|Home | About | Journals | Submit | Contact Us | Français|
To investigate the contributions of two surgical gut manipulations – exclusion of the proximal intestine from alimentary flow and exposure of the jejunum to partially digested nutrients – to body weight regulation and metabolism, we have developed a rat model of an investigational device, the endoluminal sleeve (ELS). The ELS is a 10 cm, nutrient-impermeable, flexible tube designed for endoluminal implantation. ELS devices were surgically implanted in the duodenal bulb of rats with diet-induced obesity. Body weight, food intake, stool caloric content, and glucose homeostasis were subsequently evaluated. ELS-implanted rats demonstrated a 20% reduction of body weight compared to sham-operated (SO) controls. ELS-treated animals consumed an average of 27% fewer kcal/day than SO, and there was no evidence of malabsorption. ELS treatment improved fasting glycemia and glucose tolerance after oral and intraperitoneal administration. ELS treatment improved insulin sensitivity, as demonstrated by decreased fasting and glucose-stimulated insulin levels and confirmed by calculation of homeostasis model assessment of insulin resistance. These data suggest that selective bypass of the proximal intestine with delivery of partially digested nutrients to the jejunum by ELS mimics many of the effects of RYGB on body weight and glucose metabolism. Thus, ELS implantation may be an effective treatment for obesity and diabetes. Since the ELS device is amenable to endoscopic placement, it may offer a valuable alternative to more invasive surgical approaches in selected patients with obesity and its metabolic complications.
The prevalence of obesity is increasing worldwide with profound health and economic consequences (1,2). Obese patients have substantially increased morbidity and mortality from obesity-related complications, including type 2 diabetes, cardiovascular disease, and several types of cancer (3–5). While there are many therapeutic options for obesity, including pharmacologic, dietary, and behavioral approaches, the weight loss resulting from these interventions is seldom durable. In contrast, surgical therapies commonly result in substantial and sustained weight loss, as well as improvement, and even resolution, of obesity-related comorbidities. Among the available gastrointestinal weight loss surgical (GIWLS) procedures, Roux-en-Y gastric bypass (RYGB) is currently the operation of choice. Surgical reconstruction of the GI tract during RYGB likely alters mucosal exposure to ingested macronutrients so as to differentially modulate efferent gut signals that govern energy regulation and metabolism. The result is weight loss and improved metabolic control. The precise mechanisms of these effects are unknown, although evidence is accumulating that RYGB alters neuroendocrine function in the GI tract at multiple levels (6–16).
In recent years, several animal models, including ileal transposition, jejuno-ileal bypass (JIB), RYGB, and others, have been designed to investigate the physiologic effects of GI tract reconstruction during bariatric procedures (9, 10, 14, 17–23). None of these models includes the entire complement of GI manipulations present in the RYGB, and some incorporate changes not present in the standard gastric bypass. Each surgical model is effective in reducing body weight and improving metabolism, although specific effects vary from model to model. These studies suggest that different components of RYGB contribute to its effects on energy balance and metabolism. To define the mechanisms by which RYGB regulates these processes, it is useful to consider RYGB as comprising five GI manipulations or “components,” some combination of which results in the physiologic effects of the procedure: (1) isolation of the gastric cardia, (2) exclusion of the distal stomach from the alimentary flow, (3) exclusion of the proximal intestine from the alimentary flow, (4) exposure of the jejunum to partially digested nutrients, and (5) partial vagotomy (Figure 1A). Identification of the component or combination of components responsible for the effects of RYGB on weight loss and metabolic control would help define the role of the GI tract in these functions and facilitate the development of more targeted and less invasive therapies.
