Targeting FBPase is an emerging novel approach for cancer therapy

Cancer is one of the leading causes of death worldwide [1]. In the past decade, reprogramming of energy metabolism has emerged as new hallmarks of cancer [2]. There is increasing epidemiological evidence that links the risk of cancer with metabolic disorders, such as diabetes and obesity [3]. However, the complete regulatory network of metabolism reprogramming remains to be elucidated.

Glucose homeostasis is reciprocally controlled by the catabolic glycolysis and anabolic gluconeogenesis pathways. Under normoxic conditions, normal cells process glucose first to pyruvate via glycolysis in the cytosol and thereafter to carbon dioxide in the mitochondria (oxidative phosphorylation, OXPHOS); under anaerobic conditions, glycolysis is favoured, and relatively little pyruvate is dispatched to the oxygen-consuming mitochondria [4, 5]. However, as first observed by Otto Warburg in the 1920s, some cancer cells preferentially rely on glycolysis, even in conditions of high oxygen tension (‘‘aerobic glycolysis” or “the Warburg effect’’) [6, 7]. Aerobic glycolysis was validated with the wide use of ¹⁸F-deoxyglucose positron emission tomography (FDG-PET) in most cancers [8, 9]. Aerobic glycolysis leads to several advantages for tumour growth by increasing glucose intake, enhancing lactate production and secretion, and diverting glycolytic intermediates to anabolic reactions such as ribose synthesis, serine and glycine synthesis, phosphoglycerol synthesis, and protein glycosylation [10]. Moreover, recent studies have developed promising individualized therapeutic strategies by targeting the altered energy metabolism for the treatment of cancer [11].

Though our understanding of metabolic reprogramming in cancer is progressing at an unprecedented pace, previous studies have mainly focused on glycolysis. Recently, fructose-1,6-bisphosphatase (FBPase), a rate-limiting enzyme responsible for gluconeogenesis, was found to play critical roles in tumour initiation and progression in several cancer types [12‐28]. Here, we search Pubmed, EMbase, Web of Science using keywords “FBPase” and “cancer” (the most recent published report search date: 1 August 2017) and review recent ideas and discoveries regarding the role of FBPase in cancer.

FBPase (EC 3.1.3.11), which catalyses the hydrolysis of fructose 1,6-bisphosphate to fructose 6-phosphate and inorganic phosphate, is a rate-limiting enzyme responsible for gluconeogenesis and glyconeogenesis and, more generally, for the control of energy metabolism and glucose homeostasis [29]. In mammals, two separate genes, FBP1 and FBP2, encode the liver and muscle isoforms of FBPase, respectively. The FBP1 gene, which is mainly expressed in gluconeogenic organs, such as the liver and kidneys, is located at chromosome region 9q22.3. FBP1 consists of seven exons which span more than 31 kb that encode 338 amino acids and six introns [30]. The FBP1 promoter has been well characterized since 2000 [31]. The FBP2 gene, which was initially isolated from muscle tissues and later found to be expressed in all cells, is located at chromosome region 1p36.1 [32]. FBP2 encodes a 339-amino acid protein that shares 77% identity with the FBP1 protein [33]. However, the promoter of the FBP2 gene has not yet been characterized. All FBPases are homotetrameric proteins with a molecular weight of approximately 37 kDa per subunit (Fig. 1). FBP1 has been identified as the regulatory enzyme of gluconeogenesis, but the physiological role of FBP2 is far from clear.

