ERα down‐regulates carbohydrate responsive element binding protein and decreases aerobic glycolysis in liver cancer cells

Abstract Deregulated metabolism is one of the characteristics of hepatocellular carcinoma. Sex hormone receptor signalling has been involved in the marked gender dimorphism of hepatocellular carcinoma pathogenesis. Oestrogen receptor (ER) has been reported to reduce the incidence of liver cancer. However, it remains unclear how oestrogen and ER regulate metabolic alterations in liver tumour cells. Our previous work revealed that ERα interacted with carbohydrate responsive element binding protein (ChREBP), which is a transcription factor promoting aerobic glycolysis and proliferation of hepatoma cells. Here, the data showed that ERα overexpression with E2 treatment reduced aerobic glycolysis and cell proliferation of hepatoma cells. In addition to modestly down‐regulating ChREBP transcription, ERα promoted ChREBP degradation. ERα co‐immunoprecipitated with both ChREBP‐α and ChREBP‐β, the two known subtypes of ChREBP. Although E2 promoted ERα to translocate to the nucleus, it did not change subcellular localization of ChREBP. In addition to interacting with ChREBP‐β and promoting its degradation, ERα decreased ChREBP‐α–induced ChREBP‐β transcription. Taken together, we confirmed an original role of ERα in suppressing aerobic glycolysis in liver cancer cells and elucidated the mechanism by which ERα and ChREBP‐α together regulated ChREBP‐β expression.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) usually presents substantial metabolic rearrangements. 1,2 HCC favours excessive glucose uptake and lactate production even under the condition of having oxygen, which is called aerobic glycolysis or Warburg's effect. 1 Aerobic glycolysis supplied liver cancer cells with metabolic intermediates for anabolism to support rapid cell proliferation. The level of glucose-6-phosphate, which is the first product in glycolysis, is considerably increased in HCC tissues. 3 Due to the prominent gender disparity of HCC, sex hormone receptor signalling has been involved in liver cancer pathogenesis. Oestrogen and oestrogen receptor (ER) negatively regulate initiation and progression of HCC. 4 Among the three oestrogen receptors including ERα, ERβ and G protein-coupled ER, ERα is the dominant oestrogen receptor in hepatocytes. 5,6 There are many studies exploring the mechanism by which ERα reduces liver carcinogenesis. Oestrogen inhibits transcription of hepatitis B virus (HBV) genes and reduces rates of hepatocellular carcinoma in HBV-infected women by up-regulating ERα. 7 Oestrogen reduces liver cancer risk in females by decreasing IL-6 production by Kupffer cells. 8 Oestrogen and ERα block metastasis of hepatocellular carcinoma cells by modulating glycogen synthase kinase 3β (GSK-3β) and E3 ligase (β-TrCP) expression. 9 Moreover, it was reported that 17β-oestradiol (E2) and ERα reprogrammed metabolism in terms of glucose usability in breast cancer cells. [10][11][12] However, it remains unclear how oestrogen and ER regulate aerobic glycolysis in HCC cells.
ERα and E2 suppress lipogenesis by inhibiting mRNA and protein expression of carbohydrate responsive element binding protein (ChREBP) in insulin-secreting INS-1 cells. 13 Membrane ERα signalling inhibits triglyceride synthesis through suppressing ChREBPα nuclear translocation. 14 ChREBP is one of the members for the basic helix-loop-helix/leucine-zipper (bHLH/ZIP) transcription factor family, mediating glucose-regulated gene transcription. 15 Mammalian ChREBP has two subtypes transcribed from different promoters, ChREBPα and ChREBPβ. ChREBPα binds ChREBPβ at the carbohydrate responsive element (ChoRE) site of its promoter and promotes its transcription. 16 ChREBP promotes glucose utilization of liver and lipogenesis independent of insulin. 17,18 By regulating transcription of enzyme genes in gluconeogenesis, de novo lipogenesis and glycolysis, ChREBP is involved in the oetiopathogenesis of metabolic diseases and cancers. 19 In β-cells of islets, ChREBP induces glucose-stimulated cell propagation. 20,21 ChREBP is crucial for the multiplication of liver cancer cell by facilitating aerobic glycolysis and anabolism. 22 Our former work indicated that ChREBPα interacted with both ERα and its cofactor flightless I homolog (FLII), and FLII also interacted with ChREBPβ. 23 However, it remains unknown whether the ERα-ChREBP complex regulates metabolism and proliferation of liver cancer cells. Here, we found that E2 and ERα decreased aerobic glycolysis and cell multiplication in HepG2 hepatoma carcinoma cells.
ERα co-immunoprecipitated and colocalized with both ChREBPα and ChREBPβ. ERα decreased ChREBPα-induced ChREBPβ transcription. Our results demonstrated ERα, ChREBPα and ChREBPβ are in the same complex. The ERα-ChREBP complex might be a potential target in the therapy of hepatoma carcinoma.

