Jonas B. Jensen,1,2 Elias I. Sundelin,1 Steen Jakobsen,2 Lars C. Gormsen,2

Ole L. Munk,2 Jørgen Frøkiær,2,3 and Niels Jessen1,4

[11C]-Labeled Metformin Distributionin the Liver and Small IntestineUsing Dynamic Positron EmissionTomography in Mice DemonstratesTissue-Specific TransporterDependencyDiabetes 2016;65:1724–1730 | DOI: 10.2337/db16-0032

Metformin is the most commonly prescribed oral anti-diabetic drug, with well-documented beneficial pre-ventive effects on diabetic complications. Despitebeing in clinical use for almost 60 years, the underly-ing mechanisms for metformin action remain elusive.Organic cation transporters (OCT), including multi-drug and toxin extrusion proteins (MATE), are essen-tial for transport of metformin across membranes, buttissue-specific activity of these transporters in vivo isincompletely understood. Here, we use dynamic pos-itron emission tomography with [11C]-labeled metfor-min ([11C]-metformin) in mice to investigate the role ofOCT and MATE in a well-established target tissue, theliver, and a putative target of metformin, the small in-testine. Ablation of OCT1 and OCT2 significantly re-duced the distribution of metformin in the liver andsmall intestine. In contrast, inhibition of MATE1 withpyrimethamine caused accumulation of metformin inthe liver but did not affect distribution in the small in-testine. The demonstration of OCT-mediated transportinto the small intestine provides evidence of direct ef-fects of metformin in this tissue. OCT and MATE haveimportant but separate roles in uptake and eliminationof metformin in the liver, but this is not due to changesin biliary secretion. [11C]-Metformin holds great poten-tial as a tool to determine the pharmacokinetic proper-ties of metformin in clinical studies.

Metformin is the preferred first-line drug in the treatment oftype 2 diabetes because of its beneficial effects on cardio-vascular outcomes and impressive safety profile. Metforminaction involves suppression of mitochondrial functionthrough inhibition of complex 1 in the respiratory chain(1). In the liver, metformin inhibits gluconeogenesis in micevia a decrease in energy state (2). In addition, metforminaction in the duodenum lowers hepatic glucose production inrats, and the intestine may be an important target tissue formetformin (3). A common denominator for these suggestedmechanisms of action is intracellular uptake of the drug.Therefore, understanding the biodistribution of metforminis essential to understanding its tissue-specific effects.

Metformin is hydrophilic and cannot pass cell mem-branes by passive diffusion; therefore, cellular uptakedepends on organic cation transporters (OCT). The pri-mary mediators of intestinal uptake are plasma mem-brane monoamine transporter (4) and OCT1 (5). Hepaticuptake depends on OCT1 and possibly OCT3 (6,7).Thus, knockout of OCT1 in mice reduces hepatic uptakeof metformin determined ex vivo with [14C]-labeled met-formin or high-performance liquid chromatography (8,9).In humans, reduced function alleles in SLC22A1, encodingOCT1, are associated with higher plasma levels of metfor-min (10). However, whether this translates into changesin hepatic uptake of metformin is unknown. Multidrug

1Research Laboratory for Biochemical Pathology, Department of Clinical Medi-cine, Aarhus University, Aarhus, Denmark2Department of Nuclear Medicine and PET Center, Aarhus University Hospital,Aarhus, Denmark3Department of Clinical Medicine, Aarhus University, Aarhus, Denmark4Department of Clinical Pharmacology, Aarhus University Hospital, Aarhus,Denmark

Corresponding author: Niels Jessen, niels.jessen@clin.au.dk.

Received 7 January 2016 and accepted 9 March 2016.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db16-0032/-/DC1.

© 2016 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

1724 Diabetes Volume 65, June 2016

PHARMACOLOGYAND

THERAPEUTIC

S

and toxin extrusion proteins (MATE) 1 eliminates met-formin from hepatocytes in a H+-coupled electroneutralmanner (11), and ablation of MATE1 in mice increaseshepatic metformin concentration, determined ex vivowith high-performance liquid chromatography (12). Be-cause of negligible metabolism of metformin in vivo,systemic elimination depends on renal excretion in theproximal tubules. OCT2, and to a minor degree OCT1, isresponsible for basolateral uptake, whereas MATE1 andMATE2-K contribute to luminal excretion into the urine(13).