Acute regulation of FBPase is achieved mainly through its oscillation between two conformational states: the inactive T form when in complex with AMP and the active R form [29, 36‐40]. By modulating pyruvate kinase activity, fructose-1,6-bisphosphate (F1,6BP) affects substrate cycling in the pyruvatel/phosphoenolpyruvate (PEP) substrate cycle [41]. Phosphofructokinase-1 (PFK1), the “gatekeeper” of glycolysis, converts fructose-6-phosphate (F6P) to F1,6BP and has the opposite effect of FBPase in regulating F1,6BP [42]. Fructose-2,6-bisphosphate (F2,6BP) is a potent allosteric activator of PFK1 and a competitive inhibitor of FBPase [5, 41‐44]. When the level of F2,6BP is low, the rates of gluconeogenesis are high (such as in starvation and diabetes). Conversely, when the level of F2,6BP is high, the rate of gluconeogenesis is low (such as during refeeding and insulin administration) [5]. The synthesis and hydrolysis of this regulator are catalysed by the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFK-2/FBPase-2) [45]. Phosphorylation of this enzyme by cAMP-dependent protein kinase results in inhibition of the kinase and activation of the bisphosphatase, whereas dephosphorylation has the opposite effect [36, 40]. TIGAR (TP53-induced glycolysis and apoptosis regulator), a recently identified enzyme, shares similarities with the bisphosphatase domain of PFK-2/FBPase-2 [46]. TIGAR acts to degrade intracellular F2,6BP [47]. Interestingly, FBP1 and FBP2 differ significantly in their kinetic properties. FBP2 is approximately 100 times more susceptible to the allosteric inhibitors AMP and NAD+ [48] and approximately 1000 times more sensitive to inhibition by Ca²⁺ than FBP1 [49, 50]. Crystal structures have shown that in contrast to the well-studied R form of FBP1, which is flat, the R form of FBP2 was diametrically different, with a perpendicular orientation of the upper and lower dimers [37].

Chronic regulation of hepatic glucose metabolism occurs through transcriptional and hormonal mechanisms [51]. The master hormone that promotes gluconeogenesis is glucagon. Increased glucagon stimulates gluconeogenesis via the induction of intracellular cAMP. cAMP triggers protein kinase A (PKA), leading to the phosphorylation of PFK-2/FBPase-2, which activates FBPase-2, leading to the dephosphorylation of F2,6BP and concomitant increases in FBPase activity and gluconeogenesis [45]. In addition, PKA activates cAMP response element-binding protein (CREB), the CREB coactivator CRTC2 [52], and the class II histone deacetylases (HDACs)/FOXO pathways [53, 54] and promotes the expression of key gluconeogenic genes; however, it suppresses the expression of glycolytic genes. Glucocorticoid hormones and glucagon plays essential synergistic roles in the regulation of gluconeogenesis [55]. The response to glucocorticoids is mediated by the glucocorticoid receptor (GR), which binds to glucocorticoid responsive elements (GREs) in the promoters of gluconeogenic genes [56]. In contrast, insulin acts to repress the transcription of gluconeogenic enzymes. This repression of the expression of gluconeogenic enzymes is achieved via the activation of the insulin-phosphoinositide 3-kinase (PI3K)-Akt pathway or suppression of the cAMP/PKA/CREB pathway [57] (Fig. 2).

Accumulating evidence has disclosed the role of FBPase in the carcinogenesis, development and progression of various cancer types. Lower FBPase expression frequently correlated significantly with an advanced tumour stage, a highly malignant phenotype, and worse prognoses in cancer patients. All these data implied that FBPase might be a novel biomarker and potential target for the treatment of cancer (Table 1).

The mechanism through which FBP1 influences cancer cells were summarized in Fig. 3. At present, the role of FBP1 in regulating the Warburg effect in cancer cells was best understood.

Along with increases in complex I activity and mitochondrial OCR, ROS levels increased when FBPase was expressed [12, 17, 19, 22, 25]. ROS amplify tumourigenic phenotypes, such as cancer stem cells (CSCs) [103]. Breast CSCs are enriched in cells with a CD44 high/CD24 low/EpCAM+ phenotype [104]. In BLBC cell lines, FBP1 expression significantly reduced the percentage of CD44 high/CD24 low/EpCAM+ populations and decreased tumour sphere formation. Conversely, FBP1 silencing resulted in an increased CSC-like phenotype [22]. Mechanically, increased ROS levels induced by FBP1 would shift the interaction of β-catenin from TCF4 to FOX O3a and thus inhibit tumourigenicity in vitro and tumour formation in vivo.

