Roles of the Gut in GlucoseHomeostasisDiabetes Care 2016;39:884–892 | DOI: 10.2337/dc16-0351

The gastrointestinal tract plays a major role in the regulation of postprandialglucose profiles. Gastric emptying is a highly regulated process, which normallyensures a limited and fairly constant delivery of nutrients and glucose to theproximal gut. The subsequent digestion and absorption of nutrients are associatedwith the release of a set of hormones that feeds back to regulate subsequentgastric emptying and regulates the release of insulin, resulting in downregulationof hepatic glucose production and deposition of glucose in insulin-sensitivetissues. These remarkable mechanisms normally keep postprandial glucoseexcursions low, regardless of the load of glucose ingested. When the regulation ofemptying is perturbed (e.g., pyloroplasty, gastric sleeve or gastric bypassoperation), postprandial glycemia may reach high levels, sometimes followed byprofound hypoglycemia. This article discusses the underlying mechanisms.

THE INCRETIN EFFECT

One of the ways to illustrate the role of the gut in glucose homeostasis is to comparethe fate of glucose that has been administered orally or infused intravenously. If thesame amount of glucose is given, the results may not be particularly remarkable, atleast not with relatively small amounts of glucose (25 g or roughly one-half theamount of sugar as in a can of soda). The amount of insulin secreted (estimated fromC-peptide responses) may be almost the same, and any differences in peripheralinsulin concentrations could be interpreted as indicative of differences in hepaticinsulin clearance (1). However, if the intravenous glucose infusion is adjusted so thatthe resulting plasma glucose concentrations are identical to those after oral or smallintestinal administration of glucose, substantially more insulin is secreted with oralor enteral administration, a phenomenon known as the incretin effect (2). Theincretin effect is usually interpreted to indicate secretion of insulinotropic sub-stances (incretin hormones) from the gut (3). Another way of looking at theseexperiments is to compare the amounts of infused glucose required to obtainidentical circulating glucose concentrations: for 25 g given orally, only about 19 gneed to be infused intravenously (4). This method shows that oral administrationactivates a mechanism that clears 6 of the 25 g of glucose from the circulation and issometimes referred to as gastrointestinally induced glucose disposal (GIGD), whichamounts to 24% of the oral dose in this case. With larger amounts of glucose, GIGDbecomes increasingly prominent, and with 100 g of glucose infusion, GIGD is re-sponsible, in healthy subjects, for removing about 80% of the glucose administeredorally (4), which is equivalent to ;20 g left in the circulation, comparable to theamount required tomimic the 25-g oral dose. Indeed, the plasma glucose excursionsafter 25, 50, 100, and up to 125 g of oral glucose (5) are similar, regardless of thedose, whereas excursions would be hugely different if similar amounts were infusedintravenously. GIGD, therefore, minimizes blood glucose excursions very efficiently

1The Novo Nordisk Foundation Center for BasicMetabolic Research and Department of Biomed-ical Sciences, University of Copenhagen, Copen-hagen, Denmark2Wellcome Trust-Medical Research Council Insti-tute of Metabolic Science, Addenbrooke’s Hospi-tal, Cambridge, U.K.3Discipline of Medicine, University of Adelaideand Royal Adelaide Hospital, Adelaide, SouthAustralia, Australia

Corresponding author: Jens Juul Holst, jjholst@sund.ku.dk.

Received 17 February 2016 and accepted 22March 2016.

© 2016 by the American Diabetes Association.Readersmayuse this article as longas thework isproperly cited, the use is educational and not forprofit, and the work is not altered.

See accompanying articles, pp. 857, 861,878, 893, 902, 912, 924, 934, 941, 949,and 954.

Jens Juul Holst,1 Fiona Gribble,2

Michael Horowitz,3 and Chris K. Rayner3

884 Diabetes Care Volume 39, June 2016

METABOLICSU

RGER

Y

so that they are maintained at a relativelyconstant level regardless of the amountof glucose ingested. In patients withtype 2 diabetes, things are very different:in these individuals, GIGD is greatly re-duced, even to as low as 0% (5).What mechanisms underlie GIGD? In

healthy people, increasing oral glucoseloads are associated with dose-relatedincreases in insulin secretion, and thisis a major mechanism because in pa-tients with insulin-deficient type 1 dia-betes, the GIGD is usually close to zero(6). Therefore, it may be concluded thatthe incretin effect (the amplification ofinsulin secretion by gastrointestinal fac-tors) is a major mechanism for normalglucose tolerance. Strictly speaking, theincretin effect refers to enhanced insulinsecretion in response to oral or enteralglucose, but a similar enhancement of in-sulin secretion is also observed after oralintake of both fat and protein (7). Thus,although a formal comparison betweenoral and intravenous administration can-not be made easily for mixed meals, itmay be assumed that an incretin effectplays an important role in postprandialglucose homeostasis in general. As willbe detailed later, the incretin effect isdue to hormones secreted from the gutepithelium, the most important ones be-ing glucose-dependent insulinotropicpolypeptide (GIP) and glucagon-like pep-tide 1 (GLP-1) (8). In agreement with theloss of GIGD, the incretin effect is alsogreatly reduced or entirely missing in pa-tients with diabetes (9).Recently, the existence of a decretin, a

duodenal hormone that inhibits insulinsecretion, has been proposed based onstudies in Drosophila flies. The humancounterpart is supposed tobe the peptideneuromedin U, which may inhibit insulinsecretion (10). However,more studies arerequired to support this mechanism,whichmight be relevant for themetaboliceffects of gastric bypass operations.

MECHANISMS REGULATINGGLUCOSE EXCURSION AFTERORAL INTAKE

Anticipatory EventsThe sight, smell, taste, and sensory im-pulses generated by food presentation,mastication, and swallowing of an appe-tizing meal undoubtedly result in excit-atory signals to the gastrointestinal tract(10). This has been studied in detail inrelation to gastric acid secretion, where

sham feeding (intake and chewing of anappetizing meal that is prevented fromentering the stomach, typically by spit-ting it out) elicits;65% of maximal gas-tric acid secretion (11). In experimentalanimals, a substantial cephalic phase in-sulin response canoften bedemonstrated,but in humans this is less apparent. In fact,in a recent study where glucose levelswere clamped at a permissive level(6 mmol/L [108 mg/dL]), an insulin re-sponse could not be demonstrated,whereas the secretion of pancreaticpolypeptide (known to strongly dependon vagal efferent activity [12]) wasgreatly increased (13). Similarly, dem-onstrating any clear cephalic phase forthe secretion of gut hormones involvedin the incretin effect has been difficult(13), and although it is easy to demon-strate experimentally an effect of theefferent vagus nerves on the secretionof gastric and pancreatic hormones (14)(but not gut hormones [15]), this mech-anism is not clearly activated in theinitial response to meals.

