Floriane Baraille,1,2,3,4 Sami Ayari,1,2,3,4 Véronique Carrière,1,2,3,4 Céline Osinski,1,2,3,4

Kevin Garbin,1,2,3 Bertrand Blondeau,1,2,3,4 Ghislaine Guillemain,1,2,3,4

Patricia Serradas,1,2,3,4 Monique Rousset,1,2,3,4 Michel Lacasa,1,2,3

Philippe Cardot,1,2,3,5 and Agnès Ribeiro1,2,3,4

Glucose Tolerance Is Improved in MiceInvalidated for the Nuclear ReceptorHNF-4g: A Critical Role forEnteroendocrine Cell LineageDiabetes 2015;64:2744–2756 | DOI: 10.2337/db14-0993

Intestine contributes to energy homeostasis through theabsorption, metabolism, and transfer of nutrients to theorganism. We demonstrated previously that hepatocytenuclear receptor-4a (HNF-4a) controls intestinal epithe-lium homeostasis and intestinal absorption of dietary lip-ids. HNF-4g, the other HNF-4 form highly expressed inintestine, is much less studied. In HNF-4g knockout mice,we detect an exaggerated insulin peak and improvementin glucose tolerance during oral but not intraperitonealglucose tolerance tests, highlighting the involvement ofintestine. Moreover, the enteroendocrine L-type cell line-age is modified, as assessed by the increased expressionof transcription factors Isl1, Foxa1/2, and Hnf4a, leadingto an increase of both GLP-1–positive cell number andbasal and stimulated GLP-1 plasma levels potentiatingthe glucose-stimulated insulin secretion. Using theGLP-1 antagonist exendin (9-39), we demonstrate a directeffect of GLP-1 on improved glucose tolerance. GLP-1exerts a trophic effect on pancreatic b-cells, and we re-port an increase of the b-cell fraction correlated with anaugmented number of proliferative islet cells and withresistance to streptozotocin-induced diabetes. In conclu-sion, the loss of HNF-4g improves glucose homeostasisthrough a modulation of the enteroendocrine cell lineage.

During past decades, changes in dietary habits, in-cluding increments in calorie and saturated fatty acid

intake, have occurred concomitantly with a rise ofmetabolic diseases, such as diabetes, obesity, and themetabolic syndrome (1). These metabolic disorders,which are known risk factors for cardiovascular dis-eases, represent emerging medical and public healthconcerns. Intestine, which is the first organ facing di-gestion products and which contributes to energy ho-meostasis through the transfer of nutrients to theorganism, receives little attention about its potentialrole in the onset of metabolic disorders compared withliver, pancreas, muscle, and adipose tissue.

Several transcription factors (CDX1, CDX2, GATA4/5/6) are reported to be involved in intestinal differ-entiation and regulation of gut-specific gene expression(2–5). Approaches based on transcriptome, metabolome,and bioinformatic analyses and data from in vivo experi-ments indicate that hepatocyte nuclear factor-4a (HNF-4a) is involved in the regulation of enterocyte phenotype(6–8).

HNF-4 belongs to the nuclear receptor superfamily. Inmammals, two paralog genes encode the HNF-4a andHNF-4g forms. HNF-4a is expressed in liver, kidney, pan-creas, and intestine (9), and its invalidation is lethal inmice at day 6.5 of embryogenesis (10). In human, heterozy-gous mutations in the HNF4A gene lead to maturity-onsetdiabetes of the young type 1 (11) and contribute to suscep-tibility to type 2 diabetes (12,13). Numerous studies have

1Sorbonne Universités, Université Pierre et Marie Curie, UMR_S 1138, Centre deRecherche des Cordeliers, Paris, France2INSERM, UMR_S 1138, Centre de Recherche des Cordeliers, Paris, France3Université Paris Descartes, Sorbonne Paris Cité, UMR_S 1138, Centre de Re-cherche des Cordeliers, Paris, France4Institute of Cardiometabolism and Nutrition, Pitié-Salpêtrière Hospital, Paris,France5UMR_S 1158, Neurophysiologie Respiratoire Expérimentale et Clinique, Paris,France

Corresponding authors: Agnès Ribeiro, agnes.ribeiro-pillet@crc.jussieu.fr, andPhilippe Cardot, philippe.cardot@upmc.fr.

Received 30 June 2014 and accepted 21 March 2015.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0993/-/DC1.

© 2015 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.

2744 Diabetes Volume 64, August 2015

METABOLISM

shown that HNF-4a plays pleiotropic roles in liver functionsand is a central transcription factor at the crossroads be-tween epithelial morphogenesis and functions (14,15).

HNF-4g was first identified in human kidney (16,17)and in mouse pancreas (18). It is also expressed in in-testine (19) but almost absent from liver (17,18), and ithas been shown to be highly expressed during intestinedevelopment (20). We reported previously that HNF-4a isexpressed in intestinal crypts and villi, whereas the gform is restricted to villi (21). These different distribu-tions combined with the absence of compensation byHNF-4g for some phenotypic traits of intestinal Hnf4aknockout mice, namely impairment of intestinal epithe-lium homeostasis, cell architecture, and fatty acid uptake(22–24), suggest that the a and g forms play specific anddifferent roles in intestine.

The physiological role of HNF-4g has been much lessstudied than that of HNF-4a. Hnf4g2/2 mice do notexhibit an overt phenotype. Succinctly, they presentweight gain but decreased food intake and lower energy

Table 1—Plasma parameters of fasted 6-month-old Hnf4g2/2

and Hnf4g+/+ mice

Hnf4g+/+ Hnf4g2/2 P value

Glucose (mg/dL) 98.6 6 4.09 95.84 6 3.36 0.13

Insulin (pg/mL) 342.00 6 42.15 248.17 6 38.95 0.189

Leptin (pg/mL) 715.28 6 187.9 959.85 6 269.4 0.78

GIP (pg/mL) 57.62 6 6.7 66.49 6 11.4 0.78

GLP-1 (pmol/L) 17.56 6 1.21 32.39 6 4.20 0.001

PYY (pg/mL) 90.42 6 10.38 135.88 6 13.36 0.0029

Data are mean 6 SEM. Plasma values from 6-month-old mice(n = 15–20 in each group) were analyzed after 15 h of fasting. Pvalues were determined by Mann-Whitney U test.

