Metformin Triggers PYY Secretion in Human Gut Mucosa

Abstract

Context

The antidiabetic drug metformin causes weight loss, but the underlying mechanisms are unclear. Recent clinical studies show that metformin increases plasma levels of the anorectic gut hormone, peptide YY (PYY), but whether this is through a direct effect on the gut is unknown.

Objective

We hypothesized that exposure of human gut mucosal tissue to metformin would acutely trigger PYY secretion.

Design, Setting, Participants, and Interventions

Mucosal tissue was prepared from 46 human colonic and 9 ileal samples obtained after surgical resection and ex vivo secretion assays were performed. Tissue was exposed to metformin, as well as a series of other compounds as part of our mechanistic studies, in static incubations. Supernatant was sampled after 15 minutes.

Main Outcome Measures

PYY levels in supernatant, measured using ELISA.

Results

Metformin increased PYY secretion from both ileal (P < 0.05) and colonic (P < 0.001) epithelia. Both basal and metformin-induced PYY secretion were unchanged across body mass index or in tissues obtained from individuals with type 2 diabetes. Metformin-dependent PYY secretion was blocked by inhibitors of the plasma membrane monoamine transporter (PMAT) and the serotonin reuptake transporter (SERT), as well as by an inhibitor of AMP kinase (AMPK).

Conclusions

This is a report of a direct action of metformin on the gut epithelium to trigger PYY secretion in humans, occurring via cell internalization through PMAT and SERT and intracellular activation of AMPK. Our results provide further support that the role of metformin in the treatment of metabolic syndrome has a gut-based component.

Metformin has been used to treat type 2 diabetes for more than half a century via actions to improve peripheral insulin sensitivity and decrease hepatic glucose output (1). The therapeutic effect of metformin is also thought to be mediated, in part, by actions within the gastrointestinal tract (2), including effects on the incretin system (3–5). The weight-reducing effect associated with metformin use was recognized soon after it was first used as an efficacious antidiabetic agent (6), but the underlying mechanism remains unclear. Metformin was recently shown to decrease maternal weight gain in pregnant women with obesity without type 2 diabetes (7) and has been used to achieve modest weight loss in women with polycystic ovarian syndrome (8). Furthermore, there are ongoing investigations to explore metformin’s place as an adjunct therapy to reduce weight gain secondary to atypical antipsychotic treatments (9–11). Several potential mechanisms underlying metformin-induced weight loss have been proposed, including actions to reduce plasma insulin levels secondary to improved peripheral insulin sensitivity (8). Moreover, metformin treatment could alter the composition of the gut microbiota (12), potentially inducing a shift toward a microbial population that is more metabolically favorable for the host. It has also been reported that metformin treatment induces satiety and reduces food intake in patients with obesity with or without diabetes (13, 14), which could potentially explain treatment-associated weight loss. Indeed, rodent studies have consistently shown that oral metformin treatment causes substantial weight loss in obese animals and this is likely to be due to reduced food intake (15–17). However, the mechanisms underlying metformin’s anorectic effect remained to be investigated.

The gut hormone peptide YY (PYY)_1–36 is produced in enteroendocrine L cells located primarily in the epithelia of the distal small intestine and colon, and it is secreted in a manner likely to involve both paracrine and neural signaling events (18). Secreted PYY_1–36 is rapidly cleaved by peripheral dipeptidyl peptidase IV to another active form, PYY_3–36 (19). PYY_1–36 and PYY_3–36 are important regulators of gastrointestinal functions and together with another L cell hormone, glucagon-like peptide 1 (GLP-1), are key mediators of the ileal and colonic brake, negative feedback mechanisms that inhibit the motility of the more proximal section of the gastrointestinal tract (20). PYY also inhibits secretions in the stomach and exocrine pancreas while increasing fluid and electrolyte absorption throughout the gastrointestinal tract (21). However, the primary role of PYY is its actions as a potent anorectic hormone, in which PYY_3–36 induces satiety by targeting the appetite-regulating system of the hypothalamus. PYY_3–36 is highly selective for Y2 receptors that are found on neuropeptide Y/agouti-related protein neurons in the arcuate nucleus, which suppresses the release of the orexigenic neuropeptides neuropeptide Y and agouti-related protein. This subsequently disinhibits the release of the anorexigenic α-MSH from neighboring proopiomelanocortin neurons to reduce food intake (22, 23). Postprandial levels of PYY are significantly elevated in patients after bariatric surgery and have been proposed as one of the mechanisms underlying the dramatic weight loss achieved by the procedure (24).

