Peripheral clocks are known to modulate circadian patterns of insulin secretion. GLP-1 is an incretin hormone produced by the intestinal L cell that acts as a link between the gut and pancreatic β-cell. Herein, we demonstrate the existence of a diurnal rhythm in GLP-1 secretory responses to an oral glucose load in rats, with increased release immediately preceding the normal feeding period. This profile of GLP-1 release correlated with the pattern in insulin secretion, and both rhythms were completely inverted in animals subjected to a 12-h feeding cycle disruption and abolished in rats maintained under constant light conditions. A daily variation in the insulin response to exogenous GLP-1 was also found. Consistent with these in vivo findings, we demonstrated a circadian pattern in the GLP-1 secretory response to different secretagogues in murine GLUTag L cells, as well as in the mRNA levels of several canonical clock genes. Furthermore, significant changes in the expression of several genes were demonstrated by microarray and knockdown of two of them, thyrotroph embryonic factor and protein tyrosine phosphatase 4a1, resulted in altered GLP-1 secretion. Collectively, these results indicate that an independent peripheral clock in the L cell drives a circadian rhythm in GLP-1 secretory responses.

Most living organisms have developed circadian rhythms to adjust their metabolism to defined periodic changes in the environment, primarily the light-dark cycle (1,2). Furthermore, disruption of normal circadian rhythms has been associated with the onset of metabolic disorders such as type 2 diabetes and obesity (24), whereas maintenance of robust circadian rhythms may protect the body from these perturbations (5). A number of clock genes are responsible for generating circadian rhythms in the body. Although originally described in the suprachiasmatic nucleus (SCN), clock genes are expressed in most peripheral tissues, including the pancreatic β-cell and the gastrointestinal tract. These peripheral clocks, which are synchronized not only in response to the light-dark cycle, but also to the feeding cycle, modulate daily variations in organ activity (6). The existence of clock genes has been demonstrated in pancreatic islets from humans and rodents (7,8), disruption of which leads to defective β-cell function in vitro, as well as hypoinsulinemia and hyperglycemia in vivo (9). In agreement with this observation, it is known that insulin secretory responses follow a circadian pattern (10). However, it has been shown in rodents that this rhythm is more pronounced in response to oral feeding than to an intravenous glucose load (11), suggesting the involvement of gastrointestinal factors, such as the incretin hormones. Consistent with this notion, the existence of a peripheral clock has been demonstrated in the gastrointestinal tract (12,13), and many of the physiological activities of the gut, such as digestive functions, nutrient absorption, and proliferation, are all known to demonstrate circadian rhythms (14).

GLP-1 is a peptide hormone secreted by the intestinal L cell that acts as a link between the gut and pancreatic β-cell, stimulating glucose-dependent insulin secretion (15). The relevance of GLP-1 as an incretin has been demonstrated in rodents and humans, with both GLP-1 receptor antagonism and knockout reducing the insulin response to nutrient ingestion (16,17). Moreover, GLP-1 exerts other glucose-lowering actions, including suppression of glucagon secretion, gastric emptying, and appetite (18). All of these antidiabetic properties led to the approval of GLP-1–derivative drugs for the treatment of patients with type 2 diabetes (15). Since novel therapies involving GLP-1 secretagogues are currently under consideration, the characterization of the normal GLP-1 secretory patterns throughout the day is highly relevant for the optimization of such approaches.

L-cell secretion is stimulated by food intake through a process involving initial stimulation through the vagus nerve (19) followed by a direct effect of nutrients as they pass down the lumen of the gastrointestinal tract to the distal L cell (20). Previous studies (2123) exploring daily fluctuations in GLP-1 levels have failed to find any circadian rhythm, likely because they did not compare the L-cell response at different times throughout the light-dark cycle using identical nutrient composition and caloric loads. Since the magnitude of the L-cell response is highly dependent on these parameters (24), this may obscure any observation of a circadian rhythm. However, a recent study (25) has shown that GLP-1 levels in humans are higher in the morning than in the afternoon after ingestion of identical meals, indicating a temporal variation in L-cell responses.

Since a genuine circadian rhythm in L-cell secretion has yet to be established, GLP-1 responses to identical oral glucose loads were examined in rats throughout the dark-light cycle, as well as in animals subjected to feeding cycle disruption and constant light exposure. In addition, variations in the secretory activity of the L cell in response to different secretagogues were assessed in vitro using murine GLUTag (mGLUTag) cells. We herein demonstrate the existence of a circadian variation in L-cell secretory responses to a wide variety of known secretagogues. Confirmation of this rhythm in vivo and in vitro suggests that it may be driven by an independent peripheral clock in the L cell. In support of this, daily rhythms in mRNA levels for several clock genes were found in vitro. Microarray analysis and further knock-down approaches identified two candidate mediators of this rhythm. Together, these findings demonstrate, for the first time, the existence of a circadian rhythm in GLP-1 secretion by the intestinal L cell and suggest that GLP-1 acts, at least in part, to entrain the β-cell response to nutrient intake.

In Vivo Studies

Animals

Male Wistar rats (300–350 g; Charles River Laboratories, St. Constant, Quebec, Canada) were housed two to three rats per cage under a 12 h light-dark cycle with lights on at 7:00 a.m. (zeitgeber time [ZT] 0) and controlled room temperature, with free access to tap water and regular chow, and were handled daily to avoid experimental stress. All procedures were approved by the University of Toronto Animal Care Committee.

Experiments

Food intake was determined every 2 h for 36 h. Basal blood samples (300 µL) were drawn from the tail vain every 4 h for 24 h from eight rats fed ad libitum. Samples were collected into 30 µL of 0.1 mmol/L diprotin A (Sigma-Aldrich, Oakville, Ontario, Canada), 5,000 kallikrein inhibitor units/mL aprotinin (Sigma-Aldrich), and 0.03 mol/L EDTA solution, and plasma was stored at −20 C.