| Nuclear and cytosolic fractionation
HA-ChREBPα and Flag-ERα were transfected into 293T cells. E2 was used to treat the transfected cells. The preparation of buffer A and buffer B, and the operating steps were handled as presented previously. 23 All the procedures were performed at 4℃.

| Co-immunoprecipitation
Triton X-100 buffer contained 1% Triton X-100, 100 mM NaCl, 40 mM Tris-HCl, pH 8, 1 mM EDTA and 0.5% NP-40. Using Triton X-100 buffer to lyse fresh cells, primary antibody incubation was performed overnight at 4℃, followed by an additional incubation for 2 hours together with protein A/G agarose beads (Santa Cruz) at 4℃. Triton X-100 buffer was applied to wash the beads for four times and then to boil the beads in 2 × SDS protein loading buffer. Then, the Western blotting analysis was carried out.

| Luciferase assays
Flag-ERα, HA-ChREBPα, HA-ChREBPβ, β-galactosidase, ChREBPα-Luc reporter and ChREBPβ-Luc reporter were transiently transfected into 293T cells using Lipofectamine 2000 (Invitrogen). Betagalactosidase was employed as a reference for normalizing transfection efficiency. pcDNA3 was used to normalize the total amount of transfected DNA. At 48 hours post-transfection, cells were collected and analysed by means of Beta-Gal Assay Kit (Clontech) and the luciferase reporter assay system (Promega) according to the reagent specification.

| Real-time quantitative PCR
TRIzol (Invitrogen Life Technologies) was used to isolate total RNA from cells with or without E2 treatment in accordance with the manufacturer's recommendations. The PrimeScript™ RT Reagent Kit (Takara Bio Inc) with 10 mL assay mix containing 2 mg total RNA was applied to synthesize cDNA. The reaction mixture was kept at −20℃ until the PCR analysis. β-actin was applied to the endogenous reference. The primer sequences of ChREBPα and ChREBPβ were described previously. 16 Table 1 listed other primer sequences.
Real-time PCR was analysed using a StepOnePlus™ Real-Time PCR System (Applied Biosystems). Relative quantification of mRNA amount was obtained in the light of the user manual of Applied Biosystems.

| Western blotting
Cell lysates were collected. BCA Protein Assay Kit (Pierce) was used to quantify protein concentration. Protein bands were visualized with an enhanced chemiluminescent solution (Millipore) using Amersham Imager 600 (GE).

| Metabolic assays
Cells in 10-cm dishes grew to about 80% confluence and then were trypsinized and resuspended in 3 mL of culture solution. An Oxytherm System (Hansatech) was applied to detect oxygen consumption.
Glucose uptake and lactate production were measured at 48 hours after cells were laid in 6-well plates or treated, and then collected and investigated culture medium. Lactate production and glucose uptake were analysed using Lactate Assay Kit (Sigma) and Glucose Assay Kit (Shanghai Rongsheng Biotech), respectively, according to product specification.

| Cell viability assay
A total of 2000, 4000, 6000, 10 000 or 20 000 cells were laid in 6-well plates in triplicate. After cell adherence, CCK8 reagent was added and incubated for 2 hours, followed by OD measurement at 450 nm. The standard curve was drawn according to cell numbers and OD values. The four groups are cells stably expressing GFP or ERα cDNA with or without E2 treatment. Each group of 4000 cells were plated and cultured at 37℃ with 5% CO 2 . A number of cells were measured at days 0, 2, 4 and 6 after plating.

| Immunofluorescent staining
Cells were grown on coverslips, rinsed twice with phosphate buffer saline (PBS), each for 5 minutes. Then, cells were treated with 4% paraformaldehyde for 15 minutes, rinsed twice with PBS, permeabilized with 0.3% Triton X-100 for 20 minutes at room temperature and blocked in 2% goat serum for 30 minutes at room temperature and primary antibodies were incubated overnight at 4℃. Alexa 488 and 555 Fluor ® , which were conjugated secondary antibodies, were added to the cells and reared for 30 minutes at room temperature away from light. In the final 5 minutes, DAPI (40728ES03) was added. The coverslips were mounted with nail polish after being washed twice in PBS, and images were taken using an LSM 710 laser scanning confocal microscope (Zeiss) and a confocal microscope (Leica, TCS SP8 STED).