An early clinical study showed that 3 days of metformin treatment in healthy normal-weight females was sufficient to significantly increase fasting PYY level and that 6 months of metformin treatment in women who were overweight and with polycystic ovarian syndrome elevated fasting PYY levels by 50% in more than half of the participants (25). Later studies confirmed that chronic metformin treatment is associated with elevated fasting and postprandial levels of PYY (3, 26). Intriguingly, delayed-release metformin, which is not released until it reaches the L cell–rich distal small intestine and colon, caused exaggerated postprandial PYY response in individuals with type 2 diabetes (4). This leads us to speculate whether metformin alone is capable of triggering acute PYY secretion from the intestinal mucosa, independent of neural inputs or other factors. Using an ex vivo model prepared from human intestinal mucosa, we showed that metformin causes acute PYY secretion from both ileal and colonic mucosa tissue in the absence of nutrients and that these actions require internalization of the drug by the plasma membrane monoamine transporter (PMAT) and the serotonin reuptake transporter (SERT) and the subsequent activation of AMP kinase (AMPK). This evidence supports direct actions of metformin to trigger PYY release from human L cells, which may, in part, explain the weight-loss effect of metformin.

Methods

Human tissue collection

Methods regarding human tissue collection and preparation for secretion experiments have been reported in details elsewhere (27). Briefly, ileal and colonic specimens were obtained from consented patients undergoing bowel resection for cancer or stoma reversal. Specimens used in this current study were from the same cohort of specimen donors used for a previous study (27). Clinical characteristics of specimen donors are given in Table 1. The specimens were immediately placed in ice-cold Krebs buffer (in mM, NaCl 138, KCl 4.5, CaCl₂ 2.6, NaHCO₃ 4.2, MgCl₂ 1.2, NaH₂PO₄ 1.2, HEPES 10, Glucose 5) and transported to the laboratory within 15 minutes. Mucosa was dissected from the specimen, cut into pieces, and individually weighed before they were used for secretion assay.

Table 1.

Characteristics of Tissue Donors

	Ileum Specimen	Colon Specimen
N	9	46
Age, y	72.1 ± 8.7 (57–83)	66.6 ± 12.6 (38–87)
Sex (male/female)	3/6	24/22
BMI, kg/m2	31.3 ± 6.3 (25–45)	29.3 ± 6.7 (19–55)
History of type 2 diabetes (yes/no)	3/6	11/35
Metformin treated (yes/no)	2/7	6/40

Data are means ± SD with ranges in parentheses.

Secretion experiments

Mucosal tissue pieces were incubated with 250 μL of buffer (control) or buffer containing test agents in a 96-well plate for 15 minutes and secretion experiments undertaken as previously described (27, 28). Following incubation at 37°C in 95%O₂/5%CO₂, supernatants were collected and stored in aliquots at −20°C. Total PYY levels were quantitated using a commercially available ELISA kit, according to manufacturer’s instructions (EZHPYYT66K, Merck Millipore, Danvers, MA).

Test agents

3-Isobutyl-1-methylxanthine (IBMX) and forskolin (FSK; I5879 and F6886, Sigma-Aldrich) (10 μM each) and 70 mM KCl were used as positive control stimuli. For the 70 mM KCl solution, an equimolar amount of NaCl was removed to maintain well osmolarity. The following compounds were purchased from Sigma-Aldrich: metformin (PHR1084), lopinavir (SML1222), quinine hydrochloride dihydrate (Q1125), and fluoxetine (F132). The AMPK inhibitor dorsomorphin was purchased from Merck Millipore (171260). Only samples that responded positively to at least one control stimulus (70 mM KCl or 10 μM IBMX/FSK) were included in analysis. Four colon samples did not respond to positive controls and were thus excluded from analysis.