Rats were fasted for 4 h to allow postprandial levels of GLP-1 to return to basal levels (2125) and then were administered an oral glucose tolerance test (OGTT; 3 g/kg) at six different time points in a paired fashion (i.e., ZT2/ZT14, ZT6/ZT18, and ZT10/ZT22), with each animal studied at both times (n = 8 rats/each) separated by 1 week. Blood samples were collected at t = 0–60 min. Liver and ileum were collected after euthanasia.

Animals that were fasted for 4 h received a subcutaneous injection of vehicle (NaCl 0.9%) or GLP-1 (20 µg/kg) with or without an intraperitoneal glucose load (1.5 g/kg) at ZT10 and ZT22, allowing 1–2 weeks of recovery between tests. Blood samples were collected at t = 0–30 min.

One group of rats was maintained under the normal light-dark cycle with the animals fed only during the dark period (ZT12–ZT0, control group) or only during the light period (ZT0–ZT12) for 3 weeks. Another group was kept under constant light conditions and fed only during the former dark period (ZT12–ZT0) for 12 days. OGTTs were conducted at ZT10 and ZT22, as above, allowing 5–7 days of recovery between tests. Liver and ileum were collected after euthanasia.

Glucose and Hormone Determination

Plasma glucose levels were determined by enzymatic assay (Analox Instruments, Lunenburg, MA), and insulin and active and total GLP-1 levels were determined by sandwich immunoassay (Meso Scale Discovery, Gaithersburg, MD). Although active and total GLP-1 responses followed the same pattern (Supplementary Fig. 1), total GLP-1 level was assessed in most studies because a large number of animals had undetectable active GLP-1 levels at t = 0. No daily variation in the activity of dipeptidyl peptidase IV, the enzyme that degrades active GLP-1, has been reported (25,26). Leptin levels were quantified by radioimmunoassay (Millipore, Billerica, MA).

In Vitro Studies

Cell Culture

mGLUTag cells were selected as a model of the enteroendocrine L cell since they appropriately secrete GLP-1 in response to known secretagogues (27). The cells were propagated in DMEM (Gibco, Burlington, Ontario, Canada) containing 25 mmol/L glucose and 10% FBS (Gibco). Cells were synchronized by overnight starvation in DMEM with 0.5% FBS followed by a 1 h shock with 10 μmol/L forskolin (Sigma-Aldrich) in DMEM with 10% FBS, and then washed with Hank’s balanced salt solution for secretion or expression assay.

Secretion Assay

Synchronized cells were incubated for 2 h in DMEM with 0.5% FBS alone or containing 10−3 mol/L bethanechol (Sigma-Aldrich), 10−6 mol/L phorbol myristic acid (PMA; Sigma-Aldrich), 10−7 mol/L glucose-dependent insulinotropic peptide (GIP; Bachem, Torrance, CA), or 10−7 mol/L insulin (Eli Lilly, Toronto, Ontario, Canada). Peptides were extracted from medium and cells by reversed-phase adsorption to C18 silica (Sep-Pak; Waters, Mississauga, Ontario, Canada), and total GLP-1 levels were determined by radioimmunoassay (antiserum from Affinity Research Products, Nottingham, U.K.). GLP-1 secretion was calculated as the total GLP-1 content of the medium divided by the total GLP-1 content of the medium plus the cells.

Expression Assay

Cells were synchronized as above, and total RNA was extracted every 4 h over 48 h. When the effect of different treatments on gene expression was assessed, the cells were incubated for 2 h with the respective treatments, as above.

RNA Analysis

Total RNA was extracted using an RNeasy Plus Mini Kit (Qiagen, Germantown, MD) and quantified by spectrophotometric assay. Reverse transcription was performed using Superscript II with random primers (Invitrogen, Burlington, Ontario, Canada), and semiquantitative real-time PCR using TaqMan assays (primers listed in Supplementary Table 1) with TaqMan Universal Master Mix (Invitrogen). Data were analyzed using Opticon software (Bio-Rad, Hercules, CA), and relative expression was determined following the 2-ΔΔC(t) method (28). H3f3a (liver, mGLUTag cells) or 18S (ileal mucosa) was selected as an internal control after determination that their levels did not change over the experimental period (data not shown).

Intestinal Proliferation

Crypt cell proliferation was determined using formalin-fixed paraffin-embedded ileal samples with an antibody for the mitotic marker, phosphohistone H3 (Cell Marque, Rocklin, CA). Immunopositive cells were counted to the 18th cell position from the base of the crypt, on both sides of each crypt, counting ≥80 crypts per rat.

Microarray Analysis

Microarray was conducted at the Ontario Cancer Institute Genomics Centre (Toronto, Ontario, Canada) using cDNA (1.5 ng) generated from mGLUTag cells treated for 2 h with bethanechol at 4 and 16 h. Samples were hybridized onto one Mouse WG-6 V2 BeadChip (Illumina, San Diego, CA) and scanned on an iScan (Illumina). Data were imported in GeneSpring version 12.0 for analysis.

Small Interfering RNA Transfection

Silencer Select small interfering RNA (siRNA) oligonucleotides against thyrotroph embryonic factor (Tef; 5′-GGACAAGACGUAAGAAGAAtt), protein tyrosine phosphatase 4a1 (PRL-1; 5′-CCAACCAGAUUGUCGAUGAtt), as well as Silencer Select negative control #1 (scrambled RNA) were purchased from Life Technologies (Burlington, Ontario, Canada). Transfection was conducted using 80 nmol/L of each oligonucleotide with Opti-MEM I Reduced Serum Medium and Lipofectamine 2000 (Life Technologies) following the manufacturer’s recommendations. Transfection efficiency was determined by quantitative RT-PCR. Two days after transfection, cells were synchronized for 1 h as above, and secretion and expression assays were performed as above.

