An Anti-tumorigenic Role of the Warburg Effect at Emergence of Transformed Cells
2018 Volume 43 Issue 2 Pages 171-176
Details
2018 Volume 43 Issue 2 Pages 171-176
The Warburg effect is one of the hallmarks of cancer cells, characterized by enhanced aerobic glycolysis. Despite intense research efforts, its functional relevance or biological significance to facilitate tumor progression is still debatable. Hence the question persists when and how the Warburg effect contributes to carcinogenesis. Especially, the role of metabolic changes at a very early stage of tumorigenesis has received relatively little attention, and how aerobic glycolysis impacts tumor incidence remains largely unknown. Here we discuss a novel paradigm for the effect of the Warburg effect that provides a suppressive role in oncogenesis.
Key words: Warburg effect, aerobic glycolysis, cell competition, EDAC
Dr. Otto Warburg, a German physiologist, reported a monumental study that cancer cells prefer the aerobic breakdown of glucose even when they are in the presence of abundant oxygen (Warburg, 1956). Neoplastically transformed cells rewire their metabolism to satisfy demands of growth and proliferation. This metabolic reprogramming is widely recognized as the Warburg effect (Burns and Manda, 2017; Koppenol et al., 2011; Vander Heiden et al., 2009). The Warburg effect is a robust metabolic hallmark of most tumors, thereby leading to clinical applications such as tumor imaging (Engelman et al., 2008; Gatenby and Gillies, 2007; Higashi et al., 2000). Although enhanced aerobic glycolysis is a common trait of tumors, its role in cancer development remains a subject to debate. Hence ever since its discovery, the Warburg effect is still an unresolved puzzle; Are metabolic changes drivers of cancer progression or do they just come along for the ride (Devic, 2016; Liberti and Locasale, 2016)? It is generally conceived that the Warburg effect promotes both of cell viability and metastatic potential of malignant tumors (Fantin et al., 2006; Shim et al., 1998). Despite its intense interest, much less is known about how the metabolic alteration impacts cancer development in the context of different tumor stages. To understand cancer as a metabolic disease (Seyfried et al., 2014; Seyfried and Shelton, 2010), it is necessary to uncover how the metabolic reprogramming occurs at the initial stage of carcinogenesis, in other words, at emergence of the first transformed cells. In this review, we mainly summarize recent findings on the biological consequence of the Warburg effect-like metabolic modulation in the onset of cancer incidence.
The cellular environment plays a preventive role in cancer progression. In light of this concept, it is becoming increasingly apparent that normal epithelial cells can recognize the newly emerging suboptimal cells and compare their fitness levels, leading to elimination of unfit cells. This biological phenomenon is called cell competition and is conceived to be one of essential functions to ensure tissue homeostasis (Amoyel and Bach, 2014; Claveria and Torres, 2016; Di Gregorio et al., 2016; Kajita and Fujita, 2015; Kon, 2018). Our group recently reported that the Warburg effect-like metabolic changes are induced in H-RasG12V (hereafter abbreviated as RasV12)-transformed epithelial cells when they are surrounded by normal cells and play a pivotal role in apical elimination of transformed cells as ineligible ones (Kon et al., 2017). When RasV12-transformed Madin-Darby canine kidney (MDCK) cells are co-cultured with normal MDCK cells, RasV12 cells exhibit a reduction in mitochondrial membrane potential as evidenced by decreased incorporation of tetramethylrhodamine methyl ester (TMRM) relative to RasV12 cells cultured alone. In addition, the uptake of glucose is promoted in RasV12 cells, and lactic acid fermentation, a readout of the glycolytic pathway, is enhanced through upregulation of lactate dehydrogenase A (LDHA). These observations indicate that the Warburg effect-like metabolic changes are induced in RasV12-transformed