New Lessons from the gut: studies of the role of gut peptides in weight loss and diabetes resolution after gastric bypass and sleeve gastrectomy

. It has been known since 2005 that the secretion of several gut hormones changes radically after gastric bypass operations and, although more moderately, after sleeve gastrectomy but not after gastric banding. It has therefore been speculated that increased secretion of particularly GLP-1 and Peptide YY (PYY), which both inhibit appetite and food intake, may be involved in the weight loss effects of surgery and for improvements in glucose tolerance. Experiments involving inhibition of hormone secretion with somatostatin, blockade of their actions with antagonists, or blockade of hormone formation/activation support this notion. However, differences between results of bypass and sleeve operations indicate that distinct mechanisms may also be involved. Although the reductions in ghrelin secretion after sleeve gastrectomy would seem to provide an obvious explanation, experiments with restoration of ghrelin levels pointed towards effects on insulin secretion and glucose tolerance rather than on food intake. It seems clear that changes in GLP-1 secretion are important for insulin secretion after bypass and appear to be responsible for postbariatric hypoglycemia in glucose-tolerant individuals; however, with time the improvements in insulin sensitivity, which in turn are secondary to the weight loss, may be more important. Changes in bile acid metabolism do not seem to be of particular importance in humans.


Introduction
Bariatric surgery has been performed in a variety of versions for the last 50 years or more.Initially, the purpose of surgery was to induce weight loss by malabsorption.To obtain this, patients were treated with jejuno-ileal shunting, bypassing most of the small intestine, whereby a functional short-bowel syndrome was introduced.These operations were successful in terms of weight loss, but were later abandoned because of serious side effects.The induced malabsorption often caused excessive diarrhea, nutritional and metabolic failure, and hyperoxaluria, leading to urolithiasis.Bacterial overgrowth in the excluded small intestine was thought to lead to immune complex-mediated development of inflammatory arthritis, skin diseases and the so-called "bypass enteritis syndrome".Together these side effects were thought to be responsible for cases of hepatic dysfunction and failure (1).In the 1980ies various forms of gastric banding were introduced with some success both in terms of weight loss and improved metabolism (2).However, it was clear that banding procedures, including vertical banded gastroplasty, resulted in less weight loss and poorer metabolic improvements compared with newer surgical approaches in particular gastric bypass operations and, more recently, the so-called sleeve gastrectomy.Banding also has a higher complication rate.For these reasons, vertical banded gastroplasty is less common today.Only 5% of bariatric surgeons still perform this operation (3).Particularly impressive was the effects of gastric bypass on glucose metabolism and resolution of diabetes, as reflected in the title of the study where the almost immediate resolution was first described: "Who would have thought it" (4).Today, the most used procedures are gastric bypass surgery and sleeve gastrectomy, which have been performed in numerous patients and typically elicit weight losses of 20-35 %.Both operations elicit pronounced improvements in the metabolism of the patients, and often remission of type 2 diabetes if present.A metaanalysis of 174,772 participants published in 2021 found that bariatric surgery was associated with 59% and 30% reduction in all-cause mortality among obese adults with and without type 2 diabetes, respectively (5).It was also found that median life-expectancy was 9.3 years longer for obese adults with diabetes who received bariatric surgery as compared to routine (non-surgical) care, whereas the life expectancy gain was 5.1 years longer for obese adults without diabetes (5).
These remarkable findings naturally raise the question of the mechanism underlying the beneficial effects.It was observed already in 1979 that the jejunoileal bypass operations were associated with a greatly exaggerated hormonal response to meal ingestion (6).The hormone measured was enteroglucagon, a collective designation for products of proglucagon expression in gut endocrine L-cells, covering the peptides glicentin and oxyntomodulin (7).At the time, it was not possible to assign a particular metabolic role for the exaggerated release of these two hormones.Today it is known that their secretion is paralleled by secretion of GLP-1 (8), and it may therefore be assumed that the exaggerated Lcell secretion did contribute to the benefits of the operations.It was also speculated that the exaggerated release of the L-cell hormones was due to an abnormal exposure of the distal gut, where the L cells density is higher.The mechanism behind the weight loss of the various banding procedures appeared to be different.They were not associated with an exaggerated release of any hormone (9), and it was generally accepted that the restriction provided by the operation was responsible for eliciting satiation and early termination of meals, although closer scrutiny suggests that satiation elicited via neural pathways from the distended proximal stomach may be involved (10).However, no known gut hormones appeared to be involved, not even the gastric hormone ghrelin (11).It is tempting to suggest that the explanation for the poorer effect of banding compared to bypass and sleeve gastrectomy is the missing engagement of the appetite-regulating gut hormones.
