Cancer Cell Metabolism Connects Epigenetic Modifications to Transcriptional Regulation

Abstract

Adaptation of cellular function with the nutrient environment is essential for survival. Failure to adapt can lead to cell death and/or disease. Indeed, energy metabolism alterations are a major contributing factor for many pathologies, including cancer, cardiovascular disease, and diabetes. In particular, a primary characteristic of cancer cells is altered metabolism that promotes survival and proliferation even in the presence of limited nutrients. Interestingly, recent studies demonstrate that metabolic pathways produce intermediary metabolites that directly influence epigenetic modifications in the genome. Emerging evidence demonstrates that metabolic processes in cancer cells fuel malignant growth, in part, through epigenetic regulation of gene expression programs important for proliferation and adaptive survival. In this review, recent progress towards understanding the relationship of cancer cell metabolism, epigenetic modification, and transcriptional regulation will be discussed.

Specifically, the need for adaptive cell metabolism and its modulation in cancer cells will be introduced. Current knowledge on the emerging field of metabolite production and epigenetic modification will also be reviewed. Alterations of DNA (de)methylation, histone modifications, such as (de)methylation and (de)acylation, as well as chromatin remodeling, will be discussed in the context of cancer cell metabolism. Finally, how these epigenetic alterations contribute to cancer cell phenotypes will summarized. Collectively, these studies reveal that both metabolic and epigenetic pathways in cancer cells are closely linked, representing multiple opportunities to therapeutically target the unique features of malignant growth.

Graphical Abstract

Cancer cell metabolism fuels malignant growth, in part, through epigenetic regulation of gene expression programs important for proliferation and adaptive survival. Specifically, extracellular metabolites feed metabolic pathways that promote cancer cell growth. These metabolic pathways also produce intermediary metabolites required for histone modifications, which in turn regulates transcription of genes in cell growth pathways.

The coordination of metabolic function with the environment

The interdependence of nutrient availability, cell growth and division have long been recognized [1–3]. Growth (i.e. the production of cellular building blocks, or biomass) is required for creation of a new cell during division and is directly connected to metabolic pathways, such as glycolysis, the tricarboxylic acid cycle (TCA) cycle, and oxidative phosphorylation [4] (Fig. 1). For example, yeast mutants originally identified as having defects in cell growth and division were later found to have metabolic deficiencies. Indeed, an abundance of Cell Division Cycle (CDC) genes are involved in energy metabolic and biosynthetic pathways [4]. For example, Cdc53 is a cullin and structural component of SCF complex involved in the G1-S transition, as well as a regulator of methionine biosynthesis genes that contribute to nucleotide and protein synthesis [5].

Figure 1. Epigenetic modifications linked to cancer cell metabolism.

Cancer cells have increased consumption of glucose for aerobic glycolysis that supports nucleotide and protein synthesis. G6P, glucose-6-phosphate. 3-PG, 3-phosphoglycerate. Ser and Gly is serine and glycine, respectively. One carbon metabolism produces s-adenosyl-methionine (SAM), a required cofactor for DNA methyltransferase (DNMT) and lysine methyltransferase (KMT). The methyltransferase reaction produces S-adenosyl-homocysteine (SAH). Glutamine metabolism supports lipid biosynthesis via the TCA cycle. α-ketoglutarate (α-KG) produced during glutaminolysis is a required cofactor for DNA demethylase TET enzymes and lysine demethylases (KDM). A byproduct of the demethylase reaction is succinate. Mutant isocitrate dehydrogenase (mut IDH) converts α-ketoglutarate to 2-hydroxyglutarate (2-HG), which inhibits demethylase reactions. Acetyl-CoA produced in the TCA cycle and β-oxidation is used by histone acetyltransferases (HAT). Fatty acid β-oxidation is important for membrane synthesis and also produces acyl-CoA that can be used by HATs for additional histone acylation reactions. Neoplastic cells also exhibit oxidative phosphorylation (OxPhos) to produce energy. Class III histone deacetylation (HDAC) reactions by sirtuins require NAD+ produced during oxidative phosphorylation. The deacetylation reaction produces O-acetyl ADP-ribose (OAADR) as a byproduct.

