Hypoxia-Inducible Factor Signaling in Pheochromocytoma: Turning the Rudder in the Right Direction

Abstract

Many solid tumors, including pheochromocytoma (PHEO) and paraganglioma (PGL), are characterized by a (pseudo)hypoxic signature. (Pseudo)hypoxia has been shown to promote both tumor progression and resistance to therapy. The major mediators of the transcriptional hypoxic response are hypoxia-inducible factors (HIFs). High levels of HIFs lead to transcription of hypoxia-responsive genes, which are involved in tumorigenesis. PHEOs and PGLs are catecholamine-producing tumors arising from sympathetic- or parasympathetic-derived chromaffin tissue. In recent years, substantial progress has been made in understanding the metabolic disturbances present in PHEO and PGL, especially because of the identification of some disease-susceptibility genes. To date, fifteen PHEO and PGL susceptibility genes have been identified. Based on the main transcription signatures of the mutated genes, PHEOs and PGLs have been divided into two clusters, pseudohypoxic cluster 1 and cluster 2, rich in kinase receptor signaling and protein translation pathways. Although these two clusters seem to show distinct signaling pathways, recent data suggest that both clusters are interconnected by HIF signaling as the important driver in their tumorigenesis, and mutations in most PHEO and PGL susceptibility genes seem to affect HIF-α regulation and its downstream signaling pathways. HIF signaling appears to play an important role in the development and growth of PHEOs and PGLs, which could suggest new therapeutic approaches for the treatment of these tumors.

Oxygen sensing is paramount for cell survival. The activation or inhibition of cellular oxygen-sensitive signaling pathways is involved in biological processes, such as cell development, proliferation, and growth; transformation, including tumorigenesis; and processes tightly linked to hypoxia adaptation, including erythropoietin (EPO) production and glycolysis (1–5).

Under hypoxia or pseudohypoxia (a situation where oxygen is present but cannot be processed because of an alteration in oxygen-sensing pathways), cells activate a number of adaptive responses, coordinated by various cellular pathways, most of them controlled by a common denominator, hypoxia-inducible factor (HIF) (2,6–8). Recurrent or long-lasting hypoxia by the HIF signaling pathway is currently a well-known stimulus as well as consequence of cancer development, invasion, and metastasis (9).

HIF, which acts as a heterodimer of basic helix-loop-helix PAS (bHLH-PAS) proteins, is comprised of a ubiquitously expressed α subunit (oxygen regulated) and a constitutively expressed β subunit (10). The HIF-α subunit consists of three isoforms, HIF-1α, HIF-2α, and HIF-3α (6,7). HIF-1α is expressed in all cells; HIF-2α is expressed preferentially in the endothelium, kidney, heart, lung, gastrointestinal epithelium, and neural crest cell derivatives (8,11); and HIF-3α is expressed in the thymus, Purkinje cells in the cerebellum, and the corneal epithelium of the eye (12). HIF-1α and HIF-2α are the best-studied HIF-α isoforms, and their domain architecture is highly conserved and similar. Both can serve as transcriptional activators (1). HIF-3α has several splice variants, but only some of these variants are able to interact with HIF-1α, HIF-2α, or HIF-β (13–15). The splice variant HIF-3α2, also called inhibitory PAS domain protein (IPAS), works as an HIF-1α inhibitor that is probably involved in feedback regulation because the interaction of HIF-3α2 forms an inactive complex with HIF-1α, thus acting as a dominant negative regulator of HIF-1 (12,13,16).

Under normoxic conditions, HIF-1α and HIF-2α are rapidly degraded by the ubiquitin-proteasome pathway. HIF-α degradation is controlled mostly by the hydroxylation of two specific prolyl residues by prolyl hydroxylase domain proteins (PHDs). There are three known PHD isoforms: PHD1, PHD2, and PHD3 (6,7). PHDs use oxygen and α-ketoglutarate to convert a prolyl residue to hydroxy-prolyl, producing succinate and carbon dioxide (17). Once hydroxylated, HIF-1α or HIF-2α is recognized by the von Hippel-Lindau tumor suppressor protein (pVHL) (18), which is a component of an E3 ubiquitin ligase complex known as VHL/elongin B/elongin C (VBC), and is rapidly degraded by the proteasome (7) (Figure 1). Ubiquination/degradation of HIF-1α is promoted by binding of spermidine/spermine N-acetyltransferase-2 (SSAT2), which stabilizes the interaction of pVHL and elongin C (19,20).

Figure 1.

Different hypoxia-inducible factor regulation under normoxic and hypoxic conditions. CBP = cAMP-response element-binding protein; Cul2 = cullin 2; E2 = E2 ubiquitin-conjugating enzyme; HIF = hypoxia-inducible factor; HRE = hypoxia-responsive elements; HSP90 = heat shock protein 90; p300 = histone acetyltransferase p300; PHD = prolyl hydroxylase domain protein; pVHL = von Hippel-Lindau protein; Rbx1 = ring box protein 1; SDH = succinate dehydrogenase; SSAT2 = spermidine/spermine N1-acetyltransferase 2; UQ = ubiquitin.

The other α-ketoglutarate-dependent dioxygenase is factor-inhibiting HIF-1 (FIH-1), which hydroxylates HIF-1α on the asparagine 803 residue. This blocks its interaction with the coactivators histone acetyltransferase p300 (p300) and cAMP-response element-binding protein (CBP) under normoxic conditions (21) and thus inhibits the transactivation of HIF target genes.

