Orexins Control Intestinal Glucose Transport by Distinct Neuronal, Endocrine, and Direct Epithelial Pathways Free

Male Wistar rats weighing 240–280 g (Centre Elevage Janvier, Le Genest-St-Isle, France) were caged under standard laboratory conditions with tap water and regular food provided ad libitum, on a 12-h/12-h light/dark cycle at a temperature of 21–23°C. The animals were treated in accordance with European Community guidelines concerning the care and use of laboratory animals.

Oral glucose tolerance tests (OGTTs) were performed on conscious rats after 18 h of fasting. Blood samples from fasted animals were first taken from the tail vein by 10:00 a.m. OxA or OxB (55 μg/kg) diluted in 0.9% NaCl was administered intraperitoneally 5 min before the OGTT. Controls received vehicle only. Rats in all groups were fed a 30% d-glucose solution (1 g/kg body wt), and blood samples were taken by tail bleeds at 15, 30, 60, and 120 min after glucose administration. Glucose determination in blood was run immediately using an Accu-Chek Go (Roche Diagnostics, Meylan, France). Areas under the curve were calculated according to the trapeze method and expressed in arbitrary units.

Rats were fasted 16 h with water ad libitum. Animals were killed by intraperitoneal pentobarbital overdose, and the proximal jejunum was dissected out and rinsed in cold saline solution. The mesenteric border was carefully stripped off, and the intestine was opened along the mesenteric border. Four adjacent proximal samples were mounted in Ussing chambers as described previously (11). The tissues were bathed on each side with carbogen-gassed Krebs-Ringer bicarbonate solution of the following composition: 115.4 mmol/l NaCl, 5 mmol/l KCl, 1.2 mmol/l MgCl₂, 0.6 mmol/l NaH₂PO₄, 25 mmol/l NaHCO₃, 1.2 mmol/l CaCl₂, and 10 mmol/l glucose. In the solution bathing the mucosal side of the tissue, glucose was replaced with mannitol. Mannitol was kept in the bathing solution during glucose challenge. Both solutions were gassed with 95% O₂-5% CO₂ and kept at a constant temperature of 37°C (pH 7.4).

Electrogenic ion transport was monitored continuously as short-circuit current (Isc) using an automated voltage clamp apparatus (DVC 1000; WPI, Aston, U.K.) linked through a MacLab 8 to a MacIntosh computer. Orexins were added in the serosal bath 2 min before luminal glucose challenge. Results were expressed as the difference (ΔIsc) between the peak Isc after glucose challenge (maximum measured after 3 min) and the basal Isc (measured just before the addition of glucose). Response curves to orexins were studied noncumulatively. ΔIsc response to carbachol (100 μmol/l) was used at the end of the experiment as a control.

Epithelial and nonepithelial cellular fractions of rat jejunum were obtained from everted jejunum shaken in a dispersing solution containing EDTA as described in detail previously (17,18). Nonepithelial tissues were obtained after complete removal of epithelial cells.

Total RNA was extracted from cultured cell lines (CHO/OX₁R or CHO/OX₂R) or from intestinal cells using Trizol reagent (Invitrogen, Cergy-Pontoise, France). All RNA preparations were treated with RNase free-DNase (Promega, Charbonnières, France) for 60 min at 37°C. Five micrograms of total RNA was reverse transcribed using oligo(dT) primers. Twenty-five percent of the cDNA mixture was amplified using human OX₁R sense primer (5′-CCTGTGCCTCCAGACTATGA-3′) and OX₁R antisense primer (5′-ACACTGCTGACATTCCATGA-3′); OX₂R sense primer (5′TAGTTCCTCAGCTGCCTATC-3′) and OX₂R antisense primer (5′CGTCCTCATGTGGTGGTTCT-3′); or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) sense primer (5′TGAAGGTCGGAGTCAACGGATTTGGT-3′) and GAPDH antisense primer (5′CATGTGGGCCATGAGGTCACACAC-3′). Each of the 35 cycles of amplification consisted of 94°C for 1 min, 62°C for 1 min, and 72°C for 1 min. Amplicons were separated by electrophoresis in 1% agarose gels, stained with ethidium bromide, and viewed under UV illumination.