To explore the relative contributions of selective components of RYGB to body weight and metabolic control, we developed an indwelling, endoluminal device that mimics two of them in isolation. The endoluminal sleeve (ELS) is a nutrient-impermeable, flexible tube designed to be anchored in the duodenal bulb and to extend into the proximal jejunum (Figure 1B). After implantation, ingested nutrients pass from the pylorus directly into and through the lumen of the ELS, isolating the mucosa of the proximal intestine from the nutrient-rich chyme (exclusion of the proximal intestine). As the flow of nutrients exits from the distal end of the ELS, the mucosa of the proximal jejunum is exposed to nutrients that have not been modified by either pancreaticobiliary secretions or the digestive and absorptive functions of the proximal gut (exposure of the jejunum to partially digested nutrients). Experience with a similar device in large animals (swine) and humans have demonstrated efficacy of this approach for weight loss and improved glucose metabolism (24–28). To explore the efficacy of ELS implantation on body weight and metabolic function, we developed a rat model of the ELS device and a procedure for its safe and stable implantation. In this paper, we report that ELS implantation induces substantial weight loss and improves metabolic function in rats with diet-induced obesity, suggesting an important role of exclusion of the proximal small intestine and enhanced jejunal delivery of partially digested nutrients in the mechanism of action of RYGB.
All experiments were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care. Sprague-Dawley (SD; Harlan, Indianapolis, IN) and Osborne-Mendel (OM; personal colony) were individually housed in a barrier animal facility. Animals were either fed regular rodent chow or diet with 35% calories from fat (high fat diet, [HFD]; Pharmaserv, Framingham, MA), where indicated in Study Design and text.
The ELS is a flexible, polytetrafluoroethylene tube with an expandable metal anchoring crown at the proximal end (GI Dynamics, Inc., Watertown, MA). The tube is 10 cm long with radioopaque markers at 5 and 10 cm. After overnight fasting, rats were maintained on inhaled anesthesia throughout surgical implantation of the device. After releasing the intestine from the ligament of Trietz, two enterotomies were introduced, one just distal to the pylorus and the other 10 cm from the pylorus. Next, an introduction catheter was threaded into the GI tract in a retrograde manner through the distal enterotomy to exit from the proximal duodenotomy. The flexible end of the ELS device was then sutured to the tip of the introduction catheter, and the ELS was pulled into the intestine by withdrawal of the introduction catheter. The duodenotomy was then repaired, and the crown of the ELS anchored to a surgical pledget situated within the subcutaneous compartment between skin and abdominal musculature. Finally, the distal enterotomy and laparotomy were repaired. The sham operation (SO) consisted of laparotomy with repair; release of the proximal intestine from the ligament of Trietz; duodenotomy with repair; and enterotomy with repair. Anesthesia time was standardized to 1.5 hours for both surgical groups. After surgery, animals were provided liquid diet advanced as tolerated and solid diet by postoperative day 7. Starting during post-operative week (POW) 2, animals were evaluated by fluoroscopy on a weekly basis to ensure ELS retention. Animals were restrained with a translucent animal restrainer, in the absence of anesthesia, to obtain fluoroscopic images.
SD rats were provided ad libitum access to HFD from weaning to induce obesity, as previously described (29). Upon reaching a body weight of 675–700 g, rats were randomly allocated into two groups. ELS devices were implanted in one group (n=8) while rats in the second group underwent the SO (n=8). After postoperative recovery, animals were provided the HFD ad libitum, and body weight was assessed on a weekly basis. During the POW 8, daily food intake and stool weights were measured on 7 consecutive days. Stool caloric content was determined by bomb calorimetry, and stool fat content determined by differential calorimetry with and without lipid extraction, as previously described (30). During the POW 9 and 10, glucose metabolism was evaluated, as described below. Animals were sacrificed during POW 16, and necropsy was performed to confirm maintained positioning and patency of the ELS. During the study period, one of eight implanted devices (12%) dislodged from the duodenal wall, migrated distally as detected by fluoroscopy and was excreted into the stool; this animal was removed from the data analysis. In the other seven animals, the devices remained intact and stably positioned where implanted for the duration of the study.
