The developmental Wnt signaling pathway effector β-catenin/TCF mediates hepatic functions of the sex hormone estradiol in regulating lipid metabolism

The bipartite transcription factor β-catenin (β-cat)/T cell factor (TCF), formed by free β-cat and a given TCF family member, serves as the effector of the developmental Wnt signaling cascade. β-cat/TCFs also serve as effectors of certain peptide hormones or growth factors during adulthood. We reported that liver-specific expression of dominant-negative Transcription factor 7 like 2 (TCF7L2DN) led to impaired glucose disposal. Here we show that, in this LTCFDN transgenic mouse model, serum and hepatic lipid contents were elevated in male but not in female mice. In hepatocytes, TCF7L2DN adenovirus infection led to stimulated expression of genes that encode lipogenic transcription factors and lipogenic enzymes, while estradiol (E2) treatment attenuated the stimulation, associated with Wnt-target gene activation. Mechanistically, this E2-mediated activation can be attributed to elevated β-cat Ser675 phosphorylation and TCF expression. In wild-type female mice, ovariectomy (OVX) plus high-fat diet (HFD) challenge impaired glucose disposal and insulin tolerance, associated with increased hepatic lipogenic transcription factor sterol regulatory element-binding protein 1-c (SREBP-1c) expression. In wild-type mice with OVX, E2 reconstitution attenuated HFD-induced metabolic defects. Some of the attenuation effects, including insulin intolerance, elevated liver-weight gain, and hepatic SREBP-1c expression, were not affected by E2 reconstitution in HFD-fed LTCFDN mice with OVX. Finally, the effects of E2 in hepatocytes on β-cat/TCF activation can be attenuated by the G-protein-coupled estrogen receptor (GPER) antagonist G15. Our study thus expanded the scope of functions of the Wnt pathway effector β-cat/TCF, as it can also mediate hepatic functions of E2 during adulthood. This study also enriches our mechanistic understanding of gender differences in the risk and pathophysiology of metabolic diseases.

Introduction Transcription factor 7 like 2 (TCF7L2, Tcf7l2 in rodents) and the other three T cell factor (TCF)/Tcf family members (Tcf7, Tcf7l1, and lymphoid enhancer binding factor 1 [Lef1]) are principal effectors of the canonical Wnt signalling cascade (referred to as the Wnt pathway hereafter) in mammals. When a TCF family member is partnered with the free β-catenin (βcat) molecule, the formed bipartite transcription factor β-cat/TCF regulates expression of the Wnt pathway downstream target genes, in response to the stimulation by certain canonical Wnt ligands. Investigations from different disciplines have been continuing to uncover the complexity of this signaling cascade, including the existence of multiple Wnt ligands with common or unique functions, receptors, coreceptors, membrane-bound or soluble positive and negative regulators, alternative signaling pathways, and, even more importantly, the bidirectional feature of TCF family members [1][2][3][4][5][6].
Previous investigations by our team and by others have shown that certain peptide hormones and growth factors, such as insulin, insulin-like growth factor 1 (IGF-1), and glucagonlike peptide-1 (GLP-1), may exert their functions via cross-talking with the Wnt signaling cascade. This can be achieved via stimulating β-cat phosphorylation at the C-terminal Ser675 and Ser552 residues [1,[7][8][9][10][11]. We have therefore suggested that the developmental Wnt signaling cascade effector β-cat/TCF is also among the effectors of certain peptide hormones and growth factors during adulthood [1,8].
We have reported that mouse or human hepatocytes express TCF7, TCF7L1, and TCF7L2 but not LEF1 [9]. Among them, TCF7L2 has been recognized as a type 2 diabetes (T2D) risk gene, revealed by reproducible genome-wide association studies (GWASs) [1,12,13]. Despite intensive global effort for more than a decade, we are still unable to unite our view of whether TCF7L2 variants in T2D risk subjects represent gain-of-function or loss-of-function events [1,[14][15][16]. It still remains unknown whether these T2D risk TCF7L2 single nucleotide polymorphisms (SNPs) indeed affect TCF7L2 expression in any given metabolic organs [1]. A recent study suggested that the genomic location harbouring one of the SNPs, rs7903146, may serve as a regulatory region of the gene that encodes long-chain fatty acid CoA ligase 5 (ACSL5) [17]. Nevertheless, the investigations have deepened our knowledge of metabolic functions of Wnt signaling cascade in metabolic organs in general, with the generation and assessment of various transgenic mouse models [15,[18][19][20][21][22][23].
To overcome the redundant function of TCFs and the bidirectional feature of a given TCF family member, we and others have adopted a powerful functional knockdown approach in conducting in vitro investigations as well as in generating transgenic mouse models [20,21,[24][25][26][27][28][29]. In the LTCFDN mouse model that we have generated, expression of dominant-negative TCF7L2 (TCF7L2DN) is driven by the mouse albumin promoter [26]. The LTCFDN transgenic mice fed with a regular low-fat diet (LFD) exhibited progressive impairment on pyruvate and glucose tolerance, in the absence of the development of insulin resistance. Furthermore, LTCFDN hepatocytes showed elevated glucose production and attenuated response to in vitro Wnt-3a-treatment-induced repression on glucose production [26].
Wnt signaling cascade is also involved in lipid homeostasis. Certain mutations on the Wnt ligand coreceptor lipoprotein receptor-related protein 6 (LRP6) are associated with the susceptibility of dyslipidemia [30], whereas in mice, Lrp6-deficiency-induced hyperlipidemia can be partially reversed by Wnt-3a administration [31]. The direct involvement of β-cat/TCF on lipid homeostasis or hepatic lipogenesis, however, remains largely unknown.
In the current study, we found that male LTCFDN mice exhibited elevated triglyceride (TG) levels and increased hepatic lipid deposition. Such increases were absent in female LTCFDN mice, suggesting that female hormones might be able to overcome functional attenuation on β-cat/TCF activity. Utilizing the ovariectomy (OVX) surgery procedure along with high-fat diet (HFD) feeding and the female hormone estradiol (E2) reconstitution, we obtained evidence that the attenuation effect of E2 on insulin intolerance and hepatic lipogenic program activation in response to HFD challenge requires the intact Wnt/β-cat signaling cascade in the liver.