Statistical Analysis

Data are expressed as the mean ± SE. Areas under curves (AUCs) were calculated by the trapezoid rule. Insulin resistance was indirectly calculated as the product of the insulin AUC (picomoles per liter × 60 min) and the glucose AUC (millimoles per liter × 60 min), as validated previously (29). Student t test was conducted to compare two independent groups and one-way or two-way ANOVA followed by Tukey test, as appropriate, to compare three or more groups. For the microarray data, an uncorrected t test (P < 0.01) was performed. To calculate the period of the rhythms, nonlinear curve fitting was performed using OriginPro version 8.5 software.

Circadian Variation in the GLP-1, Insulin, and Glucose Responses to an OGTT

A daily variation in basal levels of total GLP-1 and insulin was observed in ad libitum–fed rats, peaking in the dark period between ZT14 and ZT18 (Supplementary Fig. 2). In contrast, basal levels of total GLP-1 and glucose were not significantly different in animals fasted for 4 h, although basal levels of insulin continued to be significantly higher at ZT22 (Fig. 1). To circumvent the differences observed in basal insulinemia and to facilitate a direct comparison in the magnitude of the responses among different time points, the profiles of total GLP-1, insulin, and glucose responses to the OGTT are expressed as the Δ increment from the time of oral glucose administration (t = 0, Fig. 1). Plasma GLP-1, insulin, and glucose levels increased in response to the OGTT, peaking at t = 10 min for all time points studied. The highest GLP-1 responses occurred at ZT6 and ZT10 (e.g., 1300 and 1700 h), while the lowest response was detected at ZT22 (e.g., 0500 h). Accordingly, the AUC of the Δ GLP-1 response at ZT22 was significantly lower than those at ZT6 and ZT10 (P < 0.05–0.01, Fig. 2). A similar pattern was found in the insulin profiles, with the highest response observed at ZT10 and the lowest at ZT22 (Fig. 1); the AUC of the Δ insulin response at ZT10 (e.g., 1300 h) was thus significantly higher than the responses at ZT6, ZT18, and ZT22 (e.g., 1300, 0100, and 0500 h, respectively; P < 0.05–0.001; Fig. 2). When the daily patterns of GLP-1 and insulin AUC responses to the OGTT were curve fit over a 24-h period, correlation coefficients of 77% and 67%, respectively (P < 0.001 for both), were obtained. Similar patterns for both GLP-1 and insulin were observed when the AUC of the absolute values was examined (Supplementary Fig. 3). In contrast, curve fitting for the Δ glycemic responses fit to a 12-h period (R2 = 97%, P < 0.001, Fig. 2). This finding indicates that the rhythm in glucose responses is partially independent of the rhythms in GLP-1 and insulin, and is likely modulated by other factors such as daily variations in insulin sensitivity. Accordingly, the highest insulin resistance was detected at ZT10 (data not shown), coinciding in time with the greatest GLP-1 and insulin responses. Complementarily, a peak in ileal proliferative activity was detected at ZT8, concurrent with the highest L-cell secretory response (Supplementary Fig. 4). Finally, the patterns in GLP-1 and insulin responses were inversely correlated to the pattern in food intake, which fit to a 24-h rhythm (R2 = 97%, P < 0.001, Fig. 2).

Figure 1

Plasma GLP-1, insulin, and glucose profiles after an OGTT at six time points throughout the light-dark cycle in rats that were fasted for 4 h (n = 7–8). Light and dark periods are indicated by the open and solid bars, respectively. Responses are expressed as the difference from the respective time = 0 values (T0) shown at the top of each respective graph. **P < 0.01, ***P < 0.001 vs. ZT22.

Figure 1

Plasma GLP-1, insulin, and glucose profiles after an OGTT at six time points throughout the light-dark cycle in rats that were fasted for 4 h (n = 7–8). Light and dark periods are indicated by the open and solid bars, respectively. Responses are expressed as the difference from the respective time = 0 values (T0) shown at the top of each respective graph. **P < 0.01, ***P < 0.001 vs. ZT22.

Figure 2

Daily variation in the AUCs of total GLP-1 (A), insulin (B), and glucose (C) Δ responses to the OGTTs shown in Figure 1. D: Daily pattern of food intake in rats fed ad libitum and maintained under a regular 12 h light-dark cycle. Data are double plotted and separated by a broken line.

Figure 2

Daily variation in the AUCs of total GLP-1 (A), insulin (B), and glucose (C) Δ responses to the OGTTs shown in Figure 1. D: Daily pattern of food intake in rats fed ad libitum and maintained under a regular 12 h light-dark cycle. Data are double plotted and separated by a broken line.

A marked rhythm in the hepatic mRNA levels of three canonical clock genes, Bmal1, Rev-Erbα, and Per2 was found (Fig. 3), oscillating in opposing phases, as expected (6,7). Parallel oscillations in Bmal1 and Rev-Erbα mRNA were also observed in the ileal mucosa, the intestinal segment with the highest L-cell abundance (30). A rhythm in proglucagon mRNA levels was also found in the ileum, peaking at ZT22.

Figure 3

mRNA levels of Bmal1, Rev-Erbα, and Per2 in the liver, and Bmal1, Rev-Erbα, and Gcg in the ileal mucosa, after an OGTT (n = 7–8). Data are double plotted and separated by a broken line.

Figure 3

mRNA levels of Bmal1, Rev-Erbα, and Per2 in the liver, and Bmal1, Rev-Erbα, and Gcg in the ileal mucosa, after an OGTT (n = 7–8). Data are double plotted and separated by a broken line.

