Investigations on the metabolic role of the Wntsignaling pathway and hepatic transcription factor7-like 2 (TCF7L2) have generated opposing views.While some studies demonstrated a repressive ef-fect of TCF7L2 on hepatic gluconeogenesis, a recentstudy using liver-specific Tcf7l22/2 mice suggestedthe opposite. As a consequence of redundant andbidirectional actions of transcription factor (TCF)molecules and other complexities of the Wnt path-way, knockout of a single Wnt pathway componentmay not effectively reveal a complete metabolic pic-ture of this pathway. To address this, we generatedthe liver-specific dominant-negative (DN) TCF7L2(TCF7L2DN) transgenic mouse model LTCFDN. Thesemice exhibited progressive impairment in response topyruvate challenge. Importantly, LTCFDN hepatocytesdisplayed elevated gluconeogenic gene expression,gluconeogenesis, and loss of Wnt-3a–mediated re-pression of gluconeogenesis. In C57BL/6 hepatocytes,adenovirus-mediated expression of TCF7L2DN, butnot wild-type TCF7L2, increased gluconeogenesisand gluconeogenic gene expression. Our further mecha-nistic exploration suggests that TCF7L2DN-mediated in-hibition of Wnt signaling causes preferential interactionof b-catenin (b-cat) with FoxO1 and increased bindingof b-cat/FoxO1 to the Pck1 FoxO binding site, resultingin the stimulation of Pck1 expression and increasedgluconeogenesis. Together, our results using TCF7L2DNas a unique tool revealed that the Wnt signaling pathway
and its effector b-cat/TCF serve a beneficial role in sup-pressing hepatic gluconeogenesis.
Following pivotal studies indicating that the tran-scription factor 7-like 2 (TCF7L2) is an important riskgene for the development of type 2 diabetes (T2D) (1),great efforts have been made to explore its role asa Wnt signaling molecule in pancreatic b-cells andother tissues including liver (2–16). Although severalinvestigations suggested that TCF7L2 negatively reg-ulates hepatic gluconeogenesis (8,10,11,17), one re-cent study (7) reported that liver-specific knockout ofTCF7L2 reduced hepatic glucose production (HGP),while hepatic overexpression of TCF7L2 increased HGP.As a result, the controversial issues surrounding themetabolic function of TCF7L2 have been extended frompancreatic b-cells to liver and hepatocytes specifically(4–6,9,13,18–20).
TCF7L2 or other transcription factor (TCF) memberscan interact with b-catenin (b-cat), forming the bipar-tite transcription factor b-cat/TCF, which serves as theimportant effector of the Wnt signaling pathway. Therole of Wnt signaling has been intensively studied inmany tissues, including the liver, using various trans-genic animal models. Although overexpression ofa given Wnt ligand or the expression of constitutivelyactive S33Y mutant b-cat has provided solid evidencefor the involvement of this signaling cascade in liver
1Institute of Medical Science, University of Toronto, Toronto, Ontario, Canada2Division of Advanced Diagnostics, Toronto General Research Institute, UniversityHealth Network, Toronto, Ontario, Canada3Department of Physiology, University of Toronto, Toronto, Ontario, Canada
Received 28 August 2014 and accepted 1 January 2015.
This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-1329/-/DC1.
Z.C. is currently affiliated with the Department of Endocrinology and Metabolism,The Third Affiliated Hospital of Sun Yat-sen University and Guangdong ProvincialKey Laboratory of Diabetology, Guangzhou, People’s Republic of China.
development, zonation, cell proliferation, tumorigene-sis, and the susceptibility to oxidative stress (21–26),the hepatic role of Wnt signaling in metabolic homeo-stasis still remains unclear. The complexity of the Wntsignaling pathway is reflected by the existence of mul-tiple Wnt ligands, receptors, and coreceptors, as well asvarious intracellular modulating elements (27,28).Furthermore, the function of TCFs is bidirectional,depending on their interaction partners as well as theavailability and phosphorylation status of their cofactorb-cat. Finally, b-cat also interacts with FOXOs, keyeffectors of the stress and aging signaling pathway(29). The pathophysiological contribution of the inter-action between FOXOs and b-cat was demonstratedpreviously in bone diseases (30,31) and very recentlyin hepatic gluconeogenesis (32). Thus, the function ofTCF members is not only directly controlled by its co-factor b-cat, nuclear coactivators, and corepressors, butalso indirectly regulated by the stress signaling pathwayeffectors FOXOs.
Here, we generated a transgenic mouse model, LTCFDN,in which dominant-negative (DN) TCF7L2 (TCF7L2DN) wasexpressed specifically in the liver (33). TCF7L2DN has beenshown to repress Wnt target gene expression in vitro and invivo by our team and other investigators (12,34–36). Theobservations that LTCFDN mice exhibit progressive im-pairment of pyruvate and glucose tolerance along withthe upregulation of the gluconeogenic gene programand glucose output in primary hepatocytes indicate thatthe Wnt signaling cascade negatively regulates hepaticgluconeogenesis.
RESEARCH DESIGN AND METHODSAnimalsThe LTCFDN mouse model was generated by cloning themouse serum albumin 2.4-kb promoter/enhancer construct(provided by Dr. Richard Palmiter, University of Pennsylva-nia) upstream of the 1.6-kb human TCF7L2DN long-isoformcDNA sequence (12,33–35). FVB mouse zygote pronuclearmicroinjection of the linearized DNA construct and implan-tation into pseudopregnant recipients were then performedby the Toronto Centre for Phenogenomics Transgenic Core(12). Male heterozygous LTCFDN mice were consistentlybred with female wild-type (WT) FVB mice to produceheterozygous mice (labeled as LTCFDN) and control WTlittermates. Male C57BL/6 mice (at 7 weeks of age) werepurchased from Charles River Laboratories (St. Laurent, Que-bec, Canada) as previously described (37). The mice werehoused on a 12-h light-dark cycle at ambient room temper-ature with free access to normal chow diet and water. Allanimal protocols were approved by the Institutional AnimalCare and Use Committee of the University Health Network.