Male LTCFDN mice show elevated hepatic and serum TG levels and liver tissue fat deposition
To determine whether the Wnt signaling pathway effector β-cat/TCF is also involved in hepatic lipid homeostasis, we have first of all examined several lipid parameters in LTCFDN mice fed with the regular LFD. In a set of 2-week-old LTCFDN mice (regardless of their genders), we observed that hepatic TG content was increased ( Fig 1A) while their hepatic free fatty acid (FFA) level showed a trend of increase ( Fig 1B). The 2-week-old mice, however, carried no appreciable "abnormalities" on their body weight, liver weight, random glucose, serum TG, FFA, or cholesterol levels (S1 Fig).
In 12-week-old male LTCFDN mice, we found that hepatic TG content was elevated when compared with that in age-and sex-matched littermate controls (Fig 1C), although their hepatic FFA levels were comparable ( Fig 1D). Male LTCFDN (12 weeks old) also showed elevated serum TG levels during fasting but not after feeding (Fig 1E), in association with lack of a defect on intraperitoneal insulin tolerance test (IPITT) that we have reported previously [26]. Serum FFA level was increased significantly after feeding and with a trend of elevation after fasting ( Fig 1F). Serum cholesterol levels were comparable in male LTCFDN and sex-matched wild-type mice (Fig 1G). Lipid deposition in the liver tissue sections of LTCFDN male mice (24 weeks) was greatly increased compared with that in sex-and age-matched littermate controls (Fig 1H and 1I). Thus, liver-specific functional knockdown of the Wnt signaling cascade with a dominant-negative TCF molecule resulted in elevated serum and hepatic lipid contents and increased hepatic lipid deposition. We also found that hepatic glycogen content in 12-weekold male LTCFND mice showed a trend of increase, whereas in mouse primary hepatocytes (MPHs), adenovirus Ad-TCF7L2DN infection but not Ad-TCF7L2 infection increased expression of glycogen synthase 2 (Gys2) (S2 Fig).

