Mechanisms of DiabetesImprovement FollowingBariatric/Metabolic SurgeryDiabetes Care 2016;39:893–901 | DOI: 10.2337/dc16-0145

More than 20 years ago, Pories et al. published a seminal article, “Who WouldHave Thought It? An Operation Proves to Be the Most Effective Therapy for Adult-Onset Diabetes Mellitus.” This was based on their observation that bariatric sur-gery rapidly normalized blood glucose levels in obese people with type 2 diabetesmellitus (T2DM), and 10 years later, almost 90% remained diabetes free. Porieset al. suggested that caloric restriction played a key role and that the relativecontributions of proximal intestinal nutrient exclusion, rapid distal gut nutrientdelivery, and the role of gut hormones required further investigation. These find-ings of T2DM improvement/remission after bariatric surgery have been widelyreplicated, together with the observation that bariatric surgery prevents or delaysincident T2DM. Over the ensuing two decades, important glucoregulatory roles ofthe gastrointestinal (GI) tract have been firmly established. However, the physi-ological and molecular mechanisms underlying the beneficial glycemic effects ofbariatric surgery remain incompletely understood. In addition to the mechanismsproposed by Pories et al., changes in bile acid metabolism, GI tract nutrientsensing and glucose utilization, incretins, possible anti-incretin(s), and the intes-tinal microbiome are implicated. These changes, acting through peripheral and/orcentral pathways, lead to reduced hepatic glucose production, increased tissueglucose uptake, improved insulin sensitivity, and enhanced b-cell function. Aconstellation of factors, rather than a single overarching mechanism, likely medi-ate postoperative glycemic improvement, with the contributing factors varyingaccording to the surgical procedure. Thus, different bariatric/metabolic proce-dures provide us with experimental tools to probe GI tract physiology. Embracingthis approach through the application of detailed phenotyping, genomics, metab-olomics, and gutmicrobiome studies will enhance our understanding of metabolicregulation and help identify novel therapeutic targets.

REGULATION OF GLUCOSE HOMEOSTASIS: HISTORICAL INSIGHTS FROMGASTROINTESTINAL SURGERY

Impaired glucose homeostasis is characterized by a combination of insulin resis-tance and defectiveb-cell function that worsens with time. Blood glucose levels rise,and type 2 diabetes mellitus (T2DM) ensues only when b-cells are incapable of re-leasing sufficient insulin to compensate for prevailing insulin resistance (1). Genome-wide association studies have identified that b-cell dysfunction has a clear geneticcomponent (2). However, environmental factors also influence insulin resistance andb-cell function. In recent years, the remarkable effect of bariatric surgery on glucoseregulation has helped identify key glucoregulatory roles for the gastrointestinal (GI)tract.

1Centre for Obesity Research, Department ofMedicine, University College London, London,U.K.2Bariatric Centre for Weight Management andMetabolic Surgery, University College LondonHospital, London, U.K.3National Institute for Health Research, Biomed-ical Research Centre, University College LondonHospital, London, U.K.4VA Puget Sound Health Care System and Diabe-tes and Obesity Center of Excellence, Universityof Washington, Seattle, WA

Corresponding author: Rachel L. Batterham,r.batterham@ucl.ac.uk.

Received 22 January 2016 and accepted 21March 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, 884, 902, 912, 924, 934, 941,949, and 954.

Rachel L. Batterham1,2,3 and

David E. Cummings4

Diabetes Care Volume 39, June 2016 893

META

BOLIC

SURGER

Y

The notion that rerouting the GI tractalters glycemia is not new, with reports inthe 1930s of altered glucose tolerancecurves in patients afterGI surgery for pep-tic ulcer disease (3). In 1942, Evensen de-scribed alimentary hypoglycemia thatwas observed several years after pepticulcer disease surgery, and he proposedincreased insulin sensitivity as the under-lying mechanism (4). Interestingly, thesehypoglycemic accounts are strikingly simi-lar to the hypoglycemia experienced by aminority of patients after Roux-en-Y gastricbypass (RYGB).Bariatric surgical procedures were

developed in the 1950s to reduce bodyweight. Since the 1970s, however, therehave been anecdotal reports of rapidpostoperative T2DM remission. In 1984,bariatric surgery was reported to improveglucose tolerance in insulin-treated se-verely obese patients (5). In a 1992 article,“Is Type II Diabetes Mellitus (NIDDM) aSurgical Disease?”, Pories et al. reportedT2DM reversal in 78% of patients who un-derwent gastric bypass (6). However, itwas their subsequent article in 1995,“Who Would Have Thought It? An Oper-ation Proves to Be the Most EffectiveTherapy for Adult-Onset Diabetes Melli-tus,” that catalyzed research into identify-ing the mechanisms by which bariatricsurgery improves glucose homeostasisand promotes T2DM remission (7).Historically, bariatric operations were

thought to promoteweight loss by causinggastric restriction and/or malabsorption.However,newermechanistic studies, inpar-allel with establishment of the GI tract as akey regulator of energyandglucosehomeo-stasis, have made it clear that alternativemechanisms primarily mediate the weight-reducing and antidiabetes benefits of mostbariatric/metabolic operations. Discreteparts of the GI tract differentially influenceglucose homeostasis; hence, the underly-ing mechanisms contributing to improvedglucose tolerance and clinical outcomesundoubtedly differ among anatomical pro-cedures (displayed in Fig. 1). Indeed, T2DMremission rates differ according to surgerytype: lowest for laparoscopic adjustablegastric banding (LAGB) and highest forbiliopancreatic diversion (BPD) (8).

OVERVIEW OF ROLES OF THE GITRACT IN REGULATING ENERGYAND GLUCOSE HOMEOSTASIS

The presence of nutrients in the GI tracttriggers a complex series of hormonal

and neural responses that regulate en-ergy and glucose homeostasis. Gut pep-tides are synthesized and secreted fromenteroendocrine cells of the epithelialmucosa. For example, I cells and K cellsof the proximal intestine primarily pro-duce cholecystokinin and glucose-stimulated insulinotropic polypeptide(GIP), respectively,while L cells of thedistalintestine primarily produce glucagon-likepeptide 1 (GLP-1), GLP-2, oxyntomodulin,and peptide YY (PYY)dmost of whichcontribute to satiation and/or satiety.