Figure 1—Glucose homeostasis in Hnf4g2/2 and Hnf4g+/+ mice. A: OGTT of 2-month-old and 6-month-old mice after 15 h of fasting andarea under the curve of the OGTT (n = 6 for each genotype). B: Measure of blood glucose during the 10 min following glucose challenge in2-month-old mice after 15 h of fasting (n = 6 for each genotype). C: Measure of plasma insulin during the OGTT of 6-month-old mice. Theinsulinogenic index was calculated as follows: (30-min plasma insulin – fasting plasma insulin [mU/mL])/(30-min blood glucose – fastingblood glucose [mmol/L]) (n = 5–6 for each genotype). D: ITT in 6-month-old mice (n = 5–6 for each genotype). E: IPGTT in 6-month-old miceafter 15 h of fasting (n = 5 for each genotype). *P < 0.05, **P < 0.01, ***P < 0.001, compared with control mice.

diabetes.diabetesjournals.org Baraille and Associates 2745

expenditure and locomotor night activity (25). We aimedto determine the roles of HNF-4g in adult mice and toexplore the underlying mechanisms. Using the constitu-tive Hnf4g knockout model, we demonstrate that loss ofHNF-4g leads to an increased enteroendocrine L-type cellnumber and to an overproduction of GLP-1, thus gener-ating hyperinsulinemia and improving glucose toleranceduring an oral glucose tolerance test (OGTT) through theincretin effect of GLP-1.

RESEARCH DESIGN AND METHODS

Animals and TreatmentsTotal and constitutive Hnf4g gene invalidation was as de-scribed in Gerdin et al. (25). Hnf4g+/2 mice obtained fromDeltagen, Inc., on a C57BL/6J genetic background, weremated to obtain Hnf4g2/2 mice on the same genetic back-ground. In experiments, we compared Hnf4g2/2 malemice with C57BL/6J wild-type male Hnf4g+/+ micematched for age and housed in the same room.

Specific intestinal Hnf4a gene invalidation in adultmice (Hnf4aDint) was described previously (22,23,26).For experiments, 6-month-old male control Hnf4aloxP/loxP

and Hnf4aloxP/loxP/villin-CreERT2 mice received tamoxifentreatment (23), and analyses were performed 10 dayslater.

All animals were housed in the specific-pathogen freefacility of the Centre de Recherche des Cordeliers ona 12-h light/dark cycle and fed a standard diet (A03/R03;

SAFE). Experimental procedures were according to Frenchguidelines for animal studies from the Comité National deRéflexion Ethique sur l’Expérimentation Animale CharlesDarwin (Ce5/2009/045).

In Vivo StudiesAll experiments were performed on conscious mice. Afterovernight fasting (15 h), mice received a glucose load of3.6 g/kg for OGTTs or 2 g/kg for intraperitoneal glucosetolerance tests (IPGTTs). The oral glucose challenges wereadministrated using a gavage needle (20 gauge, 38 mmlong, curved, with a 21/4-mm ball end). After overnightfasting, exendin (9-39) (5 mg/mouse in NaCl 9 g/L)(Abcam) or saline (vehicle) was injected intraperitoneally30 min before the oral glucose load (3.6 g/kg) as previouslydescribed (27). For insulin tolerance tests (ITTs), mice fedad libitum were injected intraperitoneally with 1 unit/kginsulin (Actrapid; Novo Nordisk). For streptozotocin treat-ment (Sigma-Aldrich, St. Louis, MO), mice receivedstreptozotocin (150 mg/kg) or vehicle (citrate buffer pH4–4.5) intraperitoneally after overnight fasting. Blood glu-cose concentrations were measured with a glucometer(Accu-Chek Go; Roche). Blood samples (70 mL at 0, 10,30, 60, and 180 min after glucose challenge) were collectedfrom the tail into EDTA-precoated microvettes (Sarstedt)with DPP-IV inhibitor (DPP4-010 [Millipore] 4 mL/100 mLblood) to prevent inactivation of plasma GLP-1. Plasmainsulin, leptin, peptide YY (PYY), and glucose-dependentinsulinotropic polypeptide (GIP) were measured by Luminex

Figure 1—Continued.

2746 HNF-4g and Enteroendocrine Function Diabetes Volume 64, August 2015

technology (multiplex mouse gut hormones kit;Roche). Total GLP-1 was measured with an ELISA kit(Millipore).

Tissue Isolation and HistologyMice were killed. Brown and white inguinal, perirenal, andepididymal adipose tissues were removed and weighed.Pieces of jejunum or colon and pancreas tail wereimmediately fixed overnight at 4°C in formalin-aceticacid-alcohol before embedding in paraffin. Immunostainingwas performed on 5-mm paraffin sections (22). Primaryantibodies are listed in Supplementary Table 1. Horserad-ish peroxidase–labeled secondary antibodies were rabbitanti-guinea pig and goat anti-rabbit IgG (Amersham Bio-sciences). 3,39-Diaminobenzidine was used for revelation.Tissue sections were counterstained with hematoxylin. Forimmunofluorescence, the secondary antibody was AlexaFluor 546 donkey anti-rabbit (Molecular Probes). Nucleiwere stained with DAPI. Immunostaining was examined byepifluorescence microscopy (Axiophot microscope, AxioCamcamera, AxioVision 4.5 software; Zeiss). Numbers of chro-mogranin A–, GLP-1–, or PYY-positive cells per villus or percrypt of jejunum or colon were estimated in three to fiveanimals per genotype.

Morphometric analysis of pancreatic islets in the wholepancreas was as previously described (28). Surfaces occu-pied by insulin staining and total pancreatic tissue were

quantified (103 objective, Leica microscope and camera,Leica QWin software; Leica Microsystems GmbH, Wetzlar,Germany). The ratio of Ki67-positive cell number to isletarea was quantified under a light microscope.

Isolation of Intestinal Epithelial CellsAfter flushing with PBS, the jejunum was cut into smallpieces and incubated overnight (4°C) in Cell RecoverySolution (BD Biosciences) (29) containing 2% proteaseinhibitor cocktail (Sigma-Aldrich). Epithelial cell homog-enate was filtered, washed with PBS, and centrifugedto obtain villus epithelial cells homogenized for RNAextraction.