An involvement of gut hormones in response to gastric bypass operations was suggested as early as in 2005(12) when Korner et al measured plasma concentrations of PYY, which had recently been introduced as an appetite and food intake-regulating hormone (13), and ghrelin, the appetite-increasing hormone.Ghrelin levels did not rise after surgery as they would normally do in response to weight loss, and PYY levels were clearly increased compared to controls.In the laboratory of the authors of this review, GLP-1 levels were measured in a group of patients with hypoglycemia after gastric bypass (14) and were found to be extremely elevated.These findings were quickly followed up and in 2006 both Morinigo et al (from Vidal's group in Spain) and Le Roux et al ( from Bloom's group in London ) (15,16) reported increased secretion of Peptide YY (PYY) and GLP-1 after Roux-en-Y gastric bypass (see below), followed in 2007 by Korner et al who reported (9) that bypass, but not banding, was associated with massively exaggerated GLP-1 secretion.In a subsequent report, Goldfine et al could show that the GLP-1 responses were especially elevated in patients who developed hyperinsulinemic reactive hypoglycemia after gastric bypass operations (17) (see below), suggesting that the insulinotropic effects of GLP-1 might be involved.Le Roux and coworkers (18) elegantly demonstrated, using infusions of the powerfully inhibitory hormone somatostatin, that inhibition of the hormonal responses was associated with increases in food intake after gastric bypass (but not after gastric banding), suggesting that one or more of these hormonal responses were involved in the inhibition of food intake and therefore the anorexigenic effects of this operation.
The bypass operations were originally designed to provide restriction via the creation of the very small gastric pouch (often only 30 ml -see below) and malabsorption by bypassing 1.5 to 2 meters of the upper small intestine (19).However, continued research suggested that the presence of the pouch (the volume of which may be as low as 15-30 ml), rather than causing restriction, allows a greatly accelerated nutrient entry into the small intestine (20,21); and there was little if any malabsorption of ingested nutrients (22) (although we now know that there may be some malabsorption of fat (23)).Instead, the gut hormones might provide a plausible explanation for the findings.This possibility has been a major focus in our group for the last several years, and it is the purpose of the present review is to discuss some recent studies testing the hypothesis that gut hormones are at least partly responsible for the major metabolic changes after gastric bypass and sleeve gastrectomy.J o u r n a l P r e -p r o o f 2. The operations.
• The Roux-en-Y gastric bypass (RYGB) is illustrated in fig 1.The small intestine is divided around 50 cm from the pylorus.Next the distal end is anastomosed with a small pouch constructed around cardia of the stomach.The remaining stomach is closed and left in situ.Next the proximal end of the small intestine is anastomosed to the jejunum 50-150 cm further down.In this way, a Y-shaped configuration of the proximal intestine has been created, and the nutrients pass directly from the esophagus to the alimentary limb of the "Y" and further down to the common limb after the entry of the secretory limb delivering the gastric and pancreatic juices as well as the bile to the common limb.This operation results in a sustained weight loss approximating 60-70 % of the excess body weight (the difference between the preoperative weight and the ideal weight corresponding to a BMI of 25kg/m 2 ).Other beneficial effects are(5): Reduced overall mortality and fewer incidences of cardioand cerebrovascular insults; amelioration of type 2 diabetes with HbA1c reduction of 22-38 mmol/mol vs. 11-16 mmol/mol in response to conventional treatment; amelioration of hypertension, dyslipidemia, NAFLD, PCOS, sleep apnoea and osteoarthritis; and improved quality of life (24,25).Although both operative and J o u r n a l P r e -p r o o f perioperative mortality is now low (0.1-0.3%) there are other complications, including ulcers , internal herniation (although reduced after primary closure of mesenteric defects) as well as vitamin and mineral deficiency, anaemia, dumping syndrome and postprandial hypoglycaemia (23).
• Sleeve gastretomy operations consist of a permanent vertical resection of most of the stomach with a resection line running in parallel with the minor curvature (26) (see fig 1).The remaining part of the stomach is closed around a stick (a bougie) forming a tube.The resection radically reduces the volume and thus the receptive capacity of the stomach.In principle neither the pylorus nor the cardiac region is affected.The weight loss is nearly the same as that observed after bypass and the operation is easier to perform.The metabolic effects of the two operations have been compared in a randomized controlled clinical trial (27) of 109 patients with type 2 diabetes.Diabetes remission rates were significantly higher in the gastric bypass group than in the sleeve gastrectomy group, p=0•0054) after 3 years (28) and gastric bypass was associated with a greater improvement in weight-related quality of life (between group difference 9•4, 95% CI 3•3 to 15•5), less reflux symptoms (0•54, 0•17 to -0•90), greater total body weight loss (8% difference, 25% vs 17%), and a higher probability of diabetes remission (67% vs 33%, risk ratio 2•00; 95% CI 1•27 to 3 •14).In randomized trials of subjects with normal glucose tolerance, bypass surgery resulted in greater or unchanged weight loss compared sleeve gastrectomy (29,30).