Roles of the StomachBefore nutrients are absorbed into thebloodstream, they normally are retainedfor some time in the stomach. The rate ofgastric emptying highly depends on thecomposition and macronutrient contentof a meal. Fundamental differences existin patterns of gastric emptying of liquidsand solids; the emptying of digestiblesolid components of a meal comprisesan initial lag phase of 20–40 min, whensolids are ground into small particles, fol-lowed by an emptying phase, which ap-proximates an overall linear pattern (16).In contrast, low-nutrient liquids emptyin a monoexponential pattern without asignificant lag phase, which changes toa linear pattern as nutrient density in-creases. Although liquids and solids are in-gested concurrently, as in a mixed meal,liquids empty preferentially, so that;80%is emptied before solids. The rate of gastricemptying usually is relatively unaffectedby the volume of the meal. Accordingly,meals with a higher nutrient content takelonger to empty, but they empty at a sim-ilar ratewhen expressed as kilocalories perminute (17). Therefore, in experimentswith increasing amounts of glucose (25–75–125 g) ingested in the same volume,peak absorption estimated from the liquidphasemarkeracetaminophenwasdelayedfrom 30 to 90 and 180 min, respectively

(5). It is not widely appreciated that a sub-stantial interindividual variation exists ingastric emptying; in healthy persons, thisis usually 1–4 kcal/min (18).

The rate of gastric emptying is nowrecognized to be a major determinant ofnot only the early but also the overallpostprandial glucose excursion in healthyindividuals and thosewith type 2 diabetes(19,20). Consistent with the maintenanceof relatively constant postprandial glu-cose levels, irrespective of the increasingglucose load, the three glucose loads inthe study referred to previously (5) werealso associated with comparable initialrises of incretin hormones, which werethen maintained according to the doseof glucose (very briefly for the low dose,but for up to 3 h after the highest dose)(Fig. 1). In other words, the emptying ofnutrients from the stomach is adjusted sothat the entry rate of the meal into thesmall intestine is relatively constant andresults in an incretin hormone responsethat persists for as long as nutrients arestill entering the gut.

Studies involving direct infusion ofglucose into the small intestine at rateswithin the normal range of gastric emp-tying have provided important insightsinto the relationships of the glycemic,GLP-1, and GIP responses with the rateand site of glucose entry into the smallintestine (22), as discussed in REGULATION

OF GASTRIC EMPTYING. These studies demon-strate that the relationship of glycemiawith small intestinal glucose delivery isnonlinear. Both the viscosity and thephysical state of a meal influence themeal’s gastric handling, with increasingviscosity (as seen for various fermentedmilk products) directly slowing empty-ing (23). The physical state of some nu-trients may change upon entry into thestomach. Emulsified lipids (milk, mayon-naise), for instance, rapidly coalesce toform a lipid phase, which retards theiremptying (depending on posture [24]).Proteins, depending on their rate of ini-tial pepsin-mediated gastric digestionand reaction to gastric acid, may coagu-late and form solids (as seen for ca-seins), which again influences theirdigestion and emptying (25).

Regulation of Gastric EmptyingThe mechanisms that regulate gastricemptying are extremely complex, in-volving gastric and intestinal hormonesas well as long and short reflexes. In

care.diabetesjournals.org Holst and Associates 885

general, the dominant mechanisms thatregulate gastric emptying result fromthe interaction of nutrients with thesmall intestine rather than intragastricmechanisms. Small intestinal feedbackis mediated by the digestion productsof carbohydrates, fats, and proteinsand is regulated by both the lengthand the region of small intestinal expo-sure to nutrients (26). Normal gastricemptying depends on coordination ofthe contractile activity of the proximalstomach, antrum, pylorus, and uppersmall intestine through the extrinsicand intrinsic nervous systems as wellas through neurohumoral pathways.Gastric receptive relaxation (and subse-quent accommodation), which is a vagalreflex, is important because it allowsmeals to enter the stomach without in-creasing the wall tension substantially,allowing symptom-free accommodationof meals of greatly varying sizes.Pyloric motor activity is highly regu-

lated by branches of the vagus nerve(the nerve of Latarjet) in a systemof shortreflexes and hormonal mechanisms thatare stimulated by the entry of nutrientsand gastric fluids (e.g., acid) into the du-odenum. Thus, the introduction of acidinto the duodenummay halt gastric emp-tying altogether while exciting duodenalpropulsive motility (27). Ghrelin, se-creted from the stomach, can accelerategastric emptying (28), whereas secretin,

cholecystokinin (CCK), and somatostatin,secreted from theproximal small intestine,all inhibit both gastric secretion and gastricemptying. Gastric emptying may also beinfluenced by hormones from the moredistal small intestine, including GLP-1 andpeptide YY (PYY) (both the intact hormonePYY 1-36 and the 3-36 metabolite), whichare powerful inhibitors of both secretionand motility (29).

Any interference with these mecha-nisms has the potential to influence expo-sure to and, hence, rate of absorption ofnutrients from the small intestine and, ul-timately, increases in concentrations ofglucose and other nutrients in the blood-stream. The elimination or impairment ofthese powerful mechanisms accounts formany of the changes observed after bar-iatric procedures, such as Roux-en-Y gas-tric bypass and sleeve gastrectomy.