cells when they emerge in an epithelial sheet. On the molecular basis, pyruvate dehydrogenase kinase 4 (PDK4), one of four PDK isoenzymes, is non-cell autonomously upregulated in RasV12 cells and inactivates pyruvate dehydrogenase (PDH) complex by phosphorylating its E1α subunit. PDH is a gatekeeper for mitochondrial glucose oxidation by catalyzing irreversible decarboxylation of pyruvate into acetyl-CoA, thereby inhibition of PDH activity results in declined mitochondrial activity (Roche and Hiromasa, 2007; Saunier et al., 2016). The above described anti-tumorigenic function exerted by normal epithelial cells is termed as epithelial defense against cancer (EDAC) and involves a mechanical force generated by the accumulation of Filamin A, a crosslinker protein of actin filament, in normal cells adjacent to transformed cells (Kajita et al., 2014). It has also been uncovered that epithelial protein lost in neoplasm (EPLIN) is a crucial regulator for the EDAC process via transducing downstream signals such as myosin-II and protein kinase A (PKA) in transformed cells, and its localization is regulated in a spatiotemporal manner (Ohoka et al., 2015). Importantly, EDAC elevates the expression of PDK4, and EDAC-deficient cells (Filamin A- or EPLIN-knockdown) do not cause the cell competition-induced metabolic reprogramming. In addition, PDK4-knockout RasV12 cells are not eliminated and remained within the epithelia when co-cultured with normal cells. The fact that PDK4 inhibition diminishes the Warburg effect-like metabolic phenotype underscores that PDK4-mediated mitochondrial dysfunction markedly potentiates the glycolytic pathway. Taken together, these findings indicate that EDAC induces the Warburg effect-like metabolic shift, which is required for RasV12 cells to be eradicated (Fig. 1). Of particular note, PDK4 is downregulated in a wide variety of human tumors, implying that PDK4 could function as a tumor suppressor (Grassian et al., 2011). One of essential questions is whether the aberrant mitochondrial activity itself accounts for the loser status in cell competition. In the manuscript, it is demonstrated that PDH-knockdown, thereby inhibiting the entry of pyruvate into the tricarboxylic acid (TCA) cycle does not prime cells as losers (Kon et al., 2017). This result suggests that downregulated mitochondrial activity per se is not a trigger for competitive interaction.
The molecular pathways that cause the Warburg effect-like metabolic alterations in transformed cells surrounded by normal cells. EDAC from normal cells causes accumulation of EPLIN in the neighboring transformed cells, which turns to upregulate PDK4 and lead to mitochondrial dysfunction. The molecules that serve as a bridge between EPLIN and PDK4 are currently unidentified.
The functional significance of PDK4-mediated mitochondrial deregulation in the elimination of RasV12-transformed cells is also substantiated in intestinal epithelial tissues of mice. The novel mouse model in which bicistronic expression of H-RasG12V and eGFP is under the control of a floxed STOP transcriptional cassette has been established. RasV12 remains transcriptionally silent until the STOP cassette is removed by a Cre recombinase. Because the efficiency of Cre-mediated recombination in Cre-ERT2 mice is dependent on the amount of tamoxifen, injection of a low dose of tamoxifen results in the generation of genetic mosaics where only a fraction of cells undergoes a recombination event. Thus, this model not only offers the advantage of controlled expression of RasV12 mutation but also provides a platform applicable to the analysis of cell competition between normal and RasV12-transformed cells in vivo. Using this mouse model, it was observed that the vast majority of RasV12 cells were apically extruded in the intestinal lumen. Furthermore, mitochondrial membrane potential is decreased in RasV12 cells when surrounded by normal cells, and restoration of mitochondrial activity by antagonizing PDK4 compromises apical elimination of RasV12 cells (Kon et al., 2017).
