profile, but the specific amino acid motif that triggers GLP-1 secretion stimulation was not determined. Thus, receptor or peptide transporters other than PepT1 expressed in STC-1 and GLUTag cells might be involved in the peptide stimulation of GLP-1. For instance, one of the mediators suggested was the G protein-coupled receptor family C group 6 subtype A (GPRC6A)(Reference Kato, Nakanishi and Tani70) (Fig. 3).
Fig. 3. Signalling through G protein-coupled receptor family C group 6 subtype A (GPRC6A) in β- or gut cells. GPRC6A can be directly activated by amino acids and use calcium as an allosteric regulator. IP3, inositol triphosphate; PLCβ, phospholipase Cβ; GLP-1, glucagon-like peptide-1; VDCC, voltage-dependent calcium channel. ‡ Described in enterocyte L cells of the small intestine. Adapted from Wauson et al. (Reference Wauson, Lorente-Rodríguez and Cobb104).
Protein hydrolysates are also detected by the umami receptor (T1R1–T1R3 heterodimer)(Reference Raka, Farr and Kelly78) (Fig. 4) and G protein-coupled receptor 92/93 (GPR92/93)(Reference Choi, Lee and Shiu79), which leads to the release of the gut-derived satiety factor cholecystokinin. There is no direct evidence of umami stimulation and GLP-1 secretion, but the T1R1 receptors were co-expressed with GLP-1-expressing STC-1 cells(Reference Wang, Murthy and Grider80), which suggests that umami receptors play a role in GLP-1 signalling.
Fig. 4. The T1R1/T1R3 heterodimer is coupled to a heteromeric G protein, where the Gbc subunit appears to mediate the predominant leg of the signalling pathway. Ligand-binding activates Gbg, which results in activation of phospholipase Cβ2 (PLCβ2), which produces inositol trisphosphate (IP3) and diacylglycerol (DAG). IP3 activates IP3 receptor type 3 (IP3R3) which results in the release of Ca2+ from intracellular stores. AC, adenylyl cyclase; cAMP, cyclic AMP; PDE, phosphodiesterase; PIP2, phosphatidylinositol 4,5-bisphosphate. Adapted from Kinnamon(Reference Kinnamon105).
An increase in intracellular Ca has been reported to be a pathway activated by protein hydrolysates to mediate GLP-1 secretion. Pais et al. (Reference Pais, Gribble and Reimann37) reported that meat peptone-stimulated GLP-1 secretion from primary L cells was also associated with Ca influx through voltage gate Ca channels (Fig. 3). In NCI-H716 human enteroendocrine cells, tetrapeptides, but not single amino acids or any of the dipeptides, tripeptides and pentapeptides tested, were found to induce a robust and selective [Ca2+]i response associated with increased secretion of GLP-1(Reference Le Nevé and Daniel81). Moreover, these effects were not observed in either STC-1 or in GLUTag rodent cells. Interestingly, in the same paper, the authors showed that casein protein hydrolysate elicited an increase in GLP-1 without modulating intracellular Ca.
It has been suggested that GLP-1 secretion is mediated by other intracellular pathways such as extracellular signal-regulated kinase 1/2 (ERK1/2), mitogen-activated protein kinase (MAPK) and p38 MAPK, activated by peptones and mixtures of essential amino acids in NCI-H716 cells(Reference Reimer82).
Altogether, the studies show that which signalling pathways are involved in GLP-1 secretion by different peptide mixtures will depend on the peptide length, the sequences and/or the amino acid composition, and whether there are free amino acids in the mixture. Furthermore, the model studied has to be carefully considered since there are differences in the expression of key genes (such as pepT-1) and some effects might depend on the vectoriality of the system (the capacity to differentiate basolateral and apical processes).
Protein bioactivity on glucagon-like peptide-1 clearance
Like the studies on the effects of protein on GLP-1 secretion, most of the studies on the effects of protein on DPP4 inhibition are performed with protein hydrolysates. Over the past few years, bioactive peptides have shown their potential as DPP4 inhibitors, a research area that is currently expanding. In vitro simulated gastrointestinal digestion has been reported to produce DPP4-inhibitory protein hydrolysates(Reference Lacroix and Li-Chan83,Reference Mojica, Chen and de Mejía84) . Also, hydrolysis with a range of enzymes is used to release DPP4-inhibitory peptides(Reference Ishikawa, Hira and Inoue69,Reference Lacroix and Li-Chan85–Reference Connolly, Piggott and FitzGerald90) . Thus, a wide range of protein sources has been used to obtain hydrolysates, for which DPP4-inhibitory activity has been screened mainly in vitro.