To determine the effect of ELS implantation on acute body weight gain, ELS devices were implanted in lean, male OM rats. OM rats were used in this study because of their exaggerated and accelerated weight gain phenotype when provided ad libitum access to a HFD (31). To maintain a lean phenotype, OM rats were provided ad libitum access to regular chow from weaning. At 12 weeks of age, rats underwent either ELS implantation (n=15) or SO (n=15). In this experiment, ELS devices were implanted as previously described, except for the absence of the subcutaneous anchoring pledget. After postoperative recovery, animals were provided ad libitum access to HFD. Body weight and fluoroscopic ELS retention was assessed on a weekly basis and animals were sacrificed during POW 4 to confirm ELS positioning and patency. During the study period, the ELS device dislodged and migrated distally, as determined by fluoroscopy, in a subgroup of ELS-treated OM rats. This observed migration, apparently facilitated by the lack of a anchoring pledget in this series of experiments, occurred in 7 of 15 animals (46%). These animals were followed prospectively for the duration of the study to determine the effect of ELS migration on acute body weight gain. The ELS device remained appropriately positioned and patent in 8 of 15 ELS-treated rats (54%) for the duration of the study.
Metabolic studies were performed after an 18 hr overnight fast, unless otherwise indicated. Fasting blood glucose was measured from blood obtained by tail-stick, using a glucometer (Lifescan, Milpitas, CA). Glucose tolerance testing (GTT) was performed after administration of glucose (1 g/kg body weight) by oral gavage or intraperitoneal (IP) injection, as indicated in text and figures, and assessed using area under the curve (AUC) analysis of glucose excursion curves. Plasma insulin levels were determined by ELISA (CrystalChem, Downers Grove, IL). Homeostasis Model Assessment (HOMA) for insulin resistance (IR) and β-cell function (%B) were calculated as previously described (26). The acute insulin response (AIR) was calculated as previously described (33). Insulin tolerance testing (ITT) was performed after IP administration of insulin, 0.6 U/kg body weight (Eli Lilly, Indianapolis, IN), to ad libitum-fed rats and assessed using AUC analysis of glucose disappearance curves. Plasma total (acylated and de-acylated) ghrelin levels were determined by radioimmunoassay (Phoenix Pharmaceuticals, Burlingame, CA).
Data are presented as means ± SEM. Data were compared using students t test or ANOVA, were appropriate.
To determine the effect of the ELS on body weight, ELS devices were surgically implanted in the duodenum of obese SD rats (Figures 1C–F). A second group of animals underwent the same procedure omitting the ELS device (SO). Regardless of treatment (ELS or SO), all rodents exhibited an immediate postoperative loss of approximately 15–20% of their preoperative body weight (Figure 2A). By POW 2, ELS-treated animals weighed an average of 8% less than SO (583 ± 9.5 [ELS] vs. 616 ± 27 g [SO]; p<0.05), a difference that increased to 20% (552 ± 12 [ELS] vs. 671 ± 27 g [SO]; p<0.05) by POW 7 (Figure 2A). This difference persisted through the end of the experiment at POW 16 (p<0.05; Figure 2B).
To investigate the mechanism of the observed weight loss, we measured daily caloric intake and fecal caloric content during the POW 8. ELS treatment reduced daily caloric intake by 28% (63.7 ± 11.4 [ELS] vs. 88.8 ± 6.2 kcal/d [SO]; p<0.05; Figure 3, left). ELS-treated animals did not exhibit increased stool quantity or altered bowel function in comparison to SO (data not shown). Moreover, daily stool caloric content was comparable between ELS-treated and SO animals (1.88 ± 0.1 [ELS] vs. 1.49 ± 0.3 kcal/d [SO]; NS; Figure 3, right), and the percentage of daily ingested calories absorbed by the GI tracts of both ELS-treated and SO animals was comparable at approximately 98%. Furthermore, animals in both groups absorbed greater than 99% of ingested fat (data not shown).