| Statistical analysis
Experiments were done at least three times independently; one representative experiment was displayed. Data were presented as mean ± standard deviation (SD) using Prism 5 (GraphPad Software).
Student's t test was used to analyse the difference between the treatment group and the control group. P value < .05 was regarded as statistically significant.

| ERα with E2 treatment reduced aerobic glycolysis and cell multiplication in HepG2 cells
It was known that ChREBP enhanced aerobic glycolysis and cell multiplication in tumour cells. 22 Previously, we found that ERα interacts with ChREBP, 23 and we further wondered whether ERα adjusted metabolic activity and multiplication capacity of hepatoma carcinoma cells. We constructed HepG2 stable cells with Flag-tagged GFP cDNA (Flag-GFP) and Flag-tagged ERα cDNA (Flag-ERα) and compared their metabolic activity and multiplication capacity. We compared glucose uptake of Flag-GFP with E2 treatment for 24 hours and Flag-ERα with or without E2 treatment for 24 hours. HepG2 cells stably expressing Flag-ERα with E2 treatment for 24 hours showed reduced glucose uptake ( Figure 1A). HepG2 cells stably expressing Flag-ERα with E2 treatment displayed lower lactate production when compared with no E2 treatment ( Figure 1B). To investigate the effect of ERα overexpression with or without E2 treatment on liver cancer cell viability, cell number was counted at days 0, 2, 4 and 6 after plating. We found that the overexpression of ERα with E2 treatment slowed cell growth in comparison with the control ( Figure 1C). We also examined cell cycle and apoptosis of Flag-GFP with E2 treatment or Flag-ERα with or without E2 treatment. ERα with E2 treatment showed a decreased percentage of S-phase cells ( Figure 1D) and increased percentage of apoptotic cells ( Figure 1E,F). In brief, our findings indicate that ectopic expression of ERα with E2 treatment reduced aerobic glycolysis and cell multiplication in hepatoma carcinoma cells.

SMMC7721 cells stably expressing Flag-ERα with E2 treatment
showed reduced expression of glucose transporter Glut2 and Glut4 ( Figure S2B,C), which might contribute to decreased glucose uptake.

| ERα co-immunoprecipitated and colocalized with both ChREBPα and ChREBPβ
ERα and E2 can inhibit transcription and translation of ChREBPα. 13 Our previous results showed that ectopically expressed ERα interacted with ChREBPα. 23 Now, we discovered that authigenic ChREBPα and ERα protein also co-immunoprecipitated in SMMC7721 hepatoma carcinoma cells (Figure 2A). Herman et al found that ChREBP had ChREBPα and ChREBPβ isoforms. 16 We also investigated whether ERα interacted with ChREBPβ and confirmed that ERα co-immunoprecipitated with ChREBPβ and ChREBPα ( Figure 2B).

| Influence of oestrogen on the localization and expression of ERα and ChREBPα
We next analysed whether E2 affected the subcellular localization of ChREBPα and ERα. We overexpressed HA-ChREBPα and Flag-ERα in HeLa cells with or without E2 treatment and investigated the distribution of ChREBPα and ERα using immunofluorescent staining. Without E2 treatment, ERα was distributed in both cytoplasm and nucleus ( Figure 3A). However, most of ERα translocated to the nucleus after E2 treatment and nuclear colocalization of ERα and ChREBPα was increased ( Figure 3B,C and E). For ChREBPα, there was nearly no effect on its localization with or without E2 treatment ( Figure 3A,B and E). There was similar effect to ChREBPβ with E2 treatment ( Figure S1).
To confirm the findings, we performed nuclear and cytoplasm ChREBPα, ChREBPβ and ChREBP-total at 24 h in HepG2 cells stably expressing Flag-ERα with or without E2 treatment. E, Nuclear and cytosolic fractionation analysis showed that E2 promoted the translocation of ERα from cytoplasm to nucleus without affecting subcellular localization of ChREBP in 293T cells. Tubulin and PARP serve as loading controls for the cytosolic and nuclear fraction, respectively. Statistical significance was calculated by unpaired Student's t test (mean ± SD). *P < .05 Flag-tagged ERα, ERα overexpression with E2 treatment reduced mRNA levels of ChREBPα, ChREBPβ and ChREBP-total, with a more significant reduction in the ChREBPβ mRNA level ( Figure 3D).
Western blot analyses showing ERα transfection and E2 treatment down-regulated endogenous protein level of ChREBP ( Figure S2A).