In changing nutrient-rich or -limiting environments, cells have the remarkable ability to sense these dynamic environments and reprogram their energy metabolism and/or proliferative capacity accordingly. Indeed, dynamic nutrient environments are ubiquitous throughout nature and include competitive growth environments of proliferating microorganisms and tissue niches in multicellular organisms. Prime examples of metabolic plasticity can be found in biological systems that display oscillations in environmental nutrients and intracellular metabolism. For example, in metabolically-synchronized S. cerevisiae, homeostasis is observed through elegant organization of energy availability and cellular programs related to growth and division [4,6,7], called the Yeast Metabolic Cycle. Corresponding temporal oscillations in metabolite abundance were also observed as early as the 1960s [8–10]. Such organization entails temporal separation of metabolic processes, such that biogenesis (production of building blocks for a new cell) occurs before cell division. Importantly, this metabolic coordination allows cells to quickly organize and adapt their metabolic output and proliferation in synchrony with changing environmental conditions.

These Yeast Metabolic Cycles are not unlike circulating glucose and insulin oscillations observed in the sleep-wake cycles of circadian rhythms in mammals. Indeed, pancreatic β cells exhibit exquisitely synchronized fluctuations of glucose metabolism, ATP production, calcium signaling, and insulin secretion [11]. In addition, oscillations in mitochondrial respiration and glycolysis have been observed in the heart, liver, and neurons [12]. Collectively, these observations demonstrate the importance of metabolic plasticity in coordination with the nutrient environment in order to optimize cellular survival, fitness and growth programs.

Metabolic reprogramming in cancer cells

Cancer cells exhibit profound metabolic plasticity in order to survive and proliferate in diverse microenvironments in vivo, including expansion of the primary tumor, dissemination and migration into new tissue with differing nutrient availability, and growth of secondary tumors requiring high metabolic demands for proliferation [13,14].

The metabolism of cancer cells differs dramatically from that of normal cells in several ways. First observed by Otto Warburg in 1924, many cancer cells metabolize glucose via aerobic glycolysis, rather than oxidative phosphorylation, thus prioritizing biosynthesis over energy production. Later studies expanded these observations and determined that elevated rates of glucose and glutamine consumption, lipid biosynthesis, pentose phosphate metabolism, autophagy and redox maintenance are all characteristics of rapidly proliferating cancer cells [13] (Fig. 1). The metabolic signature of cancer cells is so distinguishable from that of normal cells that current clinical use of positron emission tomography (PET) measuring regional glucose metabolism has a greater than 90% accuracy in the detection of epithelial metastases [15].

Although many of the advantages of cancer cell metabolism in vivo are still being determined, there are several selective pressures that may explain the evolution of these distinct metabolic signatures [16]. For example, increased use of glycolytic pathways decreases the dependence on highly aerobic conditions that may not be available prior to stimulation of angiogenesis in surrounding tissue. Furthermore, intermediates of the glycolytic pathway can also be used for anabolic reactions, such as pyruvate for alanine and malate synthesis.

Importantly, cancer cells must cohabitate with non-cancerous surrounding tissue, thus a symbiotic relationship develops that supports maintenance of normal and growth of cancer cells. For example, cancer cells produce lactate as an end product of aerobic glycolysis. Lactate may be transported into neighboring stromal cells to regenerate and secrete pyruvate, thus refueling cancer cell metabolism [17]. In addition, ‘Reverse Warburg Effect’ has been observed and is consistent with high glycolytic activity in stromal cells that produce by-products to feed cancer cell metabolism [18]. This type of cell-to-cell metabolic cooperation is reminiscent of ‘quorum-sensing’, which enacts coherent cellular changes in response to population density. Quorum-sensing mechanisms have been observed in pathogenic and microorganisms, and have also been proposed to assist cancer cells during metastatic colonization in different tissues [19,20]. Indeed, because tumor microenvironments are extremely cell dense with signaling among cancer cells and between cancer and stromal cells, cooperative behavior that enhances metabolic function and fitness is likely to occur. These types of cooperative behaviors are particularly important given the fluctuations in nutrient availability in circulating blood.