Isolation of Pancreatic Islets and Measurement ofInsulin Secretion During Static IncubationMouse islets were isolated with collagenase solution(1 mg/mL; Sigma-Aldrich), separated on a Histopaquegradient (Sigma-Aldrich), and handpicked under micro-scope (Leica Microsystems). Batches of 50 islets weresequentially incubated for 1 h at 37°C with 2.8 mmol/Lglucose, 16.7 mmol/L glucose in the presence or ab-sence of GLP-1 (Sigma-Aldrich), and 50 mmol/L KClin Krebs-Ringer bicarbonate HEPES buffer. Total isletinsulin content was extracted in acid ethanol (1.5%volume for volume HCl in 75% volume for volume eth-anol). Islet insulin content and insulin secretion wereassayed by ELISA kit (Mercodia).

Figure 2—Effect of exendin (9-39) on glucose homeostasis in Hnf4g2/2 and Hnf4g+/+ mice. A: Measure of plasma total GLP-1 and GIP aftera glucose challenge. After 15 h of fasting, exendin (9-39) or NaCl (9 g/L) was injected intraperitoneally 30 min before an oral glucosechallenge in 6-month-old Hnf4g2/2 and Hnf4g+/+ mice. Blood glucose (B) and blood insulin (C ) during the OGTT in the presence ornonpresence of exendin (9-39). D: Measure of plasma GLP-1 and GIP 10 min after the glucose challenge in the presence of exendin(9-39). n = 5 for each condition. *P < 0.05; **P < 0.01. Ex9, exendin (9-39); ns, not significant.

diabetes.diabetesjournals.org Baraille and Associates 2747

RNA Extraction and Gene Expression AnalysisTotal RNA was isolated from jejunum or colon epithelialcells by RNA-PLUS solution (Qbiogene) and from pancre-atic islets by RNeasy Mini Kit (Qiagen). Reverse tran-scription was performed in 20 mL reaction mixture with5 mg or 400 ng RNA. Semiquantitative real-time PCR wasperformed with SYBR Green (Applied Biosystems) ina Stratagene system. Primer sequences are reported inSupplementary Table 2.

Transfection AssaysVectors encoding rat HNF-4a2 (pMT2-HNF-4a), mouse HNF-4g (pMT2-HNF-4g), and b-galactosidase (pRSV-b-Gal) wereas previously described (29). Vectors encoding the luciferasegene under the control of the 22.4-kb rat proglucagon pro-moter (pGL4-PPGLuc) (30,31) and Pax6 (pBAt14-mPax6) (32)were gifts from T. Kieffer (University of British Columbia,Vancouver, British Columbia, Canada) and M.S. German(University of California, San Francisco, San Francisco,CA), respectively. One microgram of DNA was transfectedusing X-tremeGENE HP DNA Transfection Reagent(Roche) into 12-well plates 24 h after plating COS-7 cells(0.753 105 cells/well) or in suspended GLUTag cells beforeplating at high density (2.5 3 105 cells/well) on Matrigel.

The GLUTag cell line was kindly provided by D. Drucker(Mount Sinai Hospital, Toronto, Ontario, Canada). Cellswere transfected with 200 ng pGL4-PPGLuc vector,200 ng pRSV-b-Gal plasmid, and increasing amountsof pMT2-HNF-4a, pMT2-HNF-4g, or pBAt14-mPax6 andanalyzed at 48 h posttransfection.

Statistical TestsResults are expressed as mean 6 SEM. Statistical anal-yses were performed using GraphPad Prism (GraphPadSoftware, La Jolla, CA). Two-group comparisons wereperformed using Mann-Whitney U nonparametric tests.Two-way ANOVA (Bonferroni posttest) or one-way ANOVA(Kruskal-Wallis test, Dunn posttest) test was performedfor comparisons of more than two groups. P , 0.05 wasconsidered statistically significant.

RESULTS

Hnf4g Gene Invalidation Improves Response to OGTTGrowth curves showed that the Hnf4g2/2 mice weighedmore than the Hnf4g+/+ mice, the difference becoming sig-nificant at 4 months (Supplementary Fig. 1A). AlthoughHnf4g2/2 mice did not eat more than Hnf4g+/+ mice

Figure 2—Continued.

2748 HNF-4g and Enteroendocrine Function Diabetes Volume 64, August 2015

(Supplementary Fig. 1B), they were slightly larger (4%)(Supplementary Fig. 1C) and had heavier inguinal andperirenal adipose tissues, which represent visceral adiposetissue (Supplementary Fig. 1D), although no differencewas found in the weights of brown and epididymal adi-pose tissues. We thus focused on energy metabolism, ex-ploring glucose homeostasis in 2- and 6-month-old mice(before and after weights became statistically differentbetween Hnf4g2/2 and Hnf4g+/+ mice).

Although basal blood glucose and insulin levels didnot differ significantly between 6-month-old Hnf4g2/2

and Hnf4g+/+ mice (Table 1), during an OGTT, 2- and 6-month-old Hnf4g2/2 mice had a lower glycemic peak30 min after the glucose bolus than the Hnf4g+/+ mice(Fig. 1A). Furthermore, the area under the curve wastwofold lower in Hnf4g2/2 than in Hnf4g+/+ mice. Suchan improvement in glucose tolerance could be due toa defect in intestinal glucose absorption, higher insulinsensitivity, or higher insulin secretion by pancreaticb-cells.

Glycemia did not differ between Hnf4g2/2 andHnf4g+/+ mice during the first 10 min of OGTT (Fig.1B), indicating that the enterocyte glucose absorptionwas not altered by Hnf4g gene invalidation. Althoughbasal insulinemia was similar in the two groups, thelevel of plasma insulin peak was threefold higher inHnf4g2/2 mice than in control mice during OGTT(Fig. 1C, left panel). Furthermore, the insulinogenic in-dex, which represents the first-phase insulin response toglucose challenge, was fivefold higher in Hnf4g2/2 thanin control mice (Fig. 1C, right panel). Both results in-dicate that Hnf4g gene invalidation leads to an improve-ment of oral glucose–stimulated insulin secretion.Finally, measurement of blood glucose during an ITTindicated that 2- or 6-month-old Hnf4g2/2 mice werenot more sensitive to insulin than Hnf4g+/+ mice (Sup-plementary Fig. 2A and Fig. 1D). Contrary to OGTT,during IPGTT, no difference in glycemia (SupplementaryFig. 2B and Fig. 1E) or plasma insulin (area under thecurve 42,157 6 6,378 and 40,312 6 6,548 pg/mL/min