DigestionAlthough glucose as a monosaccharide isnot a major constituent of the normaldiet, it is a major component of the car-bohydrates we ingest in the form ofstarch, lactose, and sucrose. For these nu-trients, enzymatic cleavage is required.Salivary amylase is believed to contrib-ute minimally to the digestion of car-bohydrates, which therefore primarilydepends on pancreatic amylase alongwith brush-border lactase, maltase, iso-maltase, and sucrase activities. These

mechanisms normally do not appear tobe rate limiting for glucose absorption;for example, sucrose seems to be di-gested and absorbed at a rate propor-tional to its load in the small intestine(30). For many complex carbohydrates,however, enzymatic digestion (and asdetermining factors for digestion, smallintestinal propulsive activity and seg-menting motor activity) is far more com-plex. Under normal circumstances, asubstantial amount of dietary carbohy-drate escapes absorption in the small in-testine and is presented to the colon forbacterial fermentation, even in humans(31). The colon can absorb glucose, butthis capability is probably rarely ex-ploited because fermentation proceedsrapidly. An outcome of fermentation isthe production of volatile fatty acids (ac-etate, propionate, and butyrate), whichare importantmetabolic products for co-lonic metabolism, but these fatty acidsare also absorbed and may be used forcombustion and hepatic gluconeogene-sis (particularly proprionate) (32). Gluco-neogenesis might also occur in theintestine, but its contribution to glucosehomeostasis in humans is unknown. Ofnote, increased intestinal gluconeogene-sis was proposed to account for some ofthe changes in glucose homeostasis aftergastric bypass in rats (33) (i.e., the in-creased intestinal gluconeogenesis wouldinhibit hepatic glucose production). Onthe other hand, increased glucose com-bustion within the hypertrophied gut af-ter gastric bypass in rodents has also beensuggested as an explanation (34); how-ever, glucose kinetics after bypass in hu-mans do not seem to be consistent withthis hypothesis (21). Lactase insufficiencyis, of course, limiting for lactose absorp-tion in many adults, and the extent towhich this may represent a problem inbariatric/metabolic surgery is well worthstudy.

Absorption and DepositionThe next step in intestinal glucose han-dling after digestion is transport throughthe sodium–glucose cotransporter 1(SGLT1). Fructose is transported throughGLUT-5 and partly metabolized in the in-testine before reaching the liver for fur-ther metabolism and regulation of itsglucose production. Fructose, therefore,also influences glucose metabolism, butnothing is known about its specific roleafter bariatric/metabolic surgery (35).

Figure 1—Gastric emptying (paracetamol absorption) and its impact on postprandial glucose andincretin hormone levels in healthy volunteers after oral intake of 25, 75, and 125 g of glucose. Insulinlevels (not shown) increased in proportion to emptying rates. See text for further analysis. CTRL,control; IIGI, isoglycemic intravenous glucose infusion. Modified from Bagger et al. (5).

886 Roles of the Gut in Glucose Homeostasis Diabetes Care Volume 39, June 2016

Glucose transport through SGLT1 oc-curs in cotransport with two moleculesof sodium and is very efficient. Someglucose absorption might also occur af-ter translocation of GLUT-2 to the apicalmembrane (36). At any rate, there isnormally very little malabsorption ofglucose. The absolute rate of glucoseabsorption depends on the rate of ex-posure of the small intestine to glucose(as discussed previously), the regionand length of small intestine exposed,and the number of functional entero-cytes and their expression of glucosetransporters (and, therefore, the avail-ability of luminal sodium ions). More-over, small intestinal motility and theflow of luminal content are determinantsof glucose absorption, and these arepotential targets for pharmacologicalmanipulation (37). Bariatric/metabolicsurgery is believed to induce adaptivegrowth, particularly of the alimentarylimb, and to involve the common limb(38), which facilitates absorption. Absorp-tion after surgery is extremely rapid, andcomplete absorption is obtained morerapidly than in individuals who do not un-dergo this surgery (Fig. 2) (21). Evenwhenrapid emptying of glucose from the gas-tric pouch is bypassed by infusing at acontrolled rate directly into the small in-testine, absorption of glucose is morerapid than in healthy control subjects(39).The rapid entry of nutrients from the

gastric pouch after bypass surgery (esti-mated at ;100 kcal/min for a glucosesolution [39]) often is associated with aconstellation of symptoms collectively re-ferred to as dumping, including feelingsof weakness, desire to lie down, impairedconsciousness, sweating, palpitations, andtachycardia. The syndrome is well knownfromothergastricoperationswithaccelera-ted gastric emptying (impaired retention)(40), but understanding of the pathophys-iology is still evolving (41). An increasedosmotic load in the small intestine resultsin water accumulation in the gut lumenthrough a leaky proximal intestinal epithe-lium and potentiates the usual increase inintestinal blood flow in response to nutri-ents (42), overwhelming the capacity ofthe sympathetic nervous system tocompensate and leading to a fall in bloodpressure. A second important element isthe exaggerated secretion of gut pep-tides, including GLP-1, triggered by therapid nutrient entry into the small

intestine, leading to a spike in insulin se-cretion and subsequent hypoglycemia(43), as discussed next. After Roux-en-Y gastric bypass, most patients rap-idly identify nutrients that will causedumping (e.g., soft ice or large amountsof sugar-sweetened drinks).

The increase in intestinal blood flowduring meal ingestion is also importantfor the resulting transport of glucose tothe systemic circulation. Part of the glu-cose is metabolized and used for combus-tion in the gut (as mentioned previously,this has been proposed to represent oneof the mechanisms for improved glucosetolerance after bypass [34]), but the ma-jority passes to the liver where a propor-tion is taken up and used for combustionor stored as glycogen. These processesare highly regulated by insulin. The liveris the main target for insulin action (44),and postprandially, hepatic glucose pro-duction is effectively shut down, whereashepatic glucose uptake is enhanced (21).However, a considerable part of the ab-sorbed glucose passes through the liverso that there is a rise in plasma glu-cose concentration. Simultaneously, glu-cose uptake in tissues expressing glucosetransporters (GLUT-1 to -4) increasesas a function of the increased glucoseconcentration (mass action) (45). Withrising insulin concentrations and increasedinsulin-stimulated translocation ofGLUT-4, peripheral glucose transport isfurther enhanced. Thus, after gastric by-pass surgery, glucose deposition is greatlyaccelerated and may overshoot the rateof absorption whereby glucose concen-trations fall below fasting levels (21). In-deed, in most patients who undergobypass surgery, there is a pronounced de-crease in plasma glucose concentrationsafter a glucose load, and in some, thereis significant hypoglycemia (46), whichmay worsen as insulin resistance dimin-ishes with progressive weight loss. Afterglucose ingestion, secretion of glucagonnormally falls, which, together with ris-ing glucose and insulin concentrations,accounts for the decrease in hepatic glu-cose production (which in the fastingstate is maintained by the basal levelsof glucagon reaching the liver). After gas-tric bypass, meal or glucose ingestion isassociated with an increase in peripheralglucagon concentrations (47), whichwould be expected to result in failure ofthe usual suppression of hepatic glucoseproduction. The consequences of the

paradoxical increase in glucagon for over-all glucose tolerance have not yet beenevaluated, although they could be signif-icant.