Moreover, ROS induces mitochondrial apoptosis [105]. Forced FBP1 expression in lung cancer stem cells and breast cancer or FBP2 expression in gastric cancer increased apoptosis by reducing the Bax/Bcl-2 ratio, inducing poly ADP-ribose polymerase (PARP), caspase-3 and caspase-9 activation and suppressing endogenous ROS scavenging systems such as superoxide dismutase (SOD) [12, 19, 25]. In breast cancer cells, FBP1 limited the efficient removal of diseased mitochondria and reduced the expression of hypoxia-induced factor 1α (HIF1α), BCL2/adenovirus E1B 19 kDa interacting protein 3-like (BNIP3L/NIX), and BCL2/adenovirus E1B 19 kDa interacting protein 3 (BNIP3), which disrupts BNIP3/NIX-Bcl-2 complex formation under normal conditions but promotes complex formation between Bcl-2 and Beclin 1 [19]. Annexin V+/propidium iodide (PI)− and Annexin V+/PI+ cells represent early apoptotic cells and late apoptotic/necrotic cells, respectively [106]. FBP2 expression increased both subpopulations in gastric cancer cells [25]. In brain metastatic breast cancer cells, knocking down FBP2 resulted in a significant amount of apoptotic cell death (as indicated by the increase incleaved PARP), whereas exogenous FBP2 significantly rescued cell death [21].

The ability to sustain chronic proliferation is the most fundamental trait of cancer cells. Normal tissues carefully maintain the homeostasis of cell number, normal tissue architecture and function by rigorous regulation of progression signals through the cell cycle as well as cell growth. But cancer cells, by hijacking these regulating mechanisms, become masters of their own destinies [2]. Cell cycle arrests if the numbers of cells in G2 and M phase increase, whereas it proceeds if the numbers of cells in S phase increase. Cell cycle distribution was investigated to understand the molecular mechanisms by which FBP1 suppressed colony formation and cell proliferation. In hepatocellular carcinoma and colon cancer, forced FBP1 expression was found to increase the number of G2–M phase cells but decrease the number of S phase cells [17]. In lung cancer, FBP2 accumulated in cell nuclei during the S and G2 phases and interacted with histone family members and with several proteins involved in cell-cycle regulation and RNA processing [83, 107]. All these data indicated that the growth suppression induced by forced FBP1 expression might be partly due to cell cycle arrest. However, the detailed mechanism remains unclear.

All cancer cells suffer hypoxia microenvironment due to rapid proliferation, differentiation and poorly formed tumour vasculature [108]. HIF1, a well-known transcriptional regulator, is the most important aspect of how cancer cells respond to the unfriendly microenvironment. HIF1 promotes glycolytic metabolic alterations by the activation of multiple glycolytic genes. Under conditions of persistent hypoxia, the induction of HIF1 leads to adaptive mechanisms for the reduction of ROS and re-establishment of homeostasis [109].

Knockdown of FBP1 significantly reduced growth inhibition in luminal cell lines under hypoxic condition (0.1% O₂) but not at normoxic condition (21% O₂). Similarly, FBP1 expression induced a drastic growth inhibition in BLBC cell lines under hypoxic condition but not at normoxic condition [22]. The expression of FBP1, but not of FBP2, was inversely correlated with HIF1α activity in RCC [23, 24]. FBP1 was also found to be negative correlated with HIF1α activity in HCC and breast cancer [18, 82]. Ectopic expression of FBP1 suppressed HIF activity and reduced the expression of HIF target genes, whereas loss of FBP1 enhanced HIF activity [24]. FBP1 inhibited HIF1α activity in the nucleus via a direct interaction with the HIF-inhibitory domain in an enzyme-activity-independent manner [24]. ALL these results indicated that the regulation of FBP1 are important when oxygen and glucose levels are limiting, as often occurs in solid tumours [110].