In Vivo Tolerance Tests and Glucose Production AssayFor glucose and pyruvate tolerance tests, mice were fastedfor 16 h prior to intraperitoneal injection of either glucose(2 g/kg body wt) or pyruvate (2 g/kg body wt) (38). Forthe insulin tolerance test, mice were starved for 6 h prior
to intraperitoneal insulin (0.75–1.0 units/kg body wt)injection (38). The method for the glucose productionassay has been previously described and is detailed inthe Supplementary Experimental Procedures (38).
Isolation of Mouse Primary Hepatocytes and Cell CultureMouse hepatocytes were isolated from 8- to 12-week-oldchow diet–fed male C57BL/6 mice, as has been previouslydescribed (10) and is detailed in the Supplementary Exper-imental Procedures. The culture of the mouse Hepa1-6 cellline was described previously (10).
Adenovirus ExperimentsThe WT TCF7L2 and TCF7L2DN cDNA were originallyprovided by Dr. Eric Fearon (University of MichiganMedical School) (35). The adenoviruses (Ads) Ad-greenfluorescent protein, Ad-TCF7L2WT, and Ad-TCF7L2DNwere generated using the AdEasy XL Adenoviral VectorSystem (Agilent Technologies), as detailed in the Supple-mentary Experimental Procedures.
Western BlottingWhole-cell lysates were prepared from mouse tissue orcultured cell lines and were subjected to SDS-PAGE, aspreviously described (38). Antibodies and their sourcesare listed in Supplementary Experimental Procedures.
RNA Isolation, Reverse Transcription, and QuantitativePCRIsolation of RNA, cDNA synthesis by reverse transcription,and real-time PCR were performed as previously described(38). Nonquantitative PCR was performed using the FastDNA Polymerase PCR Mix (Kapa Biosystems) followed byagarose gel electrophoresis. Specific genes were quantified byamplification using specific primers listed in SupplementaryTable 1. All instructions of the manufacturers were followed.
CoimmunoprecipitationWhole-cell lysates were prepared from Hepa1-6 cells inradioimmunoprecipitation assay buffer. Two microgramsof b-cat, FoxO1, or control IgG antibody (Santa Cruz Bio-technology) were mixed with 500 mg lysates overnight at4°C with agitation followed by incubation with ProteinA/G PLUS-Agarose beads (Santa Cruz Biotechnology) for2 h at 4°C with agitation. Beads were washed four times inradioimmunoprecipitation assay buffer by centrifugationand removal of supernatant, then were suspended in50 mL 13 sample buffer and boiled for 5 min. Sampleswere then subjected to SDS-PAGE for Western blotting.Lysates not subjected to the immunoprecipitation proce-dure were used as the input controls. The HEK293 cellline was provided by the American Type Culture Col-lection. Cells were infected by Ad-TCF7L2WT or Ad-TCF7L2DN for 36 h. Two micrograms of hemagglutinin(HA)-tagged antibody were used for immunoprecipitation,followed by Western blotting with the indicated antibody.
Chromatin ImmunoprecipitationThe chromatin immunoprecipitation (ChIP) procedurehas been described in our previous study (39) and is de-tailed in Supplementary Experimental Procedures.
1924 TCF7L2DN Elevates Hepatic Gluconeogenesis Diabetes Volume 64, June 2015
Plasmid DNA Transfection, Luciferase ReporterAnalysis, and Electrophoretic Mobility Shift AssayThe methods for plasmid DNA transfection and luciferase(LUC) reporter analyses of TOPflash and Pck1(2595/+67)-LUC were described previously, with Hepa1-6 cellscultured in six-well plates (10,34). Cells were harvested18 h after plasmid transfection (1 mg reporter plasmid or1 mg reporter plasmid plus 0.5 mg indicated cDNA orcontrol plasmid) for LUC analysis (10). The procedurefor electrophoretic mobility shift assay is detailed in theSupplementary Experimental Procedures.
Insulin MeasurementSerum insulin levels were measured using the Rat InsulinRadioimmunoassay Kit (Millipore) according to the man-ufacturer’s instructions.
Statistical AnalysisData are presented as the mean 6 SEM. Significancewas determined using the Student t test or one-wayANOVA followed by Bonferroni post hoc test as appro-priate for single or multiple comparisons, respectively.Differences were considered statistically significantwhen P , 0.05.
RESULTS
Verification of the DN Action of TCF7L2DN and theGeneration of the LTCFDN Mouse ModelThe albumin-TCF7L2DN fusion gene (Fig. 1A) was con-structed in which the expression of human TCF7L2DN(12) is driven by a mouse albumin promoter/enhancer con-struct, which was provided by Dr. Richard Palmiter (Univer-sity of Pennsylvania) (33). Lacking the b-cat interactiondomain (Fig. 1A), TCF7L2DN acts as a DNmolecule to block,in theory, the function of any TCF7L2 isoforms and otherTCFs such as TCF7 and TCF7L1, which are known to beexpressed in hepatocytes (10). In our previous in vitro andin vivo investigations (12,34), TCF7L2DN attenuated basalas well as Wnt-stimulated gut proglucagon gene expression.To verify the DN action of TCF7L2DN at the molecular level,we infected HEK293 cells with either Ad-TCF7L2WT or Ad-TCF7L2DN. The expression of each exogenous TCF7L2 pro-tein was verified by detection of the HA tag (SupplementaryFig. 1A). Both WT TCF7L2 and TCF7L2DN were bound tothe consensus TCF binding site (Supplementary Fig. 1B),and the binding in both cases was attenuated by the addi-tion of an unlabeled “cold” probe (Supplementary Fig. 1C).Precipitation of the HA tag also pulled down b-cat in
Figure 1—LTCFDN transgenic mice express TCF7L2DN exclusively in the adult liver. A: Schematic representation of the albumin-TCF7L2DN transgene. Detection of TCF7L2DN by Western blotting in 12-week-old LTCFDN mouse liver tissue (B), but not in other organs(C ). D: Detection of TCF7L2DN in 2-week-old but not in newborn LTCFDN mice. E: Axin2 mRNA expression in hepatocytes from 12-week-old LTCFDN and WT mice. n = 3 mice per group. F: mRNA expression of gluconeogenic genes in liver tissue from 12-week-old LTCFDNand WT mice. n = 4–5 mice per group. HMG, high-mobility group. *P < 0.05. Values represent the mean 6 SEM.