Gut peptides and nutrients act on pe-ripheral and central targets via the circu-lation and/or through afferent nerves.Oral glucose promotes greater insulin re-lease than does isoglycemic glucose ad-ministered parenterally, a phenomenonknown as the incretin effect, which ispredominantlymediated by the incretinsGLP-1 and GIP. These peptides enhanceglucose-stimulated insulin secretion, in-sulin action, and b-cell function. Patientswith T2DM exhibit a blunted incretineffect, coupled with attenuated GIP

Figure 1—Bariatric/metabolic operations discussed in this article. A: RYGB. The stomach isdivided into two compartments, leaving only the small upper chamber in digestive continuity.Food passes from there to the proximal jejunum, bypassing most of the stomach, the duode-num, and a small portion of jejunum. B: VSG. Most of the stomach (primarily the body andfundus) is excised, leaving a narrow sleeve along the lesser curvature. Nutrients follow thenormal route through the GI tract. C: LAGB. An inflatable silicon ring encircling the upperstomach is serially adjusted to optimize the diameter of a tight aperture that hinders foodflow. D: BPD. A large majority of the small intestine is bypassed, purposely causing malabsorp-tion. E: DJB. A modest segment of proximal intestine is bypassed, as in RYGB, without com-promising gastric capacity.

894 Antidiabetes Mechanisms of Metabolic Surgery Diabetes Care Volume 39, June 2016

action and reduced circulating GLP-1levels (9). From results based largelyon animal experiments, some investi-gators have postulated the existenceof anti-incretins: putative nutrient-stimulated GI neuroendocrine signalsemanating from the proximal gut tocounterbalance the effects of incretinsand other postprandial glucose-loweringmechanisms (10). The findings that pro-teins secreted from the small intestineof diabetic (but not nondiabetic) ro-dents induce muscle insulin resistancein cell-based assays and in vivo supportthis concept and are consistent withpreliminary human observations (11).Although specific human anti-incretinshave not yet been clearly identi-fied, a strong candidate was recentlydiscovered in Drosophila (limostatin,named a decretin by these investiga-tors) (12).From results based primarily on mech-

anistic animal studies plus complementary

associative observations in humans, bileacids (BAs) are now believed to be im-portant regulators of energy balanceand metabolism, primarily via the nu-clear farnesoid X receptor (FXR) andthe G-protein–coupled receptor TGR5(13) (Fig. 2). Postprandially, BAs are re-leased into the duodenum to mix withingested nutrients. They are then ac-tively reabsorbed from the terminalileum and returned via the portal circu-lation to the liver. A small percentageof BAs are deconjugated by gut bac-teria, forming secondary BAs, whichare reabsorbed or excreted in feces(14).

The transintestinal BA flux activatesintestinal FXR, inducing synthesis andsecretion into the circulation of theileal-derived enterokine FGF-19 (FGF-15 in mice). FGF-19 inhibits expression ofcholesterol 7 a-hydroxylase-1 (CYP7A1),the rate-limiting step of BA synthesis(13). In mice, FGF-15 can improve

glucose tolerance by regulating insulin-independent glucose efflux and he-patic glucose production. BAs acting viaTGR5 stimulate L-cell secretion of GLP-1and PYY. Directly and indirectly throughthe FXR-induced antimicrobial peptides,BAs also regulate gut microbiota com-position. This, in turn, has been linked tothe pathogenesis of obesity and T2DMin rodents, and correlative data in hu-mans are consistent with that. Recently,an orally active gut-restricted FXR agonistwas shown to restore glucose homeo-stasis in mice with diet-induced obesityand glucose intolerance by inhibit-ing hepatic glucose production (15).Thus, there is complex cross talk amongBAs, gut hormones, FGF-19, and the mi-crobiome, which in turn influences glu-cose homeostasis.

After a meal, nutrients, hormones, andneural signals inform the brain of the cur-rent nutritional status. An emerging pic-ture based on animal studies suggests

Figure 2—Diagram of some of the metabolic effects and cross talk among BAs, GLP-1, and FGF-19.

care.diabetesjournals.org Batterham and Cummings 895

that glucose homeostasis may be influ-enced by a gut–brain–liver axis in whichgut-derived signals acting centrally regu-late hepatic glucose production (16).Using intraintestinal nutrient infusionsin rodents, Lam and colleagues (17,18)demonstrated that the duodenum andjejunum sense nutrients and initiatenegative-feedback mechanisms througha gut–brain–liver neuronal axis to regu-late glycemia, mainly via reducing he-patic glucose production. Interestingly,studies comparing infusion of nutrientsinto the midjejunum compared with theduodenum inhumanswith T2DMrevealedthat jejunal infusion led to greater insulinsensitivity for glucose and fatty acids (19).

Mechanisms Mediating the Effects ofBariatric/Metabolic Surgery onGlucose HomeostasisIn their seminal article from .20 yearsago (7), Pories et al. speculated thatthe very rapid post-RYGB improvementof glucose tolerance, which typicallyoccurs before significant weight loss,might result from acute caloric restric-tion plus possible additional conse-quences of excluding ingested nutrientsfrom the proximal intestine and/orexpediting delivery of nutrients to thedistal intestine. At that time, HarveySugerman’s group published that guthormone changes were more profoundafter RYGB than the purely mechanicalvertical-banded gastroplasty, perhapshelping explain the superior weight-reducing and antidiabetes effects ofRYGB compared with vertical-bandedgastroplasty (20). Weight-independentantidiabetes effects of proximal intesti-nal bypass were subsequently demon-strated in rats in a landmark article byRubino et al. (21) on duodenal-jejunalbypass (DJB), which replicates justthe intestinal component of RYGB,and those findings have held up in nu-merous human studies. Similarly, thebeneficial effects of enhanced distal in-testinal nutrient exposure were proven inrats with ileal interposition surgery (22),and these too have translated to humans.In the 20 years since the classic article

by Dr. Pories and colleagues, mechanis-tic knowledge about bariatric/metabolicsurgery has greatly expanded, althoughmany issues remain unclear and contro-versial. This is partly related to method-ological issues in patients studied andprotocols used, and especially whether

test nutrients are administered via theGI tract or parenterally (Table 1). RYGBand vertical sleeve gastrectomy (VSG)(Fig. 1) markedly increase the rate atwhich ingested nutrients enter the smallintestine. Blood glucose levels rise rap-idly, achieving earlier and higher peaks,followed by lower nadirs associatedwith increased GLP-1 and insulin re-sponses. Furthermore, whether studiesin rodents can be extrapolated to hu-mans is unclear, given that rodents de-plete liver glycogen stores quickly andhave a greater capacity for glycogenoly-sis than humans (23). Proposed mecha-nisms underlying the glycemic effects ofbariatric surgery will now be discussed(Fig. 3).

