Pheochromocytoma/Paraganglioma: A Poster Child for Cancer Metabolism

Sergei G. Tevosian; Hans K. Ghayee


J Clin Endocrinol Metab. 2018;103(5):1779-1789. 

In This Article

Krebs Cycle Mutations

SDH and Succinate

The discovery of loss-of-function mutations in the TCA cycle enzymes SDH and fumarate hydratase (FH) in several cancers lends strong support to the proto-oncogenic function for metabolic products. Germline mutations in the SDHD gene were the first mitochondrial enzyme mutations documented in cancer.[5] SDH is one of the key enzymes in the TCA cycle. SDH is composed of four different subunits: SDHA, SDHB, SDHC, and SDHD. Two additional proteins, SDHAF1 and SDHAF2, are required for its assembly. SDH converts succinate into fumarate and its function is required for proper operation of the mitochondrial respiratory chain.[44] Mutations in SDH are found in renal carcinomas,[45] T-cell leukemia,[46] and gastrointestinal stromal tumors.[47,48]

Most importantly, mutations in SDHB, SDHC, and SDHD subunits were identified in hereditary PCC and PGLs (;[5,32,33,49–55] reviewed in 46). PCC/PGLs are rare, mostly benign, hereditary cancers of the chromaffin tissue that arise either in the adrenal medulla (PCC) or parasympathetic neuronal ganglia in the head and neck and sympathetic thoracolumbar tissue (PGL). Sympathetic PGLs produce catecholamines, whereas parasympathetic PGLs are most often nonsecretory. Mutations in SDHA and the SDH assembly factor SDHAF2 (required for flavination of SDH) are relatively more rare, but have been described in PCC/PGLs.[30,31,57–59] In summary, mutations in every component of SDH have been identified and all of them are associated with PCC/PGLs. In contrast to IDH mutations (discussed later in this article), mutations in SDH genes obey the Knudson rule for tumor suppressors, with a loss-of-function germline mutation followed by a "second hit" somatic loss of the second allele in the tumor.[60]

Lack of SDH function in mouse cells leads to increase in pyruvate consumption and the action of pyruvate carboxylase is essential to replenish aspartate in SDH-deficient cells through pyruvate carboxylation. This metabolic reprogramming presents a vulnerability that could be explored to target tumors carrying SDHx mutation.[61,62]

Inducible transcription factors [hypoxia-inducible factor (HIF)α/β dimers] are key gatekeepers of the response to low oxygen. It has been noted that tumors harboring SDH or FH mutations have a strong hypoxic signature.[39,40,63] This observation prompted investigators to explore the interactions between Krebs cycle substrates and the activation of the hypoxic pathways. Unlike IDH-related tumors [e.g., acute myeloid leukemia (AML) or gliomas], PGL/PCCs have been historically closely associated with hypoxia, because these highly vascularized tumors arise either in tissues known to be susceptible to oxygen deprivation (i.e., cells of the adrenal medulla and organ of Zuckerkandl) or in cells known to serve as oxygen sensors (i.e., neurons of the carotid body). A higher incidence of carotid body tumors was reported for those who live at high altitudes.[64–66] Hence, elucidation of hypoxic pathways can shed light on the genetic and metabolic alterations observed in these tumors and connection from metabolic alterations to tumorigenic process. Pseudohypoxia and HIFs are known to be involved in other hallmark cancer pathways that sustain tumor cell growth, vascularization, and proliferation.[67,68]

Succinate accumulation resulting from SDH deficiency is the hallmark feature of SDH-deficient cells. Similar to 2HG oncometabolite, a product of the IDH gain-of-function mutations (discussed later in this article), succinate-mediated inhibition targets α-ketoglutarate (αKG)-dependent dioxygenases (Figure 1). Although the role for pseudohypoxia and 2HG in the regulation of the prolyl-hydroxylases (PHDs) has been contested, succinate appears to be a viable candidate to serve as their inhibitor.[69,70] HIFα is continuously synthesized; however, under normoxic conditions, PHD-mediated hydroxylation marks it for degradation that involves the activity of the VHL ubiquitination complex.[71] PHD-mediated hydroxylation requires oxygen as a cofactor, so the enzyme is inactive in hypoxia and becomes ineffective in hydroxylation of HIF1α that escapes VHL recognition. Thus, hypoxia protects HIFα from being degraded and the unmodified molecule translocates to the nucleus, where it forms a transcriptionally active HIF heterodimer with a stable HIF1β subunit. The low affinity of PHDs for oxygen makes them a reliable oxygen sensor.[72] PHD inhibition by succinate leads to formation of an active HIF dimer even in the presence of abundant oxygen, a condition known as pseudohypoxia.[69,73]

Figure 1.

Biochemical pathways associated with metabolic enzyme mutations in PCC/PGL and other tumors. Loss-of-function mutations in genes encoding SDH subunits, FH, and MDH2 are shown as red stars. Gain-of-function mutations in IDH are shown in green. The mutations lead to the accumulation of the oncometabolites succinate, fumarate, and D-2HG. Oncometabolites interfere with normal function of αKG-dependent dioxygenases, such as PHDs and TET demethylases. The oncometabolite-stimulated pathways that lead to pseudohypoxia, DNA, and histone methylation are shown. MDH2, malate dehydrogenase; OAA, oxaloacetic acid; PHD, prolyl-hydroxylase; TET, ten-eleven translocation.