TCF7L2DN expression increases lipogenic gene expression
Taking advantage of possessing the TCF7L2DN adenovirus [26], we then conducted further functional knockdown experiments in vitro in the human HepG2 cell line and in MPH. Cells were infected with either the control virus green-fluorescence-protein-expressing adenovirus (Ad-GFP) or Ad-TCF7L2DN, verified by western blotting for hemagglutinin (HA) tag detection ( Fig 1J and S3A Fig). Forty-eight hours after virus infection, cells were harvested for assessing lipogenic gene expression. As shown (S3B and S3C Fig), in HepG2 cells following Ad-TCF7L2DN infection, expression of genes that encode the 2 key lipogenic transcription factors-carbohydrate-responsive element-binding protein (ChREBP, Mlxipl) and sterol regulatory element-binding protein 1-c (SREBP-1c, srebf1)-were increased. In MPH, Ad-TCF7L2DN infection also increased expression of both Mlxipl and Srebf1 (Fig 1K and 1L), accompanied by increased expression of hepatic genes that encode the 5 lipogenic enzymes (Fig 1M), known as acetyl-coA carboxylase (Acc1), enoyl-coA hydratase and 3-hydroxyacyl coA dehydrogenase (also known as peroxisomal L-bifunctional enzyme [Ehhadh]), fatty acid synthase (Fas), cytosolic malic enzyme 1 (Me1), and stearoyl-coA desaturase-1 (Scd1). Female LTCFDN mice show virtually no elevation on serum and hepatic lipid contents As the crosstalk between the Wnt/β-cat/TCF signaling cascade and the estrogen signaling has been indicated in both cancer and developmental biology fields [32,33], we determined to assess whether the defect on glucose disposal and lipid metabolism observed in male LTCFDN mice is different in female LTCFDN mice. As shown in Fig 2A, without HFD challenge, male and female LTCFDN mice had comparable fasting blood glucose levels. We then conducted the 3 common intraperitoneal metabolic tolerance tests. As shown (Fig 2B and 2C), both male and female LTCFDN mice (20 weeks) exhibited impaired glucose disposal after glucose challenge, although it appeared that the impairment was slightly worse in male mice. Impaired pyruvate tolerance was also observed in both male and female LTCFDN mice (Fig 2D and 2E). In agreement with our previous observations [26], LTCFDN mice on an LFD show a lack of insulin intolerance, assessed by IPITT, regardless of gender (Fig 2F and 2G).
We then compared hepatic and serum TG levels in male and female LTCFDN mouse littermates. As shown, male LTCFDN mice showed elevated serum ( Fig 3A) and hepatic ( Fig 3B) TG contents. However, comparable plasma and hepatic TG levels were found in age-matched female littermate mice, regardless of their genotypes (Fig 3A and 3B). Fig 3C shows that random serum FFA level in female LTCFDN mice was also not higher than that in sex-matched wild-type littermate controls. Furthermore, increased liver lipid deposition was also not observed in female LTCFDN mice (24 weeks, Fig 3D and 3E). Finally, increased hepatic ChREBP (Mlxipl) and SREBP-1c (Srebf1) expression observed in male LTCFDN mouse liver was also not observed in female LTCFDN mouse liver (Fig 3F and 3G). These observations collectively suggest that female mice are more tolerant to hepatic TCF7L2DN-expressioninduced Wnt signaling cascade functional knockdown on lipid homeostasis.
To assess the involvement of the female hormone E2, we then conducted daily E2 injection in 6-week-old male LTCFDN mice for 6 days, with age-and sex-matched wild-type littermates as controls. As shown, the injection attenuated hepatic SREBP-1c mRNA level, associated with attenuated fasting serum TG level (Fig 3H and 3I).

TCF7L2DN-induced lipogenic gene expression in hepatocytes can be attenuated by E2 treatment
The human HepG2 cell line was then infected with either the control virus Ad-GFP or Ad-TCF7L2DN. Forty-eight hours after virus infection, cells were treated with or without 100 nM E2 for 6 hours, followed by cell harvesting and quantitative reverse transcription polymerase chain reaction (qRT-PCR) analyses. As shown (S4 Fig), TCF7L2DN-induced Mlxipl and Srebf1 elevations can be attenuated by E2 treatment. In Ad-GFP-infected cells, E2 treatment generated no appreciable effect on expression of these 2 lipogenic transcription factors.

E2 treatment increases β-cat Ser675 phosphorylation and Tcf gene expression
To explore the underlying mechanism for the rescue effect of E2 treatment in response to TCF7L2DN-mediated functional knockdown; we tested the in vitro effect of E2 in MPH. We found that 6-hour E2 treatment stimulated β-cat Ser675 phosphorylation (Fig 4A), and the stimulation increased along with the treatment time within the 120-minute experimental time ( Fig 4B). E2 treatment also transiently increased cAMP response element-binding protein (CREB) phosphorylation (Ser133) (at 15 minutes, S6 Fig). Fig 4C-4E shows that in MPH, 6-hour E2 treatment resulted in elevated mRNA levels of the 3 Tcf gene members, Tcf7l2, Tcf7l1, and Tcf7, which are known to be expressed in the mouse liver [9], and the elevation sustained for 24-hour experimental period. Importantly, the elevations were associated with increased expression of axis inhibition protein 2 (Axin2), a known Wnt signaling pathway downstream target gene, especially after 24-hour E2 treatment ( Fig 4F). Finally, we found that E2 treatment increased the activity of the β-cat/TCF responsive Top-Flash luciferase (LUC) fusion gene plasmid, although the activation appeared relatively modest in the utilized 293 naïve cell system ( Fig 4G).