Until today, studies of tissue-specific uptake and elim-ination of metformin have been limited to ex vivo methods.Successful generation of [11C]-labeled metformin ([11C]-metformin) was recently demonstrated to be a powerfultool to noninvasively determine distribution of metfor-min in specific tissues (14–16). Therefore, the primaryaim of the present study was to use dynamic positronemission tomography (PET) with [11C]-metformin in miceto investigate the importance of OCT and MATE in dy-namic hepatic and small intestinal distribution of met-formin in vivo. We hypothesized that expression andfunctional impairment of these transporters affects dis-tribution of metformin in target tissues.

RESEARCH DESIGN AND METHODS

Radiochemistry[11C]-Metformin was synthesized as previously published(16). The [11C]-metformin (0.2–1.0 GBq) contained 0.1–0.5mg/mL metformin and was .98% pure.

AnimalsFemale FVB OCT1/22/2 and corresponding wild-typemice (aged 13–15 weeks) were purchased from Taconic.Female FVB mice (aged 14–17 weeks) were used for drug-drug interactions and biodistribution studies. Animals werefed standard chow and kept in a temperature- and humidity-controlled environment with a 12-h light/dark cycle. Thestudies were performed in accordance with the Danish An-imal Experimentation Act and approved by the Animal Ex-periments Inspectorate, Denmark.

MicroPET StudyWild-type (n = 7) and OCT1/22/2 mice (n = 4) underwentfunctional PET and anatomical MRI using Mediso nanoScanPET/MR (Mediso Medical Imaging Systems, Budapest,Hungary). After induction of anesthesia with 5% isoflur-ane, the animal was placed in an acrylic glass head-holder,and anesthesia was maintained with mask-deliveredisoflurane (1.8–2.0%). A bolus of [11C]-metformin (5.7 62.8 MBq/animal) was injected via a tail vein catheter,followed by 60-min dynamic PET and 30-min MRI. Bodytemperature and respiration frequency were monitored.

To pharmacologically inhibit OCT, mice were intrave-nously pretreated 5 min before the [11C]-metformin injec-tion with saline containing cimetidine (150 mg/kg) (n = 5)dissolved in 20% DMSO or pyrimethamine (5 mg/kg) (n = 4)in 40% DMSO. Control mice (n = 8) were intravenously

pretreated with corresponding vehicle in equal volume.Chemicals were from Sigma-Aldrich and used as received.

PET Image AnalysisDynamic PET data were reconstructed with a three-dimensional ordered subset expectation maximizationalgorithm (Tera-Tomo 3D; Mediso Medical Imaging Sys-tems) with four iterations and six subsets and voxel size of0.4 3 0.4 3 0.4 mm3. Data were corrected for dead-time,decay, and randoms using delayed coincidence windowwithout corrections for attenuation and scatter. The 60-mindynamic PET scans were reconstructed as 30 frames in-creasing in duration from 5 s to 10 min. Multiple regionsof interest were placed on coronal slices in the organ ofinterest using PMOD 3.5 software (PMOD TechnologiesLtd., Zurich, Switzerland) creating a volume of interest(VOI). Image-derived arterial input function was generatedby averaging images from the first 20 s and placing a circlewith a diameter of 15 pixels on the six most intensive slicesover the heart (68 mL), representing primarily the bloodpool in the left ventricle. Hepatic VOIs were drawn in theanterior part of the liver on PET images averaged from 0 to15 min in which the liver easily is identified (Fig. 1). Smallintestinal VOIs were localized on MRIs and drawn on PETimages identical to VOIs for the liver (Fig. 1). Half moon–shaped VOIs were placed in the kidney cortex on all PETimages averaged. Positioning of all VOIs was controlled ineach time frame. Time-activity curves were generated fromthe individual VOIs.

Pharmacokinetic AnalysisData are expressed as the tissue-to-blood ratio at eachtime point. The area under the curve (AUC) of the tissue-to-blood ratio reflects the tissue extraction ratio (17) andrepresents the relationship between uptake and elimina-tion from the tissue of interest.