Orexins A and B were purchased from R&D Systems (Abingdon, U.K.). Tetrodotoxin (TTX) was purchased from Alomone Labs (Jerusalem, Israel). SB334867, an OX₁R specific antagonist, was purchased from Tocris Bioscience (Bristol, U.K.). Antagonists and TTX were added in serosal bath 10 min before orexins. Ala¹¹,d-Leu¹⁵OxB, an OX₂R specific agonist, was purchased from Calbiochem (La Jolla, CA). All other chemical reagents were purchased from Sigma (St. Louis, MO).

All results are expressed as means ± SE with n = number of tissues. One-way ANOVA with Tukey-Kramer multiple comparison post test was performed using GraphPad Prism version 3.0 for Windows (GraphPad Software, San Diego, CA). The level of significance was set at P < 0.05.

To evaluate the impact of peripheral OxA and OxB on glucose homeostasis, we performed OGTTs on conscious rats after intraperitoneal administration of OxA or OxB (55 μg/kg). As depicted in Fig. 1, blood glucose concentration after glucose feeding was significantly reduced in rats receiving intraperitoneal injection of OxA. Blood glucose level at 15 min was not significantly different from that in controls, but glycemia was markedly decreased at 30 (P < 0.008) and 60 (P < 0.012) min. With OxB, blood glucose level was also decreased, but it reached statistical significance for only 15 min (Fig. 1). Areas under the curve were significantly decreased for both OxA and OxB compared with control (Fig. 1, inset). Altogether, these data indicate that peripherally administered OxA and OxB can reduce intestinal absorption of glucose in a glucose tolerance test. When glucose is given luminally, the OGTT is an index of intestinal glucose entry through the glucose transporter SGLT1 (19). Therefore, we further examined in vitro the effect of orexins on active intestinal absorption of glucose.

Luminal addition of 10 mmol/l glucose to the jejunal preparation (control) induced a rapid and marked increase in Isc (vs. basal condition before glucose challenge; Fig. 2A), which plateaued at 3 min: ΔIsc = 27.0 ± 2.3 μA/cm², n = 20. Serosal addition of OxA or OxB (10 nmol/l) 2 min before glucose challenge significantly decreased glucose-induced Isc, with ΔIsc values at a plateau of 11.6 ± 3.9 (n = 5) and 13.3 ± 2.5 μA/cm² (n = 8) for OxA and OxB, respectively (57 and 50.7% inhibition, respectively). We then studied the dose effect of OxA and OxB. As shown in Fig. 2B, maximal effects of OxA and OxB were observed at 100 nmol/l peptide concentration and represented an ∼70% inhibition of glucose-induced Isc. The half-maximal inhibitory concentration (IC₅₀) values for OxA and OxB were 0.9 and 0.4 nmol/l, respectively. To analyze the pathways involved in this inhibition, we examined the effect of the OX₁R antagonist SB334867. The response to OxA was not significantly modified by SB334867 (Fig. 3A). In sharp contrast, the OX₁R antagonist markedly shifted to the right the dose-response curve to OxB (Fig. 3B). This clearly indicates that the effect of OxB involves an OX₁ receptor, whereas the effect of OxA is mostly independent of the OX₁ receptor. In the absence of available OX₂R antagonist, we could not further explore the orexin receptor mediating the effect of OxA.