Cancer cells can achieve this metabolic reprogramming through multiple genetic alterations. For example, oncogenic mutations in RAS and BRAF, which are commonly found in melanoma and cancers of the pancreas, lung, colon and rectum, result in increased glucose uptake even in the presence of limiting nutrients [21,22]. In addition, the Myc oncogene, which is overexpressed in many cancers, regulates a transcriptional program associated with increased glucose and glutamine consumption in cancer cells [23]. Indeed, cancers that depend on Myc overexpression for survival often exhibit ‘oncogene addiction’ [24], the dependency on a single activated oncogenic protein or pathway to maintain their malignant properties [25], the origins of which may have roots in addiction to glutamine.

Epigenetic regulation facilitates metabolic reprogramming in cancer cells

Adaptive cellular responses, including metabolic adaptation, are often achieved by inducible changes in gene expression programs [26]. An ideal mechanism to achieve this is through epigenetic modification, which is rapid and reversible, and can occur through numerous enzymes, such as DNA and histone (de)methylases and (de)acetylases.

Indeed, changes in chromatin architecture are known to regulate inducible gene expression in response to intra- and extracellular signals. For example, temporal metabolic oscillations in the Yeast Metabolic Cycle are largely attributed to dynamic chromatin changes that influence the expression of metabolic genes. For example, bursts of mitochondrial respiration produce elevated intracellular acetyl-CoA levels generated by the TCA cycle [27]. A corresponding increase in acetyl-CoA-dependent histone acetylation and increased expression of genes related to cell growth and division is also observed during respiration [28,29].

This connection between metabolic pathways and histone modifications is not limited the Yeast Metabolic Cycle. Interestingly, in mammalian systems, required cofactors for several epigenetic modifications are intermediary metabolites produced during metabolic pathways, such as glycolysis and the TCA cycle (Fig. 1). Moreover, periodic histone acetylation, methylation, and chromatin-remodeling coincides with changes in expression of circadian and metabolic genes at specific times of the circadian clock in mammals [30,31]. For example, different studies demonstrate that 3–20% of genes from different mouse tissues exhibit periodic expression in tune with circadian rhythms [32]. In mouse livers, histone acetylation across the genome, and particularly on genes that regulate circadian cycles and metabolism of glucose and lipids, promotes gene expression [31,33–36]. NAD+-dependent deacetylation of circadian gene loci also exhibits periodic rhythms and helps to create a tightly-regulated circadian transcriptional program [37]. Importantly, the core circadian transcription factor, CLOCK, is an acetyltransferase itself [38], illustrating the interdependent nature of circadian metabolism, histone modification, and transcriptional regulation.

Clearly, epigenetic regulation of metabolic processes is an ideal mechanism to elicit rapid adaptive changes in gene expression, which would provide survival and proliferative advantages for cancer cells. In the subsequent sections, several metabolic pathways and their related epigenetic modifications will be discussed in relation to cancer cell metabolism.

One-carbon metabolism promotes DNA and histone methylation

One-carbon metabolism includes the folate and methionine cycles that generate one-carbon units (i.e. methyl groups). These methyl groups are donated to methionine recycling pathway, thymidylate synthesis, and purine synthesis needed for the production of new nucleotides during DNA replication (Fig 1).