CALORIC RESTRICTION ANDWEIGHT LOSS

Improvement in glucose homeostasisafter RYGB, VSG, and BPD typically be-gins within days of surgery, before sig-nificant weight loss occurs. Thus, totalbody weight loss per se is unlikely toplay a significant role in mediating earlyglycemic improvements. Further evi-dence for effects independent of weightloss stems from studies showing thatRYGB leads to a greater oral glucose tol-erance compared with patients withequivalent weight loss after LABG or ca-loric restriction (24,25). However, thetime taken to achieve a given weightreduction with caloric restriction orLABG in these studies was longer than

after RYGB, suggesting that the degreeof caloric restriction was greater afterRYGB and/or that energy expenditureis higher after that operation. Further-more, RYGB causes rapid passage of oralnutrients into the intestine, with earlyhigher peaks of blood glucose, GLP-1,and insulin, confounding direct compar-isons of oral glucose tolerance tests af-ter caloric restriction versus RYGB. Toaddress these issues, Korner and col-leagues compared RYGB patients withpatients consuming a very low-caloriediet (VLCD, 500 kcal/day) and used fre-quently sampled intravenous glucosetolerance tests rather than oral meals(26). This approach showed that VLCDand RYGB produced comparable im-provements in insulin sensitivity andb-cell function in the absence of acutelyelevated nutrient-stimulated GLP-1 lev-els. These findings suggest that acutepostoperative caloric restriction is a sig-nificant contributor and that marked en-ergy deficit exerts a glucose-loweringeffect independent of weight loss. Othershave argued that the inflammatory insultof surgery per se, which is likely to impairinsulin sensitivity, makes a direct com-parison of VLCD and surgery invalid.Studies by Lingvay et al. suggest that asurgery-related stress response occurs.They compared the effects of VLCD ver-sus RYGB in patients with T2DM, withindividuals serving as their own controls.After 10 days of VLCD, there was a sig-nificant improvement in fasting glu-cose, peak glucose, and glucose areaunder the curve during a mixed-mealchallenge test, but not after RYGB,despite a greater GLP-1 response withthe latter (27).

Patients with T2DM have increasedhepatic fat and pancreatic fat comparedwith BMI-matched normoglycemic pa-tients (28). Taylor and colleagues foundthat among patients with T2DM whounderwent a VLCD or RYGB, hepatic fatcontent decreased rapidly in parallelwith improved hepatic insulin sensitivityand normalization of fasting plasma glu-cose levels within 7 days (29). Reductionin pancreatic fat content was slowerwith both VLCD and RYGB, taking 8weeks to normalize to nondiabetic lev-els. Pancreatic fat reduction was accom-panied by restoration of the first-phaseinsulin (29,30). These studies highlightthe role of hepatic and pancreatic fatin the pathogenesis of T2DM and also

Table 1—Factors affecting studiesexamining glycemic effects of bariatricproceduresSex

Age

Ethnicity

Genetics

BMI and fat distribution

Glycemic status (normoglycemic,impaired glucose tolerance, or T2DM)

Duration of T2DM

Medications (and duration of stoppingthese before being studied)

Impact of preoperative liver-reducing diet

Stimulus used (oral or intravenous)

For oral nutrient stimuli (energy load,macronutrient composition, liquid/solid, and volume)

Method used to assess glycemic response

Time after surgery

896 Antidiabetes Mechanisms of Metabolic Surgery Diabetes Care Volume 39, June 2016

the potential of a profound energydeficit to improve T2DM. However,VLCD-induced weight loss leads to com-pensatory homeostatic changes, includingincreased hunger, increased circulatingghrelin, and reduced circulating GLP-1and PYYdchanges that are likely to con-tribute to the high degree of weightrecidivism with dieting (31). In con-trast, postprandial GLP-1 and PYY levelsincrease after RYGB and VSG, whileghrelin usually falls. These changesprobably contribute to reduced appetiteand taste changes that favor ongoingweight loss and weight-loss mainte-nance (32).The above studies suggest that b-cell

function may not improve until 8 weeksafter RYGB. However, using intravenousglucose administration, Martinussenet al. demonstrated enhanced, althoughnot normalized, first-phase insulin re-sponse and improved HOMA-b in pa-tients with T2DM within 1 week ofsurgery (33). These findings suggest thatRYGB has an early beneficial effect onb-cell function. The reduced glucotoxicity

resulting from normalized glucose levelslikely contributes to these changes; how-ever, similar studies with VLCD alone areneeded to investigate this.

Most T2DM remission occurs duringthe first 8 postoperative weeks, whereassome studies report that peripheral in-sulin resistance improves over a longertime and may be at least partly relatedto weight loss (30). In the long-term,GLP-1 or other surgery-related media-tors might possibly stimulate b-cell re-generation or hypertrophy, which couldprotect against T2DM recurrence evenwith weight regain, although this is con-troversial (see GLP-1).

The role that total weight loss plays inT2DM initial remission and longer-termmaintenance of remission is controver-sial and likely differs among bariatric/metabolic operations. A relationshipbetween the percentage of weight loss(%WL) and T2DM remission rates, whichis sometimes but not always observed, canbe interpreted in three different ways.Firstly, that %WL plays a role in mediat-ing T2DM remission; secondly, that the

presence of T2DM adversely affectsweight loss; and thirdly, that commonmechanistic factors drive both T2DMremission and weight loss. Furthermore,the relationship between T2DM remis-sion and %WL is likely to vary dependingon the duration and severity of T2DMand individual patient characteristics.