Figure 3—Enteroendocrine cells in Hnf4g2/2 and Hnf4g+/+ mice. A: Chromogranin A immunostaining (red) and DAPI nuclear staining (blue)on sections of jejunum villus to visualize enteroendocrine cells. Arrowheads indicate specific staining of enteroendocrine cells. Averagenumber of chromogranin A–positive cells per villus or per crypt of jejunum is reported in the right panel (n = 80–130 crypts and villi from fiveanimals per genotype). B: GLP-1 immunostaining (brown). Arrowheads indicate specific staining to visualize GLP-1–positive cells. Highermagnifications of the insets delimited by boxes in the upper panels are shown in the lower panels. The average number of GLP-1–positivecells per villus of jejunum is reported in the right panel (n = 150–230 villi from three animals per genotype). Semiquantitative real-time PCRsfor transcription factor gene expression of 2-month-old (C) and 6-month-old (D) Hnf4g2/2 and Hnf4g+/+ mice and for enterohormone geneexpression (E) of 6-month-old Hnf4g2/2 and Hnf4g+/+ mice. The mRNA levels were normalized by cyclophilin mRNA level. n = 5–7 for eachgenotype. *P < 0.05, **P < 0.01, ***P < 0.001, compared with control mice.

diabetes.diabetesjournals.org Baraille and Associates 2749

for Hnf4g2/2 and Hnf4g+/+, respectively) (data notshown) was observed between Hnf4g2/2 and Hnf4g+/+

mice (2 and 6 months old). Failure of pancreatic b-cellsto secrete more insulin after intraperitoneal glucosechallenge versus oral glucose challenge revealed a majorrole of intestine in glucose-stimulated insulin secretionimprovement in Hnf4g2/2 mice.

We hypothesized an increased intestinal secre-tion of enteropeptides, such as GLP-1 or GIP, inresponse to oral glucose challenge only. Enteropeptidesecretion is accompanied by amplification of insu-lin secretion through the so-called incretin effect,which is not observed upon intraperitoneal glucoseadministration.

Figure 3—Continued.

2750 HNF-4g and Enteroendocrine Function Diabetes Volume 64, August 2015

Loss of HNF-4g Affects Enteroendocrine FunctionPlasma concentrations of total GLP-1 and GIP weremeasured to determine incretin hormone levels. Be-cause of the extensive and extremely rapid degradationof the active 7-36 amide GLP-1 (2–4-min half-life) byDPP-IV in plasma, we determined the total GLP-1plasma concentration, which reflects the oral glucose-stimulated GLP-1 secretion. In 6-month-old mice,plasma GLP-1 concentration was higher in Hnf4g2/2

than in Hnf4g+/+ mice both in the fasted state (Table 1)and in response to glucose challenge (Fig. 2A, leftpanel). By contrast, plasma GIP concentrations weresimilar in both groups in the fasted state (Table 1)

and after an oral glucose challenge (Fig. 2A, rightpanel), suggesting a key role for GLP-1.

To validate a direct incretin effect of GLP-1 on theexaggerated insulin peak in Hnf4g2/2 mice, the GLP-1receptor was blocked by its antagonist exendin (9-39),which inhibits GLP-1 signaling and its effect on insulinsecretion. An OGTT was performed 30 min after intraper-itoneal injection of exendin (9-39) or vehicle. Glucosetolerance was impaired in both groups of mice in thepresence of exendin (9-39) compared with the conditionwithout exendin (9-39) (Fig. 2B). Accordingly, plasma in-sulin decreased in both groups during OGTT in the pres-ence of exendin (9-39) (Fig. 2C). However, the glucose

Figure 4—Pancreatic b-cells of Hnf4g2/2 and Hnf4g+/+ mice. A: Representative insulin immunostaining of pancreatic sections from6-month-old Hnf4g2/2 and Hnf4g+/+ mice. B: Morphometric analysis of islet density, b-cell fraction, and islet size in pancreas tail ofHnf4g2/2 and Hnf4g+/+ mice (n = 8 sections from three animals for each genotype). The b-cell fraction represents the ratio of insulin-positive cell area to total pancreatic tissue area on the entire section. C: Insulin secretion of isolated islets from 6-month-old Hnf4g2/2 andHnf4g+/+ mice in the presence of glucose or KCl. D: Insulin content of islets of Langerhans isolated from 6-month-old Hnf4g2/2 and Hnf4g+/+

mice (n = 4 for each genotype). E: Representative Ki67 immunostaining of pancreatic sections from 6-month-old Hnf4g2/2 and Hnf4g+/+ mice.Ki67-positive cells are indicated by arrowheads. Average number of Ki67-positive cells per islet area (0.1 mm2) is reported in the right panel(n = 3 for each genotype). F: Blood glucose in 6-month-old Hnf4g2/2 and Hnf4g+/+ mice injected intraperitoneally with streptozotocin150 mg/kg or vehicle (citrate) at day 0. *P < 0.05, **P < 0.01, compared with control mice. G, glucose; STZ, streptozotocin.

diabetes.diabetesjournals.org Baraille and Associates 2751

tolerance of Hnf4g2/2 mice in the presence of exendin(9-39) was similar to that of Hnf4g+/+ mice without exendin(9-39). Ten minutes after glucose challenge in the presenceof exendin (9-39) or not, plasma GLP-1 (Fig. 2D, left panel)but not plasma GIP (Fig. 2D, right panel) increasedsignificantly in Hnf4g2/2 compared with Hnf4g+/+ mice.These results demonstrate a direct effect of GLP-1 onthe improved response of Hnf4g2/2 mice to a glucosetolerance test.

Hnf4g Gene Invalidation Affects the EnteroendocrineCell LineageThe size of crypts and villi was measured in 6-month-oldHnf4g2/2 and Hnf4g+/+ mice, and no obvious differencebetween both groups was observed (Supplementary Fig.3A). The enteroendocrine cells were counted in jejunum ofHnf4g2/2 and Hnf4g+/+ mice after immunostaining ofchromogranin A, a specific marker of secretion granulesof endocrine cells. In jejunum of Hnf4g2/2 mice, theenteroendocrine cell number doubled in villi andremained unchanged in crypts compared with Hnf4g+/+

mice (Fig. 3A). Quantification of the GLP-1–positive cellsshowed that Hnf4g gene invalidation induced an increaseof GLP-1–secreting cell number in jejunum (Fig. 3B) andcolon (Supplementary Fig. 3B). PYY-positive cell numberwas also higher in colons of 6-month-old Hnf4g2/2 micecompared with Hnf4g+/+ mice (Supplementary Fig. 3C).