Daily Variation in the β-Cell Sensitivity to Circulating GLP-1

To determine whether there is a rhythm in β-cell responses to GLP-1, animals were administered GLP-1 at the demonstrated zenith (ZT10) and trough (ZT22) time points for oral glucose-induced insulin release (Fig. 2), with or without a concomitant intraperitoneal glucose load. GLP-1 treatment significantly stimulated insulin secretion compared with vehicle-treated controls under both conditions (P < 0.05, Fig. 4), although the magnitude of the insulin response promoted by GLP-1 was 8- to 10-fold higher when the peptide was administered in combination with the intraperitoneal glucose load, as expected given the glucose dependency of the GLP-1 effects on insulin secretion. Interestingly, the insulin response elicited by GLP-1 plus glucose administration was significantly (P < 0.05) higher at ZT10 compared with ZT22, suggesting enhanced β-cell sensitivity to GLP-1 during the afternoon hours.

Figure 4

AUC of the insulin response to subcutaneous administration of vehicle (0.9% NaCl) or GLP-1 (20 µg/kg) alone (A) or in combination with an intraperitoneal glucose load (B) (1.5 g/kg; n = 6–8; *P < 0.05 vs. vehicle control; #P < 0.05 vs. ZT22).

Figure 4

AUC of the insulin response to subcutaneous administration of vehicle (0.9% NaCl) or GLP-1 (20 µg/kg) alone (A) or in combination with an intraperitoneal glucose load (B) (1.5 g/kg; n = 6–8; *P < 0.05 vs. vehicle control; #P < 0.05 vs. ZT22).

Effects of Feeding Cycle Disruption and Constant Light Exposure on the Pattern of GLP-1 and Insulin Secretory Responses

Time-restricted feeding (e.g., dark only vs. light only) produced an initial drop (P < 0.001) in food consumption in the animals that were fed exclusively during the light period, without affecting body weight (Supplementary Fig. 5). However, by the day of the first OGTT (e.g., 3 weeks after the feeding regimen started), both group of animals showed comparable daily food intake and body weight. Feeding cycle disruption completely inverted the phases of hepatic mRNA levels for Bmal1, Rev-Erbα, and Per2, as expected for this nutrient-entrained tissue (1,6), but no differences were observed in the ileal mRNA levels of proglucagon between the two feeding conditions (Supplementary Fig. 6). No significant differences were detected in basal GLP-1 and insulin levels at ZT10 and ZT22 in any group of animals (Fig. 5). Both GLP-1 and insulin showed significantly greater responses to oral glucose (P < 0.05) when the load was administered at the end of the fasting period (e.g., at ZT10 in dark-fed animals and at ZT22 in light-fed rats), compared with the end of the feeding period (e.g., at ZT22 in dark-fed rats and ZT10 in light-fed animals), indicating the relevance of the feeding cycle as a zeitgeber (Fig. 5 and Supplementary Fig. 7). In addition, light-fed animals showed similar glucose responses at ZT10 and ZT22, whereas dark-fed rats exhibited significantly improved glucose tolerance at ZT22 compared with both their own ZT10 response and the response of light-fed rats at ZT22 (P < 0.05).

Figure 5

Plasma total GLP-1 (A), insulin (B), and glucose (C) responses and corresponding AUCs in response to OGTTs at ZT10 (open circles) and ZT22 (closed squares) in rats exclusively fed during the dark or light period. Time-course responses are expressed as the difference from the respective basal levels shown at the top of each graph (n = 7–8; *P < 0.05, **P < 0.01 between different time points within the same group; #P < 0.05, ##P < 0.01 for the same time point between different groups).

Figure 5

Plasma total GLP-1 (A), insulin (B), and glucose (C) responses and corresponding AUCs in response to OGTTs at ZT10 (open circles) and ZT22 (closed squares) in rats exclusively fed during the dark or light period. Time-course responses are expressed as the difference from the respective basal levels shown at the top of each graph (n = 7–8; *P < 0.05, **P < 0.01 between different time points within the same group; #P < 0.05, ##P < 0.01 for the same time point between different groups).

Unexpectedly, the variations in GLP-1 and insulin responses between ZT10 and ZT22 were completely abolished in animals under constant light exposure and fed only during the subjective night (i.e., ZT12–ZT24), with significantly (P < 0.05) higher responses of both hormones at ZT22 compared with dark-fed rats under regular light-dark cycle conditions (Fig. 6 and Supplementary Fig. 8). In addition, animals exposed to constant light showed impaired glucose tolerance, with significantly (P < 0.05) higher glucose responses at both ZT10 and ZT22.

Figure 6

Plasma total GLP-1 (A), insulin (B), and glucose (C) responses and corresponding AUCs in response to OGTTs at ZT10 and ZT22 in rats exclusively fed between ZT12 and ZT0, and maintained under regular light-dark cycle (LD; closed circles; reproduced from Fig. 5) or constant light exposure (LL; open squares). Time-course responses are expressed as the difference from the respective basal levels shown at the top of each graph (n = 7–8; *P < 0.05 between different time points within the same group; #P < 0.05 for the same time point between different groups).

Figure 6

Plasma total GLP-1 (A), insulin (B), and glucose (C) responses and corresponding AUCs in response to OGTTs at ZT10 and ZT22 in rats exclusively fed between ZT12 and ZT0, and maintained under regular light-dark cycle (LD; closed circles; reproduced from Fig. 5) or constant light exposure (LL; open squares). Time-course responses are expressed as the difference from the respective basal levels shown at the top of each graph (n = 7–8; *P < 0.05 between different time points within the same group; #P < 0.05 for the same time point between different groups).