GLP-1

It is widely accepted that postprandialcirculating GLP-1 levels are markedly el-evated after RYGB, BPD, and VSG, to-gether with elevated peak postprandialinsulin levels, whereas plasma GLP-1levels remain unaltered with caloricrestriction and LAGB (34). This exag-gerated postmeal GLP-1 secretion ispresent from day 2 postsurgery and per-sists long-term. The hindgut hypothesisasserts that enhanced delivery of nutri-ents and/or bile to the distal GI tract,as a consequence of GI anatomical re-arrangement, rapid gastric emptying,and/or other factors, leads to increasedstimulationof thedistal small intestine andcolon, with increased nutrient-stimulated

Figure 3—Schematic of potential mechanisms contributing to improved glycemia after RYGB and VSG. A: Immediate effects of RYGB and VSG due toanatomical changes. B: Potential mediators/mechanisms involved. Cross talk occurs among these factors. C: Effects on glucose homeostasis.

care.diabetesjournals.org Batterham and Cummings 897

secretion of distal gut peptides. Proof ofconcept for this theory comes from ilealinterposition studies in rodents, where asection of distal ileum is interposed to theduodenum-jejunum boundary while pre-serving the vessels and nerves supplyingthe ileal segment. Ileal interposition en-hances L-cell nutrient and BA exposure,without gastric restriction or malabsorp-tion, thereby increasing circulating GLP-1(and PYY) levels, together with im-proved glucose tolerance. The overalleffect of ileal interposition on glucosemetabolism and body weight is typicallymodest, however, suggesting this effectonly partly contributes to glycemic im-provements (35). Consistent with thehindgut hypothesis, RYGB results ingreater GLP-1 and PYY release com-pared with VSG (36). The physiologicalrelevance of increased GLP-1 secretionin mediating glycemic improvementsafter RYGB and VSG is contentiouslydebated with strong protagonists andantagonists. As discussed above, VLCDand RYGB may lead to comparableshort-term improvements in hepatic in-sulin sensitivity and b-cell functionwhen assessed using intravenous glu-cose administration in the absence ofelevated GLP-1 levels. In patients withT2DM, however, full recovery of b-cellfunction early postsurgery or in patientsin clinical remission 3 years postsurgeryis observed only with oral rather thanwith intravenous glucose administra-tion, suggesting that gut-derived fac-tors are needed for full effects (37).Studies using the GLP-1 receptor antag-onist exendin(9-39) postsurgery to in-terrogate the physiological role ofexaggerated postoperative GLP-1 levelshave yielded opposing findings (partlydue to methodological differences out-lined in Table 1), and hence, opposingconclusions. Studies using exendin(9-39) in rodents undergoing bariatricsurgery suggest a role for GLP-1 in reg-ulating glycemia after VSG (38). How-ever, studies using GLP-1 receptor–nullmice and mice with functional deletionofGLP-1 suggest that neitherGLP-1 nor itsreceptor is necessary for glycemic im-provements after RYGB or VSG (39,40).Severe postprandial hyperinsulinemic

hypoglycemia emerges in a very smallminority of patients several years afterRYGB. The onset is delayed and bears anuncanny resemblance to the alimentaryhypoglycemia reported after surgery for

peptic ulcer disease. Reports of largerpostmeal GLP-1 and insulin responsesin these patients compared with asymp-tomatic RYGB patients suggest a poten-tial pathogenic role for GLP-1 (41). Thisconcept is supported by findings thatexendin(9-39) administration eliminatesthe abnormally high insulin secretionpattern observed in these individualsand prevents hypoglycemia. Symptom-atic patients after RYGB exhibit greaterexendin(9-39) responsiveness thanasymptomatic patients after RYGB (42).Studies comparing intravenous versusoral glucose administration suggest anexaggerated pancreatic b-cell insulin re-sponse after oral but not intravenousstimulation (43). The delayed onset ofpostprandial hyperinsulinemic hypogly-cemia after RYGB and peptic ulcer dis-ease surgery is puzzling and might reflectincreased b-cell mass and/or functionresulting from heightened GLP-1 action.In rodents, GLP-1 exerts antiapoptotic ef-fects on b-cells (44), and in nondiabeticpigs, RYGB led to an increase inb-cell areaand islet number 20 days postsurgery,with increased GLP-1R immunoreactivity(45). Whether GLP-1 or RYGB alters b-cellmass in humans is controversial. Someinvestigators report that b-cell area isincreased in post-RYGB hyperinsulinemic/hypoglycemic patients compared withmatched obese nonsurgical control sub-jects (46,47), whereas others contest thisassertion (37,38).

Overall, current evidence suggeststhat GLP-1 mediates some of the postsur-gery glycemic benefits to oral nutrient in-gestion. However, other gut-derivedfactors are likely to contribute. Additionalstudies are needed to determinewhetherGLP-1 protects/enhances b-cell mass inthe longer-term. Regardless of the rela-tive importance of GLP-1 in bariatric/metabolic surgery, a key point is thateven this powerful antidiabetes inter-vention cannot reverse end-stage b-cellfailure. The strongest predictors of dia-betes nonremission postsurgery are longdisease duration, insulin usage, and lowC-peptide levelsdall likely proxies forirreversible b-cell death.

ROLES OF THE FOREGUT/REDUCED ANTI-INCRETINS

BPD and RYGB cause exclusion of duo-denum and at least part of the jejunumfrom exposure to ingestion nutrients,together with the rapid delivery of

incompletely digested food to the distalbowel. The foregut hypothesis postu-lates that proximal intestinal exclusiondiminishes/eliminates a pathophysio-logical rise in an unknown anti-incretinsignal that normally serves to counter-act incretin-mediated insulin secretionand prevent postprandial hypoglycemia.An experimental procedure, DJB, wasdeveloped to examine the role of ex-cluding the duodenum and proximal je-junum (similar to RYGB) in the absenceof gastric restriction. In Goto-Kakizakirats, a nonobese T2DM model, DJB im-proved glycemic control without reduc-ing food intake or body weight (49).Similar observations have been madeusing obese diabetic Zucker rats (50).DJB redirects and enhances nutrientflow into the midjejunum. These findingsled Lam and colleagues to investi-gate whether increased jejunal nutrientexposure affected glycemic control.They demonstrated that jejunal nutri-ent sensing was required for the earlyimprovement of glycemia induced byDJB in nonobese rodents with un-controlled diabetes. Moreover, theyidentified that DJB-enhanced jejunal-nutrient sensing lowered endogenousglucose production via a gut–brain–liver neurocircuit (18).