diabetes.diabetesjournals.org Ip and Associates 1925
HEK293 cells infected with Ad-TCF7L2WT but not inHEK293 cells infected with Ad-TCF7L2DN (SupplementaryFig. 1D and E). Thus, TCF7L2DN possessed the ability tobind to the TCF binding site on target promoters but failedto recruit the coactivator b-cat, as intended.
We generated the LTCFDN mouse model via insertionof the albumin-TCF7L2DN fusion gene. Among the fivetransgene-positive founders (Supplementary Fig. 2A), theoffspring of founder C showed germ-line transmission andsubstantial adult hepatic TCF7L2DN expression (Supple-mentary Fig. 2B and Fig. 1B), and were thus used for allfurther studies. Liver-specific expression of TCF7L2DNwas confirmed as this exogenous protein, which is 3 kDasmaller than the 78-kDa endogenous isoform, was notdetected in other adult organs (Fig. 1C). While TCF7L2DNwas also expressed in the liver of 2-week-old LTCFDNmice,it was undetectable in the liver of newborn mice (Fig. 1D).Interestingly, both WT and LTCFDN newborn mice lackedendogenous hepatic TCF7L2 protein expression (Fig. 1D),although three TCF members were detectable at the mRNAlevel in the newborn mouse liver (Supplementary Fig. 2C).In LTCFDN hepatocytes, the expression of Axin2, a knownWnt pathway downstream target gene in hepatocytes, wassignificantly reduced (Fig. 1E), indicating that TCF7L2DNexpression in hepatocytes indeed attenuated the Wnt sig-naling pathway. The expression levels of the gluconeogen-esis genes, including Pck1, G6pc, Fbp1, and Ppargc1a, wereincreased in the liver tissue of LTCFDN mice in comparisonwith WT mice (Fig. 1F). The majority of our study wasperformed in adult male mice. When analyzing newbornand 2-week-old mice, sex was ignored.
LTCFDN Mice Exhibit Progressive Defects in PyruvateToleranceOn chow diet for up to 42 weeks, LTCFDN mice showedno significant alterations in body weight (SupplementaryFig. 3A), or in fasted or fed glucose and insulin levels(Supplementary Fig. 3B and C), while 12-week-old LTCFDNmice showed a modest but significant increase in liverweight (Supplementary Fig. 3D). At the age of 9 weeks,LTCFDN mice displayed significant impairment of toler-ance to pyruvate, a major gluconeogenic precursor(Fig. 2A and Supplementary Fig. 3E). This defect was ex-acerbated at 23 and 40 weeks of age (Fig. 2B and C andSupplementary Fig. 3F and G). We did not observe animpaired tolerance to glucose challenge in 10-week-oldLTCFDN mice (Fig. 2D and Supplementary Fig. 3H). How-ever, 24-week-old LTCFDN mice were glucose intolerant(Fig. 2E and Supplementary Fig. 3I). Importantly, at 11 or41 weeks of age, we did not detect any abnormalresponses to insulin challenge (Fig. 2F and G and Supple-mentary Fig. 3J and K). Consistently, hepatocytes isolatedfrom 12-week-old LTCFDN mice produced higher levelsof glucose from gluconeogenic precursors (Fig. 2H). Whilethe canonic Wnt ligand Wnt-3a repressed glucose productionin WT hepatocytes, this effect was absent in hepatocytesisolated from LTCFDN mice (Fig. 2H and Supplementary
Fig. 3L). Furthermore, there were no defects in Akt S473phosphorylation in response to intraperitoneal insulin in-jection in the livers of LTCFDN mice (Supplementary Fig.3M and N). Finally, we examined the hepatic expressionlevels of FoxO1, S256 FoxO1, b-cat, as well as S675 b-cat.No appreciable differences were observed between LTCFDNand WT littermates, suggesting that hepatic expression ofTCF7L2DN did not directly affect the expression or phos-phorylation levels of these two components, at least in theabsence of a challenge (Supplementary Fig. 4).
Ad-Mediated Expression of TCF7L2DN but Not WTTCF7L2 Increases Gluconeogenesis In VitroWe generated Ads to elicit the expression of WT TCF7L2 orTCF7L2DN in vitro (Fig. 3A). Treatment of adult C57BL/6hepatocytes with these viruses caused .95% cell infection(Supplementary Fig. 5A). Primary hepatocytes infected byAd-TCF7L2DN, but not Ad-TCF7L2WT, showed signifi-cantly increased glucose production (Fig. 3B) and the ex-pression of a panel of gluconeogenic genes, including Pck1,G6pc, Fbp1, and Ppargc1a (Fig. 3C). In addition, hepato-cytes infected with either Ad-TCF7L2WT or Ad-TCF7L2DNshowed significantly reduced expression of Axin2, a knownWnt pathway downstream target (Fig. 3D). The repressiveeffect of WT TCF7L2 on Axin2 expression has further in-dicated the bidirectional nature of TCF members.
b-Cat and FoxO1 Exert Opposite Effects on TOPflashExpressionThe function of TCF is largely determined by its interactionpartners, including b-cat. The TOPflash reporter system isa useful tool in assessing Wnt pathway activation or b-cat/TCF activity (40). Treatment of Hepa1-6 cells with Wnt-3a,which increases nuclear b-cat content, stimulated the ex-pression of TOPflash (Fig. 4A). Although S33Y b-catcotransfection also increased TOPflash activity, FoxO1cotransfection repressed the expression of TOPflash (Fig.4B). Interestingly, FoxO1-mediated repression of TOPflashwas not attenuated by S33Y b-cat cotransfection (Fig. 4B).