Circadian Variation in GLP-1 Secretory Responses In Vitro

The existence of an independent peripheral clock in mGLUTag cells was shown by the demonstration of robust circadian fluctuations in Bmal1, Rev-Erbα, and Per2 mRNA transcript levels, with a calculated period of 28.8, 25.1, and 22.7 h, respectively (Fig. 7A). A rhythm in the GLP-1 secretory response to bethanechol was also observed in mGLUTag cells, showing peaks of secretion at 4 and 28 h after synchronization (Fig. 7B). A similar but less robust pattern was obtained for the GLP-1 secretory response to PMA (Fig. 7C). The effects of GIP and insulin on GLP-1 secretion were therefore also studied at these peak (4 h) and trough (16 h) time points. Both hormones (as well as bethanechol, positive control) induced a 3.5-fold increase in GLP-1 secretion at 4 h (P < 0.01) compared with vehicle-treated cells; in contrast, no GLP-1 response was observed at time 16 h (Fig. 7D). Proglucagon mRNA levels were also studied over a 48-h period and were found to be relatively stable throughout the study with slightly lower levels during the second 24-h period compared with the first 24-h period (Fig. 7E). In response to the different secretagogues tested, proglucagon mRNA levels were consistently higher, by 1.2- to 1.5-fold, at 4 h after synchronization, although only the response to GIP reached significance (P < 0.05, Fig. 7F).

Figure 7

A: mRNA levels of Bmal1, Rev-Erbα, and Per2 in synchronized mGLUTag cells (n = 4–8) over a period of 48 h after synchronization. The time point with the lowest average value was set as 1 for each gene. GLP-1 secretory responses to bethanechol (B) and PMA (C) in synchronized mGLUTag cells (n = 8) are shown. Results are expressed as the percentage of GLP-1 secretion with respect to t = 0. The nonlinear curve-fitting calculation is represented as a gray dotted line, with the coefficient of correlation and period depicted in the box attached to each graph. D: GLP-1 released by mGLUTag cells in response to GIP, bethanechol, and insulin at 4 and 16 h after synchronization. E: mRNA levels of proglucagon in synchronized mGLUTag cells (n = 4–8) over a period of 48 h after synchronization. F: Proglucagon mRNA levels at t = 4 h vs. 16 h in response to different treatments. Data are expressed as the fold change normalized by control group (*P < 0.05, **P < 0.01, vs. control group; ##P < 0.01 vs. t = 4 h).

Figure 7

A: mRNA levels of Bmal1, Rev-Erbα, and Per2 in synchronized mGLUTag cells (n = 4–8) over a period of 48 h after synchronization. The time point with the lowest average value was set as 1 for each gene. GLP-1 secretory responses to bethanechol (B) and PMA (C) in synchronized mGLUTag cells (n = 8) are shown. Results are expressed as the percentage of GLP-1 secretion with respect to t = 0. The nonlinear curve-fitting calculation is represented as a gray dotted line, with the coefficient of correlation and period depicted in the box attached to each graph. D: GLP-1 released by mGLUTag cells in response to GIP, bethanechol, and insulin at 4 and 16 h after synchronization. E: mRNA levels of proglucagon in synchronized mGLUTag cells (n = 4–8) over a period of 48 h after synchronization. F: Proglucagon mRNA levels at t = 4 h vs. 16 h in response to different treatments. Data are expressed as the fold change normalized by control group (*P < 0.05, **P < 0.01, vs. control group; ##P < 0.01 vs. t = 4 h).

Microarray analysis comparing samples at t = 4 and 16 h demonstrated significant differences for 517 probes, several of which are represented in Fig. 8A. PRL-1, a modulator of the extracellular signal–regulated kinase signaling pathway (31), and Tef, a transcription factor implicated in the circadian regulation of islet genes involved in insulin production and secretion (32), were among the genes with the greatest fold change (1.41 and −1.63, respectively, Fig. 8A). The expression of these two genes was therefore knocked down (by 65% and 35%, respectively, P < 0.05–0.001, Fig. 8B and C) to investigate their functional role in GLP-1 secretion. Only the cells transfected with PRL-1 siRNA showed a significant reduction (by 25%, P < 0.05) in proglucagon mRNA levels (Fig. 8D and E). However, roles for both PRL-1 and Tef in L-cell secretory activity were demonstrated in synchronized GLUTag cells (Fig. 8F). Specifically, when PRL-1 (the gene with higher expression during the peak of the GLP-1 secretory rhythm) was knocked down, the GLP-1 response to secretagogues was significantly attenuated, whereas with knockdown of Tef (the gene with higher expression during the secretory trough), GLP-1 secretion was significantly increased, both in the vehicle-treated cells and in response to secretagogues.

Figure 8

A: Relative levels of expression at t = 4 vs. 16 h for Bmal1, Rev-Erbα, Per2, PRL-1, and Tef in mGLUTag cells incubated for 2 h with bethanechol. Relative mRNA levels of the respective target gene (B and C) and proglucagon (D and E) in mGLUTag cells transfected with either PRL-1 (B and D) or Tef (C and E) siRNA. F: GLP-1 secretory response to vehicle, bethanechol, GIP, and insulin in PRL-1, Tef, and scrambled siRNA-transfected cells. Sc, scrambled. A: *P < 0.05 at 4 h compared with 16 h. C–F: *P < 0.05, ***P < 0.001 vs. scrambled control.

Figure 8

A: Relative levels of expression at t = 4 vs. 16 h for Bmal1, Rev-Erbα, Per2, PRL-1, and Tef in mGLUTag cells incubated for 2 h with bethanechol. Relative mRNA levels of the respective target gene (B and C) and proglucagon (D and E) in mGLUTag cells transfected with either PRL-1 (B and D) or Tef (C and E) siRNA. F: GLP-1 secretory response to vehicle, bethanechol, GIP, and insulin in PRL-1, Tef, and scrambled siRNA-transfected cells. Sc, scrambled. A: *P < 0.05 at 4 h compared with 16 h. C–F: *P < 0.05, ***P < 0.001 vs. scrambled control.