Statistical analysis

All statistical analyses were conducted as paired analyses, comparing responses in tissues obtained from the same individual. A paired Student t test was used for single comparisons and a paired one-way ANOVA with a Dunn post hoc test was used for multiple comparisons. Statistical significance was P < 0.05. All data are shown as means ± SEM except in Table 1, where SDs of our clinical data are provided.

Results

PYY secretions from gut epithelial tissue upon exposure to different stimuli are reported in Table 2. Exposure to high external K⁺ or to a combination of known activators of L cell secretion, FSK and IBMX, increased PYY release from colonic epithelial tissue by 2.1-fold (n = 26, P < 0.001) and 1.8-fold, respectively (n = 46, P < 0.001; Fig. 1A). Acute metformin exposure [10 µM, concentration based on therapeutic levels (29)] triggered PYY release in colonic (n = 46, P < 0.001, number of responders, 39; Fig. 1B) and ileal (n = 9, P < 0.05, number of responders, 7; Fig. 1C) tissue by 1.7- and 2.0-fold, respectively. Thus, metformin increases PYY release within 15 minutes from human colonic and ileal L cells.

Table 2.

PYY Secretion From Gut Epithelia Upon Stimulation

PYY Secretion From Gut Epithelia Upon Stimulation	PYY Secretion (ng/mL/g Tissue)
High potassium (n = 26)
Control	35.53 ± 4.37
70 mM K+	63.5 ± 7.3a
IBMX/FSK (n = 46)
Control	33.82 ± 3.36
10 μM each IBMX/FSK	62.9 ± 5.26a
Colon tissue (n = 46)
Control	33.77 ± 3.29
10 μM metformin	51.41 ± 5.52a
Ileum tissue (n = 9)
Control	5.39 ± 1.02
10 μM metformin	9.85 ± 2.23b
Metformin dose response (n = 23)
Control	35.02 ± 4.21
10 μM metformin	48.61 ± 6.42c
1 mM metformin	47.71 ± 7.73b
100 mM metformin	57.59 ± 8.63c
AMPK inhibition (n = 20)
Control	34.21 ± 4.35
10 μM metformin	47.74 ± 6.91a
10 μM metformin plus dorsomorphin	39.18 ± 5.76
Transporter blockade (n = 18)
Control	34.53 ± 5.76
10 μM metformin	49.25 ± 7.61b
10 μM metformin plus quinine	60.5 ± 13.04b
10 μM metformin plus lopinavir	40.03 ± 7.16
10 μM metformin plus fluoxetine	35.47 ± 4.72

^a

P < 0.001 compared with control.

^b

P < 0.05 compared with control.

^c

P < 0.01 compared with control.

Figure 1.

Metformin-induced secretion in human L cells. (A) Colonic epithelial preparations readily secrete PYY in response to high (70 mM) external K⁺ (n = 26) or to a combination of IBMX and FSK (n = 46). (B and C) Metformin (10 µM) increases PYY release after 15 min in epithelial tissue from human (B) colon (n = 46) and (C) ileum (n = 9). Data are means ± SEM. *P < 0.05, ***P < 0.001 compared with respective control groups.