The following hormones are of particular interest regarding the two operations: glucagonlike peptide-1, oxyntomodulin, PYY, neurotensin; ghrelin; and GIP, secretin and CCK.
3.1.GLP-1.As already mentioned, the secretion of GLP-1 (glucagon-like peptide 1) is greatly increased after gastric bypass operations (31,32).The hormone is secreted from open-type (meaning that apical cytoplasmatic processes reach the intestinal lumen) endocrine L-cells which are found in the epithelium throughout the small and large intestine.It is a product of proglucagon, which is also processed to release oxyntomodulin, which may be relevant in the context of bariatric surgery because of its high concentration after surgery (see below), and the sister hormone GLP-2, which may be important for intestinal mucosal maintenance and growth (8).GLP-1 powerfully stimulates glucose-induced insulin secretion, and there is substantial evidence that it is essential for improved insulin secretion after bypass surgery (33).Thus, blockade of the GLP-1 receptor with exendin 9-39, a potent and specific antagonist, the improvement in insulin secretion in patients with type 2 diabetes after surgery is lost, and in patients with postoperative reactive hypoglycemia exendin 9-39 effectively prevents J o u r n a l P r e -p r o o f hypoglycemia (34).Exendin 9-39 is currently being developed for therapy of postbariatric hypoglycemia.GLP-1 is also a potent inhibitor of appetite and food intake and long-acting analogs of GLP-1 have now been demonstrated to cause weight losses amounting to 15-18%.( 35) 3.2.Oxyntomodulin consists of the glucagon sequence plus a C-terminal extension identical to the one found in glicentin.It acts on the GLP-1 as well as the glucagon receptor but with only 1/100 of the potency of the cognate ligands (36).However, in agreement with anorexigenic effects of both glucagon and GLP-1, oxyntomodulin inhibits appetite and food intake, and long-acting analogs may cause considerable weight loss (37).Several analogs are in clinical development for obesity therapy and, because of glucagon's effects on hepatic lipid metabolism, also for MASH (metabolic dysfunction associated steatohepatitis, previously known as NAFLD) (38).In agreement with the increased secretion of GLP-1 and PYY from the L-cells, also oxyntomodulin (and glicentin) levels are greatly elevated after gastric bypass surgery(39).
3.3.PYY.Peptide YY is recognized as a potent inhibitor of food intake(13) but its effects are apparently difficult to harness into a clinically useful form.High doses cause side effects including nausea and vomiting.There are currently no approved analogs on the market, but several are under development.Like GLP-1, PYY is secreted from L-cells, but not from the cells in the upper small intestine, while some of the L-cells in the distal intestines do not produce GLP-1.It is secreted as a peptide of 36 amino acids, PYY 1-36, which is an agonist of the socalled Y1 receptor and stimulates appetite and food intake.However, like GLP-1, it is cleaved by the enzyme DPP-4 (40) leading to the formation of PYY 3-36 which is a selective agonist of the Y2 receptor which powerfully inhibits food intake (41).PYY 1-36 is also an agonist of the Y2 receptor and this may explain why there are no significant effects of PYY 1-36 after administration to humans (42).However, PYY released from the gut may, like GLP-1, be degraded locally leading to a higher 3-36/1-36 ratio under normal circumstances.3.4.Ghrelin is a 27 amino acid peptide produced in endocrine cells mainly located in the gastric glands (43).The cells have been variably designated as A-like, X/A-like, and Gr cells.After cleavage from its precursor, ghrelin undergoes a special posttranslational octanoylation at serine in position 3, catalyzed by the enzyme GOAT (ghrelin O-acyltransferase) (11).Without this modification the peptide is inactive.It is secreted in a pattern opposite of insulin: it increases during fasting and decreases in response to meal intake (and actually inhibits insulin secretion, see below) (44).It was isolated because of its actions on the secretagogue receptor of the growth hormone-producing cells, but its most prominent effect is to stimulate appetite and increase food intake (45).Ghrelin quickly loses its acylation after release.For plasma measurements it is therefore important to decide whether the purpose is to estimate the rate of secretion, and for that one J o u r n a l P r e -p r o o f would select an assay for total ghrelin, i. e. the sum of acylated and nonacylated ghrelin, but for estimation of the impact of secretion, measurement of the active octanoylated ghrelin would be preferable.Importantly ghrelin levels (total or intact) are lower in obesity (46), precluding that increased secretion of ghrelin contributes to the obesity pathogenesis.Importantly, a circulating antagonist of the ghrelin receptor LEAP 2 (liver-expressed antimicrobial peptide 2) (47) has been described.Not too much is established about LEAP 2, but infusions have been found to inhibit food intake in humans (48).It is possible that ghrelin measurements should always be accompanied by measurements of LEAP 2 for prediction of the possible biological impact of the combined system (49).3.5.GIP (glucose-dependent insulinotropic polypeptide) is a 42 amino acid peptide secreted from the endocrine K-cells in the upper small intestine in response to nutrient intake (50).It is one of the two incretin hormones