Peptide hormones such as GLP-1 and insulin stimulateb-cat S675 or S552 phosphorylation and hence increaseb-cat/TCF activity (10,41). We observed previously inmice that feeding increased hepatic TCF7L2 expressionwhile in vitro insulin treatment increased b-cat S675 phos-phorylation in hepatocytes (10). We show here that feedingalso increased hepatic b-cat S675 and S552 phosphoryla-tion in C57BL/6 mice (Fig. 4C and D), further supportingthe notion that b-cat/TCF mediates the hepatic function ofpeptide hormones in response to food consumption (10).
As the function of TCFs can be controlled by thecompetition between TCF members and FoxO proteins fora limited reservoir of b-cat, we verified the physical interac-tion between b-cat and FoxO1 in hepatocytes by coimmuno-precipitation. FoxO1 interacted with b-cat in mousehepatocytes as both b-cat and FoxO1 could be detected afterprecipitation of either b-cat or FoxO1 in the Hepa1-6 cell line(Fig. 4E). The above findings collectively indicate the role ofFoxO1 in interfering with b-cat/TCF activity in hepatocytes.
1926 TCF7L2DN Elevates Hepatic Gluconeogenesis Diabetes Volume 64, June 2015
Increased Binding of FoxO1 and b-Cat to the Pck1Gene Promoter in LTCFDN Hepatocytes
There is no functional TCF binding motif that has beenidentified within the 59 flanking region of the Pck1 gene,although a previous study (8) identified a TCF motifwithin the 39 region of the gene. A functional FoxO bind-ing site, however, is located within the proximal promoterregions of human and rodent Pck1 genes (Fig. 5A) (42,43).To further elucidate the mechanism underlying the in-creased levels of hepatic gluconeogenesis observed inthe LTCFDN mouse model, we assessed the binding ofFoxO1 and b-cat to the Pck1 gene promoter by ChIP.We detected binding of b-cat and FoxO1 to the Pck1 pro-moter in both WT and LTCFDN hepatocytes (Fig. 5B).Importantly, quantitative ChIP revealed that the interac-tions of FoxO1 and b-cat with the Pck1 promoter wereincreased in LTCFDN hepatocytes (Fig. 5C). These obser-vations collectively suggest that TCF7L2DN-mediated in-hibition of Wnt signaling causes preferential interactionof b-cat with FoxO1 and increased binding of b-cat/
FoxO1 to the Pck1 FoxO binding site, resulting in thestimulation of Pck1 expression.
To further verify the dual function of the b-cat mole-cule, we tested the effect of S33Y b-cat transfection andWnt-3a treatment directly on Pck1 promoter activity inHepa1-6 cells. The profound stimulatory effect generatedby b-cat cotransfection (Fig. 5D) indicates the importanceof the FoxO binding motif within the Pck1-LUC reporterconstruct. Wnt-3a treatment significantly repressed Pck1-LUC expression (Fig. 5E), which is consistent with ourobservations that Wnt-3a represses gluconeogenesis(Fig. 2H) and Pck1 mRNA levels in mouse hepatocytes(10). Together, these observations illustrate the repressiveeffect of Wnt signaling and b-cat/TCF on gluconeogene-sis, involving the competition between TCF and FoxO fortheir common cofactor b-cat as well as an undeterminedintrinsic repressive property of b-cat/TCF.
DISCUSSIONAlthough a few investigations (8,10,11,17) have suggesteda repressive effect of Wnt signaling activation or TCF7L2
Figure 2—LTCFDN mice exhibit progressive defects in HGP. LTCFDN mice exhibited progressive impairment of pyruvate tolerance, detectedat 9 (A), 23 (B), and 40 (C) weeks. LTCFDN mice demonstrated normal glucose tolerance at 10 weeks (D) but impaired glucose tolerance at 24weeks (E). LTCFDN mice showed no defect in response to insulin injection at 11 (F) and 41 (G) weeks. H: Hepatocytes from 12-week-oldLTCFDN mice exhibited increased basal glucose production and attenuation of Wnt-3a–mediated repression of glucose production. The cellswere pretreated with 2.5 nmol/L Wnt-3a for 1 h followed by the addition of the same dosage of Wnt-3a during the 2-h glucose productionassay. *P < 0.05; **P < 0.01; ***P < 0.001. Values represent the mean 6 SEM. See also Supplementary Fig. 3. n = 4–7 mice per group.
diabetes.diabetesjournals.org Ip and Associates 1927
itself on hepatic gluconeogenesis, a recent study (7) usingliver-specific TCF7L2 knockout mice and a TCF7L2 over-expression model presented an opposite view. Our studieshere were designed to address discrepancies regarding the
metabolic function of the Wnt signaling pathway as wellas the mechanistic role of TCF7L2 as a T2D risk gene. Wepropose that, because of the bidirectional function of theTCF molecules as well as the cross talk between Wnt and
Figure 3—Expression of TCF7L2DN but not WT TCF7L2 increases glucose production. A: Detection of exogenous WT TCF7L2 andTCF7L2DN in hepatocytes 36 h after Ad infection. Ad-TCF7L2DN but not Ad-TCF7L2WT infection increased glucose production (B)and gluconeogenic gene expression (C) in primary hepatocytes. D: Both Ad-TCF7L2DN and Ad-TCF7L2WT infection repressed theexpression of Axin2. *P < 0.05 in comparison with Ad-GFP. Values represent mean 6 SEM. n = 3–4 mice per hepatocyte group.