As described above, newly emerging transformed cells are apically eliminated by the surrounding normal cells through the PDK4-mediated Warburg effect-like metabolic shift. This suggests a novel mechanism for the inception of the Warburg effect-induced tumor suppressive process, and is in stark contrast to the historical view in which the Warburg effect plays a positive role in cancer progression (hereafter referred as conventional Warburg effect). However, the EDAC-induced metabolic changes share many aspects with the conventional Warburg effect. For instance, cells display an increased uptake of glucose and higher production of lactate. The aerobic glycolysis was originally viewed as a compensatory mechanism for dysfunctional respiration (Warburg, 1956), and reduced mitochondrial activity is observed in the EDAC-induced metabolic reprogramming (Kon et al., 2017). Nevertheless, it should be noted that several accumulating studies have argued that mitochondria normally function in substantial cases of cancer (Crabtree, 1929; Fantin et al., 2006; Koppenol et al., 2011; Maldonado and Lemasters, 2014; Moreno-Sanchez et al., 2007; Weinhouse, 1956, 1976), suggesting that mitochondrial impairment is not always associated with the Warburg effect phenotype. The conventional Warburg effect can be provoked through activation centered on hypoxia-inducible factor 1α (HIF-1α)-related genes such as glucose transporters (GLUTs), hexokinase 1/2 (HK1/2), pyruvate kinase M2 (PKM2), PDK1/3 and LDHA (Kim et al., 2006; Luo et al., 2011; Pouyssegur et al., 2006; Prigione et al., 2014; Semenza, 2010). Given the intimate link between HIF-1α and aerobic glycolysis, our group thoroughly investigated the involvement of HIF-1α activity in cell competition. However, there was no obvious evidence that HIF-1α activity is promoted in transformed cells when co-cultured with normal cells. Instead of PDK1/3, PDK4 was identified as one of the prime molecules to enhance aerobic glycolysis during the process of EDAC (Fig. 2). Best documented to upregulate the expression of PDK4 is peroxisome proliferator-activated receptor (PPAR) family (Abbot et al., 2005; Muoio et al., 2002; Wende et al., 2005; Zhang et al., 2006). It is very intriguing that PPARγ expression is very susceptible to mechanical stimuli as compressive forces regulate the PPARγ expression (Li et al., 2013; Tanabe et al., 2004). In addition, EPLIN has been proposed to function as a mechanosensor by sensing actomyosin fibers at adherens junctions (Taguchi et al., 2011), suggesting that mechanical forces exerted by normal cells against transformed cells would underlie the induction of metabolic reprogramming. Whether the PPAR transcriptional complex is involved in the EDAC process remains unclear at present, which should be addressed in future studies. Interestingly, upon detachment from extracellular matrix (ECM), cells show enhanced expression of PDK4, leading to a metabolic impairment (Grassian et al., 2011; Kamarajugadda et al., 2012). This implies that PDK4 activation is closely associated with a dissociation phenotype irrespective of the oncogenic status of cells. Taken together, these findings indicate that the PDK4-modulated mitochondrial activity generally influences the biological behavior of cells dissociated from an epithelial layer.
A schematic representation of the difference between EDAC-induced Warburg effect and conventional Warburg effect. In the EDAC-induced Warburg effect, a mechanical force generated from surrounding normal cells causes the Warburg effect-like metabolic changes in transformed cells via the Filamin-EPLIN-PDK4 pathway. This metabolic shift results in the elimination of transformed cells. In contrast, at the mid– to late- stage of carcinogenesis environmental stresses such as hypoxia, scarce nutrients and low pH induce the Warburg effect in cancer cells, resulting in a selective advantage for survival, invasion and metastasis.
Tumor cells are subjected to a remarkable array of pressures in a harsh condition such as hypoxia and scarce nutrients. The Warburg effect was initially considered as an adaptation to such environments. It has been postulated that the Warburg effect confers neoplastic cells with many biological advantages to sustain the uncontrolled proliferation. First, the enhanced glucose consumption serves to produce cellular building blocks (e.g., nucleotides, amino acids and lipids) to meet the requirement of rapidly proliferating cancer cells (Cairns et al., 2011; DeBerardinis et al., 2008; Levine and Puzio-Kuter, 2010; Lunt and Vander Heiden, 2011). The increase in glycolytic flux allows glycolytic intermediates such as glucose-6-phosphate or fructose-6-phosphate which can be used for nucleotide synthesis, whereas 3-phosphoglycerate and pyruvate are key precursors in the biogenesis of several amino acids. Under the condition of increased aerobic glycolysis, citrate converted from acetyl-CoA is exported from mitochondria. In the cytosol, citrate is delivered as acetyl-CoA for the synthesis of fatty acids where β-nicotinamide adenine dinucleotide 2'-phosphate, reduced (NADPH) is required for this step (Vander Heiden et al., 2009; Ward and Thompson, 2012). It is currently unknown how the production rate of these macromolecules is altered and whether it accounts for the biological consequence of the EDAC-induced Warburg effect.