The significance of one-carbon metabolism for malignant growth was identified over 60 years ago by Sydney Farber, MD, who found that disruption of the folate cycle with a folic acid antagonist resulted in temporary remission in children with acute leukemia [39]. During one-carbon metabolism, serine donates its side chain to tetrahydrofolate to drive the folate cycle, which in turn recycles methionine from homocysteine. Inhibition of serine synthesis, a major source of one-carbon units, has also been found to disrupt cancer progression [40]. Cancer cells can potentiate one-carbon metabolism through overexpression of metabolic enzymes, such as of 3-phosphoglycerate dehydrogenase (PGDH), which redirects glycolysis intermediates to the serine synthesis pathway [41].

Serine and ribose-5-phosphate can also be used for de novo purine synthesis in cancer cells [42]. The resulting ATP and methionine contributed by serine catabolism can then be used by methionine adenosyl-transferase to generate S-adenosyl-methionine (SAM) [43], the required cofactor for cellular methylation reactions, including DNA and histone methylation [44]. A by-product of this methylation reaction is S-adenosylhomocysteine (SAH), a potent inhibitor of methyltransferases [45]. In order to prevent its accumulation, SAH is hydrolyzed to homocysteine and adenine by S-adenosylhomocysteine hydrolase (AHCY) [43]. Subsequently, homocysteine can be remethylated to methionine in the folate cycle [40]. Interestingly, dietary consumption of methionine directly influences SAM/SAH ratios, resulting in corresponding changes in methylation patterns and gene expression [46].

DNA methylation, in particular, is one of the most abundant and well-studied epigenetic modification, occurring in approximately 70% of CpG di-nucleotides in the mammalian genome. Virtually all types of cancers have reported alterations in DNA methylation [47]. DNA hypomethylation, and gene expression activation, has been observed at repetitive elements and oncogenes, whereas DNA hypermethylation, and repression, is observed at tumor suppressor loci [48].

Conversely, histone methylation predominately regulates gene expression by recruiting transcriptional activators or repressors to genic loci. While the histone (de)methylases are termed ‘writers’, factors that bind these modifications are collectively termed ‘readers’. Several writers and readers have gain-of-function or loss-of-function mutations in cancer cells. For example, the enhancer of zeste homologue 2 (EZH2) methyltransferase is both mutated and overexpressed in many cancers, leading to aberrant abundance of its methylated product H3 lysine 27 (H3K27) [49]. Methylated H3K27 is important for the temporal regulation of developmental gene expression. Increased and altered localization of H3K27 methylation can result in repression of lineage-specific genes, resulting in a stem cell-like transcriptional program that supports carcinogenesis.

α-ketoglutarate is required for DNA and histone demethylation

As mentioned, cancer cells also have increased consumption of glutamine, which can contribute to different growth promoting pathways, such as amino acid synthesis, nucleotide synthesis, and TCA cycle-related biosynthesis. The first step of glutamine catabolism is its conversion to glutamate, which can promote proliferation through protein synthesis, incorporation in redox pathways, and conversion to α-ketoglutarate for replenishing TCA intermediates [50] (Fig 1).

α-ketoglutarate is also a required cofactor for Jumonji C (JmjC)-containing histone demethylases, which are 2-oxoglutarate-dependent dioxygenases. This demethylase reaction also requires oxygen and Fe(II) and is specific to removal of methyl groups from tri-methylated histones, creating succinate as a byproduct [51].

Another epigenetic modifier that is a 2-oxoglutarate-dependent dioxygenases, is the Ten-Eleven Translocation (TET) DNA demethylases. TET was originally named after a translocation between chromosomes 10 and 11, creating a fusion with the Mixed Lineage Leukemia (MLL) gene [52]. Although TET proteins were discovered quite some time ago, their function as DNA demethylases was only recently discovered to be iterative oxidation of 5-methylcytosine (5mC) to 5-hydroxymethyl cytosine (5hmC), then 5-formylcytosine (5fC), then 5-carboxylcytosine (5caC), by TET1–3 respectively [53–55].