As a stable metabolite, mitochondrial succinate appears to be the most suitable signaling molecule to communicate its own excess to cytoplasmic proteins. However, other molecules can contribute to this communication. Reactive oxygen species arise as a result of TCA cycle and respiratory chain dysfunction and can serve as transmitters of SDH loss and oxygen insufficiency to PHDs.[73–77]

Activation of HIF and the pseudohypoxia pathway appears to be particularly important in SDH-driven PGL/PCCs.[15] Gene expression profiling of the tumors confirmed the presence of hypoxia-associated gene expression. Among the activated genes are those encoding proteins involved in angiogenesis, most importantly vascular endothelial growth factor, which correlates with the highly vascular phenotype of these tumors. It was demonstrated that HIF-induced angiogenesis could be targeted clinically by using the receptor tyrosine kinase inhibitor sunitinib to suppress the angiogenic pathway with moderate tumor regression, stability, and decreased catecholamine production upon treatment.[78,79] These clinical cases provided the foundation for clinical trials to evaluate the benefit of this treatment in different groups of patients who have a mutation in genes associated with the pseudohypoxia pathway.[80]

The epithelial-to-mesenchymal transition–associated molecules such as LOXL2 or TWIST that are known to have key roles in vasculogenesis and metastasis are present in SDHx-driven cancer.[81] The dedifferentiation in these tumors provides a rationale for the improved imaging modalities using 18F-2-fluoro-2-deoxy-d-glucose. It was proposed that the increased uptake in 18F-2-fluoro-2-deoxy-d-glucose observed in pseudohypoxic tumors is due to an elevated expression of glucose transporters and other glycolytic enzymes integral to glucose metabolism.[82] Another reason for improved fluorodeoxyglucose detection in cluster 1 tumors is that HIF stabilization may directly increase glucose uptake and increased glycolysis even in the setting of normal oxygen conditions, as required by the Warburg effect. Newer, promising imaging modalities such as 68Ga DOTATE are being used that take advantage of somatostatin receptor expression in PCC/PGLs, and to identify the tumors that are prone to express these receptors.[83] It has been reported that 18F-fluoro-l-dihydroxyphenylalanine tends to be a preferred imaging modality for differentiated head and neck PGLs because they are more likely to take up radiotracers for better tumor identification compared with tumors carrying SDHx mutations.[9] In addition to SDH, genes encoding the other components of oxidative respiratory chain are dysregulated, contributing to the overall mitochondrial dysfunction.[62] In summary, the pseudohypoxia pathway is an essential component of SDH-mediated tumorigenesis.[15]

IDH1, IDH2, and D-2HG

Discovery of PCC/PGL–specific mutations in several genes associated with the Krebs cycle reinvigorated research in the metabolic makeup of tumor cells. These studies provided additional validation for Warburg's insightful work. One of the metabolic mutations identified in the PCC/PGL tumors was a mutation in the cytosolic isocitrate dehydrogenase 1 (IDH1) gene.[21] This finding was important because a whole-genome sequencing analysis performed just a year earlier identified recurrent mutations in the same IDH1 gene in glioma and AML cells.[84,85] The striking observation made by the authors was that IDH1 mutations in all cancers were largely confined to arginine R132 residing in the active site of the IDH1 protein. The mutant allele was always a missense and a wild-type IDH1 allele was retained in the tumor. Overall, this genotype did not appear to be a loss of function, but rather a dominant mutation.

It was hypothesized that, in addition to losing the ability to catalyze the conversion between isocitrate and αKG, the R132 mutants are neomorphs that acquired a new function to convert αKG (also known as 2-oxoglutarate) to a chiral compound, d-2-hydroxyglutarate (D-2HG)[86] (Figure 1). This later hypothesis quickly gained widespread acceptance. D-2HG is a rare metabolite normally only present in minute quantities; in human malignant gliomas carrying IDH1 mutations, however, markedly elevated levels of D-2HG were found.[86] High levels of D-2HG produced by cancers containing IDH1 and IDH2 mutations could be used as a tractable biomarker—a finding that is immediately applicable clinically—for diagnoses, treatment, and follow-up of these tumors.[87–90] These noninvasive methods would be advantageous over solid-tumor biopsy samples or procuring blood samples to obtain material for DNA sequencing. Metabolic screening for D-2HG also makes it possible to separate the truly neomorphic IDH mutations from polymorphic single nucleotide polymorphisms and sequencing artifacts that do not affect IDH enzyme activity. In summary, unlike changes in SDH subunits encoding genes that include truncations, insertions, and deletion mutations, IDH mutations are exclusively missense, dominant, oncogenic gain of function.