Recent morphological data showed that OX₁R immunoreactivity is present in the gut in neuronal cell bodies and mucosal epithelial cells, including enteroendocrine cells and enterocytes (5). We first examined the effect of the neuroblocker TTX (5 μmol/l) on OxB-mediated inhibition of glucose-induced Isc. As shown in Fig. 4A, TTX had no effect on OxB-induced inhibition, indicating that neuronal cells do not play a significant role in this effect. We then examined the possible relay of endocrine cells in the action of OxB. We focused our attention to endocrine I cells for three reasons: 1) CCK-producing I cells are present in the vicinity of epithelial cells in duodenum and proximal small intestine (20); 2) endocrine I cells release CCK in response to different stimuli, including orexins (6); and 3) CCK is an inhibitor of SGLT1 and subsequently an inhibitor of glucose absorption (12). To explore the possible involvement of endocrine I cells and CCK in the response to OxB, we tested the effects of a mixture of CCK1R and CCK2R antagonists in our Ussing chamber model. As shown in Fig. 4B, CCKR antagonists did not modify the effect of OxB on glucose-induced Isc. These data clearly indicate that OxB does not target enteric neurons or endocrine I cells for inhibiting glucose absorption. Therefore, our hypothesis is that OxB inhibits glucose absorption through direct interaction with OX₁R on enterocytes.

We first examined the effect of the neuronal blocker TTX on OxA action. As shown in Fig. 4C, the dose-effect curve for OxA was markedly shifted to the right in the presence of TTX. However, for high concentrations of OxA, TTX only partially blocked the effect of OxA, suggesting the existence of a TTX-insensitive component in OxA action.

The TTX-resistant effect of OxA could be due to the activation of the OX₁R identified in enterocytes (see above). We therefore ran additional experiments in which the effect of 100 nmol/l OxA in the presence of TTX was determined with or without the OX₁R antagonist SB334867. Our data clearly indicated that the OX₁R antagonist completely blocked the TTX-insensitive component in the action of OxA: control 24.3 ± 1.7 μA/cm² (n = 4), OxA 100 nmol/l alone 8.0 ± 0.8 (n = 4), OxA 100 nmol/l plus TTX 17.0 ± 4.1 (n = 3), and OxA 100 nmol/l plus TTX plus SB334867 28.3 ± 4.4 (n = 3).

The TTX-sensitive component of OxA plays a major role in its inhibitory effect on glucose absorption. The neurochemistry of neurons involved in the action of OxA remains unclear because no information is currently available regarding the expression of orexin receptors in mucosa, whereas OX₁R is reported in myenteric plexus. However, the fact that OxA was shown previously to increase CCK release (6) prompted us to investigate the role of CCK receptor antagonists on OxA action. As shown in Fig. 4D, the inhibition of glucose transport induced by OxA was reduced by a mixture of CCK1R and CCK2R antagonists. To gain further insight into the subtype of CCK receptor involved in this pathway, we set out to determine the sensitivity of a single concentration of 10 nmol/l OxA, which is close to IC₅₀, in the presence of either antagonist. As shown in Fig. 5, the inhibition of glucose transport induced by 10 nmol/l OxA was completely abolished by the CCK2R antagonist YM022 (1 nmol/l), but not by the CCK1R antagonist l-364718. These data indicate that the OxA effect on intestinal glucose absorption involves a CCK2R. Because this major component of OxA action is not mediated by the OX₁R (see Fig. 3A), it is likely that it involves the other orexin receptor, i.e., OX₂R. This is in agreement with our finding that a specific OX₂R agonist triggers an inhibition of glucose absorption (Fig. 6). In line with the effect of TTX on OxA-induced inhibition, the effect of the OX₂R agonist was sensitive to TTX. Also in accordance with the involvement of CCK2R in this pathway, inhibition induced by OX₂R agonist was found to be sensitive to CCK2R antagonist YM022 (Fig. 6).

Finally, RT-PCR analysis of OxRs in rat jejunal mucosa showed the expression of OX₁R in epithelial cells (Fig. 7, lane 5) but not in nonepithelial fraction (Fig. 7, lane 7), in line with our functional results. In contrast, OX₂R was found essentially in nonepithelial fraction (Fig. 7, lane 7), a localization in good agreement with the expression of OX₂R in neurons. A small amplicon was also found in the epithelial fraction, possibly corresponding to the presence of OX₂R in enteroendocrine cells (Fig. 7, lane 5).