The cellular abundance of these oxidized variants are much lower than that of 5mC [55], with variation in different organisms, tissues, and during development and differentiation [56]. Some technical challenges exist in discriminating 5mC from its oxidized variants in genome mapping experiments [57]. Nevertheless, initial studies reveal that 5hmC is abundant in genic regions, including promoters and exons [58]. Cancers cells exhibit both increases and decreases of 5hmC, however current trends indicate that decreases are more common, particularly in melanoma and colon cancer.

Alteration of the α-ketoglutarate metabolism pathway is also altered cancers. Specifically, mutations in isocitrate dehydrogenase 1 and 2 (IDH1 and 2), which converts isocitrate to α-ketoglutarate during glutamine metabolism, has been observed in human brain cancers and other hematological malignancies [59,60]. These mutations result in the ability of IDH1 and 2 to convert α-ketoglutarate to the oncometabolite 2-hydroxyglutarate, which inhibits DNA and histone demethylase reactions through depletion of the cofactor α-ketoglutarate, and enzymatic inhibition of demethylase activity by the oncometabolite [61] (Fig. 1). Inhibition of mutant IDH activity results in decreased cell proliferation and increased differentiation in glioma cells [61].

Acetyl Co-A is a central growth metabolite and feeds histone acetylation

As previously mentioned, the TCA cycle provides essential intermediary metabolites for nitrogen utilization in building amino acids and nitrogenous bases in nucleic acids. Also, pyruvate derived acetyl-CoA in the TCA cycle is required for biosynthesis of fatty acids, sterols and amino acids. Moreover, acetyl-CoA is also a required cofactor for histone acetylation, an abundant histone modification [62] (Fig. 1).

Recent studies demonstrate that enzymes involved in acetyl-CoA metabolism can translocate to the nucleus to supply nuclear acetyl-CoA pools [63]. This is contradictory to the long-standing assumption that restricts the location of these enzymes to the mitochondria and/or cytoplasm. Histone acetylation levels appear to be coordinated with cellular abundance of acetyl-CoA, as exogenous supplementation of acetate increases histone acetylation. Furthermore, metabolic enzymes, such as ATP-citrate lyase (ACLY) protein, which converts citrate to acetyl-CoA, and acetyl-CoA synthetase (ACCS2) are often overexpressed in cancer cells concomitant with elevated histone acetylation [64–66]. Reduction of these metabolic enzymes markedly decreases histone acetylation and tumor growth [64,65].

Histone acetyltransferases use nuclear acetyl-CoA during high glucose conditions to acetylate histones, creating a permissive state for transcription [67]. Histone acetylation also promotes the expression of genes involved in the TCA cycle [4], thereby creating a feedback mechanism that supports the acetyl-CoA production pathway. Furthermore, histone acetylation regulates the expression of genes involved in cell growth and proliferation, namely ribosome biogenesis, translation, amino acid metabolism, further demonstrating the link between high nutrient environments and promotion of cell proliferation.

Oncogene activation can also promote histone acetylation and subsequent transcriptional growth programs. For example, constitutive expression of oncogenic KRAS (G12D) activates the AKT kinase, which increases glycolytic-production of citrate, as well as ACLY phosphorylation and activation [68]. The result is global increase of histone acetylation in cancer cells expressing oncogenic KRAS.

As previously mentioned, the Myc oncogene regulates glutamine metabolism and biogenesis pathways through altered metabolic gene expression [23,69]. The oncogenic potential of Myc is closely linked to its associated chromatin-modifying abilities. When overexpressed, Myc binds ubiquitously throughout the genome at both canonical Myc binding sites (E box motifs) and intergenic regions, regulating nearly 15% of all genes [70–72]. A mechanism of Myc-induced transcriptional activation involves the recruitment of a coactivator complex that contains the GCN5 histone acetyltransferase [73], which requires acetyl-CoA. This coactivator complex is critical to the function of Myc as a transcription factor and oncogene, as siRNA-mediated GCN5 knockdown blocks Myc-induced histone acetylation and activation of target genes [74]. Myc overexpression also leads to increased production of mitochondrial acetyl-CoA [75], thereby establishing a feed-forward pathway for production and use of acetyl-CoA.