The Liver and Neural RegulationWe have discussed the paramount impor-tance of the gastrointestinal tract for post-prandial glucose dynamics as well as theimportance of retention in the stomach,small intestinal motility, enzymatic diges-tion, and transmucosal glucose transportcapacity. The importanceof the regulationof hepatic glucose production and uptakehas also been mentioned, and how thepancreatic hormones insulin and glucagonregulate this has been outlined. We nextdescribe how the gut regulates insulinand glucagon secretion, but before doingso, we may ask to what extent hepaticglucose handling is influenced by thenervous system. The human liver hasboth vagal and sympathetic fibers. Thesympathetic system is known to enhancehepatic glucose production through theactions of the catecholamines on glyco-genolysis, whereas the actions of the va-gus are more unclear. The activity of thesympathetic system is regulated by thehypothalamus as a part of the stress re-action but may also receive input fromthe gut. Glucose-sensitive nerve fibers inthe intestinal mucosa and the hepaticportal system send impulses to the brain(48). These signals must run in the affer-ent vagus or the sensory sympatheticneurons entering the medulla throughthe dorsal horns activating ascendingtracts to the brain, but evidence suggeststhat the vagal mechanisms predominate(49). These mechanisms have been stud-ied mainly in animals, and their impor-tance in humans is unknown. Severaldecades ago, an established therapy forduodenal ulcer disease was truncal va-gotomy together with a gastric drainageprocedure (usually pyloroplasty) to miti-gate the motor consequences of vagot-omy (gastric stasis) (50). Most patientstolerated this procedure well, but theirglucose metabolism was altered, with in-creased but short-lived postprandialrises and an exaggerated overshoot afterglucose challenge, but these were proba-bly consequences of the drainage proce-dure (51). Fasting glucose was littleaffected. The relevance of both efferentand afferent impulses for normal glucosetolerance cannot be determined fromthese experiments. Central nervous

care.diabetesjournals.org Holst and Associates 887

system regulation of hepatic glucose pro-duction is probably mainly exertedthrough the sympathetic innervationreaching the liver, which may be acti-vated through glucose-sensitive hypo-thalamic neurons (52); however, the

role of these systems for normal glucosehomeostasis in humans is not known. In-testinal nutrient sensing by a variety ofmechanisms has been proposed toregulate hepatic glucose productionthrough a reflex pathway comprising

ascending vagal afferents signaling tohypothalamic nuclei, which in turn regu-late hepatic glucose production (53). Ma-jor intestinal signals were believed to beCCK and leptin derived from the stomach(54,55), but the inconspicuous effects of

Figure 2—Plasma glucose concentration (A), total glucose rate of appearance (Ra) in (B) and rate of disappearance (Rd) from (C) the systemiccirculation and endogenous (D) and oral (E) Ra in the systemic circulation before and ;3 months after Roux-en-Y gastric bypass. The dotted linesmark preoperative basal levels. A primed continuous infusion of [6,6-d2]glucose was given for 2 h before determination of steady-state glucose and tracerenrichment, and then a mixed meal (200 mL [394 kcal] carbohydrate 50%, protein 15%, fat 35%) consisting of glucose (48.4 + 1.6 g [U-13C]glucose),rapeseed oil (14.1 g), and casein protein (15.2 g) was consumed slowly over 30 min (21). FFM, fat-free mass.

888 Roles of the Gut in Glucose Homeostasis Diabetes Care Volume 39, June 2016

CCK in human glucoregulation (56) andthe limited survival of leptin in the lumenof the gastrointestinal tract seem to limitthe importance of these signals. Althoughafferent sensory mechanisms undoubt-edly play an important role in the regula-tion of food intake (57), the experimentalsupport for a similar reflex mechanismto regulate hepatic glucose productionin larger mammals and humans is lacking(44).

Gut Endocrine Regulation of GlucoseMetabolismIn addition to regulating postprandialblood glucose levels by controlling thegastric emptying rate, the small intes-tine modulates appetite and pancreatichormone secretion through the releaseof gut hormones. These in turn haveprofound effects on peripheral metabo-lism. Gut endocrine control of appetiteis mediated by hormones such as GLP-1,PYY, and CCK, which are produced byenteroendocrine cells scattered withinthe intestinal epithelium. The anorexi-genic action of these hormones is prob-ably mediated through at least twodistinct pathways: altered activity of in-testinal sensory nerves (particularly theafferent vagus) relayed to the hypothal-amus through the brain stem and directsensing of circulating gut hormone lev-els at the level of the brain stem and/orhypothalamus (58). By controlling appe-tite and energy intake, gut hormones in-fluence the size of adipose tissue stores,which are themajor determinants of pe-ripheral insulin sensitivity. Gut hormonecontrol of pancreatic hormone secretionunderlies the incretin effect describedpreviously and is largely mediated bycirculating GIP and GLP-1. Both thesehormones are produced by enteroendo-crine cells, but whereas GIP is predomi-nantly found in K cells located in theduodenum and proximal small intestine,GLP-1–producing L cells are found inmuch higher density in the jejunum, il-eum, and colon (59). K and L cells res-pond to nutrient absorption rates andare able to sense a variety of digestednutritional components, including carbo-hydrates, fats, and proteins (60). By re-sponding to nutrient absorption ratherthan the mere presence of nutrients inthe gut lumen, increased gut hormonelevels provide a meaningful humoral sig-nal indicating the arrival of nutrients fromthe gut into the bloodstream. Because

GIP-producing K cells are located in theduodenum, their exposure to nutrientsand, therefore, secretion rate highly de-pend on the rate of gastric emptying (5).The more distal location of L cells addscomplexity to the postprandial profile ofGLP-1 concentrations (61). An early ini-tial peak in GLP-1 secretion triggered bythe appearance of nutrients in the jeju-num is associated with slowing of gastricemptying, and at lower nutrient flowrates, the duodenal absorption capacitycan keep up with supply, so the overspillof nutrients into the distal gut is reduced(62). In studies where glucose wasinfused directly into the duodenum,plasma concentrations of GIP increasedin approximately linear fashion with in-creasing rates of infusion. In contrast,there was minimal GLP-1 release below athreshold of 1–2 kcal/min (62) but a sub-stantial GLP-1 response at 3–4 kcal/min(22), suggesting that which of the incretinsis predominant depends on the rate ofgastric emptying. Inconsistency existsbetween studies about whether incretinsecretion is altered in obesity or type 2diabetes compared with health, and al-though decreases in GLP-1 secretionusually can be demonstrated (63), nosubstantial differences, particularly inthe early postprandial phase, have beenfound (64). However, individuals withmorbid obesity manifest accelerated glu-cose absorption from the proximal smallintestine, with a corresponding shift to-ward greater secretion of GIP and less ofGLP-1 than seen in lean individuals (65).