The effect of DJB on GLP-1 is contro-versial. Increased GLP-1 plasma levelsand a deterioration of DJB-induced gly-cemic improvements were reportedwith administration of exendin(9-39)(51), although other studies report nosuch changes, implicating duodenal nu-trient exclusion per se in improvinginsulin sensitivity, independent of in-cretins or insulin (52). Differences inpostoperative duration may accountfor this discrepancy, with GLP-1 changesreported late postsurgery but not acutely.Because VSG improves glucose homeo-stasis without duodenal exclusion,some have challenged the importanceof duodenal exclusion and anti-incretins.However, the mechanisms by which dif-ferent bariatric/metabolic procedures im-prove glucose tolerance are likely to varyand cannot be used as evidence todiscount a role for duodenal exclusion inregulating glucose tolerance.

INTESTINAL ADAPTATION

RYGB and VSG lead to marked intestinaladaptations that may contribute to im-proved glucose homeostasis. However,

898 Antidiabetes Mechanisms of Metabolic Surgery Diabetes Care Volume 39, June 2016

emerging evidence suggests that majordifferences exist between these twoprocedures regarding glucose uptake(53). After RYGB, the alimentary limbundergoes hyperplasia and hypertro-phy, together with increased expressionof glucose transporters, increased up-take of glucose into intestinal epithelialcells, and reprogramming of intestinalglucose metabolism to support tissuegrowth and increased bioenergetic de-mands. The number of cells producingGLP-1 and GIP within the alimentary limbalso increases (53). Furthermore, positronemission tomography–computed scan-ning and biodistribution analysis using2-deoxy-2-[18F]fluoro-D-glucose in ro-dents and humans show that the ali-mentary limb becomes a major site forglucose disposal (53,54). These changesare likely to contribute to improvedglycemic control. In contrast, there is noevidence of GI tract hyperplasia afterVSG. However, the number and densityof cells containing GLP-1 reportedly in-crease after VSG in rodents. Moreover,VSG reduces intestinal glucose absorp-tion, potentially contributing to im-proved glucose tolerance (53). Thesefindings yet again highlight that RYGBand VSG improve glucose homeostasisby differentdas well as overlappingdmechanisms.

BA, FGF-19, AND THE GUTMICROBIOME

Circulating BA levels increase in humansand rodents after RYGB and VSG, corre-lating with improved glucose tolerance.Similarly, circulating FGF-19 levels in-crease after RYGB and VSG, althoughthe time course of these changes andthe BA composition details are activelydebated. In contrast, neither caloric re-striction nor LAGB alters circulating BAor FGF-19 levels (55). The anatomicalrearrangements after RYGB lead to de-layed mixing of BAs with ingested foodand exposure of the ileum to digestate-freechyme, offering a plausible explanationfor increased circulating BAs and FGF-19levels. This notion is supported by the find-ing that ileal interposition, with attendantincreased BA exposure, leads to increasedcirculating BA levels (56). Although VSGcauses rapid gastric emptying, ingestednutrients and BAs mix within the duode-num; thus, mechanisms other than in-creased BA exposure are involved. Inmice post-VSG, increased ileal expression

of the apical sodium bile salt transporter isreported, potentially contributing to in-creased circulating BA levels (57).

Patients with T2DM exhibit reducedcirculating BA and FGF-19 levels com-pared with normoglycemic individu-als, regardless of BMI. Furthermore,patients with post-RYGB T2DM remissionexhibit higher circulating FGF-19 and BAlevels compared with nonremitters (58).These findings imply a link betweenT2DM/insulin resistance, FGF-19, and BAbut do not prove causality. However,catheter-mediated bile diversion to themiddistal jejunum in lean and obeserodents leads to increased circulatingBAs and improved glucose homeosta-sis independently of weight loss andfood intake (59). BA diversion is associ-ated with reduced hepatic glucoseproduction and increased intestinalgluconeogenesis within gut segmentsdevoid of bile. Administration of BA se-questrants or adding BA back to gut re-gions with BAs negates the beneficialeffects of bile diversion on glucose con-trol, suggesting that BA bioavailability iscausal.

Pattou and colleagues, using a mini-pig RYGB model, recently provided fur-ther insights into potential mechanismsoperating here (60). They showed thatintestinal uptake of ingested glucose isblunted in the BA-deprived alimentarylimb, despite intact expression of thesodium-glucose cotransporter-1. BAaddition restored alimentary limb glucoseuptake, an effect abolished by phlorizin, asodium-glucose cotransporter-1 inhibitor.Given the high concentration of sodiumin bile, they examined the effect of add-ing sodium alone to the alimentary limb(which has low sodium levels post-RYGB)and observed increased glucose uptake.Their findings suggest that BAs modulateglucose homeostasis partly by alteringsodium-glucose intestinal cotransport.Reduced alimentary limb intestinal glu-cose uptake could account for earlierfindings of increased gluconeogenesiswithin the alimentary limb. However,these observations are somewhat atodds with reports from rodents and hu-mans of increased alimentary limb glu-cose utilization. Methodological issuescould underlie these differences; alter-natively, circulating glucose, ratherthan intestinal glucose, might be thesource. Another key source of intestinalsodium is gastric sodium bicarbonate.

Thus, reduced intestinal sodium afterVSG would be anticipated and offers aplausible explanation for reduced intes-tinal glucose uptake post-VSG. In additionto altering intestinal glucose uptake, en-hanced ileal BA exposure leads to increasedcirculating levels of FGF-19 and GLP-1,with reduced hepatic glucose productionand increased tissue uptake of glucosevia insulin-dependent and -independentmechanisms.