OVX accelerates HFD-induced metabolic defects
Intensive previous investigations in rodent species have attributed the effect of E2 on attenuating HFD-induced body weight gain mainly to the function of this female hormone in adipose tissues, the muscle tissue, and the brain [34][35][36]. To investigate the involvement of the liver, we have first of all conducted a set of experiments in wild-type female mice with HFD feeding and the OVX surgery ( Fig 5A). Fig 5B shows that with the sham surgery, HFD feeding increased body weight from the second to the fourth week, but not after the fifth week. In LFD-fed mice, OVX surgery itself (without HFD challenge) increased the body weight starting from the second week. Importantly, OVX surgery significantly increased the body weight of mice that received HFD challenge, starting at the second week and continuing during the entire 6-week experimental period. These observations collectively suggest that, unlike male mice, female mice possess a mechanism in protecting body weight gain in response to HFD consumption, and the mechanism involves the ovarian gland.
IPITT and intraperitoneal glucose tolerance test (IPGTT) were then conducted during the fifth week and the sixth week, respectively. As shown, impaired glucose disposal and insulin tolerance were observed for HFD-fed mice with the OVX surgery (Fig 5C and 5D). OVX surgery itself or HFD feeding also slightly impaired glucose and insulin tolerance (Fig 5C and

The restoration effects of E2 reconstitution are impaired in LTCFDN mice
In order to determine the pathophysiological importance of the crosstalk between E2 and βcat/TCF in the liver, we have then directly compared the restoration effect of E2 administration in wild-type and LTCFDN female mice after HFD challenge and OVX surgery, with assays illustrated in Fig 6A. In wild-type littermates, E2 administration attenuated the effect of OVX surgery and HFD challenge on body weight gain, started at the second week (Fig 6B), improved glucose disposal (assessed by IPITT at the sixth week) (Fig 6C), reduced the finalday body weight, and reduced the liver as well as the abdominal fat weight (Fig 6D and 6E). In LTCFDN mice, the attenuation effect of E2 on body weight gain was found to be delayed until the third week ( Fig 6F). More importantly, E2 administration did not effectively restore the defect on glucose disposal, assessed by IPITT (Fig 6G), and had no appreciable effect on reducing the final-day body weight, although E2 administration also reduced the liver weight as well as abdominal fat volume in LTCFDN mice (Fig 6H and 6I). Fig 6J and 6K show that E2 administration did not attenuate serum or hepatic TG levels in LTCFDN as it did in the wild-type littermates, although absolute hepatic TG levels were not elevated in the transgenic mice at the 6-week experimental ending stage. Serum and hepatic FFA levels, however, were comparable in wild-type littermates and LTCFDN mice, regardless of systematic E2 administration (S8 Fig). In wild-type littermate controls, E2 administration attenuated hepatic SREBP-1c and ChREBP expression levels (Fig 6L), while such attenuation was not observed in LTCFDN female mice with OVX surgery (Fig 6M). Furthermore, we assessed expression of the 5 hepatic lipogenic enzyme genes in these mice. As shown, their expression levels were reduced in wildtype littermates but not in LTCFDN mice, following HFD challenge, OVX, and E2 administration (Fig 6N and 6O).