Ghrelin, the most potent peripherally active orexigenic hormone identified to date, has been implicated in pre-meal hunger and meal initiation (6, 34, 35). Fasting ghrelin is unchanged or decreased after RYGB-induced weight loss, suggesting that this operation blocks or bypasses a compensatory hormonal (ghrelin) response that would tend to promote weight regain (6). Fasting ghrelin levels were unchanged after ELS-induced weight loss (Figure 4).
To determine whether the ELS can prevent acute diet-induced weight gain, ELS devices were implanted in 15 lean, OM rats. After post-surgical introduction of HFD, a subgroup of ELS-treated rats (n=8) failed to gain as much weight as the SO group, as demonstrated by a 12% lower body weight at POW 4 (p<0.05 vs. SO; Figure 5). Weekly fluoroscopy demonstrated that the ELS device was intact in all animals of this subgroup for the duration of the study. Necropsy confirmed maintenance of ELS position and patency. Interestingly, several ELS-treated rats (n=7) exhibited initial weight loss followed by regain of weight to the same extent as SO animals (Figure 5). Fluoroscopic evaluation during POW 2 revealed that the ELS had dislodged from the duodenal wall, migrated distally into the GI tract, and was eventually passed in the stool. Necropsy at the end of the study confirmed that in all animals in this subgroup, the device was no longer present. Presumably, the observed weight gain in these animals was due to ELS failure and early migration.
ELS-treatment substantially decreased fasting blood glucose (78 ± 8 [ELS] vs. 90 ± 2 mg/dL [SO]; p<0.05; Figure 6A) and insulin levels (0.57 ± 0.06 [ELS] vs. 1.11 ± 0.13 ng/ml [SO]; p<0;05; Figure 6B). Calculation of the HOMA-IR demonstrated a 55% improvement of peripheral insulin resistance by ELS-treatment (2.67 ± 0.46 [ELS] vs. 5.94 ± 0.80 [SO] HOMA-IR units; p<0.05; Figure 6C). Consistent with this observation, β-cell activity, as predicted by HOMA-%B, was decreased 36% after ELS treatment (236.8 ± 25 [ELS] vs. 367.2 ± 33 HOMA-%B units [SO]; p<0.05; Figure 6D). ELS-treated rats also exhibited improved oral glucose tolerance as evidenced by a 40% decrease in the AUC analysis of glucose excursion (317.8 ± 25 [ELS] vs. 522 ± 37 [SO] AUC units; p<0.05; Figure 7A, inset). While the AIR was unchanged after oral glucose administration in ELS-treated animals (Figure 7C), glucose-stimulated insulin levels were significantly decreased (p<0.05; Figure 7B). Notably, insulin tolerance did not differ between ad libitum-fed ELS and SO rats (Figure 7D).
To determine whether the passage of endoluminal nutrients is required for the effect of ELS on glucose homeostasis, ELS-treated and SO animals underwent IP GTT. IP administration of glucose bypasses GI absorption and, presumably, the secretion of incretins and other neuroendocrine signals induced by luminal nutrients. ELS-treated animals demonstrated dramatically improved IP glucose tolerance, with an overall 43% (368 ± 62 [ELS] vs. 643 ± 111mg/dL AUC units [SO]; p<0.05) decrease in the AUC of glucose excursion (Figure 8 and inset).
