ZEB1 enhances Warburg effect to facilitate tumorigenesis and metastasis of HCC by transcriptionally activating PFKM

Abstract

Metabolic reprogramming, especially Warburg effect, is a key event in tumor initiation and progression. ZEB1 plays a vital role in metastasis of various cancers. We previously found that ZEB1 was excessively expressed in hepatocellular carcinoma (HCC) and its high expression was closely correlated with metastasis and recurrence of HCC. We want to know whether glycolytic enzymes are regulated by ZEB1 and contribute to carcinogenesis and metastasis of HCC. Methods: To explore whether ZEB1 could enhance glycolysis in HCC, we knocked down ZEB1 by short hairpin RNA (shRNA) in MHCC-97H and HCC-LM3 cells and performed glucose uptake, lactate production, ECAR and OCR assays. To investigate how ZEB1 enhances glycolysis, the protein levels of glycolytic enzymes were detected in the same cell lines using Western blot. The regulatory effect of ZEB1 on PFKM mRNA level was confirmed by RT-qPCR, luciferase report assay and ChIP assay. In order to assess the role of ZEB1-PFKM axis in cell proliferation, cell counting and CCK-8 assays were performed in MHCC-97H and HCC-LM3 cell lines knocked down for ZEB1 and further re-expressed for either ZEB1 or PFKM or not. To explored whether the ZEB1-PFKM axis also functions in HCC cell migration, invasion and metastasis, the same MHCC-97H and HCC-LM3 cell lines were performed for wound healing assays, transwell assays and colony formation assays, meanwhile, MHCC-97H cell lines were performed for orthotopic liver transplantation assays. Finally, the expression of ZEB1 and PFKM were examined in human liver cancer specimens and non-tumorous liver tissues using immunohistochemical and Western blot. Results: We found that ZEB1 transcriptionally upregulates the expression of the muscle isoform of phosphofructokinase-1 (PFKM), a rate-limiting enzyme in glycolysis. Intriguingly, a non-classic ZEB1-binding sequence in the promoter region of PFKM was identified through which ZEB1 directly activates the transcription of PFKM. Silencing of ZEB1 in MHCC-97H and HCC-LM3 cell leads to impaired PFKM expression, glycolysis, proliferation and invasion, and such impairments are rescued by exogenous expression of PFKM. Importantly, in-situ HCC xenograft assays and studies from TCGA database demonstrate that ZEB1-PFKM axis is crucial for carcinogenesis and metastasis of HCC. Conclusions: Our study reveals a novel mechanism of ZEB1 in promoting HCC by activating the transcription of PFKM, establishing the direct link of ZEB1 to the promotion of glycolysis and Warburg effect and suggesting that inhibition of ZEB1 transcriptional activity toward PFKM may be a potential therapeutic strategy for HCC.

Keywords: PFKM; ZEB1; glycolysis; intrahepatic metastasis; non-classic binding.

Figure 1

High expression of ZEB1 enhances glycolysis. (A) Western blot analysis of the ZEB1 protein level in MHCC-97H and HCC-LM3 cell lines with knockdown and further re-expression of ZEB1. E-Cadherin protein levels were detected as a positive control for ZEB1 knockdown. (B) Glucose uptake and (C) lactate production were measured in MHCC-97H and HCC-LM3 cell lines. (D) MHCC-97H cells used in (A) were detected for extracellular acidification rate (ECAR) as an indicator for deduced glycolysis flux and glycolytic capacity. (E) The oxygen consumption rate (OCR) was detected as an indicator for oxidative phosphorylation (OXPHOS) and deduced levels of basal respiration and ATP production. (F) HUH-7 and Hep3B cells were expressed for increasing doses of ZEB1, followed by detection of ZEB1 protein level, (G) glucose uptake and (H) lactate production. (I, J) HUH-7 cells used in (F) were detected for ECAR as an indicator of glycolysis flux and glycolytic capacity (I), and OCR as an indicator of OXPHOS and deduced levels of basal respiration and ATP production (J). The data in Figure 1 except B and F are shown as means±SD of three independent experiments (*P<0.05, **P<0.01, ***P<0.001). The Statistical analyses in (B-E) and in (G-J) are performed using unpaired Student's t test and one-way ANOVA, individually.