Caloric restriction, bile diversion,RYGB, and VSG lead to gut microbiomechanges, with modulation from an obesebacterial profile, with a high ratio ofFirmicutes to Bacteroidetes, to a leanerbacterial profile. Bacteroidetes play akey role in bile acid deconjugation;hence, this change could affect BA com-position. A recent longitudinal studyfound a biphasic increase in total fastingplasma BA levels post-RYGB. The earlypeak, 1 month postsurgery, was due tobacterially derived secondary BAs such asursodeoxycholic acid. A later peak 2 yearspost-RYGB reflected increased primaryBAs and the secondary BAs deoxycholicacid and glycodeoxycholic acid. Circulat-ing FGF-19 increased, but not until severalmonths postsurgery after the more rapidmetabolic improvements had occurred(61). The early changes in ursodeoxy-cholic acid and its metabolites may con-tribute to early improvements in insulinsensitivity after RYGB.

Three lines of evidence suggest a rolefor the microbiome in contributing tothe beneficial effects of bariatric sur-gery. Firstly, Kaplan and colleaguesfound that germ-free mice treated withfecal transplants from RYGB-treatedmice lost weight, whereas similar recip-ients given fecal transplants fromweight-matched sham-operated micegained weight. Alterations in micro-bially produced short-chain fatty acidswere proposed as potential mediators(62). Secondly, studies undertaken byRyan et al., using global FXR knockout(FXRKO) mice, suggest a key role for theBA–FXR–microbiome pathway in medi-ating the weight-reducing and glycemiceffects of VSG. After VSG, circulating BAchanges, weight loss, and glycemic im-provements were attenuated in globalFXRKO mice compared with wild-typemice, together with altered microbiomecomposition (63). However, whetherthese microbiome changes relate tometabolic differences or FXR deficiency

care.diabetesjournals.org Batterham and Cummings 899

per se is unclear. The use of global FXRKOmice makes interpreting these findingsdifficult due to the phenotype of thoseanimals, which includes increased circu-lating BAs, altered adaptive thermogen-esis, insulin resistance, and resistance todiet-induced obesity (64). Future stud-ies in mice with intestine-specific FXRdeletion will help clarify underlyingmechanisms. Thirdly, Tremaroli et al.showed that RYGB and VSG led tolong-term comparable alterations inthe human gut microbiome that wereindependent of BMI. Fecal transplantfrom RYGB and VSG patients reducedadiposity in recipient mice (65).Overall, these studies show that RYGB

and VSG increase circulating BA, FGF-19,and GLP-1 levels, also altering intestinalglucose utilization and the gut micro-biome. These changes favor improvedglucose tolerance and weight loss andare likely to contribute to the antidiabe-tes effects of RYGB and VSG.

FUTURE PERSPECTIVES

The impressive antidiabetes effects ofbariatric/metabolic surgery in T2DM pa-tients impel ongoing efforts to furtherclarify mechanisms mediating thesebenefits. This will not be easy, becauseit is becoming increasingly apparentthat surgery engages a constellation ofinterrelated peripheral and centralchanges that together improve glycemiccontrol. However, increasing evidencethat RYGB and VSG influence glycemiccontrol through differential mechanismshighlights the opportunity to use differ-ent GI interventional procedures as toolsto gain insights into the glucoregulatoryeffects of various regions of the GI tract.A further relevant complexity is the di-verse underlyingbiologyofpatient groupsstudied; for example,men versuswomen,insulin-treated T2DM versus newly diag-nosed, and class I obesity versus class III,etc. Indeed, the ultimate challenge andopportunity lie in tailoring the most ef-fective therapeutic options to individualpatients and identifying the optimumtime to intervene.

Acknowledgments. The authors would liketo thank Karima Yousseif (University CollegeLondon) for the illustrations.Funding. R.L.B. is supported by grants from theRosetrees Trust and Stoneygate Trust. D.E.C. issupported by National Institute of Diabetesand Digestive and Kidney Diseases grants

RO1-DK-103842, RO1-DK-084324, RO1-DK-089528, and U34-DK-107917.Duality of Interest. D.E.C. is a principal in-vestigator on the Comparison of Surgery vs.Medicine for Indian Diabetes (COSMID) trial,which is funded by Johnson & Johnson, and theAlliance of Randomized Trials of Medicinevs Metabolic Surgery in Type 2 Diabetes(ARMMS-T2D) trial, which is funded by Johnson& Johnson and Covidien, in conjunction withthe National Institutes of Health. No otherpotential conflicts of interest relevant tothis article were reported.

References1. Kahn SE, Cooper ME, Del Prato S. Pathophys-iology and treatment of type 2 diabetes: per-spectives on the past, present, and future.Lancet 2014;383:1068–10832. Rutter GA. Dorothy Hodgkin Lecture 2014.Understanding genes identified by genome-wide association studies for type 2 diabetes.Diabet Med 2014;31:1480–14873. Barnes CG. Hypoglycaemia following partialgastrectomy; report of three cases. Lancet1947;2:536–5394. Evensen OK. Alimentary hypoglycemia afterstomach operations and influence of gastricemptying on glucose tolerance curve. ActaMed Scand 1942;110:143–1535. Herbst CA, Hughes TA, Gwynne JT, BuckwalterJA. Gastric bariatric operation in insulin-treatedadults. Surgery 1984;95:209–2146. Pories WJ, MacDonald KG Jr, Flickinger EG,et al. Is type II diabetes mellitus (NIDDM) a sur-gical disease? Ann Surg 1992;215:633–642; dis-cussion 6437. Pories WJ, Swanson MS, MacDonald KG,et al. Who would have thought it? An operationproves to be the most effective therapy foradult-onset diabetes mellitus. Ann Surg 1995;222:339–350; discussion 350–3528. InabnetWB3rd,Winegar DA, Sherif B, SarrMG.Early outcomes of bariatric surgery in patients withmetabolic syndrome: an analysis of the bariatricoutcomes longitudinal database. J Am Coll Surg2012;214:550–556; discussion 556–5579. Baggio LL, Drucker DJ. Biology of incretins:GLP-1 and GIP. Gastroenterology 2007;132:2131–215710. Rubino F, Gagner M. Potential of surgeryfor curing type 2 diabetes mellitus. Ann Surg2002;236:554–55911. Salinari S, Debard C, Bertuzzi A, et al. Jejunalproteins secreted by db/db mice or insulin-resistant humans impair the insulin signalingand determine insulin resistance. PLoS One2013;8:e5625812. Alfa RW, Park S, Skelly KR, et al. Suppressionof insulin production and secretion by a decretinhormone. Cell Metab 2015;21:323–33313. Penney NC, Kinross J, Newton RC, PurkayasthaS. The role of bile acids in reducing the metaboliccomplications of obesity after bariatric surgery: asystematic review. Int J Obes 2015;39:1565–157414. Sayin SI, Wahlstrom A, Felin J, et al. Gutmicrobiota regulates bile acid metabolism byreducing the levels of tauro-beta-muricholicacid, a naturally occurring FXR antagonist. CellMetab 2013;17:225–23515. Fang S, Suh JM, Reilly SM, et al. IntestinalFXR agonism promotes adipose tissue browning