To further identify the molecular mechanisms un-derlying the increase in number of GLP-1–producing cells,the expression of transcription factors known to affectenteroendocrine cell differentiation was quantified. Theexpression of Foxa1 and Foxa2 was increased by 1.97-

and 2.68-fold, respectively, in 2-month-old Hnf4g2/2

mice (Fig. 3C). At 6 months, only Isl1 expression was in-creased in these mice (1.63-fold) (Fig. 3D). Of note, themRNA level of Hnf4a was twofold higher in jejunum (Fig.3C and D) and colon (Supplementary Fig. 3D) of Hnf4g2/2

mice than in Hnf4g+/+ mice.To check whether HNF-4a or HNF-4g could regulate

GLP-1 transcription directly, we analyzed their effect ona luciferase reporter gene under the control of the pro-glucagon promoter (30). In COS-7 cells or enteroendo-crine GLUTag cells, no direct activation of theproglucagon promoter by HNF-4a or HNF-4g was seen,whereas Pax6 transactivated this promoter, as expected(32,33) (Supplementary Fig. 4). The results demonstratethat no direct connection exists between an increase inHnf4a gene expression and proglucagon gene expression.

Of note, the mRNA levels of the enterohormonesproglucagon (Gcg), which gives rise to GLP-1 after post-translational maturation; Gip; and Cck were not modifiedin jejunum of Hnf4g2/2 mice compared with Hnf4g+/+

mice (Fig. 3E). Thus, the higher plasma GLP-1 level reliedon the increased GLP-1–positive cell number.

Loss of HNF-4g Increases Insulin-ProducingPancreatic b-Cell NumberBy its trophic effect on pancreatic b-cells, an increasedcirculating GLP-1 level could also account for an increasedb-cell mass and function, leading to the exaggerated in-sulin peak during OGTT. Having demonstrated an in-creased number of GLP-1–positive cells and a directincreased incretin effect of GLP-1 in Hnf4g2/2 mice, wehypothesized an augmented GLP-1 trophic effect on

Figure 4—Continued.

2752 HNF-4g and Enteroendocrine Function Diabetes Volume 64, August 2015

pancreatic b-cells. The b-cell fraction of 6-month-oldHnf4g2/2 and Hnf4g+/+ mice was measured after immu-nostaining of insulin (Fig. 4A). The morphometric analysisshowed that the islet density did not differ in both groups(Fig. 4B, left panel). However, the b-cell fraction (Fig. 4B,middle panel), the islet mean size (Fig. 4B, right panel),and the density of larger-sized islets (Supplementary Fig.5A) were significantly higher in Hnf4g2/2 than in Hnf4g+/+

mice. The ex vivo insulin release in response to 2.8 or 16.7mmol/L glucose did not differ significantly between iso-lated islets from Hnf4g+/+ or Hnf4g2/2 mice (Fig. 4C).However, isolated islets from Hnf4g2/2 mice containedtwofold more insulin than those isolated from Hnf4g+/+

mice (Fig. 4D). Pancreatic b-cells from Hnf4g2/2 mice couldalso be more sensitive to GLP-1 and/or express moreGLP-1 receptor. We performed a GLP-1 dose-responseassessment of glucose-stimulated insulin secretion inisolated islets (Supplementary Fig. 5C) and did not ob-serve any modification of insulin secretion by isletsfrom Hnf4g2/2 compared with Hnfg+/+ mice. TheGlp1r mRNA level was not modified in islets ofHnf4g2/2 compared with Hnf4g+/+ mice (SupplementaryFig. 5B). These data suggest that the in vivo exaggeratedinsulin peak in response to glucose in Hnf4g2/2 micedepends on the pancreatic b-cell abundance rather than

on b-cell insulin secretory capacity or GLP-1 sensitivity.We did not observe a modification in the expression ofPdx1, NeuroD, Glut2, Ins1, Ins2, and Gck as well as ofHnf4a mRNA in islets from Hnf4g2/2 mice versusHnf4g+/+ mice, suggesting that b-cell function was notaffected by the HNF-4g invalidation (SupplementaryFig. 5B). Furthermore, the number of Ki67-positive pro-liferative cells was fourfold higher in Hnf4g2/2 thanin Hnf4g+/+ pancreatic islets (Fig. 4E). Finally, the mas-sive hyperglycemia provoked by streptozotocin, ab-cytotoxic drug, was delayed by 24 h in Hnf4g2/2 com-pared with Hnf4g+/+ mice (Fig. 4F). Such a result wasconsistent with the increased number of insulin-producingb-cells in the pancreas of Hnf4g2/2 mice.

DISCUSSION

The results demonstrate that the nuclear receptor HNF-4g plays a critical role in glucose homeostasis and thatloss of HNF-4g affects the enteroendocrine cell lineage,inducing increases in GLP-1–positive cell numbers andbasal and glucose-stimulated GLP-1 plasma levels. Thisleads to an exaggerated glucose-induced insulin secretionthat improves glucose tolerance of Hnf4g2/2 micethrough an augmentation of the GLP-1 incretin effectand a trophic impact on pancreatic b-cell mass (Fig. 5).