Figure 2

ZEB1 stimulates glycolysis by upregulating PFKM. (A) Immunoblot of glycolysis enzymes in ZEB1 knockdown and re-expression MHCC-97H cell lines. (B-D) MHCC-97H cells were knocked down for ZEB1 and further expressed for exogenous ZEB1 or PFKM, followed by determination of glucose uptake (B), lactate production (C) and expression levels of indicated proteins (D). (E-G) HUH-7 cells with low endogenous expression of ZEB1 was overexpressed for Flag-tagged ZEB1 and further knocked down for PFKM, followed by measurement of glucose uptake (E), lactate production (F) and protein levels (G). (H) MHCC-97H cells used in (B) were detected for ECAR to indicate glycolysis flux and glycolytic capacity. (I) The OCR was detected to indicate basal respiration and ATP production. (J-K) Relative abundances of F6P (J) and FBP (K) in the same MHCC-97H cell lines as in (B) were measured using LC-MS. (L) The protein levels of ZEB1 and PFKM in a series of HCC cell lines were detected (upper panel) and analyzed for their correlation (lower panel). The data in (B-C, E-F and H-K) are shown as means±SD of three independent experiments and analyzed using Student's t test (**P<0.01, ***P<0.001, N.S.: P ≥ 0.05).

Figure 3

Three E2-box-like sequences in the promoter region of PFKM are not required for its transactivation by ZEB1. (A) MHCC-97H cell line was knocked down and further re-expressed for ZEB1. Then relative mRNA levels of ZEB1 and PFKM were determined using RT-qPCR. (B) HEK-293T cells were transfected with PFKM-Luc and increasing doses of ZEB1. 24 h post-transfection, luciferase activity was determined. The data are presented as means±SD of three independent experiments. Statistical analysis was performed by one-way ANOVA. (C) Schematic diagram showing wildtype PFKM promoter and its various mutants with three E2-box-like sequences mutated alone or in combinations. For each E2-box-like sequence, CACCT(G) was mutated to CAAAT(G). (D) Luciferase reporter vectors carrying either wildtype PFKM promoter or any of its mutants was transfected into HEK-293T cells together with ZEB1. After 24 h of transfection, relative luciferase activity was determined and normalized to cells transfected with wildtype PFKM promoter alone (column 1). The data in (A) and (D) are shown as means±SD of three independent experiments (***P<0.001, N.S.: P ≥ 0.05, unpaired Student's t test).

Figure 4

ZEB1 activates PFKM transcription through a non-canonic binding sequence. (A) Schematic diagram showing wildtype and truncated PFKM promoters. (B-C) Luciferase reporter vectors of wildtype PFKM promoter or its truncations were co-transfected with ZEB1 into HEK-293T cells, followed by determination of luciferase activity. The relative luciferase activity was normalized to cells transfected with wildtype PFKM promoter alone (column 1). (D-E) ChIP assay was performed in MHCC-97H cell line by using rabbit IgG and anti-ZEB1 antibody. CDH1 promoter serves as a positive control and ACTB promoter as a negative control (D). The data in (D) were quantified and normalized to input, and are presented as means±SD of three independent experiments (E) (***P<0.001, Student's t test).

Figure 5

PFKM plays a key role in ZEB1-stimulated tumorigenesis and intrahepatic metastasis of HCC. (A-B) Cell proliferation were determined in ZEB1 KD MHCC-97H cells with or without further expression of ZEB1 and PFKM employing cell counting (A) and CCK-8 assays (B). (C-I) The same MHCC-97H cell lines as in (A) were performed for wound healing assays (C), transwell assays (D), colony formation assays (E) and orthotopic liver transplantation assays (F-I). Typical pictures showing mouse livers with tumor lesions (red circles indicate the primary tumor formed in the injection site and red arrows point to the intrahepatically metastatic tumor nodes on the surface of liver) (F). The primary tumor weight (G), primary tumor volume (H) and the number of metastatic tumor nodes (I) were determined. The scale bars in (C), (D), (E) and (F) represents 200 µm, 200 µm, 5 mm and 1 cm, individually. The data in (A, C-E) are shown as means±SD of three independent experiments, in (B) are shown as means±SEM of three independent experiments in quadruplication (*P<0.05, **P<0.01, ***P<0.001). The data in (G-I) are means±SD (n=5) (*P<0.05, **P<0.01, ***P<0.001). The data in this Figure are analyzed statistically using unpaired Student's t test where necessary.

Figure 6

Interrelated high expression of ZEB1 and PFKM is correlated with poor prognosis of HCC. (A-B) Representative pictures (A) and relative optical density of IHC staining of ZEB1, PFKM and PCNA in HCC tissues and the corresponding adjacent normal tissues (B). The scale bars represent 100 µm. The data are shown as means±SD (n=7, ***P <0.001, Student's t test). (C-D) The protein levels of ZEB1 and PFKM in HCC tissues and the corresponding adjacent normal tissues (n=20) were determined using Western blot (C), followed by analysis of their correlation (D). (E) The comparison of either ZEB1 (upper panel) or PFKM (lower panel) expression between normal liver tissues and HCC tissues. Data are publicly available in Oncomine (the MAS liver). (F) The Kaplan-Meier curves of overall survival basic relating to ZEB1-PFKM axis in HCC. Data are publicly available in the Kaplan-Meier Plotter.