| ChREBPα or ChREBPβ with ERα and E2 inhibited ChREBPβ transcription
Herman et al reported that one isoform of ChREBP, ChREBPα, transcriptionally activated the other isoform, ChREBPβ. 16 We next investigated whether ERα regulated the ability of ChREBPα to transcriptionally activate ChREBPβ. The luciferase assay showed that ERα and ChREBPα or ChREBPβ with E2 treatment sharply decreased ChREBPβ reporter transcriptional activity ( Figure 4A,C and F), but there was no significant effect on the transcriptional activity of ChREBPβ mutant reporter ( Figure 4B). Moreover, ERα and ChREBPα or ChREBPβ with E2 treatment did not affect ChREBPα reporter transcriptional activity ( Figure 4D,E). To assess whether there was specificity that ERα and ChREBPα or ChREBPβ with E2 treatment reduced ChREBPβ reporter transcriptional activity, we co-transfected ERα and FLII to determine their effect on ChREBPβ reporter transcriptional activity. The results revealed there was no obvious difference for the effect of ERα and FLII on ChREBPβ reporter transcriptional activity with or without E2 treatment ( Figure 4G).
We found that ERα weakened the interaction between ChREBPα and ChREBPβ, and ERα with E2 treatment further weakened the association ( Figure 5). These results suggested that ERα, ChREBPα and ChREBPβ might be in the same complex.
In summary, ERα with E2 treatment reduced aerobic glycolysis and inhibited cell multiplication in liver cancer cells. One of the underlying mechanisms is that ERα suppressed ChREBP activity, including ChREBPα-mediated ChREBPβ transcription ( Figure 6).

| D ISCUSS I ON
ERα and oestrogen negatively regulate liver cancer cell proliferation. 4 Here, we provided a novel mechanism by showing that ERα with E2 treatment reduced aerobic glycolysis by suppressing ChREBP activity. We found ERα and oestrogen could regulate metabolism and multiplication of hepatoma carcinoma cells by reducing ChREBP protein levels. ChREBP was discovered to be a major regulator of vital genes related to glycolysis, lipogenesis and gluconeogenesis in metabolic tissues. 15,17,18,[24][25][26][27][28][29] Additionally, ChREBP facilitated the proliferation of liver and colorectal cancer cells. 22 It was known ChREBP F I G U R E 4 ERα and E2 reduced ChREBPα or ChREBPβ-induced ChREBPβ transcription. A, Luciferase reporter assay showed that co-expression of ERα and ChREBPα with E2 treatment decreased the transcriptional activity of ChREBPβ promoter. B, Luciferase reporter assay showed that co-expression of ERα and ChREBPα with E2 treatment did not decrease the transcriptional activity of mutant ChREBPβ promoter, which did not contain ChoRE sites. C, Luciferase reporter assay showed that co-expression of ERα and ChREBPβ with E2 treatment decreased the transcriptional activity of ChREBPβ promoter. D, Luciferase reporter assay showed that co-expression of ERα and ChREBPα with E2 treatment did not change the transcriptional activity of ChREBPα promoter. E, Luciferase reporter assay showed that coexpression of ERα and ChREBPβ with E2 treatment did not change the transcriptional activity of ChREBPα promoter. F, Luciferase reporter assay showed that co-expression of ERα, ChREBPα and ChREBPβ with E2 treatment decreased the transcriptional activity of ChREBPβ promoter. G, Luciferase reporter assay showed that co-expression of ERα and GFP-FLII with E2 treatment hardly changed the transcriptional activity of ChREBPβ promoter. Statistical significance was calculated by unpaired Student's t test (mean ± SD). *P < .05 has two subtypes: ChREBPα and ChREBPβ; ChREBPβ is a target gene of ChREBPα. 16 ChREBP reported in previous studies is actually ChREBPα. Our finding showed that ERα suppressed ChREBPα-mediated ChREBPβ transcription provided another regulatory mechanism for the two isoforms of ChREBP.
ChREBP was down-regulated in human breast tumour in comparison with vicinal normal tissues. 30 Moreover, ChREBP significantly correlated with increased survival in breast cancer. 31,32 The study suggested that ChREBP might play different roles in regulating cell proliferation in breast and liver cancers. Therefore, it is worthwhile to find out why the ERα-ChREBP axis plays distinct roles in breast and liver cancers. Taken together, we uncovered a novel mechanism for ERα decreasing aerobic glycolysis and cell multiplication in liver cancer cells.
The data showed that ERα, ChREBPα and ChREBPβ coexisted in a complex, and ERα inhibited ChREBPα-mediated ChREBPβ transcription. The ERα-ChREBP axis might be a potential target in the therapy of liver cancer.

ACK N OWLED G EM ENTS
This work was supported by grants from the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.