responsible for the amplification of insulin secretion that occurs upon oral as opposed to intravenous administration of nutrients (51).In experiments with antagonists of the GIP receptor it has been demonstrated that GIP is responsible for most (2/3) of the incretin effect in heaaethy subjects (52).It is of great current interest because of the unprecedented effects of tirzepatide, an agonist of both the GLP-1 and the GIP receptor, on food intake and body weight (up to 25 % weight loss) as well as type 2 diabetes, with apparent remission observed in >50 % of type 2 diabetes patients (53,54).The weight effects are surprising because GIP infusion has not been reported to affect appetite or food intake in infusions studies in humans(55).3.6.Cholecystokinin is the classical appetite-inhibiting hormone which is released after particularly protein and fat intake from the endocrine I-cells of duodenal and proximal small intestinal mucosa.Whereas its motor effects on the bile tree, promoting flow of bile, and on pancreatic enzyme secretion are well established with the help of specific CCK1 Receptor antagonists (56), its precise role in regulation of food intake is uncertain, and agonists for therapy of obesity have not yet been successfully developed.3.7.Neurotensin secretion is markedly increased after bypass operations(57), but its metabolic role is uncertain, see below.3.8.Secretin may also show increased secretion after bypass operations and is currently being evaluated for effects on food intake(58).
4. The role of gut hormones following Gastric Bypass.

Their role in weight loss
As discussed above, the role of the gastrointestinal hormones for the inhibition of appetite, food intake and body weight after gastric bypass was suggested from the changes in their plasma concentrations observed after the operation.Marked increases were observed (16, J o u r n a l P r e -p r o o f 59) particularly in the peptides from the distal small intestine, including GLP-1and PYY, but also GLP-2, oxyntomodulin, and neurotensin.The increases were demonstrated to be related to the accelerated intestinal nutrient exposure and correspondingly elevated rates of nutrient absorption (60,61).The hormones from the proximal small intestine were less affected: Increased as well as decreased GIP responses were reported, while CCK and secretin secretion were moderately elevated (59).Ghrelin concentrations were moderately lowered, but meal-induced suppression was preserved.As mentioned, in experiments involving somatostatin, which inhibits the secretion of the gastrointestinal hormones, food intake would increase in operated subjects, supporting a role for the gastrointestinal hormones (18).Somatostatin, however, has numerous effects on not only the gastrontestinal tract, but also on the brain and the nervous system (62), and these results, therefore, must be classified as supportive but not conclusive.Given the powerful effects of the hormones GLP-1 and PYY on appetite and food intake and their massive increases after bypass, it was reasonable to assign special importance to those hormones.Indeed, it had been demonstrated that the two hormones apparently potentiated the inhibitory actions of each other on food intake (63,64).However, surprisingly, the GLP-1 receptor antagonist exendin 9-39, infusion of which causes increases in food intake in obese individuals before a planned gastric bypass operation, had no effect on food intake in steady state after the operation(65).However, Svane et al were able to explain this phenomenon and to provide strong evidence for a role for two hormones (65).They hypothesized that, because of the potentiation, it was necessary to block the actions of both hormones simultaneously (as in the somatostatin experiments) to see an effect.For PYY 3-36 there is no specific antagonist for human use, but it is possible to block the essential conversion of PYY 1-36 to PYY 3-36 (discussed above) with inhibitors of the responsible enzyme, DPP-4(66).Svane et al, therefore, studied ad libitum food intake after combined blockade with exendin 9-39 and a DPP-4 inhibitor, sitagliptin (65).Through accurate measurements of the concentrations of total and intact GLP-1 as well as PYY 1-36 and 3-36 it was possible to demonstrate exactly what happened in the experiments.Looking first at plasma concentrations of intact GLP-1 and PYY 3-36, it was clear that exendin 9-39 greatly increased the responses of both hormones (Fig. 2); thus, the increases in PYY probably prevented effects on food intake of the GLP-1 receptor blockade.The reason for the increases is probably the negative feed-back effect of GLP-1 on L-cell secretion via local somatostatin production (67).The DPP-4 inhibition, on the other hand, effectively prevented the conversion of PYY 1-36 to 3-36, but at the same time greatly the increased responses of intact GLP-1, which is also a substrate for DPP-4, thus preventing an effect on food intake of the prevented conversion of PYY.A combination of the two treatments resulted in strongly reduced PYY 3-36 responses (Fig. 2) and very high intact GLP-1 responses, but this time they were prevented from having effects on food intake because of the exendin 9-39 blockade.As a result of these changes food intake remained unaffected by each of the individual interventions, whereas the combination led to a significant 20 % increase in food intake, consistent with an important role for the two hormones in the regulation of food intake.