Figure 4—b-cat and FoxO1 exert opposite effects on TOPflash expression. A: Wnt-3a treatment increased the expression of TOPflash activity.B: S33Y b-cat and FoxO1 cotransfection generated stimulatory and repressive effects on TOPflash activity, respectively. Panels A and B arerepresentative data of three independent experiments performed with triplicate samples. C: Increased b-cat S675 and S552 phosphorylation inthe livers of fed C57BL/6 mice. D: Densitometric quantification of panel C. n = 3 fasted animals and n = 5 fed animals. E: b-Cat and FoxO1coimmunoprecipitate in Hepa1-6 cells. Representative data are from three independent experiments. IP, immunoprecipitation. *P < 0.05.
1928 TCF7L2DN Elevates Hepatic Gluconeogenesis Diabetes Volume 64, June 2015
the stress FoxO signaling pathway, knockout of a singleTCF member is an inadequate strategy to reveal the com-plete role of the Wnt signaling pathway in metabolic ho-meostasis. We thus conducted our in vitro and in vivoinvestigations primarily using a DN blocker, TCF7L2DN,and explored two other components in this pathway,b-cat and FoxO1. There are several advantages of ourLTCFDN mouse model. First, the albumin promoter isnot expressed in the fetus (Fig. 1D), and thus the poten-tial effect of TCF7L2DN on embryonic liver developmentis avoided. Second, the functional knockdown approachcircumvents the potential problems caused by the bidirec-tional and redundant actions of TCF7L2 and other TCFmembers, which would be experienced using regular over-expression or knockout approaches. Finally, this approach
allowed us to dissect the contribution of b-cat and FoxO1,as TCF7L2DN lacks the interaction motif for b-cat. Nev-ertheless, a relatively lower level of transgene expressionin our system is a potential drawback.
The LTCFDN transgenic mice generated for this studydeveloped a progressive impairment of pyruvate tolerancein the absence of appreciable whole-body or hepaticinsulin resistance (Fig. 2F and G and Supplementary Fig.3M and N), suggesting that liver-specific TCF7L2DN ex-pression may directly cause the elevation of hepatic glu-coneogenesis. Based on these results and those obtainedfrom the use of Ad-TCF7L2WT and Ad-TCF7L2DN inmouse primary hepatocytes, we conclude that the Wntsignaling pathway effector b-cat/TCF negatively regulateshepatic gluconeogenesis. It is necessary to note that, in
Figure 5—Increased binding of FoxO1 and b-cat to the Pck1 gene promoter in LTCFDN hepatocytes. A: Schematic of the mouse Pck1promoter, conserved FoxO binding site, and design of PCR primers used after ChIP. B: ChIP PCR showing binding of b-cat or FoxO1 to thePck1 FoxO binding site region but not the Pck1 intron 1 region. C: Quantitative ChIP shows that the interaction of b-cat and FoxO1 with theFoxO binding site on the Pck1 promoter is increased in LTCFDN hepatocytes. Pck1(2595/+67)-LUC activity is increased by S33Y b-cat (D)but is reduced by Wnt-3a (E). *P < 0.05. Values represent the mean 6 SEM. F: Schematic representation of the role of b-cat/TCF and Wntligands in regulating gluconeogenic gene expression. During fasting, TCF7L2 levels are relatively low, while glucagon stimulates gluco-neogenesis via increasing FoxO1 nuclear translocation and b-cat COOH-terminal phosphorylation. After feeding, TCF7L2 expressionincreases, and b-cat COOH-terminal phosphorylation is stimulated by insulin. This, along with the repressive effect of insulin on FoxO,leads to reduced gluconeogenesis. In addition, canonical Wnt ligands, such as Wnt-3a, repress gluconeogenesis via an undeterminedmechanism.
diabetes.diabetesjournals.org Ip and Associates 1929
the absence of pyruvate injection, basal plasma glucoselevels were similar in LTCFDN and WT littermates. Fur-thermore, LTCFDN mice remained glucose tolerant at 10weeks old (Fig. 2D). This is likely because of the existenceof compensatory mechanisms in controlling glucose ho-meostasis in the absence of challenge. Nevertheless, thenegative regulation of gluconeogenesis represents a novelmechanism by which metabolic hormones control hepaticglucose metabolism in response to physiological or nu-tritional changes. During fasting, glucagon, which waspreviously shown to increase hepatic b-cat S675 phos-phorylation (10), stimulates the stress signaling pathwayeffector FoxO1 and gluconeogenic gene expression whilehepatic TCF7L2 levels are relatively low. In response tofood intake, the level of insulin elevates. Insulin not onlyinactivates FoxO1 via the PI3K/Akt signaling cascade, butalso increases the expression of TCF7L2 and b-cat S675and S552 phosphorylation, leading to the repression ofhepatic gluconeogenic gene expression and gluconeogen-esis (Fig. 5F).