The metabolic effects of E2 in cultured hepatocytes can be blocked by the G-protein-coupled estrogen receptor antagonist
In addition to the nuclear receptors ERα and ERβ, estrogens, including E2, may utilize G-protein-coupled estrogen receptor (GPER; also known as G-protein-coupled receptor 30 [GPR30]) to exert its regulatory effects [37]. The nongenomic effects of E2, mediated by GPER, include adenylyl cyclase (AC) activation and cAMP elevation in the liver [38]. We have therefore tested the possible involvement of GPER with a GPER antagonist and an agonist. As Underlying numerical values can be found in S1 Data. AUC, area under the curve; ChREBP, carbohydrate-responsive element-binding protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HFD, high-fat diet; i.p., intraperitoneal; IPGTT, intraperitoneal glucose tolerance test; IPITT, intraperitoneal insulin tolerance test; LFD, low-fat diet; OVX, ovariectomy; p-ACC, phosphorylated acetyl-coA carboxylase; PKB, protein kinase B; SREBP-1c, sterol regulatory element-binding protein 1-c.
https://doi.org/10.1371/journal.pbio.3000444.g005 β-cat/TCF mediates metabolic function of estradiol in the liver shown in HepG2 cells, E2 treatment (100 nM for 120 minutes)-stimulated β-cat Ser675 phosphorylation can be blocked by the GPER antagonist G15 (Fig 7A and 7B). In conducting this experiment, we also observed a trend increase of CREB Ser133 phosphorylation level by 120-minute E2 treatment, while G15 treatment attenuated the increase (S9 Fig). In HepG2 cells as well as in MPH, E2 treatment increased cellular cAMP levels, while the increase was blocked by G15 treatment (Fig 7C and 7D). Furthermore, the GPER agonist G1 treatment showed a similar stimulatory effect on cellular cAMP elevation to that of E2 treatment (Fig 7C  and 7D).
In response to a challenge, E2 may control energy intake and improve insulin signaling via its effects in the brain, adipose tissue, skeletal muscles, pancreatic islets, macrophages, and elsewhere, as documented in the literature. In hepatocytes, functional attenuation of β-cat/TCF activity leads to lipogenic program activation. E2, however, can activate β-cat, possibly via GPER, followed by protein kinase A (PKA) activation and increased β-cat Ser675 phosphorylation. E2 can also stimulate expression of the 3 TCF family members, via yet to be determined mechanisms with the participation of estrogen receptor α/β (ERα/β) or GPER, or both. The combined effect on β-cat post-translational modification and TCF gene expression leads to the repression of lipogenic transcription factor expression and reduced hepatic lipogenesis.
We then tested whether E2-stimulated TCF expression can be attenuated by G15 pretreatment. In the HepG2 cell line, G15 itself had no appreciable effect on TCF7L2 or TCF7 mRNA expression, while it blocked the stimulatory effect of E2 treatment (Fig 7E and 7F). Similarly, E2-stimulated Axin2 expression was also blocked by G15 pretreatment (Fig 7G). The blockage effect of G15 on E2-stimulated Tcf7l2 was then observed in MPH (Fig 7H). These observations collectively suggest the involvement of GPER in mediating the effect of E2 on β-cat/TCF activity, although we cannot exclude the potential involvement of ERα/β at this stage.
https://doi.org/10.1371/journal.pbio.3000444.g006 β-cat/TCF mediates metabolic function of estradiol in the liver Following the discovery that TCF7L2 is a T2D risk gene by GWAS [12,13], great efforts have been made to define its metabolic functions in various metabolic organs, including pancreatic islets, liver, brain, and adipose tissues. A few studies have suggested that T2D risk TCF7L2 SNPs represent gain-of-function events and TCF7L2 may stimulate hepatic gluconeogenesis [14,15,19]. Other investigations, including a few of them conducted by our team, however, presented a completely opposite view. Utilizing both cell culture systems and various mouse models, these investigators concluded that Wnt pathway activation represses hepatic gluconeogenesis [9,16,22,26,47,48]. We have suggested previously that β-cat/TCF mediates the function of postprandial insulin elevation on repressing hepatic gluconeogenesis [9]. Severe discrepancy in the literature exists regarding the assessment of pancreatic islet Tcf7l2 knockout mouse models as well [15,21,23,25,49,50].
We suggest that the discrepancy is mainly due to the redundant and overlapping functions of TCF family members and the existence of multiple isoforms for each of the members [1,26]. More importantly, a given TCF is only a "half" effector of Wnt signaling cascade activation. The genetic regulatory function of TCF7L2 or other TCF family members is actually bidirectional, depending on the availability and post-translational modifications of its partner, the βcat molecule [1]. Thus, to simply knock out a given component of this complicated signaling cascade in mice may not always reveal the exact function of this signaling cascade at physiological and pathophysiological settings.