Figure 5—Schematic representation of the specific effects of HNF-4g on glucose homeostasis. The expression of HNF-4a and HNF-4galong the crypt-villus axis is shown in wild-type and Hnf4g2/2 mice. The Hnf4a gene invalidation in intestine does not affect glucosehomeostasis. In Hnf4g2/2 mice in which Hnf4a expression is increased in villi, the L-cell number and plasma GLP-1 level are increased,leading to expansion of pancreatic islet size and exaggerated insulin peak in response to glucose. Consequently, Hnf4g gene invalidationimproves glucose homeostasis. In conclusion, the HNF-4a/HNF-4g expression balance has to be tightly regulated to maintain the ho-meostasis of intestinal epithelium and glucose.

diabetes.diabetesjournals.org Baraille and Associates 2753

Mice invalidated for Hnf4g gene expression werereported to present a higher weight and a lower food intake,associated with lower night energy expenditure, than wild-type mice (25). Control of food intake plays a major role inenergy homeostasis and weight in mammals. Among hor-mones that control food intake, leptin, which is secreted byadipose tissue, and PYY, which is secreted by enteroendo-crine cells, are anorexigens (34). The current results confirmthat Hnf4g2/2 mice are heavier than Hnf4g+/+ mice, but wedid not observe difference in food intake, despite anincreased PYY plasma level in Hnf4g2/2 mice (Table 1).One hypothesis is that the level of leptin compensatesfor the increased level of PYY in Hnf4g2/2 mice, but theplasma leptin level was similar in the two groups of mice(Table 1). Another hypothesis is that increased PYY andGLP-1 levels enhance intestinal transit time and nutrientabsorption efficiency, which would in turn lead to weightgain.

The weight gain in Hnf4g2/2 mice was likely due totheir larger size and an increased amount of white adiposetissue. The visceral fat augmentation suggested metabolicdisorders and prompted us to investigate the insulin sen-sitivity of Hnf4g2/2 mice. Of note, disruption of theHnf4g gene led to an increased glucose tolerance thatwas due to neither the inhibition of glucose absorptionnor the higher insulin sensitivity but, rather, to increasedinsulinemia. These increases of glucose tolerance and in-sulin secretion in response to glucose were under thecontrol of intestinal functions because they were not ob-served when intestinal glucose absorption and signalingwere bypassed during an intraperitoneal glucose challenge.The specificity of Hnf4g gene disruption for this phenotypewas assessed in mice with specific intestinal invalidationof Hnf4a (22). During OGTT, neither glucose tolerancenor plasma insulin levels differed between 6-month-oldHnf4alox/lox and Hnf4aDint mice (Supplementary Fig. 6).

In response to glucose, incretin hormones, such asGLP-1 and GIP, which are secreted by enteroendocrinecells, potentiate insulin secretion (35). Only the basal andglucose-stimulated plasma GLP-1 levels were increased inHnf4g2/2 mice compared with Hnf4g+/+ mice. The increasedmean islet size and the higher density of large pancreaticislets in Hnf4g2/2 compared with control mice were consis-tent with the role of GLP-1 in pancreatic b-cell mass main-tenance (36–38) and proliferation (39,40). HNF-4g lossaffected b-cell abundance but not b-cell insulin secretorycapacity. Finally, the larger islets led to a resistance tostreptozotocin, a b-cell cytotoxic drug, suggesting thatHNF-4g is involved in susceptibility to diabetes.

HNF-4a is expressed in pancreas and induces insulingene transcription (41) and insulin secretion (42). A com-pensatory effect of the loss of HNF-4g by HNF-4a couldexplain the hyperinsulinemia in response to glucose inpancreas. However, the expression of Hnf4a was not mod-ified in pancreas of Hnf4g2/2 mice (Supplementary Fig.5B), supporting the impact of HNF-4g loss and the role ofintestine on glucose homeostasis through a direct effect

of GLP-1, as assessed by the use of the GLP-1 antagonistexendin (9-39). In accordance with the increased GLP-1plasma level, we show that Hnf4g gene invalidation leadsto modifications of enteroendocrine cell lineage character-ized by an increased number of GLP-1–producing cells injejunum and colon.

The expression of Foxa1, Foxa2, and Isl1 was en-hanced, suggesting that modifications of the transcrip-tion factor network favored the GLP-1–secreting celllineage (43,44). FOXA1 and FOXA2 are known to con-trol the expression of Hnf4a (45,46). Furthermore, theyare considered pioneer factors that engage chromatinbefore other transcription factors (47). One may hy-pothesize that in the absence of HNF-4g, Foxa1/2 levelsincrease during development to young adulthood andallow an overexpression of HNF-4a throughout life.Whether the observed effects were the direct conse-quence of the loss of HNF-4g or resulted indirectlyfrom the HNF-4a increment in Hnf4g2/2 mouse intes-tine is difficult to assert.

HNF-4a and HNF-4g, which are encoded by two dif-ferent genes, share high homology in their DNA-bindingor ligand-binding domains (16,18). In vitro, HNF-4a andHNF-4g display the same apolipoprotein A-IV promotertransactivation capacity and binding affinity to hormone-responsive elements (29). A chromatin immunoprecipita-tion sequencing study failed to discriminate betweenHNF-4a and HNF-4g binding sites (48). However, thesetwo transcription factors 1) have a different spatial dis-tribution along the crypt-villus axis, with HNF-4a beingexpressed along the crypt-villus axis and HNF-4g beingrestricted to the differentiated villus compartment, and 2)have poor homology in their transactivation and regula-tion domains. These observations suggest that these twoforms of HNF-4 play specific roles. After an inducible andintestine-specific knockout of the Hnf4a gene, we demon-strated that HNF-4a controls the intestinal epitheliumhomeostasis and cell architecture (22) and the uptake offatty acids by enterocytes (23). HNF-4a is also requiredfor absorptive function and ion transport in colon(49,50). Despite an ectopic expression of HNF-4g incrypts of Hnf4aDint mice (22), a potential compensatoryeffect of HNF-4g on proliferation was weak becauseHnf4aDint mice display tumors in colon 4–6 months afterinjection of azoxymethane, a potent carcinogen in colon(24). Of note, the expression of Hnf4a was increased injejunum and colon of Hnf4g2/2 mice compared withHnf4g+/+ mice. Such Hnf4a overexpression could affectthe absorptive function and water status in colon.

We previously showed that the enteroendocrine cellpopulation (chromogranin A–positive cells) as well as thelevel of proglucagon and Gip-encoding mRNA decrease invilli of Hnf4aDint mice compared with control mice (22).Altogether, these results suggest that HNF-4g is a negativemodulator of the enteroendocrine cell lineage, whereasHNF-4a is a positive one. Thus, we conclude that thebalance between the expression of HNF-4a and HNF-4g

2754 HNF-4g and Enteroendocrine Function Diabetes Volume 64, August 2015

must be finely regulated to maintain enteroendocrine celllineage and glucose homeostasis (Fig. 5).

