Introduction
Understanding of the role of the intestine in glucose homeostasis is progressing with the increasing adoption of bariatric surgery in the management of morbid obesity. Bariatric surgery, originally intended to treat obesity, improves glycaemic control and/or reverses diabetes [
1]. Roux-en-Y gastric bypass (RYGB), considered the standard technique, involves creating a small gastric pouch using a surgical stapler. This pouch is divided from the gastric remnant and anastomosed to the jejunum, allowing ingested food to bypass 95% of the stomach, the duodenum and the first portions of jejunum [
1]. Alternatives to RYGB include sleeve gastrectomy (SG), in which a large portion of the stomach is removed along the greater curvature [
2].
In a systematic review, Meijer and others report a type 2 diabetes mellitus reversal rate of 83% in patients undergoing the RYGB procedure [
3]. Substantial weight loss occurs during the postoperative weeks and months, but the main metabolic changes occur rapidly and independently of weight loss [
4]. Both insulin secretion and the improvement of insulin sensitivity are involved in the process. The increase in insulin sensitivity is a major early postoperative outcome of RYGB, whereas insulin secretion becomes affected at a later stage, possibly as an adaptation to the weight loss [
5]. As with any intervention to treat diabetes, the residual beta cell function is a strong predictor of remission after bariatric surgery [
6].
The intestine is a large organ and is responsible for glucose absorption. Among the mechanisms involved in the reversal of diabetes after surgery, the role of the entero–insular axis remains controversial [
4,
7]. Two mechanisms are considered mainly responsible for the early improvement in glucose control: an increase in hepatic insulin sensitivity following energy restriction and an exaggerated secretion of glucagon-like peptide 1 (GLP-1) due to a more rapid exposure of nutrients to the distal jejunum [
8].
Changes in the metabolic activity of the intestine after bariatric surgery remain unclear. Quantitative studies in humans are lacking, presumably because anatomical considerations limit non-invasive access to the organ. Positron emission tomography (PET) is a unique method of measuring metabolic rates in a specific tissue; the use of fluorine-18-labelled fluoro-deoxyglucose ([
18F]FDG) in quantitative studies has been validated in several tissues and organs, such as the liver [
9], skeletal muscle [
10] and brain [
11]. [
18F]FDG is transported into cells, where it undergoes phosphorylation, but it cannot be further metabolised, enabling its accumulation rate in specific regions to be measured. We have recently validated the method to quantify intestinal glucose uptake (GU) against autoradiography and tissue samples, and demonstrated that within the intestine, most of the FDG accumulates in the mucosal layer [
12]. Using this method, we observed that the intestinal mucosa is insulin resistant in human obesity. It is not known whether insulin resistance of the intestinal mucosa is reversed by bariatric surgery. The aim of this study was to evaluate the effects of insulin on intestinal mucosal GU before and after bariatric surgery in individuals with and without diabetes.
Discussion
To our knowledge, these are the first results demonstrating quantitative changes in the glucose metabolism of the intestine in response to bariatric surgery in humans. Hyperinsulinaemia does not enhance GU from the circulation in morbidly obese insulin-resistant individuals, in contrast to its effects in healthy individuals. This indicates insulin resistance of intestinal enterocytes mucosa in the morbidly obese [
12]. When obese individuals were re-studied 6 months after the operation, a small but significant improvement in jejunal, but not duodenal, glucose fluxes were observed.
Whether the difference in results is related to the type of operation performed and consequent anatomical and metabolic alterations or to the actual divergence in glucose metabolism between the intestinal segments is not immediately evident. Our data failed to show a significant divergence in intestinal GU between the two types of surgery, although jejunal GU tended to be higher in the patients who underwent SG.