Fatty acid oxidation and histone acylation

It is widely known that fatty acids are used by proliferating cancer cells to promote membrane synthesis. Fatty acid oxidation is also another pathway to produce acetyl-CoA during low glucose conditions. By products of β-oxidation include acetyl-CoA and other acyl-CoAs depending on the carbon chain length of the fatty acid substrate. These acyl-CoA include propionyl-, malonyl-, crotonyl-, butyryl-, succinyl-, glutaryl, 2-hydroxy-isobutyryl-, and β-hydroxy-butyryl-CoA. Interestingly, expression of fatty acid metabolism genes and corresponding histone modifications have also been observed in cultured cells, indicating that, much like histone acetylation, other histone acylation levels are closely linked to the abundance of required cofactors [76,77]. Indeed, known acetyltransferases have been identified to facilitate the histone acylation [78], again suggesting common modes of regulation between acetylation and other acylations.

In cultured cells with ample abundance of glucose and glutamine, these acylations were found to be in low abundance compared to acetyl-CoA [79]. Furthermore, the functions of many acylations are still under investigation. However, initial studies suggest that histone crotonylation, in particular, play roles in transcriptional activation based on in vitro experiments where crotonate was exogenously added to culture media of mammalian cells [80,81]. Furthermore, in the Yeast Metabolic Cycle that spontaneously undergoes distinct periodic expression of metabolic enzymes associated with the TCA cycle and fatty acid oxidation pathway, a transcriptionally repressive role for histone crotonylation was identified [77]. Specifically, histone crotonylation and acetylation were temporally segregated in relation to the gene activation and repression. During oxidative phosphorylation and acetyl-CoA production, histone acetylation and gene activation at ‘growth gene’ loci was observed. However, during fatty acid oxidation, histone crotonylation increases at these same loci, concomitant with decreased expression. Exogenous addition of crotonic acid also reduces expression of ‘pro-growth’ genes. Collectively, these results suggest that histone acylations may have diverse roles in different cellular and organismal contexts.

Clearly, additional research is needed to clarify the roles of histone acylations in transcriptional regulation and their impact in cancer cells. However, initial studies demonstrate that disruptions in fatty acid metabolism from dietary lipid uptake have been linked to cancer progression [82]. Indeed, because histone acylations are extremely responsive to cellular metabolic status, it is likely that they play important roles in cancer cells that need to rapidly adapt to changing nutrient conditions. For example, initial studies demonstrate that histone crotonylation is decreased in hepatocellular carcinoma, and increasing crotonylation inhibits cell motility and proliferation [83].

Oxidative phosphorylation and histone deacetylation

Oxidative phosphorylation produces energy in the form of ATP for numerous cellular processes important for proliferating cancer cells. In the process NADH is converted to NAD+, which are required cofactors for class III histone deacylases, namely sirtuins [84,85] (Fig. 1).

Sirtuins (SIRT) are implicated in aging, diabetes, cancer, cardiovascular disease, inflammatory disease, and neurodegenerative disease, the origins of which have many links to metabolic dysfunction [86]. An initial connection to cancer cell metabolism was identified when SIRT1 was found to deacetylate and repress the activity of the tumor suppressor TP53 [87,88]. These results suggest an oncogenic function for SIRT1, however, because of its diverse roles in different cellular processes, sirtuins have also been associated with tumor suppressor function [89]. Specifically, SIRT2, 3 and 4 can maintain genome integrity through regulation of the cell cycle and/or mitochondrial metabolism [90–92]. SIRT6 or SIRT4 deletion has been observed in several cancers and leads to increased expression of glycolytic and glutamine metabolism genes concomitant with tumor growth in mice [92,93].