Enteroendocrine cells detect rates ofnutrient absorption through a variety ofmechanisms. Mirroring SGLT1-mediatedglucose absorption across the intestinalbrush border, SGLT1-dependent glucoseuptake by K and L cells directly generatesan electrical signal through the concomi-tant uptake of Na+ ions, in turn triggeringvoltage-gated Ca2+ channel opening andCa2+-dependent hormone release (66).Rates of GLP-1 and GIP secretion by Kand L cells thereby are linked to the ratesof glucose absorption by neighboringenterocytes. Detection of ingested fat,by contrast, seems to largely dependon G-protein–coupled receptors (GPRs)that detect the triacylglycerol digestionproducts, fatty acids (sensed by GPR40and GPR120), and monoacylglycerols(sensed by GPR119) (60). Of note, GPR40-dependent fatty acid detection by L cellsalso seems to depend on the rate of

nutrient absorption, as intestinal perfusionexperiments have indicated that the recep-tor is accessible from the basolateral ratherthan luminal direction (67). Although closeapposition of enteroendocrine cells withthe terminals of enteric neurons has beendescribed, the physiological role of neuro-nal signals in regulating gut hormone secre-tion remains uncertain (15). Expression ofsweet taste receptors on gut endocrinecells, in particular L cells, has been reportedas well as activation of these associatedwith increased secretion (68). Sweet tastereceptor expression is not observed inmouse primary cells (69,70), however, andinhumans, artificial sweetenersdonotelicitincretin hormone secretion (71).

Some of most prominent metaboliceffects of GLP-1 and GIP (in addition tothemotility effects of GLP-1) aremediatedthrough their direct actions on insulin-producing pancreatic b-cells. Receptorsthat detect GLP-1 (GLP1R) and GIP arehighly expressed on the surface ofb-cells,and binding of either hormone to its re-ceptor triggers an intracellular signalingcascade resulting in elevated cytoplasmicconcentrations of cAMP that in turnenhance rates of insulin release (72). Afundamental feature of GLP-1– and GIP-triggered insulin secretion is their glucosedependence: unless glucose concentra-tions are at or above normal levels, thecAMP signal in b-cells is largely ineffec-tive. The translational importance of thisfinding is seen in people with type 2 di-abetes treated with GLP-1 mimetics,whichdunlike sulphonylureas or insulindare not associated with an increase inthe risk of hypoglycemia.

At the same time as they increase in-sulin secretion, GLP-1 and GIP modulatethe release of the other pancreatic islethormones glucagon and somatostatin.Whereas GIP enhances glucagon release(73), GLP-1 suppresses it. The suppres-sion of glucagon secretion may contrib-ute to the beneficial metabolic effect ofGLP-1 because of the consequent reduc-tion in hepatic glucose output (74). Themechanisms by which GLP-1 suppressesglucagon release remains a mystery be-cause GLP1R is only detected on ;10%or less of glucagon-producing a-cells(75). An indirect pathway involvingGLP1R-dependent stimulation of so-matostatin secretion seems likely be-cause somatostatin receptors generallyare coupled to inhibition of their targetcells. This concept is supported by

care.diabetesjournals.org Holst and Associates 889

studies on perfused pancreas thatshowed a loss of GLP-1–inhibited gluca-gon release in the presence of somato-statin receptor inhibitors (76).Exaggerated secretion of GLP-1 result-

ing in enhanced insulin secretion appearsto be one of the important mechanismsunderlying the resolution of diabetes of-ten seen after gastric bypass surgery, asillustrated in experiments involving theGLP1R antagonist exendin 9-39, whicheliminates the effect of the operation oninsulin secretion and impairs glucose tol-erance (77). The power of thismechanismis also illustrated by cases of reactive hy-poglycemia after the operation, whichmay also be prevented by the antagonist(78), or by feeding through gastrostomycatheter, which eliminates the exagger-ated GLP-1 response (79); the latter is inkeeping with the principle that hypoglyce-mia results from extremely rapid carbohy-drate entry into the small intestine (39).

Gut Microbiota and Glucose HandlingAlterations to the gut microbiota havebeen observed in numerous diseases, in-cluding human metabolic diseases suchas obesity, type 2 diabetes, and irritablebowel syndrome, and some animal exper-iments have suggested causality (80).However, few studies have validatedcausality in humans, and the underlyingmechanisms largely remain to be eluci-dated. In studies of themicrobiome of pa-tients randomly assigned to Roux-en-Ygastric bypass or vertical banded gastro-plasty, similar anddurable changes on thegut microbiome resulted in altered levelsof fecal and circulating metabolites (81).In germ-free mice colonized with stoolsfrom the patients, reduced fat depositionwas found. The results suggest that thegut microbiota play a direct role in thereduction of adiposity observed afterbariatric surgery (81) and that thechanges were mainly due to weight loss.On the other hand, elimination of colonicmicrobiota has little or no effect on glu-cose metabolism in humans (82,83).