Research has shown that the amino acid sequence plays a much greater role in DPP4-inhibitory activity than other physico-chemical parameters such as length, isoelectric point, hydrophobicity and net charge(Reference Lacroix and Li-Chan91,Reference Nongonierma and Fitzgerald92) . DPP4 preferentially cleaves substrates that bear proline or alanine at their P1 position (Xaa-Pro and Xaa-Ala; where Xaa represents any amino acid) and also acts on substrates that bear other residues, such as glycine, serine, valine and leucine(Reference Lambeir, Durinx and Scharpé93). Hydrophobic and basic residues at the P2 position enhance the affinity for cleavage compared with acidic residues(Reference Power, Nongonierma and Jakeman94). The presence of tryptophan residue at the N-terminal position increases the susceptibility to cleavage. Although the residues at the N-terminal position may have a major impact by inhibiting DPP4, the authors pointed out that the C-terminal amino acid also affects the potency of DPP4 because it is involved in the interaction with the enzyme(Reference Lan, Ito and Ohno95).
To date, some studies have been carried out on the in vivo DPP4-inhibitory effects of the hydrolysates and peptides from dietary proteins. Peptides derived from milk and bean proteins, which have been shown to inhibit the activity of DPP4 in vitro, were also found to have glycaemic effects in mice(Reference Tominaga, Yokota and Tanaka96,Reference Uchida, Ohshiba and Mogami97) as plasma glucose levels decreased after an OGTT. A β-casein-derived peptide LPQNIPPL found in Gouda-type cheese with in vitro DPP4-inhibitory effects has also been tested with animal models. Oral administration of this octapeptide resulted in 1·8-fold lower postprandial glucose AUC; however, insulin plasma levels did not differ(Reference Uenishi, Kabuki and Seto98). In these studies, the authors did not measure plasma DPP4 activity, so it is not known whether the lower blood glucose was caused by inhibition of DPP4 activity. Chicken feet hydrolysates with DPP4-inhibitory activity in vitro improved hyperglycaemia in diet and aged models of glucose homeostasis impairment(Reference Casanova-Martí, Bravo and Serrano99).
As well as hydrolysates from milk and bean protein, in in vivo models hydrolysate from the egg protein lysosyme has also shown a 25 % reduction in blood serum DPP4 activity and a trend towards higher serum GLP-1 levels after 90 min in diabetic rats undergoing chronic treatment(Reference Wang, Landheer and van Gilst100). Streptozotocin-induced diabetic rats were used to evaluate the effects of porcine skin gelatin hydrolysates(Reference Huang, Hung and Jao48), Atlantic salmon skin gelatin(Reference Hsieh, Wang and Hung47), and halibut and tilapia skin gelatin(Reference Wang, Hsieh and Hung49). In all these studies, diabetic animals showed reduced blood glucose levels during OGTT, increased plasma insulin and active GLP-1 levels, and reduced plasma DPP4 activity after a chronic treatment with these proteins compared with water. Diabetic rats treated for 42 d with a daily dose of 300 mg/kg of porcine skin gelatin showed their plasma glucose AUC reduced from 30 000 to 28 000 mg × min/dl (1665 to 1554 mmol × min/l), insulin levels increased 2-fold, active GLP-1 levels reduced from 15 to 13·5 pm and DPP4 activity reduced by half(Reference Huang, Hung and Jao48). In another study in which the animals were treated for 35 d with a daily dose of 300 mg/kg of Atlantic salmon skin gelatin hydrolysate, blood glucose levels were reduced to less than 200 mg/dl (11·1 mmol/l) during OGTT, insulin levels increased 3-fold, active GLP-1 levels increased 1·6-fold and DPP4 activity was reduced from 115·5 to 82·6 % (lower than in normal rats)(Reference Hsieh, Wang and Hung47). When these animals received a 30 d treatment involving a daily dose of 750 mg/kg of halibut (HSGH) or tilapia skin gelatin hydrolysate (TSGH) the plasma glucose was lower than 200 mg/dl (11·1 mmol/l) in the TSGH-treated group. When TSGH was administered, insulin levels were 1·56 g/l, higher than that of HSGH (1·14 g/l) and the diabetic control group (0·43 g/l). The active GLP-1 plasma levels of the diabetic control rats (5·14 pm) were lower than those for TSGH-treated group (13·32 pm) and for HSGH-treated group (7·37 pm) and