Acknowledgments. The authors thank A. Benkhoui (Centre deRecherche des Cordeliers) for hormone dosage by Luminex, S. Saint-Just(Centre de Recherche des Cordeliers) for mouse line maintenance, C. Ayassamy(Centre d’Explorations Fonctionnelles) for mouse care, and C. Amorin (Centre deGénotypage et de Biochimie) for mouse genotyping.Funding. This work was supported by INSERM, France, and Université Pierreet Marie Curie (UPMC), Paris, France. F.B. received a doctoral fellowship fromCancéropôle, Île-de-France. S.A. received a doctoral fellowship from UPMC.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. F.B. and A.R. researched data and contributed tothe discussion and writing, review, and editing of the manuscript. S.A. and P.S.researched data and contributed to the discussion. V.C., B.B., G.G., M.L., and P.C.researched data and contributed to the discussion and review and editing of themanuscript. C.O. researched data. K.G. researched data and contributed to thereview and editing of the manuscript. M.R. contributed to the discussion andwriting, review, and editing of the manuscript. A.R. is the guarantor of this workand, as such, had full access to all data in the study and takes responsibility forthe integrity of the data and the accuracy of the data analysis.Prior Presentation. Part of these data was presented at the Congress ofthe Société Francophone du Diabète, Paris, France, 11–14 March 2014.

References1. Shikany JM, White GL Jr. Dietary guidelines for chronic disease prevention.

South Med J 2000;93:1138–11512. Benoit YD, Paré F, Francoeur C, et al. Cooperation between HNF-1alpha,

Cdx2, and GATA-4 in initiating an enterocytic differentiation program in a normal

human intestinal epithelial progenitor cell line. Am J Physiol Gastrointest Liver

Physiol 2010;298:G504–G5173. Beuling E, Baffour-Awuah NY, Stapleton KA, et al. GATA factors regulate

proliferation, differentiation, and gene expression in small intestine of mature

mice. Gastroenterology 2011;140:1219–12294. Hryniuk A, Grainger S, Savory JG, Lohnes D. Cdx function is required for

maintenance of intestinal identity in the adult. Dev Biol 2012;363:426–4375. Stringer EJ, Duluc I, Saandi T, et al. Cdx2 determines the fate of postnatal

intestinal endoderm. Development 2012;139:465–4746. Li X, Madison BB, Zacharias W, Kolterud A, States D, Gumucio DL. De-

convoluting the intestine: molecular evidence for a major role of the mesenchyme

in the modulation of signaling cross talk. Physiol Genomics 2007;29:290–3017. Mariadason JM, Nicholas C, L’Italien KE, et al. Gene expression profiling of

intestinal epithelial cell maturation along the crypt-villus axis. Gastroenterology

2005;128:1081–10888. Stegmann A, Hansen M, Wang Y, et al. Metabolome, transcriptome, and

bioinformatic cis-element analyses point to HNF-4 as a central regulator of gene

expression during enterocyte differentiation. Physiol Genomics 2006;27:141–1559. Benoit G, Cooney A, Giguere V, et al. International Union of Pharmacology.

LXVI. Orphan nuclear receptors. Pharmacol Rev 2006;58:798–83610. Duncan SA, Nagy A, Chan W. Murine gastrulation requires HNF-4 regulated

gene expression in the visceral endoderm: tetraploid rescue of Hnf-4(-/-) em-

bryos. Development 1997;124:279–28711. Gupta RK, Kaestner KH. HNF-4alpha: from MODY to late-onset type 2 di-

abetes. Trends Mol Med 2004;10:521–52412. Love-Gregory LD, Wasson J, Ma J, et al. A common polymorphism in the

upstream promoter region of the hepatocyte nuclear factor-4 alpha gene on

chromosome 20q is associated with type 2 diabetes and appears to contribute to

the evidence for linkage in an Ashkenazi Jewish population. Diabetes 2004;53:

1134–1140

13. Silander K, Mohlke KL, Scott LJ, et al. Genetic variation near the hepatocytenuclear factor-4 alpha gene predicts susceptibility to type 2 diabetes. Diabetes2004;53:1141–114914. Battle MA, Konopka G, Parviz F, et al. Hepatocyte nuclear factor 4alphaorchestrates expression of cell adhesion proteins during the epithelial trans-formation of the developing liver. Proc Natl Acad Sci U S A 2006;103:8419–842415. Hwang-Verslues WW, Sladek FM. HNF4a—role in drug metabolism andpotential drug target? Curr Opin Pharmacol 2010;10:698–70516. Drewes T, Senkel S, Holewa B, Ryffel GU. Human hepatocyte nuclear factor4 isoforms are encoded by distinct and differentially expressed genes. Mol CellBiol 1996;16:925–93117. Plengvidhya N, Antonellis A, Wogan LT, et al. Hepatocyte nuclear factor-4gamma: cDNA sequence, gene organization, and mutation screening in early-onset autosomal-dominant type 2 diabetes. Diabetes 1999;48:2099–210218. Taraviras S, Mantamadiotis T, Dong-Si T, et al. Primary structure, chro-mosomal mapping, expression and transcriptional activity of murine hepatocytenuclear factor 4gamma. Biochim Biophys Acta 2000;1490:21–3219. Bookout AL, Jeong Y, Downes M, Yu RT, Evans RM, Mangelsdorf DJ. An-atomical profiling of nuclear receptor expression reveals a hierarchical tran-scriptional network. Cell 2006;126:789–79920. Li X, Udager AM, Hu C, Qiao XT, Richards N, Gumucio DL. Dynamic pat-terning at the pylorus: formation of an epithelial intestine-stomach boundary inlate fetal life. Dev Dyn 2009;238:3205–321721. Sauvaget D, Chauffeton V, Citadelle D, et al. Restriction of apolipoproteinA-IV gene expression to the intestine villus depends on a hormone-responsiveelement and parallels differential expression of the hepatic nuclear factor 4alphaand gamma isoforms. J Biol Chem 2002;277:34540–3454822. Cattin AL, Le Beyec J, Barreau F, et al. Hepatocyte nuclear factor 4alpha,a key factor for homeostasis, cell architecture, and barrier function of the adultintestinal epithelium. Mol Cell Biol 2009;29:6294–630823. Frochot V, Alqub M, Cattin AL, et al. The transcription factor HNF-4a: a keyfactor of the intestinal uptake of fatty acids in mouse. Am J Physiol GastrointestLiver Physiol 2012;302:G1253–G126324. Saandi T, Baraille F, Derbal-Wolfrom L, et al. Regulation of the tumorsuppressor homeogene Cdx2 by HNF4a in intestinal cancer. Oncogene 2013;32:3782–378825. Gerdin AK, Surve VV, Jönsson M, et al. Phenotypic screening of hepatocytenuclear factor (HNF) 4-gamma receptor knockout mice. Biochem Biophys ResCommun 2006;349:825–83226. Hayhurst GP, Lee YH, Lambert G, Ward JM, Gonzalez FJ. Hepatocyte nuclearfactor 4alpha (nuclear receptor 2A1) is essential for maintenance of hepatic geneexpression and lipid homeostasis. Mol Cell Biol 2001;21:1393–140327. Lamont BJ, Li Y, Kwan E, Brown TJ, Gaisano H, Drucker DJ. PancreaticGLP-1 receptor activation is sufficient for incretin control of glucose metabolismin mice. J Clin Invest 2012;122:388–40228. Stolarczyk E, Le Gall M, Even P, et al. Loss of sugar detection by GLUT2affects glucose homeostasis in mice. PLoS One 2007;2:e128829. Archer A, Sauvaget D, Chauffeton V, et al. Intestinal apolipoprotein A-IVgene transcription is controlled by two hormone-responsive elements: a role forhepatic nuclear factor-4 isoforms. Mol Endocrinol 2005;19:2320–233430. Jin T, Drucker DJ. Activation of proglucagon gene transcription througha novel promoter element by the caudal-related homeodomain protein cdx-2/3.Mol Cell Biol 1996;16:19–2831. Fujita Y, Chui JW, King DS, et al. Pax6 and Pdx1 are required for productionof glucose-dependent insulinotropic polypeptide in proglucagon-expressingL cells. Am J Physiol Endocrinol Metab 2008;295:E648–E65732. Sander M, Neubüser A, Kalamaras J, Ee HC, Martin GR, German MS. Ge-netic analysis reveals that PAX6 is required for normal transcription of pancreatichormone genes and islet development. Genes Dev 1997;11:1662–167333. Hill ME, Asa SL, Drucker DJ. Essential requirement for Pax6 in control ofenteroendocrine proglucagon gene transcription. Mol Endocrinol 1999;13:1474–1486