Implantation of an endoluminal sleeve device in the duodenum and proximal jejunum mimics many of the effects of RYGB. In SD rats with diet-induced obesity, ELS implantation causes modest weight loss and substantially improves glucose homeostasis. Presence of this device in the proximal intestine also prevents the development of diet-induced obesity in otherwise susceptible animals. Despite creating a barrier between ingested nutrients and the duodenal mucosa, the ELS does not cause significant malabsorption. Rather, the observed weight loss results primarily from reduced calorie intake. In addition, ELS implantation appears to improve glucose homeostasis by multiple mechanisms. Both fasting blood glucose and insulin levels decrease after this procedure, which is reflected in a 55% decrease in HOMA-IR and is consistent with significant improvement of basal peripheral insulin resistance. Since hepatic glucose output is the major determinant of the fasting glucose level, these data suggest that ELS implantation leads to decreased hepatic glucose output, likely as a result of improved insulin signaling in hepatocytes. Improved oral glucose tolerance with a concurrent decrease in glucose-stimulated insulin levels further underscores an overall improvement in peripheral insulin sensitivity. The similar effect of ELS implantation on IP glucose tolerance suggests a global improvement in glucose homeostasis that is independent of post-prandial signals generated by endoluminal nutrient passage and also consistent with overall improvement of insulin sensitivity. The improvement in glucose tolerance suggests that ELS implantation also enhances glucose disposal. We observed no effect of the device on insulin tolerance, which is predominantly a marker of insulin-stimulated glucose disposal in muscle and adipose tissue. However, insulin tolerance testing is a relatively insensitive measure of peripheral insulin sensitivity, so these data are consistent with a modest effect of the device on glucose disposal in these tissues.
Although the ELS induces only about two-thirds as much weight loss as RYGB in the same strain of rats, the effect on glucose homeostasis is similar to that observed after RYGB (9, 10,13,14 20, 22). This observation suggests that manipulations of the small bowel leading to duodenal exclusion and enhanced delivery of partially digested nutrients to the jejunum are primary mechanisms by which RYGB alters glucose metabolism. In contrast, the more limited effect of ELS on body weight suggests that other components of RYGB (e.g., gastric manipulation and/or partial vagotomy) likely make a significant contribution to the weight loss induced by this procedure. The differential ability of ELS implantation to mimic some of the effects of RYGB on body weight and glucose homeostasis underscores the complexity of GI regulation of these metabolic functions. It suggests that different components of RYGB contribute to the myriad effects of this operation and that multiple regions of the GI tract participate in the normal regulation of metabolic function. By examining selected components of this complex operation in isolation, we have attempted to dissect the contributions of these manipulations to the regulation of body weight, food intake, and glucose homeostasis. The differential effect of ELS on glucose homeostasis underscores the basic validity of this approach.
Of course, ELS is not merely a component of RYGB, and there are fundamental differences in the manner in which these two manipulations may affect GI anatomy and function. Although the ELS device prevents interaction between the nutrient stream and the mucosa of the proximal intestine, these regions of the gut are not completely bypassed as they are after RYGB. With ELS, nutrients pass through the lumen of the device and retain the opportunity to affect local mechanoreceptors and motor function. In addition, the presence of an endoluminal device in the duodenum (including the fixed anchor at the proximal end) may affect gastric emptying and gastric motility, two biological processes known to acutely regulate satiety (36, 37) more generally. We have observe by fluoroscopic assessment that gastric emptying seems to be somewhat slowed after ELS implantation. However, gastric emptying has not been formally measured; thus, a possible contribution of gastric motility effects to the observed weight loss, decreased food intake and improved glucose homeostasis cannot be excluded. Similarly, the potential contribution of increased energy expenditure to ELS-induced weight loss is currently unknown. We have observed the resting energy expenditure is increased after RYGB in rats, and pair-feeding (PF) studies suggest that this effect is a significant contributor to RYGB-induced weight loss. Given the similarities in the metabolic effects of the two procedures, we predict that ELS-treatment also affects energy expenditure and this can be experimentally addressed through PF studies and indirect calorimetry.