Given the important role of histone acylation in regulating growth transcriptional programs, deacylases are expected similarly be regulated by cancer cell metabolism and control proliferative capacity.

Chromatin remodelers promote cancer cell metabolism

Not only are histone modifications directly linked to energy metabolism, but chromatin remodelers are as well. Chromatin remodelers use the energy of ATP to alter the contacts between histones and the DNA to reposition or edit nucleosome composition [94]. Chromatin remodelers have diverse roles in many DNA-templated processes, such as transcription, DNA repair, and replication [95,96]. The first characterized ATP-dependent remodeling complex was the S. cerevisiae SWI/SNF (switch/sucrose non-fermenting) complex [97,98], the subunits of which were originally identified as transcriptional regulators of metabolic pathways fueled by alternative fermentable carbon sources [99,100].

The SWI/SNF complex is highly conserved and regulates energy metabolism in both yeast and mammals. Mammalian SWI/SNF complexes are a family of BRG-/BRM-associated factor (BAF) and Polybromo-associated BAF (PBAF) complexes. The link between BAF/PBAF and mammalian disease has been repeatedly demonstrated, as loss of function contributes to developmental disorders and cancer [101–103].

Intriguing similarities exist between the metabolism of cancer cells and S.cerevisiae, in that both are optimized for rapid proliferation in diverse nutrient environments. S. cerevisiae have also evolved metabolic diversity in carbon catabolism pathways. Specifically, in glucose-rich environments, budding yeast preferentially utilize glycolysis followed by fermentation. When glucose is limiting, cells shift their energy metabolism to respiration [104]. Growth in high glucose results in “glucose repression,” which is characterized by transcriptional repression of genes involved in alternate carbon source metabolism, including those in respiration [105]. The state of “glucose repression” is not unlike that of the “Warburg effect”, where cancer cells utilize aerobic glycolysis to feed growth pathways, such as lipid and protein biogenesis, over energy production via respiration [106].

One type of cancer that is clearly dependent on the Warburg effect is clear cell renal cell carcinoma (ccRCC), named after its cellular histological appearance caused by elevated glycogen and lipid storage resulting from increased glycolysis. Approximately, 46% of ccRCCs have mutations in Polybromo-1 (PBRM1) [107], a subunit of the PBAF chromatin-remodeling complex. It is the second most mutated gene after the Von Hippel-Lindau (VHL) tumor suppressor gene, which is mutated in 48% of ccRCCs. A mouse model of ccRCC caused by inactivation of both PBRM1 and VHL recapitulated the histological features of patient-derived ccRCCs [108]. Importantly, increased glycogen storage and decreased expression of genes in the oxidative phosphorylation pathway are dependent on loss of PBRM1 and VHL [109], demonstrating that BAF/PBAF complexes plays a critical role in the regulation of cancer cell metabolism.

BAF/PBAF are not the only chromatin remodelers known to regulate energy metabolism. The INO80 complex, originally identified for its role in inositol metabolism [110,111], also regulates glycolytic and respiration pathways in budding yeast [112]. Similar to the SWI/SNF complex, the INO80 complex regulates “glucose repression” in yeast. In glucose-rich environments, mutants of the INO80 complex display increased expression of nearly every gene involved in the respiration pathway, while genes in glycolysis are decreased [112]. Accordingly, mitochondrial potential and oxygen consumption are increased in INO80 mutants.

Subunits of the INO80 complex exhibit multiple alterations in a variety of cancers, with notable amplification of the ACTL6A subunit in lung squamous cell carcinomas [107]. Additional studies demonstrate that overexpression of the INO80 subunits are needed to maintain proliferation and anchorage-independent growth of lung cancer cells [113]. Expression analysis also reveal that several INO80 subunits exhibit increased expression in metastatic melanoma compared to primary melanoma and benign nevi [114]. Silencing of INO80 subunit expression impairs melanoma cell growth in culture and in mouse xenografts. Furthermore, in a mouse model for colon cancer, Ino80 haplo-insufficiency decreases tumor formation and increases survival [115]. Taken together, these results suggest that, unlike the BAF/PBAF remodeler, the INO80 complex has oncogenic roles during carcinogenesis.