Ingestion of nutrients is the primary stimulus for L-cell secretion, with the highest GLP-1 levels achieved during the immediate postprandial period and a return to basal values within 3–4 h (2125). Furthermore, L-cell responses are highly dependent on meal composition and caloric content (24). In the current study, therefore, to enable direct comparison of GLP-1 secretory responses over a 24-h period, rats were fasted for 4 h and then administered an identical glucose load. Using this protocol, we demonstrate the existence of a diurnal rhythm in GLP-1 secretory responses in rats that is synchronized with insulin secretion. This rhythm is entrained by nutrient ingestion in animals kept under regular light-dark cycle conditions, but is disrupted by changes in the feeding cycle as well as by exposure to constant light. Complementarily, we also show a circadian rhythm in L-cell responses to a variety of secretagogues in vitro, as well as rhythmic expression of clock genes. Knockdown of two of the detected cycling genes in the L cells resulted in alterations in GLP-1 release. Taken together, our data demonstrate the existence of a diurnal pattern in L-cell secretion that may serve to tightly coordinate intestinal function and insulin release with nutrient ingestion. GLP-1 is thus a possible mediator of the metabolic disorders triggered by the disruption of the regular sleep-wake cycle (3,4).

A daily pattern in insulin secretion in response to an OGTT has previously been demonstrated in humans and rats (10,33). In agreement with these studies, we found the highest insulin responses at the end of the inactive/light period, when the animals begin to eat, a change that was highly correlated with the GLP-1 responses in the same animals. The diurnal variation in GLP-1 responses was limited to the first phase of the secretory response (e.g., 10–30 min after oral glucose administration). Since L cells are mostly concentrated in the distal ileum and the colon (30), and glucose has not reached this area of the gastrointestinal tract when GLP-1 secretion is triggered (34), the first phase of GLP-1 release is believed to be vagally mediated (19,35). Consistent with this notion, we herein demonstrate the existence of a circadian rhythm in the mGLUTag L-cell response to bethanechol, a nonspecific muscarinic agonist that acts via activation of protein kinase C (35). Moreover, a circadian rhythm in the GLUTag cell response to PMA, a diacylglycerol mimetic that activates protein kinase C, was also observed. Thus, the diurnal variation in GLP-1 responses to oral glucose observed in rats could be a result of a daily rhythm in the L-cell response to parasympathetic stimulation. However, microarray analysis failed to detect significant differences in the expression levels of the components of these pathways, including the M1 and M2 receptors, between the established peak and trough times of mGLUTag cell secretory activity (e.g., at 4 vs. 16 h postsynchronization; data not shown). Furthermore, we also found that mGLUTag cell responses to GIP and insulin, which are known to stimulate GLP-1 secretion by activating cAMP and Erk1/2 signaling, respectively (36,37), showed a similar pattern, indicating that the daily variation in the L-cell responsiveness to stimulus is independent of the signaling pathways being activated. The molecular mechanisms underlying the circadian rhythm in L-cell activity may therefore be a result of a complex integration of several signals. Herein, we provide evidence that PRL-1 and Tef, two genes whose expression was found to oscillate in opposite phases in the GLUTag cells, play significant roles in GLP-1 secretion. However, further studies are required to provide a better understanding of the molecular mechanisms underlying the regulation of L-cell circadian activity by these effectors.

Many of the physiological activities of the gut are known to follow circadian rhythms (14), consistent with the expression of peripheral clock genes in this tissue (12,13). We now demonstrate, for the first time, the existence of self-sustained oscillators (the canonical clock genes Bmal1, Rev-Erbα, and Per2) in an L-cell model, which may modulate the diurnal variation in L-cell activity. Although, under normal conditions, the circadian activity of peripheral clocks is coordinated by the SCN, other factors, such as the feeding cycle, are also able to synchronize the activity of these peripheral clocks (1,6,12,13). Consistent with this, we demonstrated that inversion of the feeding regimen for 3 weeks completely inverted the GLP-1 and insulin responses to an OGTT. These data suggest that rhythmic expression of the clock genes by the L cell is entrained by the feeding rhythm, as we found for these genes in the liver. However, in animals kept under constant light exposure and fed exclusively during the subjective dark period for 10 days, the rhythmicity in GLP-1 and insulin responses to oral glucose was lost, with significantly higher responses of both hormones during the subjective nadir. This finding is consistent with previous studies, which have shown several metabolic alterations in response to constant light exposure (38). It is known that constant light conditions greatly impair the rhythmic activity in the neurons of the SCN (38), and disruptions in the master clock may thus also impact entrainment of the molecular clock of the L cell. Ultimately, these findings also suggest that a disruption in the circadian variation of GLP-1 responses may precede the metabolic disorders caused by alterations in the sleep-wake cycle, a hypothesis that can be tested. Moreover, although the β-cell is regulated by its own independent clock genes (79), our observation that β-cell sensitivity to GLP-1 demonstrates a diurnal rhythm further supports the tight association between the GLP-1 and insulin secretory responses to nutrient ingestion. The coordination in the rhythmic activity of both the L cell and the β-cell may be particularly important to compensate for changes in insulin sensitivity, as we have observed that the GLP-1–stimulated insulin release was enhanced when the lowest insulin sensitivity was detected in our rats.

We have also demonstrated a diurnal variation in the proliferation of the ileal mucosa, with a peak in mitotic activity at ZT8, as previously reported (39). Interestingly, this pattern closely matched the rhythm in GLP-1 responses. Since GLP-2, which is cosecreted with GLP-1, stimulates intestinal proliferation (40), the rhythm in crypt proliferation may reflect the changes in GLP-2 cosecretion by the intestinal L cell. Because long-acting GLP-2 analogs have recently been approved by the Food and Drug Administration for treatment of patients with short bowel syndrome, further analyses of potential rhythms in GLP-2 action are clearly warranted.

In summary, our data demonstrate the existence of a daily variation in the L-cell response to stimuli in rats as well as in a murine L-cell line in vitro. This rhythm is entrained by the feeding cycle and appears to be produced by an independent molecular clock in the L cell. Furthermore, this rhythm may be instrumental in the daily pattern of insulin secretory responses as well as in gut homeostasis. Future studies are required to establish whether this pattern in L-cell secretion also occurs in humans, including shift workers, to determine whether their increased risk for metabolic disorders (2,3) is associated with alterations in their GLP-1 secretory rhythm. These findings may also have clinical implications since secretagogues that would enhance the levels of GLP-1 at specific times of day may be therapeutically useful in conditions such as type 2 diabetes.