Within the lumen of the proximal intestine, duodenal, pancreatic, and biliary secretions further digest nutrients previously emulsified in the stomach. The macromolecular components of this chyme are then capable of interacting with chemoreceptors and mechanoreceptors in the mucosa and bowel wall to generate neurohumoral signals of nutrient intake, composition, and availability (6, 8, 11, 15, 16, 38, 39). Candidate pathways mediating this communication include: (1) peptide hormones secreted by mucosal enteroendocrine cells; (2) non-peptide signals (e.g., lipids and carbohydrates) transported across the mucosal epithelium and released into the portal and lymphatic circulation; (3) enteric neuronal circuits that communicate with other regions of the gut directly or via reflex arcs; and (4) autonomic nervous pathways that communicate with the central nervous system. End-organ targets including other regions of the GI tract, the central nervous system (e.g., hypothalamus, brainstem and reward centers), the liver, and the pancreas (8, 11 15, 38, 39) normally coordinate a biological response to these gut-derived signals through the appropriate, compensatory regulation of energy intake, storage, and utilization. Implantation of the ELS appears to alter the luminal microenvironment of the duodenum and proximal jejunum by segregating ingested nutrients both from luminal secretions and from the mucosa of the proximal intestine. Minimally digested nutrients and GI secretions are then delivered to the jejunum. As a result, nutrient binding to chemoreceptors of the proximal intestine is prevented, ingested nutrients pass beyond the region of the ELS to potentially interact with regulatory chemoreceptors of the jejunum, and GI secretions untitrated by their usual substrates are exposed to the mucosa of the proximal small bowel. These alterations of the luminal environment likely modulate efferent gut signals to end-organ targets, thereby contributing to the observed effects of the device on feeding behavior, body weight, and glucose homeostasis.
The specific gut pathways activated in response to ELS-treatment are unknown. We and others have demonstrated that bypass surgery affects gut hormone secretion, particularly hormones implicated in the regulation of feeding behavior, satiety, glucose homeostasis, and the so-called “ileal brake” mechanism (9, 10, 17, 19, 20). Neutralizing antibodies to peptide YY (PYY) reverse the inhibitory effects of jejunoileal bypass on food intake in thin rodents, suggesting a causal role for gut hormones in at least some of the effects of GIWLS (10). In addition, ghrelin, a potent peripherally active orexigenic hormone, has been implicated in a compensatory hormonal response to weight loss induced by food restriction or restrictive bariatric procedures (6). As with RYGB (6), ELS treatment does not lead to increased ghrelin levels. This finding suggests that weight loss induced by intestinal exclusion and/or accelerated jejunal delivery of ingested nutrients does not stimulate a compensatory stimulus to ghrelin secretion. To date, the roles of specific gut hormones in the physiological effects of RYGB or ELS have not been established. Because of the limited nature of ELS implantation, exploration of the relative effects of ELS and RYGB on hormonal and neuronal signaling from the gut will reveal specific contributions of the proximal small bowel to the regulation of appetite, body weight and metabolic function.
In summary, exclusion of the proximal intestine from ingested nutrients and accelerated nutrient delivery to the jejunum induced by ELS mimic many of the effects of RYGB on body weight, food intake and glucose homeostasis. The substantial effect of ELS on glucose homeostasis compared to weight loss suggests an important role for the duodenum and jejunum in the regulation of this metabolic function. The more limited effect of ELS implantation on food intake and body weight suggests that other components of RYGB, particularly manipulation of the stomach, likely provide a substantial contribution to GI regulation of body weight and ingestive behavior. The degree to which all of these regulatory effects are mediated through the pancreas, liver, or central nervous system remains unknown but is an exciting and interesting area for further study. In the long term, understanding these mechanisms will likely lead to new targets and approaches for the treatment of obesity and diabetes. More immediately, the efficacy of the ELS device in this rat model brightens the prospect that its endoscopically-placed human analog will contribute to the treatment of these important disorders.
The authors thank Andy Levine for invaluable help in the development of the rat ELS model and Philip Davis and Lynn Wartschaw for excellent technical assistance. We also thank Morris White and Daniel Podolsky for their helpful review of the manuscript.
These studies were supported by NIH grants DK057478, DK046200, DK043351 (L.M.K.); training awards DK007191 (V.A.) and HD052961 (N.S.); and a research grant from GI Dynamics (Watertown, MA).
Writing assistance: None
Research support from GI Dynamics, Inc.