diabetes.diabetesjournals.org Baraille and Associates 2755

34. Perry B, Wang Y. Appetite regulation and weight control: the role of guthormones. Nutr Diabetes 2012;2:e2635. Phillips LK, Prins JB. Update on incretin hormones. Ann N Y Acad Sci 2011;1243:E55–E7436. Buteau J, El-Assaad W, Rhodes CJ, Rosenberg L, Joly E, Prentki M.Glucagon-like peptide-1 prevents beta cell glucolipotoxicity. Diabetologia 2004;47:806–81537. Hui H, Nourparvar A, Zhao X, Perfetti R. Glucagon-like peptide-1 inhibitsapoptosis of insulin-secreting cells via a cyclic 59-adenosine monophosphate-dependent protein kinase A- and a phosphatidylinositol 3-kinase-dependentpathway. Endocrinology 2003;144:1444–145538. Kim SJ, Winter K, Nian C, Tsuneoka M, Koda Y, McIntosh CH. Glucose-dependent insulinotropic polypeptide (GIP) stimulation of pancreatic beta-cellsurvival is dependent upon phosphatidylinositol 3-kinase (PI3K)/protein ki-nase B (PKB) signaling, inactivation of the forkhead transcription factor Foxo1,and down-regulation of bax expression. J Biol Chem 2005;280:22297–2230739. Liu Z, Habener JF. Glucagon-like peptide-1 activation of TCF7L2-dependentWnt signaling enhances pancreatic beta cell proliferation. J Biol Chem 2008;283:8723–873540. Perfetti R, Zhou J, Doyle ME, Egan JM. Glucagon-like peptide-1 induces cellproliferation and pancreatic-duodenum homeobox-1 expression and increasesendocrine cell mass in the pancreas of old, glucose-intolerant rats. Endocrinology2000;141:4600–460541. Bartoov-Shifman R, Hertz R, Wang H, Wollheim CB, Bar-Tana J, Walker MD.Activation of the insulin gene promoter through a direct effect of hepatocytenuclear factor 4 alpha. J Biol Chem 2002;277:25914–25919

42. Miura A, Yamagata K, Kakei M, et al. Hepatocyte nuclear factor-4alpha is

essential for glucose-stimulated insulin secretion by pancreatic beta-cells. J Biol

Chem 2006;281:5246–525743. Schonhoff SE, Giel-Moloney M, Leiter AB. Minireview: development and

differentiation of gut endocrine cells. Endocrinology 2004;145:2639–264444. Ye DZ, Kaestner KH. Foxa1 and Foxa2 control the differentiation of

goblet and enteroendocrine L- and D-cells in mice. Gastroenterology 2009;

137:2052–206245. Duncan SA, Navas MA, Dufort D, Rossant J, Stoffel M. Regulation of

a transcription factor network required for differentiation and metabolism. Sci-

ence 1998;281:692–69546. Si-Tayeb K, Lemaigre FP, Duncan SA. Organogenesis and development of

the liver. Dev Cell 2010;18:175–18947. Kaestner KH. The FoxA factors in organogenesis and differentiation. Curr

Opin Genet Dev 2010;20:527–53248. Neph S, Stergachis AB, Reynolds A, Sandstrom R, Borenstein E,

Stamatoyannopoulos JA. Circuitry and dynamics of human transcription

factor regulatory networks. Cell 2012;150:1274–128649. Ahn SH, Shah YM, Inoue J, et al. Hepatocyte nuclear factor 4alpha in the

intestinal epithelial cells protects against inflammatory bowel disease. Inflamm

Bowel Dis 2008;14:908–92050. Darsigny M, Babeu JP, Dupuis AA, et al. Loss of hepatocyte-nuclear-

factor-4alpha affects colonic ion transport and causes chronic in-

flammation resembling inflammatory bowel disease in mice. PLoS One

2009;4:e7609

2756 HNF-4g and Enteroendocrine Function Diabetes Volume 64, August 2015