During the first minutes after PET tracer injection, theefflux in the liver can be assumed to be much smaller thanthe influx (17). Tracer supply through the portal veinwithin this time is negligible. Consequently, we used anirreversible single-compartment model to calculate the ini-tial rate of hepatic [11C]-metformin uptake, influx rateconstant (mL blood/mL tissue/min) (18), from 0 to 60 susing the image-derived arterial input function.

Biodistribution StudyAnesthetized mice were administered [11C]-metformin(6.7 6 2.6 MBq/body weight) through the tail vein. At15 and 60 min after injection, the gallbladder, small in-testinal wall, gastric wall, liver, and blood were harvestedfrom five mice per time point. Tissue radioactivity was de-termined using a well-crystal scintillation detector (PackardBioScience) and expressed relative to blood radioactivity.All measurements were corrected for decay.

Statistical AnalysisData are expressed as mean 6 SE. Distribution was testedusing the Shapiro-Wilk normality test. Normally distributeddata were compared using the Student t test or one-way

diabetes.diabetesjournals.org Jensen and Associates 1725

ANOVA. The Mann-Whitney rank sum test was used fordata with unequal variance. Significance was assumed atP , 0.05. SigmaPlot 11.0 software (Systat Software) wasused for all analyses.

RESULTS

During scannings, most of the metformin was detectedin the urinary bladder, kidneys, and small intestine. Whencoregistering PET with anatomical MRIs, hepatic andsmall intestinal distributions were well visualized from0 to 15 min and primarily the small intestine from 15 to60 min (Fig. 1A and B). We did not observe spillover fromthe kidneys to the intestine and liver. Time-activity curvesfrom blood, liver, small intestine, and kidney during the60 min are shown in Supplementary Figs. 1–4.

Nonspecific Inhibition of OCT Concomitantly InhibitsHepatic Uptake and Elimination of MetforminPretreatment with cimetidine, a nonspecific OCT inhibi-tor (19), lowered the initial uptake of metformin in theliver, reflected as a significant reduction in influx rateconstant (Fig. 2B). This was associated with an increasedhepatic AUC for metformin from 2 to 60 min, reflectingpronounced inhibition of hepatic elimination.

Ablation of OCT1 and -2 Impairs Hepatic Distributionof MetforminDynamic distribution of metformin in the liver of OCT1/22/2 mice revealed stable influx rates during the first minuteafter tracer injection, with no significant differences betweengroups (Fig. 2C and D). From 2 to 60 min, hepatic distribu-tion was severely lowered in OCT1/22/2 mice (Fig. 2D).

Inhibition of MATE1 Reduces Hepatic Eliminationof MetforminPretreatment with pyrimethamine, in doses that specif-ically inhibit MATE1 (14), caused hepatic accumulation ofmetformin (Fig. 2E). The influx rate constant did not vary

significantly between groups, whereas hepatic distributionfrom 2 to 60 min was significantly increased by pyrimeth-amine pretreatment (Fig. 2F).

Nonspecific OCT Inhibition Impairs Uptake ofMetformin in the Small IntestineDistribution of metformin in the small intestine wascalculated as AUC of the small intestine–to–blood ratiofrom 0 to 60 min. Pretreatment with cimetidine signifi-cantly reduced uptake of metformin (Fig. 3A) and theAUC for metformin (Fig. 3B).

Inhibition of OCT1/2 but Not MATE1 Lowers Uptake ofMetformin in the Small Intestine in a Cimetidine-LikeMannerAs shown in Fig. 3C and D, uptake of metformin in thesmall intestine was severely reduced in OCT1/22/2 mice,resulting in significant reduction in AUC. Pretreatmentwith pyrimethamine was not associated with statisticallysignificant effects (Fig. 3F).

Metformin Is Not Eliminated by Biliary Excretion[11C]-Metformin activity in the small intestinal wall was sig-nificantly higher than in the gallbladder at 15 and 60 minafter tracer administration (Fig. 4). Hepatic [11C]-metforminactivity was significantly higher than gallbladder activity15 min after injection. In fact, [11C]-metformin activity ofthe gallbladder resembled the gastric wall, where only minor[11C]-metformin activity was observed on PET images.