CONCLUDING REMARKS

This review analyzed important factorsof the gastrointestinal tract for glucoseregulation in humans. Some are intrinsicto the gut and are essential parts of theabsorptive process, including propulsiveactivity, increasing nutrient exposureto the mucosal surface, digestive andabsorptive mechanisms, and intestinal

blood flow. The emptying of the stom-ach, in particular, perhaps representsthe most powerful of all mechanismsregulating postprandial glucose concen-trations. Gastrointestinal function inturn is regulated by metabolic, endo-crine, and neuronal signals generatedin the gut or associated with vagal activ-ity; these signals influence the secretionof gut hormones that regulate appetite(and thereby food intake), gastrointesti-nal motility, and pancreatic endocrinefunctions. Among the gut hormonesare the incretin hormones, which ensurethat postprandial glucose excursions arekept low and relatively constant, despitevariable amounts of ingested carbohy-drates, through actions on insulin se-cretion and gut motility. We view theprandial regulation of hepatic glucoseproduction as resulting mainly fromthe actions of portal glucose and thepancreatic islet hormones, althoughstudies in experimental animals havesuggested that central regulation mayalso play a role. New studies show thatthe intestinal microbiota may also influ-ence metabolism (e.g., through produc-tion of short-chain fatty acids and bileacid metabolism).

Duality of Interest. No potential conflicts ofinterest relevant to this article were reported.Author Contributions. J.J.H., F.G., M.H., andC.K.R. contributed to the preparation of themanuscript.

References1. Madsbad S, Kehlet H, Hilsted J, Tronier B.Discrepancy between plasma C-peptide and in-sulin response to oral and intravenous glucose.Diabetes 1983;32:436–4382. Perley MJ, Kipnis DM. Plasma insulin re-sponses to oral and intravenous glucose: studiesin normal and diabetic subjects. J Clin Invest1967;46:1954–19623. McIntyre N, Holdsworth CD, Turner DS. Newinterpretation of oral glucose tolerance. Lancet1964;2:20–214. Nauck MA, Homberger E, Siegel EG, et al.Incretin effects of increasing glucose loads inman calculated fromvenous insulin and C-peptideresponses. J Clin Endocrinol Metab 1986;63:492–4985. Bagger JI, Knop FK, Lund A, Vestergaard H,Holst JJ, Vilsbøll T. Impaired regulation of theincretin effect in patients with type 2 diabetes.J Clin Endocrinol Metab 2011;96:737–7456. Hare KJ, Vilsbøll T, Holst JJ, Knop FK. Inap-propriate glucagon response after oral com-pared with isoglycemic intravenous glucoseadministration in patients with type 1 diabe-tes. Am J Physiol Endocrinol Metab 2010;298:E832–E837

7. Lindgren O, Carr RD, Deacon CF, et al. Incre-tin hormone and insulin responses to oral ver-sus intravenous lipid administration in humans.J Clin Endocrinol Metab 2011;96:2519–25248. Holst JJ. On the physiology of GIP and GLP-1.Horm Metab Res 2004;36:747–7549. Nauck M, Stockmann F, Ebert R, CreutzfeldtW. Reduced incretin effect in type 2 (non-insulin-dependent) diabetes. Diabetologia 1986;29:46–5210. Alfa RW, Park S, Skelly KR, et al. Suppressionof insulin production and secretion by a decretinhormone. Cell Metab 2015;21:323–33311. Stenquist B, Knutson U, Olbe L. Gastric acidresponses to adequate andmodified sham feed-ing and to insulin hypoglycemia in duodenal ul-cer patients. Scand J Gastroenterol 1978;13:357–36212. Schwartz TW, Stenquist B, Olbe L. Cephalicphase of pancreatic-polypeptide secretion stud-ied by sham feeding in man. Scand J Gastroen-terol 1979;14:313–32013. Veedfald S, Plamboeck A, Deacon CF, et al.Cephalic phase secretion of insulin and otherenteropancreatic hormones in humans. Am JPhysiol Gastrointest Liver Physiol 2016;310:G43–G5114. Holst JJ, Schwartz TW, Knuhtsen S, JensenSL, Nielsen OV. Autonomic nervous control ofthe endocrine secretion from the isolated, per-fused pig pancreas. J Auton Nerv Syst 1986;17:71–8415. Hansen L, Lampert S, Mineo H, Holst JJ.Neural regulation of glucagon-like peptide-1 se-cretion in pigs. Am J Physiol Endocrinol Metab2004;287:E939–E94716. Horowitz M, Dent J, Fraser R, Sun W,Hebbard G. Role and integration of mechanismscontrolling gastric emptying. Dig Dis Sci 1994;39(Suppl.):7S–13S17. Jones KL, O’Donovan D, Russo A, et al. Ef-fects of drink volume and glucose load on gastricemptying and postprandial blood pressure inhealthy older subjects. Am J Physiol Gastroint-est Liver Physiol 2005;289:G240–G24818. Brener W, Hendrix TR, McHugh PR. Regula-tion of the gastric emptying of glucose. Gastro-enterology 1983;85:76–8219. Horowitz M, Edelbroek MA, Wishart JM,Straathof JW. Relationship between oral glu-cose tolerance and gastric emptying in normalhealthy subjects. Diabetologia 1993;36:857–86220. Jones KL, Horowitz M, Carney BI, WishartJM, Guha S, Green L. Gastric emptying in earlynoninsulin-dependent diabetes mellitus. J NuclMed 1996;37:1643–164821. Jacobsen SH, Bojsen-Møller KN, Dirksen C,et al. Effects of gastric bypass surgery on glucoseabsorption and metabolism during a mixedmeal in glucose-tolerant individuals. Diabetolo-gia 2013;56:2250–225422. Ma J, Pilichiewicz AN, Feinle-Bisset C, et al.Effects of variations in duodenal glucose load onglycaemic, insulin, and incretin responses intype 2 diabetes. Diabet Med 2012;29:604–60823. Lentle RG, Janssen PW.Manipulating diges-tion with foods designed to change the physicalcharacteristics of digesta. Crit Rev Food Sci Nutr2010;50:130–14524. Horowitz M, Jones K, Edelbroek MA, SmoutAJ, Read NW. The effect of posture on gastric