Our results demonstrate that OxA and OxB acutely inhibit the active absorption of luminal glucose mediated by SGLT1. This delineates a new function for orexins in the short-term control of enterocyte glucose absorption and provides an original insight into the role of orexins as a link between peripheral energy balance and the CNS (21,22). Because glucose is both an energy-rich molecule and a signal molecule, the initial arrival of glucose in the intestinal lumen and its absorption into enterocytes through SGLT1 is a major event in the energy homeostasis process. The nature of the peripheral neural circuitry through which homeostatic pathways may be integrated into the regulation of energy balance, appetite, locomotor control, and body weight is not yet fully understood. Many recent studies have shown that orexin neurons are sensitive to changes in nutritional status. In normal mice, orexin expression negatively correlates with changes in blood glucose (23). Fasting or insulin-induced hypoglycemia upregulates rat prepro-orexin (1) and orexin mRNA (24) and stimulates c-Fos expression in orexin neurons (25), indicating that changes in circulating glucose concentration can activate orexin neurons. OxA secretion from the endocrine pancreas is also stimulated by low glucose levels (7). Our finding that orexin A and B can induce a short-term modulation of the incoming flux of glucose gives a new insight into the involvement of orexins in glucose regulation.

We found that both orexins significantly reduced blood glucose when administered by intraperitoneal route. In previous studies, it was found that OxA, when given intravenously, had no effect on glycemia (5) or increased it (7). Whether difference in the dose delivered and/or in the route of administration (intraperitoneal vs. intravenous) is concerned remains to be determined.

Uptake of alimentary glucose by SGLT1 is a fundamental process regulated by several hormones and peptides. These peptides can lead to a decrease (i.e., leptin [11,26] or CCK [12]) or an increase (i.e., adenosine [27], epidermal growth factor [13], glucagon-like peptide 2 [14], GIP [28], or insulin [15]) in glucose transport. The activity of SGLT1 and the resulting glycemia occur in the context of a balance between the activating and inhibitory hormones, peptides, and neuropeptides that modulate the activity of the glucose transporter according to food availability. The relevance of the effect of endogenous orexins on glycemia should be considered in this physiological context. This should explain why orexin knockout mice exhibit normal blood glucose (29).

Another important finding is the distinct mode of action of orexins A and B on enterocyte function. Such marked differences in the mechanism of action of orexin peptides have not been reported previously. Either OxA or OxB was found to be active in vivo and in vitro with IC₅₀ values in the low nanomolar range and significant activities observed in the picomolar range. In this context, we were able to dissect the pathways involved in orexins actions, leading us to propose a novel scheme to illustrate how endogenous OxA and OxB inhibits SGLT1-mediated glucose absorption across enterocytes (Fig. 8). We found that OxA was sensitive to TTX (indicating the involvement of neuronal cells) and also sensitive to the CCK2R antagonist (indicating the involvement of endocrine I cells). These results may be explained by several potential mechanisms. First, neurosecretion of OxA could activate OX₂R present on as yet uncharacterized neuronal cell types, which, in turn, release an unknown neurotransmitter to activate neuroendocrine I cells to secrete CCK (Fig. 8, pathway 1). The nature of the neuronal cells involved in this pathway is unknown and requires further investigation. Second, OxA could directly activate endocrine I cells via OX₂R (Fig. 8, pathway 2). This is consistent with a previous report on neuroendocrine STC-1 cells, which express OX₂R and release CCK in response to orexin A (6). Third, OxA could also directly activate efferent CCK-releasing neuronal cells via OX₂R (Fig. 8, pathway 3). Such efferent CCK cells have been described in submucosal networks of enteric neurons (30). In our scheme, these pathways are not mutually exclusive. In all cases, CCK released by endocrine I cells is certainly a major mediator of OxA action in the inhibition of SGLT1 activity. However, the fact that the action of OxA is not completely inhibited by the CCKR antagonist cocktail (Fig. 4C) suggests the possible involvement of minor mediators in this action, the nature of which remains to be established. Here, we further demonstrate that this CCK-mediated effect of OxA involved a CCK2R subtype. This finding is in agreement with previous data showing that inhibition by leptin of glucose absorption in rat jejunum is mediated by CCK (11).