Genome-wide association studies revealed that certainTCF7L2 single nucleotide polymorphisms are stronglyassociated with susceptibility to T2D (1,44–46). Follow-ing this milestone discovery in diabetes research, themetabolic function of TCF7L2 has been intensively in-vestigated. Initial studies were primarily conducted inpancreas or pancreatic b-cells as impaired GLP-1–inducedinsulin secretion was observed in TCF7L2 T2D risk singlenucleotide polymorphism carriers (3). Many studies (3–5,13,36,47) suggested that TCF7L2 exerts beneficialeffects on b-cell proliferation, viability, and insulin secre-tion. Boj et al. (7), however, demonstrated very recentlythat b-cell–specific deletion of TCF7L2 in mice generatedno deleterious effects on these parameters, which is incontrast to the findings of a study presented by da SilvaXavier et al. (9), showing an impairment of glucose ho-meostasis and b-cell function in mice with b-cell–specificTcf7l2 deletion. This discrepancy can be attributed to sub-tle experimental details, such as the usage of differentb-cell–specific promoters (Pdx1 vs. Ins2) to drive Crerecombinase–mediated excision of LoxP-flanked Tcf7l2sequences in the b-cell lineage. However, we cannot ex-clude the possibility that, because of the extreme com-plexity of the Wnt signaling pathway, knockout oroverexpression of a given member may not always besufficient to reveal the true function of this signalingcascade. TCF7L2 and other TCF members may possesscertain compensatory functions as they carry very similarDNA binding and b-cat interaction domains. In addition,all TCF members can recruit either nuclear coactivators,such as CBP, or nuclear corepressors, such as Groucho andCtBP (48–50).
We have previously expressed TCF7L2DN in vitro aswell as in vivo to assess the role of Wnt signaling in gutand brain proglucagon gene expression, GLP-1 produc-tion, and function (12,34). Here, we further explored thecharacteristics of the TCF7L2DN protein in binding to the
consensus TCF binding motif by electrophoretic mobilityshift assay and the inability to recruit b-cat by coimmu-noprecipitation. In mouse pancreatic islets, liver, and adi-pocytes, there are two major isoforms of TCF7L2 of size78 and 58 kDa, both of which possess common DNA andb-cat interaction domains. Further functional explorationof these two major isoforms in hepatic metabolic homeo-stasis is necessary. Very recently, Takamoto et al. (36)generated a transgenic mouse model in which the shortisoform of TCF7L2DN (58 kDa) was driven by the Ins2gene promoter. These transgenic mice showed impairedglucose homeostasis and reduced b-cell mass and insulinsecretion. Thus, TCF7L2DN indeed represents a powerfulalternative approach to explore the metabolic functions ofb-cat/TCF and the Wnt signaling pathway.
We and others have investigated the metabolic func-tion of hepatic TCF7L2 during the past few years. Nortonet al. (8) found that the silencing of TCF7L2 in hepato-cytes induced a marked increase in basal glucose produc-tion and gluconeogenic gene expression, while TCF7L2overexpression reversed this phenotype and reduced glu-cose production. An in vivo study presented by Oh et al.(11) also suggested a crucial role of TCF7L2 in reducingHGP. We demonstrated the expression of three TCFmembers in mouse hepatocytes and revealed that hepaticTCF7L2 expression was stimulated by feeding in C57BL/6mice or by insulin treatment in hepatocytes in vitro, whileWnt-3a treatment repressed gluconeogenesis in mouseprimary hepatocytes (10). Most recently, Neve et al.(17) replicated our findings on the repressive effect ofTCF7L2 on gluconeogenic gene expression. Evidencefrom the current study using the unique LTCFDN mousemodel and Ad-TCF7L2DN expression in vitro furtherprompts us to conclude that Wnt and b-cat/TCF represshepatic gluconeogenesis in response to nutrient intake.
It is worth highlighting that our observations alsosuggest that the cofactor b-cat, but not TCFs, serves asa limiting factor in regulating hepatic Wnt activity, at leastin certain settings. Overexpression of WT TCF7L2 producedno appreciable effect, while TCF7L2DN generated a stimu-latory effect on gluconeogenic gene expression in primaryhepatocytes. Furthermore, although Ad-TCF7L2DN re-pressed Axin2 expression, Ad-TCF7L2WT also repressedAxin2 expression, supporting the notion of bidirectionalfunction of TCF members. Indeed, in the absence ofb-cat, TCFs are known to recruit nuclear corepressors(48–50).
Although Norton et al. (8) located a TCF binding motifwithin the Pck1 gene, this potential cis element is posi-tioned in the 39 region of the gene, and its contribution toPck1 gene expression remains unknown. Hence, we de-cided to conduct our investigation on the role of Wntsignaling and b-cat/TCF on Pck1 expression by assessingthe contribution of the known FoxO binding site withinthe proximal promoter region. The stimulatory effect ofS33Y b-cat and the repressive effect of Wnt-3a on Pck1promoter activity further indicate that b-cat cooperates
1930 TCF7L2DN Elevates Hepatic Gluconeogenesis Diabetes Volume 64, June 2015
with FoxO1 in stimulating hepatic gluconeogenesis (32).In addition, we speculate that Wnt activation by Wnt-3aexerts an intrinsic repressive effect on gluconeogenesisvia an undetermined atypical TCF binding site withinthe proximal promoter region, a hypothesis that requiresfurther exploration. Indeed, starvation reduced the inter-action between TCF7L2 and b-cat, and blunted the ex-pression of a panel of Wnt targets (32). In furthersupport of this notion, we demonstrated that b-cat accu-mulation was increased on the Pck1 promoter in LTCFDNhepatocytes. Interestingly, FoxO-mediated repression ofTOPflash reporter activity was not attenuated by S33Yb-cat cotransfection (Fig. 4B), suggesting that othermechanisms exist besides the competition between TCFand FoxO1 for the common cofactor b-cat (30,31). Onemay speculate that b-cat could recruit FoxO to the TCFbinding site, followed by nuclear corepressor recruitment.Alternatively, the repressive effect of FoxO may simply beoverwhelming in this particular in vivo reporter genetransfection system.
In summary, we applied a novel concept and a uniqueanimal approach to address an important yet controver-sial issue. Using TCF7L2DN as a tool for both in vivo andin vitro investigations, we demonstrated that the blockageof Wnt signaling led to increased gluconeogenesis andgluconeogenic gene expression. This observation was atleast partially attributed to the increased occupancy of thegluconeogenic gene promoter Pck1 by b-cat/FoxO1.