DISCUSSION

The current study demonstrates the importance of OCTand MATE in the dynamic distribution of metformin inthe liver and small intestine using PET imaging with[11C]-metformin in vivo. Absence or inhibition of OCT1and -2 reduce distribution of metformin to the liver andsmall intestine, whereas MATE1 inhibition impairs elim-ination of metformin from the liver.

Figure 1—Whole-body distribution of [11C]-metformin in mouse. Coronal whole-body PET coregistered with T1-weighted MRI sequence ina wild-type mouse from 0 to 15 min (A) and from 15 to 60 min (B). The projection is anterior to the kidneys. C: Regions of interest (ROI) in thesmall intestine and liver. PET images were averaged from 0 to 15 min for defining multiple ROI on coronal slices in liver and small intestine.Scale bar to the left represents standard uptake value 0–4.

1726 Metformin in Liver and Small Intestine Diabetes Volume 65, June 2016

Figure 2—Hepatic metformin distribution. Time course of [11C]-metformin distribution in the liver of cimetidine-pretreated (Cim) (A), OCT1/22/2 (C), and pyrimethamine (Pyr)-pretreated mice (E). Data are expressed as liver-to-blood ratio by dividing the hepatic concentration of[11C]-metformin by the blood concentration at each time point for each animal. B, D, and F: Influx rate constant and AUC from 2 to 60 min ofrespective liver-to-blood ratios of [11C]-metformin. Data represent the mean 1 SE. Error bars that are not visible are contained within thesymbols. *P < 0.05 and **P < 0.001.

diabetes.diabetesjournals.org Jensen and Associates 1727

Our data support an emerging role of the intestines asa target for metformin action. In rodents, the glucose-lowering effect of metformin, when administrated orally,is superior to that in intravenous and portal administration

(20), and metformin action in the duodenum lowers hepaticglucose production in rats (3). We observed a rapid andsignificant uptake of metformin in the small intestineafter intravenous administration that was dependent on

Figure 3—Distribution of metformin in the small intestine. Time course of [11C]-metformin distribution in the small intestine of cimetidine-pretreated (Cim) (A), OCT1/22/2 (C), and pyrimethamine (Pyr)-pretreated mice (E). Data are expressed as small intestine-to-blood ratio bydividing small intestinal concentration of [11C]-metformin by the blood concentration at each time point for each animal. B, D, and F: AUCfrom 0 to 60 min of respective small intestine-to-blood ratios of [11C]-metformin. Data represent the mean 1 SE. Error bars that are notvisible are contained within the symbol. *P < 0.05 and **P < 0.001.

1728 Metformin in Liver and Small Intestine Diabetes Volume 65, June 2016

expression of OCT1 and -2. In contrast, MATE1 inhibitiondid not cause major differences in metformin uptake inthe small intestines after intravenous administration, butwhether MATE1 is involved in metformin uptake afteroral administration remains to be determined. Biliarysecretion of metformin has been suggested as a majorroute of elimination (17,21). This could form the basisfor enterohepatic cycling and thereby explain metforminin the small intestines after intravenous administration.However, we observed only negligible [11C]-metformincontent in the gallbladder 15 min after tracer administra-tion, but a significant uptake in the small intestine wasdetected at the same time. Biliary elimination is thereforeunlikely to be a major contributor for intestinal uptakeafter intravenous administration. Instead, the metforminuptake in the small intestine under these conditions mayreflect basolateral transport capacity.

Hepatic distribution of metformin was severely low-ered in OCT1/22/2 mice, and this is in accordance withprevious findings (22). Interestingly, the initial influx ratewas normal in OCT1/22/2 mice but was significantlylower after pretreatment with cimetidine. This could in-dicate additional transporter capacity in OCT1/22/2 mice.Recent data show decreased antiglycemic effects of met-formin in OCT32/2 mice (23), and our data could indicatea minor hepatic OCT3-mediated uptake of metformin.Inhibition of MATE1 with pyrimethamine did not affecthepatic uptake of metformin, consistent with a recent studyby Shingaki et al. (15). To further test the role of MATE1,we pretreated two OCT1/22/2 mice with pyrimethamine

and determined metformin distribution. Although thenumber of animals was insufficient to draw firm conclu-sions, we did not observe further reduction in hepatic up-take under these conditions, indicating an insignificant roleof MATE1 in hepatic uptake of metformin. Instead, inhi-bition of MATE1 profoundly affected hepatic elimina-tion of metformin, which supports previous reports frompyrimethamine-treated mice (24). These prominent effectsopen the possibility to potentiate metformin action bycotreatment with MATE1 inhibitors.