See accompanying article, p. 3584.

Funding. This work was supported by an operating grant from the Canadian Diabetes Association (2973) and an equipment grant from the Canadian Foundation for Innovation and Ontario Research Fund (19442). M.G.-L. was supported by fellowships from the Canadian Institutes of Health Research Sleep and Biological Rhythms Training Program, University of Toronto; E.L.M. and W.K.W. by summer studentships from the Banting and Best Diabetes Centre, University of Toronto; S.A.R. by a summer studentship from the Canadian Association of Gastroenterology; and P.L.B. by the Canada Research Chairs Program.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.G.-L. contributed to the concept and design of the study; the acquisition, analysis, and interpretation of the data; and the drafting and revision of the manuscript. E.L.M., W.K.W., and S.A.R. contributed to the acquisition, analysis, and interpretation of the data. P.L.B. contributed to the concept and design of the study, obtained funding, provided study supervision, and contributed to the drafting and revision of the manuscript. P.L.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 73rd Scientific Sessions of the American Diabetes Association, Chicago, IL, 21–25 June 2013.

1.
Asher
G
,
Schibler
U
.
Crosstalk between components of circadian and metabolic cycles in mammals
.
Cell Metab
2011
;
13
:
125
137
[PubMed]
2.
Froy
O
.
Metabolism and circadian rhythms—implications for obesity
.
Endocr Rev
2010
;
31
:
1
24
[PubMed]
3.
Suwazono
Y
,
Sakata
K
,
Okubo
Y
, et al
.
Long-term longitudinal study on the relationship between alternating shift work and the onset of diabetes mellitus in male Japanese workers
.
J Occup Environ Med
2006
;
48
:
455
461
[PubMed]
4.
Shi
SQ
,
Ansari
TS
,
McGuinness
OP
,
Wasserman
DH
,
Johnson
CH
.
Circadian disruption leads to insulin resistance and obesity
.
Curr Biol
2013
;
23
:
372
381
[PubMed]
5.
Hatori
M
,
Vollmers
C
,
Zarrinpar
A
, et al
.
Time-restricted feeding without reducing caloric intake prevents metabolic diseases in mice fed a high-fat diet
.
Cell Metab
2012
;
15
:
848
860
[PubMed]
6.
Damiola
F
,
Le Minh
N
,
Preitner
N
,
Kornmann
B
,
Fleury-Olela
F
,
Schibler
U
.
Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus
.
Genes Dev
2000
;
14
:
2950
2961
[PubMed]
7.
Pulimeno
P
,
Mannic
T
,
Sage
D
, et al
.
Autonomous and self-sustained circadian oscillators displayed in human islet cells
.
Diabetologia
2013
;
56
:
497
507
[PubMed]
8.
Mühlbauer
E
,
Wolgast
S
,
Finckh
U
,
Peschke
D
,
Peschke
E
.
Indication of circadian oscillations in the rat pancreas
.
FEBS Lett
2004
;
564
:
91
96
[PubMed]
9.
Marcheva
B
,
Ramsey
KM
,
Buhr
ED
, et al
.
Disruption of the clock components CLOCK and BMAL1 leads to hypoinsulinaemia and diabetes
.
Nature
2010
;
466
:
627
631
[PubMed]
10.
Jarrett
RJ
,
Baker
IA
,
Keen
H
,
Oakley
NW
.
Diurnal variation in oral glucose tolerance: blood sugar and plasma insulin levels morning, afternoon, and evening
.
BMJ
1972
;
1
:
199
201
[PubMed]
11.
Kalsbeek
A
,
Strubbe
JH
.
Circadian control of insulin secretion is independent of the temporal distribution of feeding
.
Physiol Behav
1998
;
63
:
553
558
[PubMed]
12.
Sládek
M
,
Rybová
M
,
Jindráková
Z
, et al
.
Insight into the circadian clock within rat colonic epithelial cells
.
Gastroenterology
2007
;
133
:
1240
1249
[PubMed]
13.
Hoogerwerf
WA
,
Hellmich
HL
,
Cornélissen
G
, et al
.
Clock gene expression in the murine gastrointestinal tract: endogenous rhythmicity and effects of a feeding regimen
.
Gastroenterology
2007
;
133
:
1250
1260
[PubMed]
14.
Pácha
J
,
Sumová
A
.
Circadian regulation of epithelial functions in the intestine
.
Acta Physiol (Oxf)
2013
;
208
:
11
24
[PubMed]
15.
Dong
CX
,
Brubaker
PL
.
Ghrelin, the proglucagon-derived peptides and peptide YY in nutrient homeostasis
.
Nat Rev Gastroenterol Hepatol
2012
;
9
:
705
715
[PubMed]
16.
Edwards
CM
,
Todd
JF
,
Mahmoudi
M
, et al
.
Glucagon-like peptide 1 has a physiological role in the control of postprandial glucose in humans: studies with the antagonist exendin 9-39
.
Diabetes
1999
;
48
:
86
93
[PubMed]
17.
Scrocchi
LA
,
Brown
TJ
,
MaClusky
N
, et al
.
Glucose intolerance but normal satiety in mice with a null mutation in the glucagon-like peptide 1 receptor gene
.
Nat Med
1996
;
2
:
1254
1258
[PubMed]
18.
Drucker
DJ
.
The biology of incretin hormones
.
Cell Metab
2006
;
3
:
153
165
[PubMed]
19.
Rocca
AS
,
Brubaker
PL
.
Role of the vagus nerve in mediating proximal nutrient-induced glucagon-like peptide-1 secretion
.