Funding. W.I. is a recipient of the Canadian Institutes of Health ResearchDoctoral Canada Graduate Scholarship, Ontario Graduate Scholarship, and theBanting & Best Diabetes Centre-University Health Network Graduate Award. Thiswork was supported by operating grants from the Canadian Institutes of HealthResearch to M.B.W. (MOP-102588) and T.J. (MOP-89987 and MOP-97790).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. W.I. designed and conducted the experiments,analyzed the data, and wrote the manuscript. W.S., Z.S., and Z.C. conducted theexperiments. M.B.W. provided research material, assisted with the experimentaldesign, and edited the manuscript. T.J. designed and conducted the experimentsand wrote the manuscript. T.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. Grant SF, Thorleifsson G, Reynisdottir I, et al. Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet 2006;38:320–3232. Lyssenko V, Lupi R, Marchetti P, et al. Mechanisms by which commonvariants in the TCF7L2 gene increase risk of type 2 diabetes. J Clin Invest 2007;117:2155–21633. Schäfer SA, Tschritter O, Machicao F, et al. Impaired glucagon-like peptide-1-induced insulin secretion in carriers of transcription factor 7-like 2 (TCF7L2)gene polymorphisms. Diabetologia 2007;50:2443–24504. Shu L, Sauter NS, Schulthess FT, Matveyenko AV, Oberholzer J, Maedler K.Transcription factor 7-like 2 regulates beta-cell survival and function in humanpancreatic islets. Diabetes 2008;57:645–6535. da Silva Xavier G, Loder MK, McDonald A, et al. TCF7L2 regulates lateevents in insulin secretion from pancreatic islet beta-cells. Diabetes 2009;58:894–905
6. Savic D, Ye H, Aneas I, Park SY, Bell GI, Nobrega MA. Alterations in TCF7L2expression define its role as a key regulator of glucose metabolism. Genome Res2011;21:1417–14257. Boj SF, van Es JH, Huch M, et al. Diabetes risk gene and Wnt effectorTcf7l2/TCF4 controls hepatic response to perinatal and adult metabolic demand.Cell 2012;151:1595–16078. Norton L, Fourcaudot M, Abdul-Ghani MA, et al. Chromatin occupancy oftranscription factor 7-like 2 (TCF7L2) and its role in hepatic glucose metabolism.Diabetologia 2011;54:3132–31429. da Silva Xavier G, Mondragon A, Sun G, et al. Abnormal glucose toleranceand insulin secretion in pancreas-specific Tcf7l2-null mice. Diabetologia 2012;55:2667–267610. Ip W, Shao W, Chiang YT, Jin T. The Wnt signaling pathway effector TCF7L2is upregulated by insulin and represses hepatic gluconeogenesis. Am J PhysiolEndocrinol Metab 2012;303:E1166–E117611. Oh KJ, Park J, Kim SS, Oh H, Choi CS, Koo SH. TCF7L2 modulates glucosehomeostasis by regulating CREB- and FoxO1-dependent transcriptional pathwayin the liver. PLoS Genet 2012;8:e100298612. Shao W, Wang D, Chiang YT, et al. The Wnt signaling pathway effectorTCF7L2 controls gut and brain proglucagon gene expression and glucose ho-meostasis. Diabetes 2013;62:789–80013. Shu L, Zien K, Gutjahr G, et al. TCF7L2 promotes beta cell regeneration inhuman and mouse pancreas. Diabetologia 2012;55:3296–330714. Chiang YT, Ip W, Shao W, Song ZE, Chernoff J, Jin T. Activation of GLP-1signaling attenuates impaired hepatic glucose disposal in aged male p21-activated protein kinase-1 knockout mice. Endocrinology 2014;155:2122–213215. Liu Z, Habener JF. Wnt signaling in pancreatic islets. Adv Exp Med Biol2010;654:391–41916. Liu Z, Habener JF. Stromal cell-derived factor-1 promotes survival ofpancreatic beta cells by the stabilisation of beta-catenin and activation of tran-scription factor 7-like 2 (TCF7L2). Diabetologia 2009;52:1589–159817. Neve B, Le Bacquer O, Caron S, et al. Alternative human liver transcripts ofTCF7L2 bind to the gluconeogenesis regulator HNF4a at the protein level. Dia-betologia 2014;57:785–79618. McCarthy MI, Rorsman P, Gloyn AL. TCF7L2 and diabetes: a tale of twotissues, and of two species. Cell Metab 2013;17:157–15919. Nobrega MA. TCF7L2 and glucose metabolism: time to look beyond thepancreas. Diabetes 2013;62:706–70820. Murtaugh LC, Law AC, Dor Y, Melton DA. Beta-catenin is essential forpancreatic acinar but not islet development. Development 2005;132:4663–467421. Lehwald N, Tao GZ, Jang KY, Sorkin M, Knoefel WT, Sylvester KG. Wnt-beta-catenin signaling protects against hepatic ischemia and reperfusion injury inmice. Gastroenterology 2011;141:707–718, 718.e1–e522. Lehwald N, Tao GZ, Jang KY, et al. b-Catenin regulates hepatic mitochon-drial function and energy balance in mice. Gastroenterology 2012;143:754–76423. Tao GZ, Lehwald N, Jang KY, et al. Wnt/b-catenin signaling protects mouseliver against oxidative stress-induced apoptosis through the inhibition of forkheadtranscription factor FoxO3. J Biol Chem 2013;288:17214–1722424. Zhang XF, Tan X, Zeng G, et al. Conditional beta-catenin loss in micepromotes chemical hepatocarcinogenesis: role of oxidative stress and platelet-derived growth factor receptor alpha/phosphoinositide 3-kinase signaling. Hep-atology 2010;52:954–96525. Debies MT, Gestl SA, Mathers JL, et al. Tumor escape in a Wnt1-dependentmouse breast cancer model is enabled by p19Arf/p53 pathway lesions but notp16 Ink4a loss. J Clin Invest 2008;118:51–6326. Nejak-Bowen K, Kikuchi A, Monga SP. Beta-catenin-NF-kB interactions inmurine hepatocytes: a complex to die for. Hepatology 2013;57:763–77427. Clevers H, Nusse R. Wnt/b-catenin signaling and disease. Cell 2012;149:1192–120528. Rajagopal J, Carroll TJ, Guseh JS, et al. Wnt7b stimulates embryonic lunggrowth by coordinately increasing the replication of epithelium and mesenchyme.Development 2008;135:1625–1634