and reduces obesity and insulin resistance. NatMed 2015;21:159–16516. Scarlett JM, SchwartzMW. Gut-brainmech-anisms controlling glucose homeostasis.F1000Prime Rep 2015;7:1217. Wang PY, Caspi L, Lam CK, et al. Upper in-testinal lipids trigger a gut-brain-liver axis toregulate glucose production. Nature 2008;452:1012–101618. Breen DM, Rasmussen BA, Kokorovic A,Wang R, Cheung GW, Lam TK. Jejunal nutrientsensing is required for duodenal-jejunal bypasssurgery to rapidly lower glucose concentrationsin uncontrolled diabetes. Nat Med 2012;18:950–95519. Salinari S, Carr RD, Guidone C, et al. Nutri-ent infusion bypassing duodenum-jejunum im-proves insulin sensitivity in glucose-tolerant anddiabetic obese subjects. Am J Physiol EndocrinolMetab 2013;305:E59–E6620. Kellum JM, Kuemmerle JF, O’Dorisio TM,et al. Gastrointestinal hormone responses tomeals before and after gastric bypass and verti-cal banded gastroplasty. Ann Surg 1990;211:763–770; discussion 770–77121. Rubino F, Marescaux J. Effect of duodenal-jejunal exclusion in a non-obese animal modelof type 2 diabetes: a new perspective for an olddisease. Ann Surg 2004;239:1–1122. Strader AD, Vahl TP, Jandacek RJ, Woods SC,D’Alessio DA, Seeley RJ. Weight loss through ilealtransposition is accompanied by increased ilealhormone secretion and synthesis in rats. Am JPhysiol Endocrinol Metab 2005;288:E447–E45323. 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;15:656–66424. Laferrere B, Teixeira J, McGinty J, et al. Ef-fect of weight loss by gastric bypass surgeryversus hypocaloric diet on glucose and incretinlevels in patients with type 2 diabetes. J ClinEndocrinol Metab 2008;93:2479–248525. Plum L, Ahmed L, Febres G, et al. Compar-ison of glucostatic parameters after hypocaloricdiet or bariatric surgery and equivalent weightloss. Obesity (Silver Spring) 2011;19:2149–215726. Jackness C, KarmallyW, Febres G, et al. Verylow-calorie diet mimics the early beneficialeffect of Roux-en-Y gastric bypass on insulinsensitivity and b-cell function in type 2 diabeticpatients. Diabetes 2013;62:3027–303227. Lingvay I, Guth E, Islam A, Livingston E.Rapid improvement in diabetes after gastric by-pass surgery: is it the diet or surgery? DiabetesCare 2013;36:2741–274728. Taylor R. Banting Memorial lecture 2012:reversing the twin cycles of type 2 diabetes.Diabet Med 2013;30:267–27529. Lim EL, Hollingsworth KG, Aribisala BS, ChenMJ, Mathers JC, Taylor R. Reversal of type 2 di-abetes: normalisation of beta cell function in as-sociation with decreased pancreas and livertriacylglycerol. Diabetologia 2011;54:2506–251430. Steven S, Hollingsworth KG, Small PK, et al.Weight loss decreases excess pancreatic triacyl-glycerol specifically in type 2 diabetes. DiabetesCare 2016;39:158–16531. Sumithran P, Prendergast LA, Delbridge E,et al. Long-term persistence of hormonal

900 Antidiabetes Mechanisms of Metabolic Surgery Diabetes Care Volume 39, June 2016

adaptations to weight loss. N Engl J Med 2011;365:1597–160432. Manning S, Pucci A, BatterhamRL. Roux-en-Ygastric bypass: effects on feeding behavior andunderlying mechanisms. J Clin Invest 2015;125:939–94833. Martinussen C, Bojsen-Møller KN, DirksenC, et al. Immediate enhancement of first-phaseinsulin secretion and unchanged glucose effec-tiveness in patients with type 2 diabetes afterRoux-en-Y gastric bypass. Am J Physiol Endocri-nol Metab 2015;308:E535–E54434. Manning S, Pucci A, Batterham RL. GLP-1: amediator of the beneficial metabolic effects ofbariatric surgery? Physiology (Bethesda) 2015;30:50–6235. Ramzy AR, Nausheen S, Chelikani PK. Ilealtransposition surgery produces ileal length-dependent changes in food intake, body weight,gut hormones and glucose metabolism in rats.Int J Obes 2014;38:379–38736. Yousseif A, Emmanuel J, Karra E, et al. Differ-ential effects of laparoscopic sleeve gastrectomyand laparoscopic gastric bypass on appetite,circulating acyl-ghrelin, peptide YY3-36 and activeGLP-1 levels in non-diabetic humans. Obes Surg2014;24:241–25237. Dutia R, Brakoniecki K, Bunker P, et al. Lim-ited recovery of b-cell function after gastricbypass despite clinical diabetes remission. Dia-betes 2014;63:1214–122338. Chambers AP, Jessen L, Ryan KK, et al.Weight-independent changes in blood glucosehomeostasis after gastric bypass or verticalsleeve gastrectomy in rats. Gastroenterology2011;141:950–95839. Chambers AP, Smith EP, Begg DP, et al. Reg-ulation of gastric emptying rate and its role innutrient-induced GLP-1 secretion in rats aftervertical sleeve gastrectomy. Am J Physiol Endo-crinol Metab 2014;306:E424–E43240. Mokadem M, Zechner JF, Margolskee RF,Drucker DJ, Aguirre V. Effects of Roux-en-Y gas-tric bypass on energy and glucose homeostasisare preserved in two mouse models of func-tional glucagon-like peptide-1 deficiency. MolMetab 2013;3:191–20141. 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 ClinEndocrinol Metab 2007;92:4678–468542. Salehi M, Gastaldelli A, D’Alessio DA. Block-ade of glucagon-like peptide 1 receptor corrects