890 Roles of the Gut in Glucose Homeostasis Diabetes Care Volume 39, June 2016

emptying and intragastric distribution of oil andaqueous meal components and appetite. Gas-troenterology 1993;105:382–39025. Bohl M, Bjørnshave A, Rasmussen KV, et al.Dairy proteins, dairy lipids, and postprandial li-pemia in persons with abdominal obesity(DairyHealth): a 12-wk, randomized, parallel-controlled, double-blinded, diet interventionstudy. Am J Clin Nutr 2015;101:870–87826. Little TJ, Doran S, Meyer JH, et al. The re-lease of GLP-1 and ghrelin, but not GIP and CCK,by glucose is dependent upon the length ofsmall intestine exposed. Am J Physiol EndocrinolMetab 2006;291:E647–E65527. Allescher HD, Daniel EE, Dent J, Fox JE,Kostolanska F. Neural reflex of the canine pylo-rus to intraduodenal acid infusion. Gastroenter-ology 1989;96:18–2828. Levin F, Edholm T, Schmidt PT, et al. Ghrelinstimulates gastric emptying and hunger in normal-weight humans. J Clin Endocrinol Metab 2006;91:3296–330229. Wettergren A, Schjoldager B, MortensenPE, Myhre J, Christiansen J, Holst JJ. TruncatedGLP-1 (proglucagon 78-107-amide) inhibits gas-tric and pancreatic functions in man. Dig Dis Sci1993;38:665–67330. Gray GM, Ingelfinger FJ. Intestinal absorp-tion of sucrose in man: interrelation of hydroly-sis and monosaccharide product absorption.J Clin Invest 1966;45:388–39831. Wong JM, Jenkins DJ. Carbohydrate digest-ibility and metabolic effects. J Nutr 2007;137(Suppl.):2539S–2546S32. Puertollano E, Kolida S, Yaqoob P. Biologicalsignificance of short-chain fatty acid metabo-lism by the intestinal microbiome. Curr OpinClin Nutr Metab Care 2014;17:139–14433. Troy S, Soty M, Ribeiro L, et al. Intestinalgluconeogenesis is a key factor for early meta-bolic changes after gastric bypass but not aftergastric lap-band in mice. Cell Metab 2008;8:201–21134. Saeidi N, Meoli L, Nestoridi E, et al. Reprog-ramming of intestinal glucose metabolism andglycemic control in rats after gastric bypass. Sci-ence 2013;341:406–41035. Bantle AE, Wang Q, Bantle JP. Post-gastricbypass hyperinsulinemic hypoglycemia: fruc-tose is a carbohydrate which can be safely con-sumed. J Clin Endocrinol Metab 2015;100:3097–310236. Kellett GL. The facilitated component of in-testinal glucose absorption. J Physiol 2001;531:585–59537. Thazhath SS, Marathe CS, Wu T, et al. Theglucagon-like peptide 1 receptor agonist exena-tide inhibits small intestinal motility, flow, tran-sit, and absorption of glucose in healthysubjects and patients with type 2 diabetes: arandomized controlled trial. Diabetes 2016;65:269–27538. Rhee NA, Wahlgren CD, Pedersen J, et al.Effect of Roux-en-Y gastric bypass on thedistribution and hormone expression of small-intestinal enteroendocrine cells in obese pa-tients with type 2 diabetes. Diabetologia 2015;58:2254–225839. Nguyen NQ, Debreceni TL, Bambrick JE,et al. Rapid gastric and intestinal transit is a ma-jor determinant of changes in blood glucose,intestinal hormones, glucose absorption and

postprandial symptoms after gastric bypass.Obesity (Silver Spring) 2014;22:2003–200940. Dirksen C, Jørgensen NB, Bojsen-Møller KN,et al. Gut hormones, early dumping and restingenergy expenditure in patients with good andpoor weight loss response after Roux-en-Y gas-tric bypass. Int J Obes 2013;37:1452–145941. Berg P, McCallum R. Dumping syndrome: areview of the current concepts of pathophysiol-ogy, diagnosis, and treatment. Dig Dis Sci 2016;61:11–1842. Trahair LG, Horowitz M, Hausken T, Feinle-Bisset C, Rayner CK, Jones KL. Effects of exoge-nous glucagon-like peptide-1 on the blood pressure,heart rate, mesenteric blood flow, and glycemicresponses to intraduodenal glucose in healthyolder subjects. J Clin Endocrinol Metab 2014;99:E2628–E263443. Goldfine AB, Mun EC, Devine E, et al. Pa-tients with neuroglycopenia after gastric bypasssurgery have exaggerated incretin and insulinsecretory responses to a mixed meal. J Clin En-docrinol Metab 2007;92:4678–468544. Ramnanan CJ, Edgerton DS, CherringtonAD. Evidence against a physiologic role for acutechanges in CNS insulin action in the rapid regu-lation of hepatic glucose production. Cell Metab2012;15:656–66445. Schwartz MW, Seeley RJ, Tschop MH, et al.Cooperation between brain and islet in glucosehomeostasis and diabetes. Nature 2013;503:59–6646. Patti ME, McMahon G, Mun EC, et al. Se-vere hypoglycaemia post-gastric bypass requir-ing partial pancreatectomy: evidence forinappropriate insulin secretion and pancreaticislet hyperplasia. Diabetologia 2005;48:2236–224047. Jørgensen NB, Jacobsen SH, Dirksen C, et al.Acute and long-term effects of Roux-en-Y gas-tric bypass on glucose metabolism in subjectswith type 2 diabetes and normal glucose toler-ance. Am J Physiol Endocrinol Metab 2012;303:E122–E13148. Niijima A. Neural mechanisms in the controlof blood glucose concentration. J Nutr 1989;119:833–84049. Holzer HH, Turkelson CM, Solomon TE,Raybould HE. Intestinal lipid inhibits gastricemptying via CCK and a vagal capsaicin-sensitiveafferent pathway in rats. Am J Physiol 1994;267:G625–G62950. Kronborg O. Clinical results 6 to 8 years af-ter truncal vagotomy and drainage for duodenalulcer in 500 patients. Acta Chir Scand 1975;141:657–66351. Plamboeck A, Veedfald S, Deacon CF, et al.Characterisation of oral and i.v. glucose han-dling in truncally vagotomised subjects with py-loroplasty. Eur J Endocrinol 2013;169:187–20152. Seoane-Collazo P, Fernø J, Gonzalez F, et al.Hypothalamic-autonomic control of energy ho-meostasis. Endocrine 2015;50:276–29153. Duca FA, Bauer PV, Hamr SC, Lam TK. Glu-coregulatory relevance of small intestinal nutri-ent sensing in physiology, bariatric surgery, andpharmacology. Cell Metab 2015;22:367–38054. Rasmussen BA, Breen DM, Luo P, et al. Du-odenal activation of cAMP-dependent proteinkinase induces vagal afferent firing and lowersglucose production in rats. Gastroenterology2012;142:834–843