The tracer doses of metformin used in these experi-ments are far below therapeutic metformin concentrations.Thus, high-affinity, low-capacity transporter proteins couldtheoretically affect distribution of [11C]-metformin withoutaffecting metformin at therapeutic levels. However, con-comitant treatment with metformin in therapeutic dosesdoes not affect [11C]-metformin distribution in pigs (16).Consequently, no data suggest that distribution of [11C]-metformin disassociates from therapeutic use of metformin.

In conclusion, dynamic tissue-specific distribution ofmetformin can be determined in vivo by [11C]-metforminfunctional PET imaging. The present data demonstrate thatOCT1/2 are important for normal distribution of metfor-min in the liver and small intestine, whereas MATE1 isnecessary for hepatic elimination. Furthermore, MATE1eliminates hepatic metformin primarily to the systemic cir-culation. [11C]-Metformin holds great potential to deter-mine unsolved pharmacokinetics properties of metforminin clinical PET studies.

Funding. This work was supported by Danish Council for IndependentResearch (Det Frie Forskningsråd) (grant DFF–4183-00384) and a Novo NordiskFoundation Excellence Project grant to N.J.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. J.B.J., E.I.S., S.J., and N.J. participated indesigning the study. J.B.J., E.I.S., S.J., and O.L.M. conducted experiments andanalyzed the data. J.B.J., E.I.S., S.J., L.C.G., O.L.M., J.F., and N.J. contributed towriting the manuscript. N.J. is the guarantor of this work and, as such, had fullaccess to all the data in the study and takes responsibility for the integrity of thedata and the accuracy of the data analysis.

References1. Owen MR, Doran E, Halestrap AP. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratorychain. Biochem J 2000;348:607–6142. Foretz M, Hébrard S, Leclerc J, et al. Metformin inhibits hepatic gluco-neogenesis in mice independently of the LKB1/AMPK pathway via a decrease inhepatic energy state. J Clin Invest 2010;120:2355–23693. Duca FA, Côté CD, Rasmussen BA, et al. Metformin activates a duodenalAmpk-dependent pathway to lower hepatic glucose production in rats. Nat Med2015;21:506–5114. Zhou M, Xia L, Wang J. Metformin transport by a newly cloned proton-stimulated organic cation transporter (plasma membrane monoaminetransporter) expressed in human intestine. Drug Metab Dispos 2007;35:1956–19625. Müller J, Lips KS, Metzner L, Neubert RH, Koepsell H, Brandsch M. Drugspecificity and intestinal membrane localization of human organic cation trans-porters (OCT). Biochem Pharmacol 2005;70:1851–1860

Figure 4—Biodistribution of [11C]-metformin. Mice were intrave-nously administered [11C]-metformin and killed 15 (n = 5) or 60 min(n = 5) after injection. Tissues were harvested, and radioactivityconcentrations were measured and are expressed in relation toblood radioactivity concentrations. The data represent mean 1SE. Error bars that are not visible are contained within the symbol.*P < 0.05.