J o u r n a l P r e -p r o o f These observations do not exclude a contribution from some of the other hormones mentioned.The concentrations of oxyntomodulin are sufficiently elevated to be relevant, and the appetite inhibitory effects of this hormone, particularly in combination with GLP-1 and PYY, have been documented several times (68,69).The elevations of CCK may also contribute although the rise is not huge, and CCK responses were -for unknown reasonsactually greater in individuals with poorer weight responses (70).Currently, a contribution of the moderately increased release of secretin is hypothetical (58), but effects on food intake have been reported (71).As mentioned, GIP response to meals do not show consistent changes (both small decreases and increases have been reported) (72) and the effects of these on food intake and body weight are unclear.Neurotensin is an interesting candidate, and its plasma levels have been demonstrated to increase dramatically after RYGB (57).
Although an inhibitory effect of this hormone has been observed in rodents (73), neurotensin infusions in humans have so far been without effect, and the implications of this exaggerated release are so far unknown.
After gastric bypass, a subset of patients (around 5-10%) never obtain a weight loss of more than 50 % of the excess BMI and are often designated as primary weight loss failures.Bojsen-Møller et al. recently investigated this phenomenon in more detail (74) hypothesizing that changes in GLP-1 and PYY secretion might be involved, as suggested in previous studies (18,70).Bojsen-Møller et al, therefore, investigated meal responses in 20 women with inadequate and 20 individually matched women with adequate weight loss with or without somatostatin analogue injections to block hormone secretion.Surprisingly, in this study postprandial, GLP-1, PYY and CCK responses were similar.Fasting ghrelin were lower,

J o u r n a l P r e -p r o o f
as expected because of the lesser weight loss( 45), but meal-induced suppression was similar.Interestingly, meal intake was completely unaffected by somatostatin in those with inadequate weight loss, whereas the expected increase (+ 23 %) was observed in the patients with successful weight loss.The inadequate weight loss, therefore, was not caused by changes in appetite regulating gut hormones, but by the ability of the hormones to inhibit food intake.On this background, a genetic variant analysis for GLP-1 and PYY and their receptors was also conducted, but no differences could be found (74).Although the genetic analysis was unsuccessful, it must be assumed that errors in the receptive mechanisms underly the inadequate weight loss, and further genetic studies are therefore clearly warranted(74).
4.2.The role of the hormones in glucose metabolism.
Numerous studies have addressed the role of the gut hormones for the improvements in glucose metabolism after gastric bypass (20,45,(75)(76)(77)(78)(79).The primary candidate for this role is GLP-1; secretion of GIP, the other incretin hormone, is not consistently influenced by bypass, and except for the 100-fold weaker hormone, oxyntomodulin, the other hormones do not appreciably stimulate insulin secretion or glucose turnover.Indeed, in studies involving GLP-1 receptor antagonism with exendin 9-39, GLP-1 has consistently been found to be responsible for a major part of the enhancing effect of the operation on postprandial insulin secretion (33,78,80).However, it is equally clear that the improvements in insulin sensitivity elicited by the operation play a major role for the improvements in glucose tolerance.The improvement in hepatic insulin sensitivity which is related to the disappearance of excess hepatic fat is observed already few days after the operation and, after weeks to months, improvements in peripheral insulin sensitivity also appear, following the overall weight loss (81).Indeed, after several months the improvement in insulin sensitivity exceeds that observed in individuals with a similar weight loss obtained by dieting (82).The reason for this is unknown.It has been discussed whether the improvement in glucose tolerance is also a consequence of GLP-1 actions.Positive findings were reported by Jorgensen et al (33) whereas Vetter et al claimed that the importance was limited (83).The massive improvement in hepatic and subsequently peripheral insulin sensitivity probably represents an important confounding factor in the interpretation of the antagonist studies; furthermore exendin 9-39 may not be an ideal tool in the exploration of glucose tolerance because of its hyperglycemic effect even in the fasting state (without gut hormone release).There are also observations indicating that the importance of the increased GLP-1 secretion for glucose tolerance dwindles with time; at least, the effect of GLP-1 blockade with exendin 9-39 diminished with time and was not significant after 2 years; it is possible that the improvements in insulin sensitivity have a predominating effect on glucose tolerance at this time (84).