Gemcitabine is the first-line chemotherapy for pancreatic cancer [111]. However, gemcitabine fails to significantly improve prognosis of pancreatic carcinoma patients due to acquisition of chemoresistance in patients. It is well documented that gemcitabine treatment results in undesirable activation of the RAS-RAF-MAPK pathway [112]. IQ motif-containing GTPase-activating protein 1 (IQGAP1) is a MAPK scaffold that directly regulates the activation of RAF, MEK, and extracellular signalling-regulated kinases (ERKs) [113]. In pancreatic cancer, FBP1 expression impeded gemcitabine-induced ERK activation through inhibition of the IQGAP1-ERK1/2 signalling axis in a manner independent of its enzymatic activity. Co-treatment of FBP1-derived small peptide inhibitor FBP1 E4 enhanced the anti-cancer efficacy of gemcitabine in pancreatic cancer [27]. However, the molecular mechanisms by which FBP1 antagonizes cancer chemoresistance in pancreatic cancer warrant further investigation. Moreover, whether FBP1 exerts similar effect in chemoresistance in other cancer types needs to be elucidated.

The mechanisms for the aberrant expression of FBP1 are mainly focused on the epigenetic regulation of the FBP1 promoter. However, recently, ubiquitin-mediated degradation and copy number loss of FBP1 were also shown to explain the loss of FBPase in cancer (Table 2).

Copy number loss of FBP1 was observed in ccRCC cases [86]. Copy number loss of FBP1 was significantly associated with lower FBP1 expression in HCC [15]. These data indicated that genomic alterations were also responsible for FBP1 inhibition.

Recently, the E3 ubiquitin ligase TRIM28 has been found to play a critical role in regulating FBP1 protein levels through a post-translational mechanism in HCC. TRIM28 was found to directly bind to and promote the ubiquitination and degradation of FBP1. MAGE-A3 and MAGE-C2, which are known to specifically bind to TRIM28 [94, 116], can enhance TRIM28-dependent degradation of FBP1 by forming ubiquitin ligase complexes with TRIM28 [80].

Several chemical inhibitors have been shown to successfully restore the expression of FBP1. Drugs targeting promoter methylation, HDAC and the upstream regulator of FBPase, have been proven efficacious. Their interactions are detailed below (Fig. 4).

Recently, 5AZA, a hypomethylating agent [126], together with HDAC inhibitors, such as trichostatin A (TSA), sodium butyrate (NaBu) [127], SAHA [128] and LBH589 [129], have been approved by the Food and Drug Administration (FDA) of the USA for cancer therapy in myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML) [130]. A combination of DNA methyltransferase inhibitors and HDAC inhibitors has also shown promising synergistic effects in the treatment of MDS, AML and non-Hodgkin’s lymphoma [131]. Treatment of cells with 5AZA resulted in a significant increase in the expression of FBPase mRNA in breast cancer, gastric cancer and lung cancer [13, 25, 75], while treatment with sodium butyrate, SAHA and LBH589 upregulated FBP1 protein and mRNA expression in HCC [81]. Restoration of FBP1 expression by HDAC inhibitors led to a switch from glycolysis to gluconeogenesis, altered energy metabolism and inhibition of tumour growth [81]. All these data indicated that the silencing of FBP1 can be a target of methyltransferase inhibitors and HDAC inhibitors for the potential treatment of cancer. However, we are still a long way from the clinical application of these epigenetic drugs for solid tumours [132].

LSD1, the first histone demethylase identified, converts active H3K4me2/3 to the less active H3K4me1 mark, leading to gene activation [133]. Meanwhile, LSD1 also converts the inactive H3K9me3 into the less repressive H3K9me1 or H3K9me2 marks, thereby leading to gene derepression [134]. Thus, the functional outcome of LSD1 activity depends on the balance between the modification of H3K4me2/3 or H3K9me3. Tranylcypromine (TCP) is a potent inhibitor of the demethylation activity of LSD1 [135]. FBP1 was robustly and quickly induced by TCP treatment in HepG2 cells [114]. However, before using LSD1 inhibitors, their pleiotropic actions should be taken into careful consideration.