To identify whether the effect of metformin on PYY secretion was altered in human obesity or type 2 diabetes, we examined responses to metformin (10 µM) in our colonic preparation across body mass index (BMI) and in samples obtained from patients with type 2 diabetes. Neither the basal release of PYY (Fig. 2A) nor the degree of metformin-stimulated PYY release (Fig. 2B) correlated with BMI (n = 46). BMI of patients with type 2 diabetes is significantly higher than that of nondiabetic individuals (nondiabetic vs type 2 diabetes: 27.4 ± 0.84 vs 33.9 ± 2.57 kg/m², P < 0.01), but no difference was seen in either basal (Fig. 2C) or stimulated PYY release (Fig. 2D) between tissue obtained from nondiabetic individuals (n = 35) or individuals with type 2 diabetes (n = 11). Thus, basal PYY secretion from colonic L cells and their responses to metformin do not change across BMI and are unrelated to diabetic status. Metformin significantly triggered PYY secretion from samples obtained from metformin-treated donors, and the magnitude of the response is not significantly different from that of metformin-naive donors (fold increase of metformin-induced PYY secretion: naive vs treated, 1.7 ± 0.29 vs 1.6 ± 0.11).

Figure 2.

Metformin-induced colonic L cell secretion does not change in obesity or diabetes. (A and B) No relationship exists between BMI and (A) basal and (B) stimulated PYY release (n = 46). (C) Basal release and (D) stimulated PYY release are similar in tissues from nondiabetic subjects (ND, n = 35) and subjects with type 2 diabetes (T2D, n = 11). Data are means ± SEM.

We then investigated the mechanism by which metformin triggers L cell secretion. Increasing concentrations of metformin up to 100 mM caused significant PYY secretion (n = 23, P < 0.01; Fig. 3A). AMPK has been associated with metformin action (30), and inhibiting AMPK activity by dorsomorphin (10 μM) blocked metformin-induced PYY secretion (n = 20; Fig. 3B). We then used a series of membrane transporter antagonists to identify the mechanism of metformin internalization in human colonic L cells. Quinine [organic cation transporter 1 (OCT1) inhibitor] had no effect on metformin-induced PYY release, whereas lopinavir (PMAT inhibitor) and fluoxetine (SERT inhibitor) both blocked metformin-induced PYY release (n = 18; Fig. 3C).

Figure 3.

Mechanisms controlling metformin-induced secretion in human colonic L cells. (A) Metformin concentrations (10 µM, 1 mM, and 100 mM) all increased PYY release significantly (n = 23). (B) Metformin-induced PYY release was blocked by the AMPK inhibitor dorsomorphin (n = 20). (C) Metformin-induced PYY release was blocked by the PMAT inhibitor lopinavir and the SERT inhibitor fluoxetine (n = 18). Data are means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 compared with respective control groups.

Discussion

This is a report of acute metformin exposure directly triggering PYY release within human intestinal epithelium, as well as mechanisms by which it occurs. Others have previously shown that chronic oral metformin treatment increases fasting and postprandial PYY levels in humans (3, 4, 25, 26) but could not exclude potential effects of metformin administration on the autonomic and enteric nervous systems, or on the reabsorption and metabolism of bile acids, which are potent stimulants for PYY secretion (31, 32). We also showed that this effect was not altered in obesity or in patients with type 2 diabetes. Our results, therefore, further support the use of oral metformin in patients in which weight loss or minimal weight gain is desired.

The cellular actions of metformin are mediated by activation of AMPK at sites including the liver (30), skeletal muscles (33), and adipose tissue (34). We therefore hypothesized that metformin’s action on L cells to stimulate PYY secretion also required AMPK activation. We suspected that cellular internalization of metformin was also required, and based on current knowledge of membrane transporters involved in intestinal uptake of metformin (35), we tested whether OCT1, PMAT, and SERT were involved in this process. In line with our hypothesis, we observed significant attenuation of metformin-stimulated PYY release when AMPK, PMAT, and SERT were pharmacologically inhibited. However, despite the link between OCT1 polymorphisms and metformin responses (36, 37), we did not observe any changes to metformin-induced PYY secretion in the presence of the OCT1 inhibitor quinine, suggesting that OCT1 polymorphisms may affect metformin responses through alternate, indirect mechanisms.