postprandial hypoglycemia after gastric bypass.Gastroenterology 2014;146:669–680.e243. Patti ME, Li P, Goldfine AB. Insulin responseto oral stimuli and glucose effectiveness in-creased in neuroglycopenia following gas-tric bypass. Obesity (Silver Spring) 2015;23:798–80744. Chambers AP, Wilson-Perez HE, McGrath S,et al. Effect of vertical sleeve gastrectomy onfood selection and satiation in rats. Am J PhysiolEndocrinol Metab 2012;303:E1076–E108445. Lindqvist A, Spegel P, Ekelund M, et al.Gastric bypass improves b-cell function andincreases b-cell mass in a porcine model. Dia-betes 2014;63:1665–167146. Service GJ, Thompson GB, Service FJ,Andrews JC, Collazo-Clavell ML, Lloyd RV. Hy-perinsulinemic hypoglycemia with nesidioblas-tosis after gastric-bypass surgery. N Engl J Med2005;353:249–25447. Patti ME, McMahon G, Mun EC, et al. Severehypoglycaemia post-gastric bypass requiring par-tial pancreatectomy: evidence for inappropriateinsulin secretion and pancreatic islet hyperplasia.Diabetologia 2005;48:2236–224048. Meier JJ, Butler AE, Galasso R, Butler PC.Hyperinsulinemic hypoglycemia after gastricbypass surgery is not accompanied by islethyperplasia or increased beta-cell turnover.Diabetes Care 2006;29:1554–155949. Rubino F, Forgione A, Cummings DE, et al.The mechanism of diabetes control after gastro-intestinal bypass surgery reveals a role of theproximal small intestine in the pathophysiologyof type 2 diabetes. Ann Surg 2006;244:741–74950. Rubino F, Zizzari P, Tomasetto C, et al. Therole of the small bowel in the regulation of cir-culating ghrelin levels and food intake in theobese Zucker rat. Endocrinology 2005;146:1745–175151. Kindel TL, Yoder SM, Seeley RJ, D’AlessioDA, Tso P. Duodenal-jejunal exclusion improvesglucose tolerance in the diabetic, Goto-Kakizakirat by a GLP-1 receptor-mediated mechanism.J Gastrointest Surg 2009;13:1762–177252. Salinari S, le Roux CW, Bertuzzi A, Rubino F,Mingrone G. Duodenal-jejunal bypass andjejunectomy improve insulin sensitivity inGoto-Kakizaki diabetic rats without changes inincretins or insulin secretion. Diabetes 2014;63:1069–107853. Cavin JB, Couvelard A, Lebtahi R, et al. Dif-ferences in alimentary glucose absorption and in-testinal disposal of blood glucose after Roux-en-Y

gastric bypass vs sleeve gastrectomy. Gastro-enterology 2016;150:454–464.e954. Saeidi N, Meoli L, Nestoridi E, et al. Reprog-ramming of intestinal glucose metabolism andglycemic control in rats after gastric bypass. Sci-ence 2013;341:406–41055. Sachdev S,Wang Q, Billington C, et al. FGF 19and bile acids increase following Roux-en-Y gastricbypass but not after medical management in pa-tients with type 2 diabetes. Obes Surg 201556. Kohli R, KirbyM, Setchell KD, et al. Intestinaladaptation after ileal interposition surgery in-creases bile acid recycling and protects againstobesity-related comorbidities. Am J PhysiolGastrointest Liver Physiol 2010;299:G652–G66057. Ding L, Yang L, Wang Z, Huang W. Bile acidnuclear receptor FXR and digestive system dis-eases. Acta Pharm Sin B 2015;5:135–14458. Gerhard GS, Styer AM, Wood GC, et al. Arole for fibroblast growth factor 19 and bileacids in diabetes remission after Roux-en-Ygastric bypass. Diabetes Care 2013;36:1859–186459. Kohli R, Setchell KD, Kirby M, et al. A surgi-cal model in male obese rats uncovers protec-tive effects of bile acids post-bariatric surgery.Endocrinology 2013;154:2341–235160. Baud G, Daoudi M, Hubert T, et al. Bile di-version in Roux-en-Y gastric bypass modulatessodium-dependent glucose intestinal uptake.Cell Metab 2016;23:547–55361. Albaugh VL, Flynn CR, Cai S, Xiao Y, TamboliRA, Abumrad NN. Early increases in bile acidspost Roux-en-Y gastric bypass are driven byinsulin-sensitizing, secondary bile acids. J ClinEndocrinol Metab 2015;100:E1225–E123362. Liou AP, Paziuk M, Luevano JM Jr,Machineni S, Turnbaugh PJ, Kaplan LM. Con-served shifts in the gut microbiota due to gastricbypass reduce host weight and adiposity. SciTransl Med 2013;5:178ra4163. Ryan KK, Tremaroli V, Clemmensen C, et al.FXR is a molecular target for the effects of ver-tical sleeve gastrectomy. Nature 2014;509:183–18864. Kuipers F, Groen AK. FXR: the key to benefitsin bariatric surgery? Nat Med 2014;20:337–33865. Tremaroli V, Karlsson F, Werling M, et al.Roux-en-Y gastric bypass and vertical bandedgastroplasty induce long-term changes onthe human gut microbiome contributing tofat mass regulation. Cell Metab 2015;22:228–238

care.diabetesjournals.org Batterham and Cummings 901