On the other hand, post-bariatric hypoglycemia provides a powerful illustration of the effects of GLP-1 for both insulin release and plasma glucose.In these patients, the risk of postprandial hypoglycaemia can be completely avoided by blockade of the GLP-1 receptors (34,85).
Glucagon secretion has been found to be variably affected by bariatric surgery, but it now appears that the high levels of proglucagon derived peptides from the gut in these patients J o u r n a l P r e -p r o o f have cross-reacted in most assays for glucagon.In agreement with the high postprandial levels of glucose, GLP-1 and insulin (which all would be expected inhibit glucagon secretion) measurements with the most specific sandwich ELISAs (developed to solve this problem) show decreasing levels of glucagon(86-88).
5. The role of gut hormones following Sleeve Gastrectomy.
As mentioned, sleeve gastrectomy has effects on body weight that are similar or only slightly weaker than those of RYGB (26,76), whereas the effects on glucose metabolism are less pronounced and durable when determined in a controlled clinical trial (27).Controlled clinical trials are not easily organized regarding surgical procedures, but the Oseberg study was a randomized controlled comparison between gastric bypass and sleeve gastrectomy (27,28).With respect to the hormones, the most notable differences were significantly higher GLP-1 responses (and greater beta cell sensitivity to glucose which is likely an effect of GLP-1( 89)) whereas ghrelin levels were lower in sleeve gastrectomy.It is therefore possible that the differences between the two operations might be due to differences in GLP-1 secretion.However, we also performed another comparison between RYGB and sleeve gastrectomy (90).This was a study of 12 patients with sleeve gastrectomy and 12 with RYGB, who were carefully matched for preoperative BMI, age, sex, and postoperative weight loss (!), as well as 12 weight matched controls.Nutrient absorption and metabolism were also studied and it was confirmed that secretion followed rates of absorption of the nutrients in both groups (90).However, despite identical weight losses, GLP-1 and PYY responses were much higher after RYGB than after sleeve (where responses were higher than in controls).Major differences were also observed in ghrelin responses, where RYGB was associated with a normal concentration profile but at an overall lower level, whereas sleeve gastrectomy showed generally very low concentrations without mealinduced suppression.GIP concentrations showed an early short-lasting peak after RYGB, and elevated (compared to controls) and prolonged responses after sleeve gastrectomy.From these results it was concluded that the mechanism behind the weight losses after RYGB and SG had to be different, and the focus was turned to GIP and ghrelin.

The role of GIP
In order to get a better understanding if the relative importance of GLP-1 and GIP on postprandial glucose tolerance after the two operations, we studied, in both unoperated controls, SG-operated and RYGB-operated people, the effects of blocking separately the hormone receptors with specific receptor antagonist, exendin 9-39 as already mentioned above, and GIP 3-30NH2, which is a potent and specific antagonist of the human GIP receptor (91).GIP 3-30NH2 has previously been used for determination of the importance of GIP for the incretin effect (52).The pattern of hormone secretion in the patients was like that already described: GLP-1 highest in RYGB and higher in SG than in controls.GIP concentrations were lowest in RYGB.The study showed that GLP-1 blockade reduced betacell sensitivity to glucose and increased postprandial glucose responses in both patient J o u r n a l P r e -p r o o f groups, but the blockade had little effect in controls.GIPR blockade, on the other hand, reduced beta cell sensitivity and increased glucose concentrations in the control and SG groups, but had no effects in RYGB.The results support that GIP is the most important incretin in controls, while GIP and GLP-1 are equally important in SG, and GLP-1 is the most important incretin hormone after RYGB(92).

The role of ghrelin.