55. Rasmussen BA, Breen DM, Duca FA, et al.Jejunal leptin-PI3K signaling lowers glucose pro-duction. Cell Metab 2014;19:155–16156. Baum F, Nauck MA, Ebert R, et al. Role ofendogenously released cholecystokinin in de-termining postprandial insulin levels in man: ef-fects of loxiglumide, a specific cholecystokininreceptor antagonist. Digestion 1992;53:189–19957. Cummings DE, Overduin J. Gastrointestinalregulation of food intake. J Clin Invest 2007;117:13–2358. Grill HJ, Hayes MR. Hindbrain neurons as anessential hub in the neuroanatomically distrib-uted control of energy balance. Cell Metab2012;16:296–30959. Sjolund K, Sanden G, Hakanson R, Sundler F.Endocrine cells in human intestine: an immuno-cytochemical study. Gastroenterology 1983;85:1120–113060. Gribble FM, Reimann F. Enteroendocrinecells: chemosensors in the intestinal epithelium.Annu Rev Physiol 2016;78:277–29961. Miholic J, Orskov C, Holst JJ, Kotzerke J,Meyer HJ. Emptying of the gastric substitute,glucagon-like peptide-1 (GLP-1), and reactivehypoglycemia after total gastrectomy. Dig DisSci 1991;36:1361–137062. Schirra J, Katschinski M, Weidmann C, et al.Gastric emptying and release of incretin hor-mones after glucose ingestion in humans. JClin Invest 1996;97:92–10363. Færch K, Torekov SS, Vistisen D, et al. GLP-1response to oral glucose is reduced in prediabe-tes, screen-detected type 2 diabetes, and obe-sity and influenced by sex: the ADDITION-PROstudy. Diabetes 2015;64:2513–252564. Nauck MA, Vardarli I, Deacon CF, Holst JJ,Meier JJ. Secretion of glucagon-like peptide-1(GLP-1) in type 2 diabetes: what is up, what isdown? Diabetologia 2011;54:10–1865. Nguyen NQ, Debreceni TL, Bambrick JE,et al. Accelerated intestinal glucose absorptionin morbidly obese humans: relationship to glu-cose transporters, incretin hormones, and gly-cemia. J Clin Endocrinol Metab 2015;100:968–97666. Gribble FM, Williams L, Simpson AK,Reimann F. A novel glucose-sensing mechanismcontributing to glucagon-like peptide-1 secre-tion from the GLUTag cell line. Diabetes 2003;52:1147–115467. Christensen LW, Kuhre RE, Janus C,Svendsen B, Holst JJ. Vascular, but not luminal,activation of FFAR1 (GPR40) stimulates GLP-1secretion from isolated perfused rat small intes-tine. Physiol Rep 2015;3:368. Jang HJ, Kokrashvili Z, Theodorakis MJ,et al. Gut-expressed gustducin and taste re-ceptors regulate secretion of glucagon-likepeptide-1. Proc Natl Acad Sci U S A 2007;104:15069–1507469. Parker HE, Habib AM, Rogers GJ, GribbleFM, Reimann F. Nutrient-dependent secretionof glucose-dependent insulinotropic polypep-tide from primary murine K cells. Diabetologia2009;52:289–29870. Tolhurst G, Reimann F, Gribble FM. Nutri-tional regulation of glucagon-like peptide-1 se-cretion. J Physiol 2009;587:27–3271. Maersk M, Belza A, Holst JJ, et al. Satietyscores and satiety hormone response after

care.diabetesjournals.org Holst and Associates 891

sucrose-sweetened soft drink compared withisocaloric semi-skimmed milk and with non-caloric soft drink: a controlled trial. Eur J ClinNutr 2012;66:523–52972. Drucker DJ. The biology of incretin hor-mones. Cell Metab 2006;3:153–16573. Lund A, Vilsbøll T, Bagger JI, Holst JJ, Knop FK.The separate and combined impact of the intes-tinal hormones, GIP, GLP-1, and GLP-2, on gluca-gon secretion in type 2 diabetes. Am J PhysiolEndocrinol Metab 2011;300:E1038–E104674. Hvidberg A, Nielsen MT, Hilsted J, Orskov C,Holst JJ. Effect of glucagon-like peptide-1 (pro-glucagon 78-107amide) on hepatic glucose pro-duction in healthy man. Metabolism 1994;43:104–10875. Richards P, Parker HE, Adriaenssens AE, et al.Identification and characterization of GLP-1 re-ceptor-expressing cells using a new transgenicmouse model. Diabetes 2014;63:1224–1233

76. de Heer J, Rasmussen C, Coy DH, Holst JJ.Glucagon-like peptide-1, but not glucose-dependent insulinotropic peptide, inhibitsglucagon secretion via somatostatin (receptor sub-type 2) in the perfused rat pancreas. Diabetologia2008;51:2263–227077. Jørgensen NB, Dirksen C, Bojsen-Møller KN,et al. Exaggerated glucagon-like peptide 1 re-sponse is important for improved b-cell func-tion and glucose tolerance after Roux-en-Ygastric bypass in patients with type 2 diabetes.Diabetes 2013;62:3044–305278. Salehi M, Gastaldelli A, D’Alessio DA.Blockade of glucagon-like peptide 1 receptorcorrects postprandial hypoglycemia after gas-tric bypass. Gastroenterology 2014;146:669–68079. McLaughlin T, Peck M, Holst J, Deacon C.Reversible hyperinsulinemic hypoglycemia aftergastric bypass: a consequence of altered

nutrient delivery. J Clin Endocrinol Metab2010;95:1851–185580. Wu H, Tremaroli V, Backhed F. Linking mi-crobiota to human diseases: a systems biologyperspective. Trends Endocrinol Metab 2015;26:758–77081. Tremaroli V, Karlsson F, Werling M, et al.Roux-en-Y gastric bypass and vertical bandedgastroplasty induce long-term changes on thehuman gut microbiome contributing to fatmass regulation. Cell Metab 2015;22:228–23882. Mikkelsen KH, Frost M, Bahl MI, et al. Effectof antibiotics on gut microbiota, gut hormonesand glucose metabolism. PLoS One 2015;10:e014235283. Vrieze A, Out C, Fuentes S, et al. Impact oforal vancomycin on gut microbiota, bile acidmetabolism, and insulin sensitivity. J Hepatol2014;60:824–831

892 Roles of the Gut in Glucose Homeostasis Diabetes Care Volume 39, June 2016