the DPP4 activity reduced from 115·5 in the diabetic group to 86·6 and 71·6 % in the HSGH- and TSGH-treated groups, respectively(Reference Wang, Hsieh and Hung49). Moreover, rodents receiving halibut and tilapia skin gelatin hydrolysates also showed increased total GLP-1 levels. Therefore, the findings of this study suggest that these hydrolysates exert their anti-hyperglycaemic effect via dual actions of DPP4 inhibition and GLP-1 secretion enhancement. Similarly, the ileal administration of zein protein hydrolysate to rats was found to potentiate the incretin effect when administered before an intraperitoneal glucose tolerance test, resulting in decreased glucose concentration, increased insulin levels, decreased plasma DPP4 activity, and increased total and active GLP-1 secretion compared with water(Reference Mochida, Hira and Hara101). Rice-derived peptides were likewise found to act via dual action. Oral administration increased plasma GLP-1 levels compared with water during an intraperitoneal glucose tolerance test, and ileal administration reduced plasma DPP4 activity and increased the ratio of active GLP-1 to total GLP-1(Reference Ishikawa, Hira and Inoue69) in rats. In vitro studies also showed dual mechanisms for protein hydrolysates; both enhanced GLP-1 secretion and inhibited DPP4, as has been shown for the cuttlefish (Sepia officinalis) viscera protein hydrolysate and bovine Hb hydrolysate(Reference Cudennec, Balti and Ravallec72,Reference Caron, Cudennec and Domenger77) , whey proteins(Reference Power-Grant, Bruen and Brennan56) and chicken feet hydrolysate(Reference Casanova-Martí, Bravo and Serrano99). Therefore, these two mechanisms might also take part in vivo for some protein sources, leading to an increase in active GLP-1 and improve glycaemia.
Human studies, although limited, offer some evidence that food-derived peptides, mostly from dairy protein, act as DPP4 inhibitors(Reference Horner, Drummond and Brennan102). It was shown that a whey preload, consumed before the breakfast meal, reduced glucose levels by 28 % and increased insulin and total GLP-1 levels by 105 and 141 %, respectively, compared with water. Nevertheless, no significant differences in plasma DPP4 activity were found. This could be interpreted as whey protein acting as an endogenous inhibitor of DPP4 in the proximal small intestine, but not in the plasma (intestinal DPP4 activity was not assessed)(Reference Jakubowicz, Froy and Ahrén60). Further studies are needed to examine the potential of casein- and whey-derived peptides, as well as peptides derived from other sources, to act with DPP4 inhibitors in human subjects.
Conclusions
Food proteins target the enteroendocrine system. They directly enhance GLP-1 release from enteroendocrine cells. Current studies suggest that the source of the protein might lead to differences in GLP-1 secretion, although there is not enough literature to enable the different proteins to be compared. The effect of gastrointestinal digestion can also enhance or decrease GLP-1-secreting capacity depending on the protein type. Thus, it is important to consider this digestion when discussing the effects of protein on GLP-1 secretion in vitro. In addition, peptides with DPP4-inhibitory effects can be released during the digestion process, which could modulate the life span of target enterohormones. However, whether this hydrolysis remains important after intestinal digestion in vivo remains to be clarified. Thus, the use of protein/protein hydrolysates to ameliorate situations of glucose derangements is promising, but more research, specifically human studies, is required to define the most effective sources/treatments.
Acknowledgements
The present review was supported by grant no. AGL2017-83477-R from the Spanish government. A. C.-M. received doctoral research grants from the Universitat Rovira i Virgili. M. P. and X. T. are Serra Húnter fellows. The funding providers had no role in the design, analysis or writing of this article.
M. P. conceived the idea, reviewed the literature and drafted and scripted the basis of the manuscript. A. M.-G and A. C.-M. had a role in the design of the tables and writing of the article. All authors critically reviewed the manuscript and approved the final version.
There are no conflicts of interest.