Because of the low concentrations of ghrelin after SG, presumably due to the removal of a large part of the stomach, we hypothesized that ghrelin deficiency would contribute to the lower appetite and food intake after SG.One way of testing this was to try to restore preoperative concentrations of ghrelin in operated patients (93).We included 12 participants scheduled for SG and subjected them, both before and three months after surgery, to a mixed-meal test for characterization followed by an ad libitum meal test with concomitant primed infusions of acyl-ghrelin (1 pmol/kg/min) or placebo.Infusions began 60 minutes prior to meal intake to reach a steady state before the mixed-meal and continued throughout the study day.We also made additional infusions of acyl-ghrelin (at 10 pmol/kg/min and 0.25 pmol/kg/min) three months after surgery.Restoration of ghrelin levels (compared to controls) was successfully obtained with the low infusion rate.However, surprisingly ad libitum meal intake (6-11% increase) did not differ significantly during ghrelin administration compared with placebo, despite the ad libitum intake was reduced by about 60 % after surgery.In contrast, postprandial glucose concentrations increased dosedependently during ghrelin infusions compared to placebo infusions, while both basal and postprandial insulin secretion rates were inhibited.Ghrelin also lowered measures of β-cell function but had no impact on insulin sensitivity.Thus, unexpectedly, we could demonstrate that the changes in ghrelin probably contribute to the improved postoperative glucose metabolism but could not demonstrate a role of the peptide for food intake in this way.
6.The role of bile salts.
In 2009 Patti et al (94) reported that plasma concentrations of bile acids were elevated in individuals operated with RYGB and hypothesized that this might be of importance for changes in energy and glucose metabolism.This hypothesis subsequently gained strong support from experimental studies in rodents (95) when Ryan et al reported that vertical sleeve gastrectomy (VSG) was associated with increases in bile acids and that genetic deletion of the bile acid receptor, FXR, assumed to transmit the metabolic signals of the bile acids (96), largely abolished the weight lowering effects of VSG (although the deletion itself lead to a marked weight loss which complicated the interpretation).We were able to confirm the elevated peripheral venous concentrations of bile acids after RYGB, but also found that the changes were not apparent until at least 3 months after the operation, at a time when the improvements in glucose metabolism had already taken place and body weight was decreasing steeply (97).It was concluded that the early changes were not related to the increases in peripheral bile acids concentrations.But could the bile acids be responsible for some of the observed changes anyway?The bile acids have been demonstrated to powerfully stimulate the secretion of GLP-1 via the TGR5 receptor localized J o u r n a l P r e -p r o o f to the L-cells (98).Indeed, oral intake of the primary bile acid, chenodeoxycholic acid ( 99), increased GLP-1 secretion in healthy humans, and single-dose administration of sevelamer, a bile acid sequestrant binding the bile acids in the lumen of the gastrointestinal tract, eliminated bile acid-mediated GLP-1 secretion in patients with type 2 diabetes; this could be explained by reduced bile acid stimulation of the basolaterally localized TGR5 on enteroendocrine L cells (100).In further studies, the effect of 1250 mg of chenodeoxycholic acid was studied in 11 subjects 32 months after RYGB, which resulted in a weight loss of 37 kg.Chenodeoxycholic acid administration resulted in significant increases in plasma concentrations of GLP-1, C-peptide, glucagon, PYY, neurotensin, total bile acids, and fibroblast growth factor 19 (a hormone which transmits the feed-back effects of bile acid signaling in the gut to the liver).Ursodeoxycholic acid (a secondary bile acid) was without any effect (101).However, in subsequent crossover studies (102), eight RYGB operated participants first ingested CDCA 1.25 g or CDCA 1.25 g + the bile acid sequestrant colesevelam 3.75 g on separate days, and secondly twelve RYGB participants ingested on separate days a mixed meal with addition of 1) CDCA 1.25 g, 2) COL 3.75 g or 3) COL 3.75 g × 2, or 4) no additions.The first part of the study confirmed that oral intake of CDCA increased circulating bile acids, GLP-1, C-peptide, glucagon, and neurotensin.Importantly, addition of colesevelam nearly abolished all responses.In other words, it would be possible to remove the activity of luminal bile salts with this dose of the bile acid sequestrant.In the second part of the study, addition of CDCA to the meal increased the meal-induced plasma responses of GLP-1, glucagon and FGF-19 and lowered plasma glucose and C-peptide concentrations; however, adding colesevelam to the meal lowered circulating bile acids (showing that it did limit exposure and absorption of endogenous bile acids) but had no influence on meal-induced changes in plasma glucose or any of the gastrointestinal hormones.These findings suggest a limited role for endogenous bile acids in the acute regulation of postprandial gut hormone secretion and glucose metabolism after RYGB.