From the very beginning, the ability of the mutant IDH to convert αKG into D-2HG led to the idea that enzymes relying on αKG for their activity could be affected. Several reports have provided evidence that 2HG can inhibit various αKG-dependent dioxygenase enzymes.[91] Specifically, it has been proposed that 2HG is an inhibitor of PHD2, which targets HIF1α for degradation and that the increase in HIF1α level in IDH-null cells is the cancer culprit.[92,93] However, there is no clear hypoxic signature in IDH-driven tumors, and a link between D-2HG, PHD2 inhibition, and HIF1α stabilization was generally not supported.[84,94–98] Only the l isoform of 2HG appears to inhibit PHD activity.[94,98] Still, hypoxia should be considered one of the important factors that can increase the presence of oncometabolites (including D-2HG) and contribute to carcinogenesis. Although PHD-related/HIF stabilization mechanisms were proposed for all three oncometabolites discussed here—2HG, succinate, and fumarate (discussed later in this article)—it remains unclear whether D-2HG consistently uses these pathways in tumors.[99] In contrast, in PCC/PGL tumors, mutations in SDHx completely overlap with the pseudohypoxic subtype.[15]

Given the close relationship among succinate, isocitrate, αKG, and 2HG, it is not surprising that there is commonality among the pathways affected by these molecules. The propensity of 2HG elevation to inhibit another family of αKG-dependent enzymes, the ten-eleven translocation (TET) family, was confirmed in several independent studies.[92,100,101] TET proteins produce 5′-hydroxymethylcytosine, an intermediate in DNA demethylation reaction (Figure 1). TET inhibition by 2HG is supported by the observation that the mutations in IDH1 or IDH2 were mutually exclusive with TET2 loss-of-function mutations in a large AML cohort.[100,101]

Understanding the mechanism for biochemical fallout of the IDH mutations provides several opportunities for novel clinical approaches to IDH-driven tumors. The most straightforward strategy aims to inhibit the neomorphic activity of mutant IDH and its production of 2HG oncometabolite. Several inhibitors of IDH1 and -2 are in clinical trials.[102–104]

FH and Fumarate

Another TCA cycle enzyme, FH, converts fumarate into malate. Homozygous FH mutations are considered an inborn error of metabolism and lead to fumaric aciduria, with patients presenting with dysmorphia, infantile encephalopathy, and brain malformations.[105] In contrast, loss of heterozygosity in patients carrying germline FH mutations causes hereditary leiomyomatosis and renal cell cancer.[106,107]FH is also downregulated in renal carcinomas,[108] Leydig cell tumors,[109] and, not dissimilar to IDH, deleted in neural tumors, neuroblastomas.[110] Most importantly, FH mutations have been isolated in PCC/PGLs,[16,111] underscoring the propensity of the mutations in TCA-cycle enzymes to produce these tumors.

Mutations in FH that result in clinical outcomes localize to conserved regions responsible for either the catalytic activity or the folding and stability of the enzyme, leading to abnormal accumulation of fumarate.[112–115] Research into FH deficiency illustrates an important point: Loss-of function mutation does not simply create a backlog of the lone precursor (fumarate), it also leads to a comprehensive reorganization of the cell metabolism as a whole. For example, FH mutant cells use significantly more glutamine and upregulate genes required for heme synthesis and bilirubin excretion. Despite the defective TCA cycle, FH-deficient cells maintain adequate mitochondrial NADH production and membrane potential by compensatory increase in glucose consumption and lactate synthesis.[113] Furthermore, the increase in unprocessed fumarate in FH-deficient cells forces the reversal of the urea cycle enzyme argininosuccinate lyase, driving the production of argininosuccinate from fumarate and arginine instead of aspartate and citrulline.[114,115] This makes FH-deficient cells auxotrophic for arginine, and depleting arginine from the medium reduces their proliferation and viability.[115]

Some oncogenes (e.g., RAS) produce tumors mostly when they are mutated; others (e.g., MYC) are rarely mutated and the mere overexpression (excess) of the wild-type protein is sufficient to drive oncogenic transformation. Similarly, although the definition of oncometabolite is most commonly applied to a normally absent 2HG, fumarate and succinate, the omnipresent components of the TCA cycle, cause a similar oncogenic effect when they accumulate in excess. Akin to 2HG and succinate, fumarate's role in cancer is thought to include an epigenetic mechanism. In renal cancer (where the consequences of FH deficiency are best understood), fumarate inhibits histone and DNA (e.g., TETs) demethylases and affects the epigenetic landscape.[116–118] It was also proposed that fumarate could inhibit other αKG-dependent dehydrogenases, specifically PHD, and induce pseudohypoxia acting through HIFs.[67,119]

Malate Dehydrogenase 2

Recently, Cascón et al.[120] described a patient who carried a mutation in yet another Krebs cycle enzyme, malate dehydrogenase 2 (MDH2). These tumors, deficient in the malate-oxidizing activity, accumulate fumarate. MDH2-mutated tumors have a global transcriptional profile similar to that of SDH-related tumors. All three metabolites—succinate, fumarate, and malate—can inhibit prolyl hydroxylation of HIFα.[121,122] A high fumarate-to-succinate ratio is common to the FH- and MDH2-mutated tumors, in contrast to SDHx-mutated tumors, and suggests that an alternative or additional mechanism can trigger oncogenesis in these patients.