There are differences in the physiology of the duodenum and jejunum. GLUT2 has been shown to be present in the apical membrane of jejunal enterocytes of diabetic individuals [
19], and this has not been observed in duodenal biopsies [
20]. Intestinal glucose absorption is largely regulated by changes in the location of GLUT2 in epithelial cells [
21]. GLUT2 is rapidly translocated to the brush border membrane in response to dietary glucose. GLUT2 internalisation stimulated by insulin is the limiting factor for GU from the intestine [
21]. Insulin-resistant humans have GLUT2 in a permanently apical location, which favours blood-to-lumen glucose flux during fasting hyperglycaemia and remarkably high GU from the lumen after a sugar-rich meal [
19]. Our results support this hypothesis. The obese individuals would present with a permanently apical GLUT2 leading to a pathological blood-to-lumen glucose flux and decreased accumulation of glucose in the enterocytes. Based on this, bariatric surgery would lead to a healthier glucose metabolism in the jejunum by enabling the insulin-inducible internalisation of GLUT2 which would thus explain the higher postoperative jejunal GU. Duodenal GU responds to insulin in healthy individuals. As GLUT2 is absent in the apical membrane of duodenal enterocytes, their insulin-responsive glucose use is mediated by other pathways. Intestinal gluconeogenesis is most likely one of the major factors, as demonstrated by recent studies in mice and obese humans. This diet-induced intestinal gluconeogenesis has been proposed to mediate food intake and glucose homeostasis (e.g. endogenous glucose production via portal glucose sensing) [
22,
23].
The main mechanisms suggested to account for early diabetes resolution after bariatric surgery include increased hepatic insulin sensitivity due to energy restriction and improved beta cell function. This is mainly a consequence of a more substantial secretion of GLP-1 from the distal small intestine [
7]. Whether the increase in jejunal GU seen in our data directly promotes incretin secretion requires further study. In our results, patients were stratified according the change (or lack of change) in diagnostic category, from diabetes or IGT/IFG to normal glucose tolerance, but no significant difference in the respective changes in intestinal GU was found. The sample size of these subgroups was insufficient to detect differences based on categories. However, significant correlations were seen between intestinal and femoral muscle GU after surgery, which may indicate that intestinal insulin sensitivity helps promote the resolution of diabetes in an indirect fashion (i.e. via the improvement in whole-body GU). In our data, there was no correlation between intestinal and muscle GU in the healthy individuals, most probably because of the relatively low number of participants and the small variation between them.
PET is a highly sensitive method of quantitatively measuring tissue-specific metabolic rates. We have recently validated its use for the intestinal tract [
12]. Some limitations still remain. The duodenum is fairly fixed in its location, but more distal intestinal segments shift with changes in abdominal pressure. This was addressed by selecting easily identifiable vertical segments of the intestine and confirming the location on PET and MRI while drawing VOI on the jejunum. Second, the transaxial resolution of PET in conjunction with the thinness of the intestinal wall might affect the results via two phenomena, spillover and partial volume. However, these effects proved modest in our previous validation study [
12]. In the present study, the number of participants was relatively small because of the demanding protocol and limited scanning capacity. In addition, MR images from a separate day were used as an anatomical reference, causing some variability in anatomy between the imaging modalities. The obese individuals had a mean weight of 121 kg. As the most morbidly obese patients eligible for bariatric surgery were excluded, our results may have underestimated the metabolic effects of obesity. The participants were also already well treated with regard to diabetes, which explains the small difference in glucose levels between the obese and the healthy groups (Table
1). However, participants were found to have variability in regional and systemic insulin sensitivity.
In conclusion, this study shows that intestinal insulin resistance is ameliorated after bariatric surgery. In our data, obese individuals had intestinal insulin resistance regardless of their glycaemic status, suggesting that intestinal insulin resistance may occur before, and independently of, systemic insulin resistance. The results highlight that insulin is a potent stimulator of GU in the intestine, and that bariatric surgery induces significant metabolic changes in the gut. It is likely that persistent changes in intestinal glucose metabolism may have an influence on both local processes in the gut and systemic glucose homeostasis.
Acknowledgements
The study was conducted within the Finnish Centre of Excellence in Molecular Imaging in Cardiovascular and Metabolic Research, supported by the Academy of Finland, University of Turku, Turku University Hospital and Åbo Akademi University. We wish to thank R. M. Badeau, University of Turku, for the linguistic editing of this manuscript.
Some of the data have been previously presented at the 47th EASD Annual Meeting in Lisbon, Portugal in September 2011.