In a detailed study of the possible origin of the elevated plasma concentrations of bile acids after gastric bypass and sleeve gastrectomy, 15 RYGB, 10 sleeve gastrectomy and 15 control subjects underwent 99 Tc-mebrofenin cholescintigraphy combined with intake of a highfat 111 In-DTPA-labelled meal and frequent blood sampling.A 75 Se-HCAT test was used to assess bile acid retention.After RYGB, gallbladder filling decreased while basal flow of bile into the small intestine increased.Bile acid retention was increased, and basal bile acid plasma concentrations elevated.During the meal, foods passed unimpeded through the gastric pouch resulting in almost instant postprandial mixing of bile and foods, but the postprandial rise in plasma bile acids was brief.After SG, bile flow and retention were largely unaltered, but gastric emptying was variably accelerated (p < 0.001) causing earlier mixture of bile and foods also in this group, but neither basal nor postprandial bile acid concentrations differed between individual with SG and controls (103).Taken together the experiments suggest that the importance of the role of altered bile acid secretion and metabolism after these bariatric interventions may not be as big as hitherto assumed.

J o u r n a l P r e -p r o o f
So how do we answer the big question of the mechanism behind weight loss after Roux-en-Y gastric bypass?The immediate answer would be that gut hormones indeed are importantly involved and that the strongest candidates are GLP-1 and PYY.We have the necessary requirements: that an increased secretion of these hormones can be demonstrated, and that administration of these hormones lead to inhibition of food intake -in the case of GLP-1, the effect is so obvious that obesity is now being treated with long-acting GLP-1 analogs, the effect of which is approaching that of bariatric surgery.A critical experiment is to estimate what would happen if these hormones were eliminated.Rodent studies of knockout animals have not been helpful, but have limited value, since their reactions to surgery differs importantly from that of humans: in the experimental animals, energy expenditure increases after the operation and the operation effect does not last, whereas in humans energy expenditure falls while the effects on body weight and food intake lasts for decades.But we have the pharmacological elimination experiments with the broadly inhibiting hormone somatostatin and the combination of specific eliminations of the effects of GLP-1 and PYY.Thus, obesity operations seem to be a model for understanding the extent of the regulation of appetite, food intake and body weight by these hormones.The role of ghrelin is currently unclear and clearly deserves further investigation, and the importance of changes in bile acid turnover is also unclear but currently seems limited.Some malabsorption of particularly lipids may also play a role.
Regarding glucose metabolism, the picture is less clear.The most important factor may very well be the improvements in insulin sensitivity, first in the liver, later in the peripheral tissues caused by the initial negative energy balance and subsequently the weight loss.It seems that early on and during the first year, the grossly exaggerated secretion of GLP-1 is important for the changes in insulin secretion, but there are indications that the effect wanes with time, although still being of importance), perhaps as a reflection of beta cell adaptation to the chronically improved insulin sensitivity.However, the complete prevention of the very frequent complication of bariatric hypoglycemia with a GLP-1 receptor antagonist strongly supports the important involvement of exaggerated GLP-1 secretion for the changes in glucose tolerance.In addition, the increase in insulin sensitivity appears to be secondary to the weight loss, which in turn may result from changes in gut hormone secretion.
For sleeve gastrectomy, the mechanisms are even less clear.The accelerated nutrient passage into the small intestine after the operation probably explains the increased secretion of GLP-1 and PYY as after RYGB, but the discrepancy between the release pattern and the weight outcomes compared to RYGB, suggest that something is missing, although the greater variation in the gastric emptying rate could play a role.A changed secretion of GIP may contribute to the improvement in glucose tolerance but could hardly explain the weight loss effects.Changes in ghrelin secretion could explain part of the beneficial effect on insulin secretion and glucose tolerance but did not seem to be of importance for the weight changes in a study where ghrelin levels were restored by infusions.It may be that the hormonal changes observed after RYGB are greater than those required for the weight loss effects, so that the smaller responses after SG are sufficient to explain the similar weight J o u r n a l P r e -p r o o f loss, but this is speculation.Perhaps an additional factor, engaged by sleeve gastrectomy, is still missing.

Fig. 1
Fig. 1 Roux-en-Y gastric bypass and gastric sleeve operations.See text for description.

Fig. 2 .
Fig. 2. Changes in the plasma concentrations of intact GLP-1 and PYY 3-36 in response to a fixed meal (at time = 0) and an ad libitum meal (at time 240 min) in individuals with previous gastric bypass operation.PYY conversion was inhibited with a DPP-4 inhibitor (sitagliptin) and GLP-1 action was prevented with exendin 9-39.Adapted from Svane et al(65).