Bortezomib, a 26S proteasome inhibitor, has been approved by the FDA for the treatment of multiple myeloma and mantle cell lymphoma [136‐139]. In HCC, a TRIM28-induced decrease in FBP1 protein levels (but not mRNA levels) was completely inhibited by bortezomib treatment. Meanwhile, bortezomib-induced decreases in glucose consumption, lactate levels and cell growth inhibition were largely diminished by the knockdown of FBP1. These data indicated that bortezomib could regulate the Warburg effect by inhibiting the proteasome-dependent degradation of FBP1 [80], which might be harnessed in combination therapy [140], but not bortezomib monotherapy, for cancer [141].

In addition, drugs targeting upstream regulators of FBPase have shown efficacy in restoring FBPase expression. Ras, the first oncogene isolated from human tumour cells, was aberrantly activated in many cancers [142]. In gastric cancer, Ras-induced FBP1 downregulation was reversed after the inhibition of NF-ΚB activity by either a chemical inhibitor of NF-ΚB, BAY11-7085, or a genetic suppressor of NF-ΚB, IkB-alpha M [26, 143]. Dexamethasone, an active form of synthesized glucocorticoids, restored the expression of gluconeogenesis genes, including FBP1, thereby antagonizing the Warburg effect and leading to therapeutic efficacy in the treatment of hepatocarcinoma [89]. Nevertheless, considering the complex regulatory networks of these regulators, adverse effects and toxicity caused by poor specificity may always be potential problems in their clinical applications.

Research on metabolic reprogramming in cancer is progressing at an unprecedented pace. FBPase, one of the rate-limiting enzymes responsible for gluconeogenesis, is usually found to be downregulated in many cancers and is treated as a tumour-suppressor gene. Meanwhile, these results should be interpreted with caution due to only one to two publications in some cancer types such as in colon cancer, pancreatic cancer and small intestinal neuroendocrine tumor. In most cancers, FBPase goes far beyond its enzymatic function, as it is located inside the cell nucleus and is efficacious through a catalytic-activity-independent mechanism via direct interaction with other genes. In our opinion, to maintain rapid proliferation and differentiation, all cancer cells will confront a microenvironmental energy crisis, including the intrinsic shortage of metabolic substrates and energy as well as the external deterioration of the microenvironment due to factors such as hypoxia, acidosis and hypoglycaemia. To overcome all these unfavourable growth conditions, cancer cells have to reprogram their metabolic and epigenetic phenotypes through the activation of oncogenes and inactivation of tumour-suppressor genes, altering the pattern of epigenetic modification and leading to the aberrant expression of numerous genes, including those involved in metabolic rate-limiting, metastasis or differentiation. Therefore, it is not surprising to find an aberrant epigenetic modification of FBPase in cancer. Furthermore, it is also not surprising to find a close correlation between FBPase expression and transcriptional factors such as HIF, EMT transcriptional factors and cancer differentiation. However, the complete mechanism underlying the interplay of FBPase and other genes is complicated and unclear. Does noncoding RNA play an important role in the regulation of FBPase expression? What is the meaning and intrinsic mechanism of the interplay between aberrant FBPase and other cancer hallmarks? Investigating the role of FBPase in CSCs and apoptosis, for example, will promote our understanding of cell differentiation and programmed cell death and reveal new clues for combating cancer.

Several problems remain to be solved for the use of FBPase as a target in cancer treatment. One main obstacle is the identification of distinct regulatory markers exclusive to aberrant FBPase. Without this knowledge, drugs targeting aberrant FBPase with higher specificity but fewer side effects can’t be found or designed. Another problem is identifying the optimal timing for the use of FBPase-targeting drugs, as these drugs may have completely different effects at different stages of progression of different types of cancer. These drugs may induce a switch from aerobic glycolysis to OXPHOS and inhibit the aggressive phenotypes of some cancers; meanwhile, in other circumstances, they may help promote the survival of cancer cells and sustain metastasis. Co-treatment of drugs targeting FBPase with conventional chemotherapy is promising, as expression of FBPase promotes differentiation and apoptosis and inhibits the chemoresistance of cancer cells. However, the efficacy of cancer drugs targeting FBPase warrants further investigation in the real world.