diabetes.diabetesjournals.org Ip and Associates 1931
29. Essers MA, de Vries-Smits LM, Barker N, Polderman PE, Burgering BM,Korswagen HC. Functional interaction between beta-catenin and FOXO in oxi-dative stress signaling. Science 2005;308:1181–118430. Manolagas SC, Almeida M. Gone with the Wnts: beta-catenin, T-cell factor,forkhead box O, and oxidative stress in age-dependent diseases of bone, lipid,and glucose metabolism. Mol Endocrinol 2007;21:2605–261431. Iyer S, Ambrogini E, Bartell SM, et al. FOXOs attenuate bone formation bysuppressing Wnt signaling. J Clin Invest 2013;123:3409–341932. Liu H, Fergusson MM, Wu JJ, et al. Wnt signaling regulates hepatic me-tabolism. Sci Signal 2011;4:ra633. Pinkert CA, Ornitz DM, Brinster RL, Palmiter RD. An albumin enhancer lo-cated 10 kb upstream functions along with its promoter to direct efficient, liver-specific expression in transgenic mice. Genes Dev 1987;1:268–27634. Yi F, Brubaker PL, Jin T. TCF-4 mediates cell type-specific regulation ofproglucagon gene expression by beta-catenin and glycogen synthase kinase-3beta. J Biol Chem 2005;280:1457–146435. Kolligs FT, Hu G, Dang CV, Fearon ER. Neoplastic transformation of RK3E bymutant beta-catenin requires deregulation of Tcf/Lef transcription but not acti-vation of c-myc expression. Mol Cell Biol 1999;19:5696–570636. Takamoto I, Kubota N, Nakaya K, et al. TCF7L2 in mouse pancreatic betacells plays a crucial role in glucose homeostasis by regulating beta cell mass.Diabetologia 2014;57:542–55337. Shao W, Wang Z, Ip W, et al. GLP-1(28-36) improves b-cell mass andglucose disposal in streptozotocin-induced diabetic mice and activates cAMP/PKA/b-catenin signaling in b-cells in vitro. Am J Physiol Endocrinol Metab 2013;304:E1263–E127238. Ip W, Shao W, Chiang YT, Jin T. GLP-1-derived nonapeptide GLP-1(28-36)amide represses hepatic gluconeogenic gene expression and improves py-ruvate tolerance in high-fat diet-fed mice. Am J Physiol Endocrinol Metab 2013;305:E1348–E135839. Wang P, Wang Q, Sun J, et al. POU homeodomain protein Oct-1 functionsas a sensor for cyclic AMP. J Biol Chem 2009;284:26456–26465
40. Korinek V, Barker N, Morin PJ, et al. Constitutive transcriptional activation bya beta-catenin-Tcf complex in APC-/- colon carcinoma. Science 1997;275:1784–178741. Liu Z, Habener JF. Glucagon-like peptide-1 activation of TCF7L2-dependent Wntsignaling enhances pancreatic beta cell proliferation. J Biol Chem 2008;283:8723–873542. Hall RK, Yamasaki T, Kucera T, Waltner-Law M, O’Brien R, Granner DK.Regulation of phosphoenolpyruvate carboxykinase and insulin-like growth factor-binding protein-1 gene expression by insulin. The role of winged helix/forkheadproteins. J Biol Chem 2000;275:30169–3017543. Nakae J, Kitamura T, Silver DL, Accili D. The forkhead transcription factorFoxo1 (Fkhr) confers insulin sensitivity onto glucose-6-phosphatase expression.J Clin Invest 2001;108:1359–136744. Sladek R, Rocheleau G, Rung J, et al. A genome-wide association studyidentifies novel risk loci for type 2 diabetes. Nature 2007;445:881–88545. Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide associationstudy of type 2 diabetes in Finns detects multiple susceptibility variants. Science2007;316:1341–134546. Saxena R, Voight BF, Lyssenko V, et al.; Diabetes Genetics Initiative ofBroad Institute of Harvard and MIT, Lund University, and Novartis Institutes ofBioMedical Research. Genome-wide association analysis identifies loci for type 2diabetes and triglyceride levels. Science 2007;316:1331–133647. Shu L, Matveyenko AV, Kerr-Conte J, Cho JH, McIntosh CH, Maedler K.Decreased TCF7L2 protein levels in type 2 diabetes mellitus correlate withdownregulation of GIP- and GLP-1 receptors and impaired beta-cell function.Hum Mol Genet 2009;18:2388–239948. Brannon M, Brown JD, Bates R, Kimelman D, Moon RT. XCtBP is a XTcf-3co-repressor with roles throughout Xenopus development. Development 1999;126:3159–317049. Brantjes H, Roose J, van De Wetering M, Clevers H. All Tcf HMG boxtranscription factors interact with Groucho-related co-repressors. Nucleic AcidsRes 2001;29:1410–141950. Roose J, Molenaar M, Peterson J, et al. The Xenopus Wnt effector XTcf-3interacts with Groucho-related transcriptional repressors. Nature 1998;395:608–612
1932 TCF7L2DN Elevates Hepatic Gluconeogenesis Diabetes Volume 64, June 2015