β-cat serves as the common partner of the 4 TCF family members via interacting with their common β-cat interaction domain. It is well known that β-cat N-terminal phosphorylation (at Ser/Thr 33, 37, 41, 45 residues) lead to its proteasome-mediated degradation, involving glycogen synthase kinase-3 (GSK-3) and another protein kinase casein kinase Iα (CKIα) [51]. The latter one serves as the priming kinase [51]. β-cat can also undergo post-translational modifications on other amino acid residues by various means, and these modifications can be controlled by metabolic hormones, growth factors, and other regulatory molecules, resulting in various functional outcomes [1,8,52]. The utilization of TCF7L2DN, lacking of the β-cat interaction domain, should result in functional knockdown of β-cat/TCF activity in general, avoiding the drawbacks generated by simply knocking out a given TCF family member [1,24,27]. By assessing our LTCFDN mouse model, we have reported previously the role of hepatic β-cat/ TCF in down-regulating glucose production [26] and present here that hepatic β-cat/TCF also controls hepatic lipid metabolism. More importantly, our current study with the LTCFDN mouse model suggests the existence of crosstalk between β-cat/TCF and female hormones.
https://doi.org/10.1371/journal.pbio.3000444.g007 β-cat/TCF mediates metabolic function of estradiol in the liver the ovarian gland. When both OVX surgery and HFD challenge were applied, E2 reconstitution effectively restored the metabolic defects in wild-type mice but not in the LTCFDN mice (Fig 6). Thus, intact hepatic β-cat/TCF activity is required for systematic E2 administration to protect female subjects in the development of metabolic defects in response to challenge, such as HFD feeding. Our observations thus open a new window in understanding gender differences in the risk of diabetes and other metabolic diseases.
Susceptibilities to metabolic diseases, including T2D, fatty liver disease, and cardiovascular diseases, vary between the genders [53]. With increased life expectancy, women are now spending over 3 decades of life of post menopause, associated with increased susceptibility to metabolic diseases [54]. Menopause hormone therapy has been suggested to delay the onset of T2D [55][56][57]. Mechanistic explorations to date have mainly implicated the role of estrogens and their receptors (ERα and ERβ) in regulating fuel homeostasis in response to HFD challenge to the suppression of energy intake, adipose tissue accumulation, and improvement of insulin secretion or its sensitivity in the muscle [34][35][36]. Much less effort has been made on the contribution of the liver, although the nuclear receptor of E2, ERα, was shown to regulate expression of hepatic genes that are involved in de novo lipogenesis (DNL) [58].
In the current study, we observed elevated hepatic and plasma lipid contents in male LTCFDN mice, indicating the hepatic function of the Wnt signaling cascade or its effector βcat/TCF on lipid homeostasis, in agreement with intensive studies on the implication of genetic mutations on the Wnt ligand coreceptor LRP6 in the susceptibility of dyslipidemia [30,59].
Increased hepatic TG content was observed in LTCFDN mice at the age of 2 weeks, without the development of defects on other assessed parameters. In this model, TCF7L2DN expression is driven by the albumin promoter, and the transgene cannot be detected in the newborn mice [26]. Thus, increased hepatic TG content in these infant mice is directly caused by β-cat/TCF functional knockdown with hepatic TCF7L2DN expression within the first 2 weeks after their birth. This finding is in agreement with previous investigations on familial combined hyperlipidemia with LRP6/Lrp6 mutations [31]. Importantly, our observations revealed for the first time the existence of the crosstalk between the β-cat/TCF signaling cascade and female hormones on hepatic lipid homeostasis, although their convergence has been suggested in cancer and other research fields [32,33].
The function of the β-cat molecule as the Wnt signaling pathway effector was initially recognized in colon cancer studies [42]. Although men and women show comparable frequency in colon cancer development, the incidence rate appears significantly lower in women with menopause hormone therapy [60]. Kouzmenko and colleagues found that ERα and β-cat precipitate within the same immune complexes, and they reciprocally recruited to the cognate response elements in their target promoters and affected their transcriptions [32]. Gupta and colleagues found that in breast cancer cells, β-cat knockdown resulted in reduced ERα levels, associated with reduced ERα target expression [61]. More recently, Wu and colleagues found that estrogen enhances the activity of Wnt signaling cascade during osteogenesis [33]. We revealed here that estrogen facilitates both β-cat Ser675 phosphorylation and TCF expression in hepatocytes and attribute these as the underlying mechanism for repressing hepatic lipogenesis by E2 in mice upon HFD challenge, as E2 also blocked the stimulation of TCF7L2DN on expression of Mlxipl and Srebfl, as well as their downstream target genes in vitro. More importantly, certain metabolic defects in HFD-fed LTCFDN mice with OVX cannot be effectively restored with E2 