In addition to these biosynthetic functions, the rewiring of metabolic status also affects the generation of reactive oxygen species (ROS). There is enormous evidence that large amounts of glycolytic intermediates are diverted to the pentose phosphate pathway (PPP) to produce reducing equivalents in the form of NADPH. NADPH is a major cellular antioxidant which maintains glutathione in a reduced state to secure the redox balance. The electron transport chain (ETC) is a major source of ROS production as leaky electrons react with oxygen to produce superoxide across the respiratory chain. Given that transformed cells are inherently under increased oxidative stress as a result of a higher rate of proliferation, downregulated mitochondrial activity as an antioxidant mechanism has been proposed to function to lower oxidative burden, which potentiates cell viability (Brand and Hermfisse, 1997; Fantin and Leder, 2006; Lee and Yoon, 2015). In this regard, it is tempting to assume that the PDK4-mediated decrease in mitochondrial membrane potential is beneficial for cells to negate oxidative stress induced by EDAC. However, the opposite model of mitochondrial function in that it counteracts ROS through NADPH production by isocitrate dehydrogenase 2 (IDH2), a TCA cycle enzyme, has been recently put forward (Hawk et al., 2018; Jiang et al., 2016). In addition, it has been implicated that ROS is a vital regulator for cellular activity, as not only does ROS production cause DNA damage, but also function as a signaling molecule (D’Autreaux and Toledano, 2007; Liberti and Locasale, 2016). For instance, ROS inactivates phosphatases such as phosphatase and tensin homolog (PTEN) and protein tyrosine phosphatases. From this perspective, EDAC-induced ROS might act as a mediator to transduce downstream signal(s). Future studies aimed at delineating how the redox status is changed in transformed cells during cell competition are necessary.
In mitochondrial oxidative phosphorylation (OXPHOS), oxidation of one glucose generates 36 molecules of adenosine 5'-triphosphate (ATP), whereas glycolysis in cytosol produces a net gain of 2 ATP. Thus, aerobic glycolysis appears at first glance to be an inefficient means to generate ATP as the mitochondrial oxidative phosphorylation can maximize ATP production (Burns and Manda, 2017; Locasale and Cantley, 2011). However, ATP production by glycolysis is up to 100 times faster than that of OXPHOS (Pfeiffer et al., 2001; Shestov et al., 2014). Hence if extracellular glucose is abundant, the metabolic shift to glycolysis would allow cells to meet acute energy demand (DeBerardinis et al., 2008; Guppy et al., 1993; Liberti and Locasale, 2016; Locasale and Cantley, 2011). In line with this concept, our group found that the intracellular ATP level is profoundly higher in RasV12 cells surrounded by normal cells compared to RasV12 cells cultured alone based on the analysis of FRET-based ATP imaging (Kon et al., 2017). The previous findings that myosin-II-driven contraction, PKA activation and enhanced endocytosis in transformed cells are required to force them out of epithelia (Anton et al., 2014; Hogan et al., 2009; Saitoh et al., 2017) suggest that rapid ATP production might provide free energy for transformed cells to sustain those biological reactions (Fig. 3).
A proposed model for the biological consequence of the Warburg effect-like metabolic changes at emergence of transformed cells. EDAC causes a reduction in mitochondrial membrane potential in the neighboring transformed cell, coupling with increased glycolysis. Higher production of intracellular ATP generated by the metabolic shift could be utilized for various biological reactions to leave away from an epithelial sheet.
In contrast to known pro-tumorigenic effects of the Warburg effect, the EDAC-induced Warburg effect-like metabolic shift acts in a tumor-suppressive capacity. Molecularly non-overlapped yet enhanced aerobic glycolysis establishes the diverse role of the Warburg effect in carcinogenesis; the fate of transformed cells is different depending on tumor stages. The tumor stage-specific key regulators of the Warburg effect should be identified and would present inviting target(s) for cancer treatment or prevention. Given that tumor is kind of the top of a mountain consisting of a multitude of transformed cells, a key question is whether the metabolic signature which is tagged in the evolutionary origin of cancer can be traced or vanished. Future works will help to fully understand when and how the metabolic signature is switched in cancer evolution.
This work is financially supported by The Sumitomo Foundation (170705).