Although multiple actions of metformin on metabolic tissues are likely to contribute to its weight loss effects, our results support the notion that some of the effects are partly mediated by augmented release of PYY. Such PYY secretion could then enter the circulation to act centrally on the hypothalamic feeding circuit to induce satiety and reduce food intake. Although this hypothesis has not yet been directly tested, Kim et al. (16) showed that oral metformin administration significantly increased c-Fos immunoreactivity within the nucleus tractus solitarius of obese mice, which is an important target for peripherally administered PYY_3–36 to reduce food intake (38). It would be informative to investigate whether metformin-induced weight loss and reduced food intake were attenuated in Pyy knockout mice. We recently reported that acute metformin exposure also significantly triggered the secretion of the incretin GLP-1 in an AMPK-dependent manner (27). Therefore, it is possible that PYY and GLP-1, which are often colocalized in and cosecreted by L cells (39–41), mediate some of the beneficial weight-loss effects of metformin in a synergistic fashion (42). However, in contrast to PYY, metformin-induced GLP-1 release from the human colon was not sensitive to PMAT blockade (27), which suggests that the mechanisms underlying metformin internalization are not identical between GLP-1– and PYY-secreting cells. Although the co-release of the two hormones is often observed (39), there are instances where the secretion of the two hormones is discordant (43), which could be attributed to differential effects on subpopulations of L cells that synthesize only either GLP-1 or PYY (40, 41). Also note that metformin attenuates the release of the orexigenic gut hormone ghrelin from primary rat gastric cell cultures in an AMPK-dependent fashion (44). These synergistic actions of metformin on anorexigenic PYY and orexigenic ghrelin may act to further reduce food intake in clinical settings. Increased food intake is one of the major reasons why many patients gain significant body weight as a result of atypical antipsychotic treatments (45, 46). Metformin’s stimulatory effect on PYY secretion may explain why it is an efficacious option in limiting atypical antipsychotic-induced weight gain (9–11). However, note that although increased postprandial PYY levels are thought to contribute to the weight loss effects of metabolic surgeries (24), and although acute PYY infusion significantly reduced food intake in healthy and obese volunteers (22, 42, 47), direct weight-loss effects of prolonged PYY treatment in humans remains unexplored, likely due to its nauseating effects at supraphysiological levels (48). One limitation to our experimental setup is the inability to exclude the possibility that metformin could act on absorptive enterocytes or other enteroendocrine cells to trigger L cell secretion in a paracrine fashion. Although clinical studies have shown that metformin treatment of 7 and 10 days significantly increased fasting PYY levels in treated subject (25, 26), we did not study the effect of metformin exposure on PYY secretion beyond 15 minutes, and thus it remains to be determined whether metformin treatment of longer duration than 15 minutes can also increase PYY biosynthesis or upregulate PYY expression. Owing to the ad hoc nature of the availability of surgical specimens, it is difficult to control and adjust for all clinical characteristics of the specimen donors, which may partially explain the marked interpatient variation in their basal PYY secretion and metformin response.

In summary, we have shown that metformin is a secretagogue of the anorectic gut hormone PYY in human intestinal tissue and shown that metformin triggers PYY secretion in the absence of nutrient stimulation and independent of any neural inputs or luminal factors such as bile acids or the gut microbiota. Such augmented release of PYY has the potential to be of substantial metabolic benefit, as it is implicated as a driver of the marked weight loss achieved by bariatric surgery (24).

Abbreviations:

Abbreviations:
 
• AMPK

  AMP kinase

 
• BMI

  body mass index

 
• FSK

  forskolin

 
• GLP-1

  glucagon-like peptide 1

 
• IBMX

  3-isobutyl-1-methylxanthine

 
• OCT1

  organic cation transporter 1

 
• PMAT

  plasma membrane monoamine transporter

 
• PYY

  peptide YY

 
• SERT

  serotonin reuptake transporter

Acknowledgments

Financial Support: This work was supported by Australian Research Council Grant LP150100419 (to D.J.K.) and National Health and Medical Research Council Grant APP1088737 (to D.J.K.).

Disclosure Summary: The authors have nothing to disclose.