diabetes.diabetesjournals.org Jensen and Associates 1729

6. Schweifer N, Barlow DP. The Lx1 gene maps to mouse chromosome 17 andcodes for a protein that is homologous to glucose and polyspecific trans-membrane transporters. Mamm Genome 1996;7:735–7407. Green RM, Lo K, Sterritt C, Beier DR. Cloning and functional expression of amouse liver organic cation transporter. Hepatology 1999;29:1556–15628. Shu Y, Sheardown SA, Brown C, et al. Effect of genetic variation in theorganic cation transporter 1 (OCT1) on metformin action. J Clin Invest 2007;117:1422–14319. Wang DS, Jonker JW, Kato Y, Kusuhara H, Schinkel AH, Sugiyama Y.Involvement of organic cation transporter 1 in hepatic and intestinal distributionof metformin. J Pharmacol Exp Ther 2002;302:510–51510. Shu Y, Brown C, Castro RA, et al. Effect of genetic variation in the organiccation transporter 1, OCT1, on metformin pharmacokinetics. Clin Pharmacol Ther2008;83:273–28011. Hiasa M, Matsumoto T, Komatsu T, Moriyama Y. Wide variety of locationsfor rodent MATE1, a transporter protein that mediates the final excretion step fortoxic organic cations. Am J Physiol Cell Physiol 2006;291:C678–C68612. Tsuda M, Terada T, Mizuno T, Katsura T, Shimakura J, Inui K. Targeteddisruption of the multidrug and toxin extrusion 1 (mate1) gene in mice reducesrenal secretion of metformin. Mol Pharmacol 2009;75:1280–128613. Gong L, Goswami S, Giacomini KM, Altman RB, Klein TE. Metforminpathways: pharmacokinetics and pharmacodynamics. Pharmacogenet Genomics2012;22:820–82714. Hume WE, Shingaki T, Takashima T, et al. The synthesis and biodistributionof [(11)C]metformin as a PET probe to study hepatobiliary transport mediated bythe multi-drug and toxin extrusion transporter 1 (MATE1) in vivo. Bioorg MedChem 2013;21:7584–759015. Shingaki T, Hume WE, Takashima T, et al. Quantitative evaluation ofmMate1 function based on minimally invasive measurement of tissue concen-tration using PET with [(11)C]metformin in mouse. Pharm Res 2015;32:2538–2547

16. Jakobsen S, Busk M, Jensen JB, et al. A PET tracer for renal organic cationtransporters, 11C-metformin: radiosynthesis and preclinical proof-of-conceptstudies. J Nucl Med 2016;57:615–62117. Testa A, Zanda M, Elmore CS, Sharma P. PET tracers to study clinicallyrelevant hepatic transporters. Mol Pharm 2015;12:2203–221618. Maguire RP, Leenders KL. Cerebral blood flow–single-tissue-compartmentmodel. In PET Pharmacokinetic Course Manual. Maguire RP, Leenders KL, Eds.University of Groningen, Groningen, The Netherlands and McGill University,Canada 2003, Chapter 4, p. 29–3219. Ito S, Kusuhara H, Yokochi M, et al. Competitive inhibition of the luminalefflux by multidrug and toxin extrusions, but not basolateral uptake by organiccation transporter 2, is the likely mechanism underlying the pharmacokineticdrug-drug interactions caused by cimetidine in the kidney. J Pharmacol Exp Ther2012;340:393–40320. Stepensky D, Friedman M, Raz I, Hoffman A. Pharmacokinetic-pharmacodynamicanalysis of the glucose-lowering effect of metformin in diabetic rats re-veals first-pass pharmacodynamic effect. Drug Metab Dispos 2002;30:861–86821. Otsuka M, Matsumoto T, Morimoto R, Arioka S, Omote H, Moriyama Y. Ahuman transporter protein that mediates the final excretion step for toxic organiccations. Proc Natl Acad Sci U S A 2005;102:17923–1792822. Higgins JW, Bedwell DW, Zamek-Gliszczynski MJ. Ablation of both organiccation transporter (OCT)1 and OCT2 alters metformin pharmacokinetics but hasno effect on tissue drug exposure and pharmacodynamics. Drug Metab Dispos2012;40:1170–117723. Chen EC, Liang X, Yee SW, et al. Targeted disruption of organic cationtransporter 3 attenuates the pharmacologic response to metformin. Mol Phar-macol 2015;88:75–8324. Ito S, Kusuhara H, Kuroiwa Y, et al. Potent and specific inhibition ofmMate1-mediated efflux of type I organic cations in the liver and kidney bypyrimethamine. J Pharmacol Exp Ther 2010;333:341–350

1730 Metformin in Liver and Small Intestine Diabetes Volume 65, June 2016