reconstitution. Our observations obtained with the utilization of the GPER antagonist G15 and the GPER agonist G1 suggest that the effects on β-cat Ser675 phosphorylation and Tcf7l2 expression are at least partially mediated by GPER (Fig 7A-7E). In the rat liver, it has been shown that the GPER/AC/PKA signaling cascade mediates the function of estrogen on cholestasis [38]. Observations made during the past decade caused us to believe that GPER mediates the nongenomic effects of E2 in the liver and elsewhere via increasing cAMP levels [62], consistent with our observation that E2 rapidly activated CREB Ser133 phosphorylation in MPH, followed by increased β-cat Ser675 phosphorylation levels within the 120-minute experimental period. It is worth mentioning that E2 utilized in a previous study on stimulating cAMP elevation a rodent hepatic cell line was as high as 50 μM [38]. We reproduced the results in both HepG2 and MPH with 100 nM E2 (Fig 7B and 7C). At this dosage, E2 treatment can also stimulate Tcf gene expression (Fig 4C-4E).
β-cat C-terminal Ser675 phosphorylation was initially reported by Hino and colleagues via PKA activation [63]. Such phosphorylation was then demonstrated in different cell lineages by insulin, GLP-1, and IGF-1. We and others have shown that p21-activated protein kinase (Pak1) mediates the effect of insulin or IGF-1 on β-cat S675 phosphorylation [7,10,11,64,65], while PKA is the downstream protein kinase of the incretin hormone GLP-1 [7,66]. Our current study has expanded the spectrum of hormones that exert their functions via post-translational modifications of β-cat, to include the steroid hormone E2.
It is worth mentioning that the stimulatory effect of E2 treatment on the β-cat/TCF responsive Top-flash activity was relatively modest with the utilization of the 293 naïve cell system ( Fig 4G). This, along with the lack of a strong activation on Axin2 in hepatocytes received 6 hour E2 treatment (Fig 4F) prompted us to suggest that the modest stimulation on β-cat Ser675 itself by E2 is insufficient to activate β-cat/TCF activity robustly. The robust activation can be achieved via the increase on both β-cat Ser675 phosphorylation and TCF expression, occurring in hepatocytes 24 hours after E2 treatment but not in the 293 naïve cell system for conducting our LUC reporter assay (Fig 4G).
Observations obtained in assessing our LTCFDN mouse model also prompted us to speculate whether β-cat/TCF is a "missing link" for understanding the paradox of hepatic insulin function and insulin resistance [67][68][69]. Postprandial insulin elevation represses hepatic gluconeogenesis but increases hepatic lipogenesis. Paradoxically, subjects with insulin resistance show elevations on both gluconeogenesis and lipogenesis. On a LFD without developing insulin resistance, our LTCFDN mice showed increased gluconeogenesis as well as lipogenesis, suggesting that β-cat/TCF may serve as a "paradoxical effector" of hepatic insulin signaling.
Fig 7F summarizes our current understanding on the role of E2 on anti-obesity and lipid metabolism in response to Wnt pathway attenuation, which can occur along with aging [1,70,71]. It has been suggested that E2 works with its receptors ERα and ERβ in regulating energy homeostasis in response to HFD challenge via suppressing energy intake and adipose tissue accumulation, increasing insulin secretion, and improving insulin sensitivity, involving the brain, adipose tissue, skeletal muscle, pancreatic islets, and possibly the immune system and other organs [34][35][36]72,73]. E2 may also utilize the nuclear receptors ERα and ERβ as well as GPER to repress abnormal hepatic lipogenic program activation in response to HFD consumption. This is achieved via facilitating the function of the Wnt signaling pathway effector β-cat/TCF. Thus, our current study expanded the scope of hormonal factors that utilize β-cat/ TCF to exert their metabolic functions during adulthood, to include the steroid hormone E2, in addition to previously recognized peptide hormones and growth factors [8] (Fig 7F). Whether E2-stimulated expression of TCF7L2 involves ERα/β as well, or only involves the nongenomic effects of E2 with the participation of GPER in the liver [62], requires further investigations.
Gender differences in the risk, pathophysiology, and complications of T2D and other metabolic diseases have received intensive attention [34]. We learned the impact of testosterone deficiency on the development of visceral obesity and insulin resistance in men, and a recent study revealed a potential mechanism, involving extranuclear actions of the androgen receptor in regulating insulin secretion [74,75]. Our current study suggests that in addition to the regulation of glucose homeostasis via repressing hepatic gluconeogenesis, hepatic β-cat/TCF also participates in lipid homeostasis. Further exploration on how this developmental signaling pathway effector mediates the effect of various metabolic and sex hormones during fasting and after feeding, as well as during health and diseases, may lead to the emergence of novel approaches for metabolic disease treatment and prevention.