Endocrinology
1999
;
140
:
1687
1694
[PubMed]
20.
Roberge
JN
,
Brubaker
PL
.
Regulation of intestinal proglucagon-derived peptide secretion by glucose-dependent insulinotropic peptide in a novel enteroendocrine loop
.
Endocrinology
1993
;
133
:
233
240
[PubMed]
21.
Elliott
RM
,
Morgan
LM
,
Tredger
JA
,
Deacon
S
,
Wright
J
,
Marks
V
.
Glucagon-like peptide-1 (7-36)amide and glucose-dependent insulinotropic polypeptide secretion in response to nutrient ingestion in man: acute post-prandial and 24-h secretion patterns
.
J Endocrinol
1993
;
138
:
159
166
[PubMed]
22.
Orskov
C
,
Wettergren
A
,
Holst
JJ
.
Secretion of the incretin hormones glucagon-like peptide-1 and gastric inhibitory polypeptide correlates with insulin secretion in normal man throughout the day
.
Scand J Gastroenterol
1996
;
31
:
665
670
[PubMed]
23.
Mingrone
G
,
Nolfe
G
,
Gissey
GC
, et al
.
Circadian rhythms of GIP and GLP1 in glucose-tolerant and in type 2 diabetic patients after biliopancreatic diversion
.
Diabetologia
2009
;
52
:
873
881
[PubMed]
24.
Vilsbøll
T
,
Krarup
T
,
Sonne
J
, et al
.
Incretin secretion in relation to meal size and body weight in healthy subjects and people with type 1 and type 2 diabetes mellitus
.
J Clin Endocrinol Metab
2003
;
88
:
2706
2713
[PubMed]
25.
Lindgren
O
,
Mari
A
,
Deacon
CF
, et al
.
Differential islet and incretin hormone responses in morning versus afternoon after standardized meal in healthy men
.
J Clin Endocrinol Metab
2009
;
94
:
2887
2892
[PubMed]
26.
Ahrén
B
,
Landin-Olsson
M
,
Jansson
PA
,
Svensson
M
,
Holmes
D
,
Schweizer
A
.
Inhibition of dipeptidyl peptidase-4 reduces glycemia, sustains insulin levels, and reduces glucagon levels in type 2 diabetes
.
J Clin Endocrinol Metab
2004
;
89
:
2078
2084
[PubMed]
27.
Brubaker
PL
,
Schloos
J
,
Drucker
DJ
.
Regulation of glucagon-like peptide-1 synthesis and secretion in the GLUTag enteroendocrine cell line
.
Endocrinology
1998
;
139
:
4108
4114
[PubMed]
28.
Livak
KJ
,
Schmittgen
TD
.
Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method
.
Methods
2001
;
25
:
402
408
[PubMed]
29.
Tran
TT
,
Gupta
N
,
Goh
T
, et al
.
Direct measure of insulin sensitivity with the hyperinsulinemic-euglycemic clamp and surrogate measures of insulin sensitivity with the oral glucose tolerance test: correlations with aberrant crypt foci promotion in rats
.
Cancer Epidemiol Biomarkers Prev
2003
;
12
:
47
56
[PubMed]
30.
Eissele
R
,
Göke
R
,
Willemer
S
, et al
.
Glucagon-like peptide-1 cells in the gastrointestinal tract and pancreas of rat, pig and man
.
Eur J Clin Invest
1992
;
22
:
283
291
[PubMed]
31.
Bai
Y
,
Luo
Y
,
Liu
S
, et al
.
PRL-1 protein promotes ERK1/2 and RhoA protein activation through a non-canonical interaction with the Src homology 3 domain of p115 Rho GTPase-activating protein
.
J Biol Chem
2011
;
286
:
42316
42324
[PubMed]
32.
Allaman-Pillet
N
,
Roduit
R
,
Oberson
A
, et al
.
Circadian regulation of islet genes involved in insulin production and secretion
.
Mol Cell Endocrinol
2004
;
226
:
59
66
[PubMed]
33.
Hara
E
,
Saito
M
.
Diurnal changes in plasma glucose and insulin responses to oral glucose load in rats
.
Am J Physiol
1980
;
238
:
E463
E466
[PubMed]
34.
Dahlqvist
A
,
Thomson
DL
.
The digestion and absorption of sucrose by the intact rat
.
J Physiol
1963
;
167
:
193
209
[PubMed]
35.
Anini
Y
,
Brubaker
PL
.
Muscarinic receptors control glucagon-like peptide 1 secretion by human endocrine L cells
.
Endocrinology
2003
;
144
:
3244
3250
[PubMed]
36.
Simpson
AK
,
Ward
PS
,
Wong
KY
, et al
.
Cyclic AMP triggers glucagon-like peptide-1 secretion from the GLUTag enteroendocrine cell line
.
Diabetologia
2007
;
50
:
2181
2189
[PubMed]
37.
Lim
GE
,
Huang
GJ
,
Flora
N
,
LeRoith
D
,
Rhodes
CJ
,
Brubaker
PL
.
Insulin regulates glucagon-like peptide-1 secretion from the enteroendocrine L cell
.
Endocrinology
2009
;
150
:
580
591
[PubMed]
38.
Coomans
CP
,
van den Berg
SA
,
Houben
T
, et al
.
Detrimental effects of constant light exposure and high-fat diet on circadian energy metabolism and insulin sensitivity
.
FASEB J
2013
;
27
:
1721
1732
[PubMed]
39.
Al-Dewachi
HS
,
Wright
NA
,
Appleton
DR
,
Watson
AJ
.
Studies on the mechanism of diurnal variation of proliferative indices in the small bowel mucosa of the rat
.
Cell Tissue Kinet
1976
;
9
:
459
467
[PubMed]
40.
Rowland
KJ
,
Trivedi
S
,
Lee
D
, et al
.
Loss of glucagon-like peptide-2-induced proliferation following intestinal epithelial insulin-like growth factor-1-receptor deletion
.
Gastroenterology
2011
;
141
:
2166
2175.e7
[PubMed]

Supplementary data