Conversely, our data indicate that OxB produced and secreted by neuronal cells in the mucosa could directly activate the OX₁R on enterocytes (Fig. 8, pathway 4). We found that OxB-induced inhibition was not sensitive to TTX, indicating the absence of a neuronal intermediate between the local release of neuropeptide and its action on epithelial cells. Because STC-1 cells have been shown to express both types of orexin receptors, it is conceivable that these cells may also be responsive to OxB. The absence of significant modifications in the effect of OxB in the presence of CCKR antagonists (Fig. 4B) indicates that this effect does not involve CCK-releasing cells but rather suggests that it is a direct effect on enterocytes. The inhibitory effect of OxB appears to be mediated by OX₁R, as indicated by the sensitivity of the OxB effect to the OX₁R antagonist SB334867 (Fig. 3B). Analysis of OxR transcripts in the epithelial and nonepithelial compartments of mucosa gave results that were consistent with this global scheme. OX₁R was found only in the epithelial fraction that corresponds to enterocytes. In contrast, OX₂R expression was found mainly in the nonepithelial fraction of the mucosa containing neuronal cells, with also a small signal in epithelial fraction that may be ascribed to the presence of endocrine cells within the jejunal epithelium. Another analysis of OXR distribution in the gut by RT-PCR showed the expression of OX₁R mRNA in submucosal and myenteric plexus of the guinea pig duodenum and rat ileum (31). Whether this discrepancy is related to a difference in species or organ distribution remains to be determined.

The nature of the peripheral neural circuitry through which signals from the homeostatic pathways may be integrated into the regulation of energy balance, appetite, locomotor control, and body weight is not yet fully understood. Orexin neurons have been recently demonstrated to sense blood glucose levels in both central and peripheral areas (10). Our finding that orexins have a peripheral effect on glucose entry and its consequences on glucose level in vivo suggests a possible link between intestinal entry of glucose and orexin glucosensing neurons in the submucosal plexus that may respond to and integrate signals involved in the regulation of energy homeostasis, resulting in negative feedback and adjustment of energy levels. In line with their recently demonstrated role in glucose sensing in the CNS, orexins appear to be important players in the physiological regulation of peripheral glucose levels. Orexins have been shown to be involved in the central regulation of wakefullness and locomotor activity that support food seeking. There are increasing data to support the concept that reduced food availability has a global stimulatory effect on reward perception (32). By decreasing neuronal input from adiposity signals, energy restriction increases responses to rewarding stimuli. A model has been proposed in which peripheral signals reflecting negative energy balance, such as reduced plasma glucose, induce fasting-related arousal by triggering increased activity of orexin neurons (23). We propose that orexins released by neurons and endocrine cells in the gut, and acting via distinct pathways, are connected to glucose transporter activity to allow sensing and control of incoming levels of glucose (as a part of a glucose “sensing” cell network that includes the pancreatic cells [7] and duodenal SGLT3 sensing [33]). Modulation of glucose levels through this postulated loop, in close relationship with other hormones and neuropeptides (i.e., leptin, CCK, and ghrelin), could in turn result in an increase in neuronal activity (32) and in production/secretion of peripheral orexin (7).

In conclusion, our data provide a rational basis for orexin-induced short-term regulation of intestinal glucose transport. Together with the recent findings that orexin neurons are specialized cells for sensing blood glucose in the hypothalamus (10), our results indicate that orexins can also modulate peripheral glucose disposal. In addition, they suggest that orexins are also involved in a regulatory loop between enteric glucose entry and central evaluation of peripheral available energy levels, enlarging the group of gut peptides involved in energy homeostasis (34). Further studies are needed to evaluate this possible role of orexins in pathologies associated with anomalies in energy homeostasis, including obesity and diabetes.

A.E.F. has received an MRT fellowship from the Ministère de l’Education et de la Recherche.

We thank Emma Pilling for re-reading the manuscript and Jean-Pierre Laigneau for drawing figures.

Parts of this study were presented at the 21st meeting of the European Intestinal Transport Group, Oberwiesenthal, Germany, 3–6 March 2007.