Ethics statement
Male and female mice in FVB genetic background were utilized in this study. Animals were handled in strict accordance with good animal practice as defined by national and/or local animal welfare bodies, and all animal work was approved by the Institutional Animal Care and Use Committee of the University Health Network (Protocol number 1560.28). Before conducting the OVX surgery, meloxicam was utilized for analgesia, along with isoflurane. After the surgery, the antibiotic baytril was added to the drinking water for preventing infection.

Animals, OVX and E2 reconstitution
The generation of LTCFDN was described previously in which the expression of the human long isoform (75 kDa) TCF7L2DN was under the control of the 2.4 kb mouse serum albumin promoter/enhancer element [26]. Male heterozygous LTCFDN mouse were always bred with wild-type female mice for the generation of heterozygous male and female mice and the control wild-type littermates [26]. Bilateral OVX or sham surgery was performed under isoflurane anesthesia. E2 reconstitution was performed via subcutaneous E2 injection (2 μg/mice in peanut oil; 100 μL) every 4 days for 6 weeks, based on the literature [76], or with daily subcutaneous injection for 6 days. OVX was performed with a mid-dorsum approach. In the sham procedure, a similar incision was made and the ovary visualized, but no tissue material was removed.

Cell cultures
The human hepatic cell line HepG2 (ATCC) and MPHs were prepared and cultured as we have previously described [26]. MPHs were isolated with liver perfusion of 10-to 12-week-old male C57BL6/J mice. HepG2 or MPHs were starved overnight with serum-free DMEM prior to a given treatment. The AD-293 cell line (Agilent Technologies) utilized for adenovirus production was cultured per manufacturer instruction. The 293T cell line (ATCC) was utilized for LUC report gene analysis.

Metabolic tolerance tests
For IPGTTs and intraperitoneal pyruvate tolerance tests (IPPTTs), both male and female mice were starved for 16 hours prior to the intraperitoneal injection of pyruvate (2 g/kg body weight) or glucose (2 g/kg body weight). For IPITTs, male or female mice were starved for 6 hours prior to the injection of insulin (0.75 U/kg body weight). Assays were conducted as we have previously presented [26].

Adenovirus infection, RNA isolation, RT-PCR, and real-time RT-PCR
The control adenovirus (Ad-GFP) and TCF7L2DN expressing adenovirus (Ad-TCF7L2DN or Ad-TCFDN) were generated using the AdEasy XL adenovirus vector (Agilent Technologies), as we have previously described [26]. PCR primers and their nucleotide sequences are listed in S1 Table. LUC reporter gene analysis The β-cat/TCF responsive Top-flash reporter gene construct and the control Fop-flash one, as well as the methods for 293 cells transfection and LUC reporter analysis, were described in our previous studies [9,26].

Western blotting
Whole-cell lysates were prepared for HepG2 cells or MPHs, or mouse liver tissue for western blotting, as we have described previously [9]. Antibodies utilized in this study are listed in S2 Table.

TG, FFA, and cholesterol measurements
To measure TG contents, liver tissue samples (30 mg) were homogenized in 5% NP-40/H 2 O, followed by centrifugation for supernatant collection. The TG content in the liver supernatant or mouse serum (10 μl) was determined by the utilization of a Serum Triglyceride Determination Kit (TR0100, Sigma Aldrich, Canada). FFA contents were quantified in the liver tissue (10 mg) or serum (2 μl) using the Free Fatty Acid Quantification Kit (Sigma-Aldrich, Canada), following the instructions of the manufacturer. Total cholesterol was measured in serum (2 μl) using the Total Cholesterol E Assay (Wako Diagnostics, USA), following the manufacturer's instructions. For the measurement of hepatic glycogen content, 20 mg of liver tissue was homogenized in 2 mol/l HCl and boiled for 1 hour followed by neutralization with NaOH. Hydrolyzed glycogen was then determined by measuring glucose using the Glucose (GO) Assay Kit (Sigma Aldrich, Canada).

Cellular cAMP level measurement
Cellular cAMP levels in the human HepG2 cell line and MPHs were determined utilizing the commercial kit (510040) from Cayman Chemical (Ann Arbor, MI), following the manufacturer's instruction.

Statistical analyses
Results are expressed as mean ± SEM. For comparison of 2 groups, the Student t test was used. Comparisons between multiple groups were determined by one-way or two-way ANOVA followed by Bonferroni post hoc tests. P < 0.05 was considered to indicate a statistically significant difference.
Supporting information S1 Fig. (Related to Fig 1). Two-week-old LTCFDN mice carry no appreciable abnormalities on their body weight, liver weight, random glucose, serum TG, FFA, or cholesterol levels.
(A) Body weight, (B) liver weight, (C) random glucose levels, (D) serum TG levels, (E) serum FFA levels, and (F) serum cholesterol levels. � p < 0.05. N = 3-4 for WT mice and N = 8 for LTCFDN mice, regardless of the sex. Values represent mean ± SEM. Underlying numerical values can be found in S1 Data. (TIF) S2 Fig. (Related to Fig 1). TCF7L2DN expression results in increased glycogen synthesis.  4). CREB S133 phosphorylation level was increased after E2 treatment. (A) CREB S133 phosphorylation level treated with 100 nM E2 for indicated time.  Fig. (Related to Fig 7). Western blot shows the effect of E2 and G15 treatment on CREB S133 phosphorylation. Representative blot of 3 independent experiments, with densitometrical analysis results presented in the bottom panel. Cells were pretreated with or without G15 (10 nM) for 45 minutes, followed by E2 (100 nM) or vehicle treatment (as control) for another 120 minutes. N = 3 for each treatment. Values represent mean ± STD. (TIF) S1 Table. (Related to Figs 1, 3, 4, 6 and 7). Primers utilized in this study. (XLS) S2 Table. (Related to Figs 1, 4, 5, 6 and 7). Antibodies utilized in this study. (XLS) S1 Data. Numerical values underlying the summary data displayed in this study.