The roles of INO80 in cancer development may be multi-faceted. For example, alterations in cancer gene expression are observed in cells with increased INO80 subunit expression [113,114]. DNA damage checkpoint activation has also been observed in Ino80 haplo-insufficient cells [115]. However, yeast studies hint that INO80 function may also be related to metabolic reprogramming. As mentioned, loss of INO80 results in decreased glycolytic metabolism [112]. In addition, INO80 deficiency results in uncontrolled proliferation in low nutrient (i.e. metabolically unfavorable) environments [116]. Interestingly, yeast genetic screens identify novel connections between INO80 and the Target of Rapamycin (TOR) signaling pathway [117], which is responsible for coordinating stress and growth responses with environmental cues in both yeast and mammals. The mammalian TOR (mTOR) kinase is deregulated in numerous metabolic disorders and cancers [118]. Recent research reveals that INO80 regulates TOR-dependent transcriptional pathways [116], and suggests that INO80 promotes histone acetylation on growth genes downstream of TOR signaling [117].

Future Directions

The intersection of epigenetics and metabolism has just emerged in the last decade. Thus, knowledge of how these processes are co-opted to fuel cancer evolution has only begun to be collected. Nevertheless, growing evidence demonstrates that the specialized metabolism of cancer cells elicits corresponding changes in the epigenome to promote growth and survival transcriptional programs. Future research may be focused on imposing controlled metabolic fluctuations in mammalian cells and unbiased profiling of corresponding histone marks and gene expression changes. In addition, how chromatin remodelers may be influenced by dynamic changes in metabolism and corresponding histone modifications has yet to be investigated. These studies are critical to understanding the influence of chromatin remodelers in disease and may reveal novel epigenetic therapies to combat carcinogenesis.

Because cancer cells depend on metabolic reprogramming for survival, targeting of cancer cell metabolism has proven successful to limit tumor growth in mouse studies [13,119]. For example, PIK3CA mutant-derived lung adenocarcinomas in mice respond to combinatorial therapies that inhibit both PI3K activity and mTOR, a central regulator of cellular response to environmental nutrients and stress [120]. Also, inhibition of lactate dehydrogenase A in mouse models of non-small cell lung cancer alters pyruvate metabolism in cancer cells and inhibits tumorigenesis [121]. Inhibition of acetyl Co-A catabolism in energy metabolism pathways also reduces tumor burden in mouse models [64]. As siRNA and CRISPR technologies have increased abilities for genetic screens in mammalian cells, several metabolic pathways, including glucose and serine metabolism, have been found to be essential for survival and proliferation of a number of cancer cells [122–124].

As many therapies have already been developed to treat other metabolic disorders, such as heart disease and diabetes, opportunities exist to repurpose them for cancer therapy. Indeed, several have proven effective in limiting growth and proliferation of cancer cells, particularly in combination with other chemotherapies [119], and are currently being tested in clinical trials [125]. The influence of these metabolic therapeutics on epigenetic modifications and transcriptional growth programs are largely unknown, yet studies are growing in number. Likewise, thorough examinations of metabolic alteration in response to epigenetic therapies is lacking. Future research will be needed to resolve these questions and many other in order to expand our understanding of the links between cancer cell metabolism and epigenetic modification.

Abbreviations:

TCA

tricarboxylic acid

acetyl-CoA

acetyl coenzyme A

NAD

nicotinamide adenine dinucleotide

SAM

s-adenosyl-methionine

SAH

s-adenosylhomocysteine

TET

Ten-Eleven Translocation

5mC

5-methylcytosine

IDH

isocitrate dehydrogenase

BAF

BRG-/BRM-associated factor

PBAF

Polybromo-associated BAF

TOR

Target of Rapamycin