Under hypoxic (or pseudohypoxic) conditions, HIF-α becomes stabilized, heterodimerizes with HIF-β, recruits coactivators, and binds to the core DNA sequence RCTCG at hypoxia-responsive elements (HREs) in target genes, where it activates their transcription (7,22,23). HIF-1α and HIF-2α activate transcription of target genes that overlap, but some of them are distinct for HIF-1α and for HIF-2α (Table 1) (8,22,24,25). In addition, HIF-2α is relatively resistant to FIH-1–mediated inactivation, in contrast with HIF-1α (26). HIF-α expression or stabilization can also be regulated by many mechanisms other than hypoxia, all allowing unique crosstalk between different signaling pathways, resulting in cell transformation and cancer development (Figure 1) (7,27,28).

Table 1.

Major target genes regulated by HIF-1α and HIF-2α*

Target gene	Gene product (protein)	Function	HIF-1α	HIF-2α
SLC2A1	GLUT1	Glucose transport	+	+
PLIN2	ADRP	Lipid metabolism	+	+
CA12	CAXII	pH homeostasis	+	+
FLG	Filaggrin	Cytoskeletal structure	+	+
IL6	IL-6	Immune cytokine	+	+
ADM	Adrenomedullin	Angiogenesis	+	+
VEGFA	VEGFA	Angiogenesis	+	+
BNIP3	BNIP3	Autophagy and apoptosis	+	−
HK1	hexokinase 1	Glycolysis	+	−
HK2	hexokinase 2	Glycolysis	+	−
PFK	Phosphofructokinase	Glycolysis	+	−
ALDOA	ALDA	Glycolysis	+	−
PGK1	PGK1	Glycolysis	+	−
LDHA	LDHA	Glycolysis	+	−
NOS2	iNOS	Nitric oxide production	+	−
ABL2	ARG	Inhibitor of nitric oxide production	−	+
EPO	erythropoietin	Erythropoiesis	−	+
POU5F1	OCT4	Stem cell identity	−	+
SCGB3A1	secretoglobin 3A1	Cytokine inhibiting growth?	−	+
TGFA	TGFα	Growth factor	−	+
CCND1	cyclin D1	Cell cycle progression	−	+
DLL4	DLL4	NOTCH signaling and endothelial cells branching	−	+
ANGPT2	angiopoietin 2	Blood vessel remodeling	−	+
EDN1	endothelin 1	Maintaining vascular tone, comitogenic activity	+	+
EGLN3	PHD3	Cellular oxygen sensing, HIFα hydroxylation	+	+
EGLN1	PHD2	Celluar oxygen sensing, hydroxylation of HIF	+	NR
TF	transferrin	Iron transporter, stimulating cell proliferation	+	+
TGFB	TGFβ3	Proliferation, differentiation	+	+
FLT1	VEGFR1	Angiogenesis	+	+

* ABL2 = Abelson tyrosine-protein kinase 2; ADM = adrenomedullin; ADRP = adipose differentiation-related protein; ALDA = fructose-bisphosphatealdolase A; ALDOA = fructose-bisphosphatealdolase A; ANGPT = angiopoietin 2; ARG = Abelson-related gene protein; BNIP3 = BCL2/adenovirus E1B 19kDa protein-interacting protein 3; CA12 = carbonic anhydrase 12; CAXII = carbonic anhydrase 12 protein; CCND1 = G1/S-specific cyclin-D1; DLL4 = delta-like 4 protein; EDN = endothelin 1; EGLN1 = egl nine homolog 1 gene; EGLN3 = egl nine homolog 3 gene; EPO = erythropoietin gene; FLG = fillagrin gene; FLT1 = Fms-like tyrosine kinase 1 precursor; GLUT1 = glucose transporter 1; HK1 = hexokinase 1; HK2 = hexokinase 2; IL6 = interleukin 6 gene; IL-6 = interleukin 6; iNOS = inducible nitric oxide synthase; LDHA = lactate dehydrogenase A; NOS2 = nitric oxide synthase; NOTCH = translocation-associated notch protein; NR = not reported; OCT4 = octamer-binding protein 4; PFK = phosphofructokinase gene; PGK1 = phosphoglycerate kinase 1 gene; PHD2 = prolyl hydroxylase domain 2; PHD3 = prolyl hydroxylase domain 3; PLIN2 = perilipin 2; POU5F1 = POU class 5 homeobox 1; SCGB3A1 = secretoglobin family 3A member 1 precursor; SLC2A1 = solute carrier family 2 (facilitated glucose transporter), member 1; TF = transferrin gene; TGFA = transforming growth factor alpha gene; TGFα = transforming growth factor alpha; TGFB = transforming growth factor beta gene; TGFβ = transforming growth factor beta; VEGFA = vascular endothelial growth factor A; VEGFR1 = vascular endothelial growth factor receptor 1 [for a review see (2,8,24,25)].

HIF-1α and HIF-2α have been shown to play an important role in neural crest development and differentiation. HIF-1α signaling is critical for the development of the cardiovascular system and neural tube, whereas high HIF-2α expression was observed in developing paraganglia, and HIF-2α has been shown to be necessary for catecholamine production (29–31). This implies that HIF-2α plays an important role in the development or function of the adrenal medulla and paraganglia, origins of pheochromocytoma (PHEO) or paraganglioma (PGL), respectively.

HIF and Cancer

Because there is an interaction of HIF with many pivotal signaling pathways and (pseudo)hypoxia is found in almost all cancers, the HIF signaling pathway has been recently considered as the essential checkpoint (gatekeeper) in tumorigenesis (Figure 2) (7,23, 28,32–34). Tumorigenesis is a multistep process that manifests with genetic or epigenetic changes in malignant cells, resulting in six previously described alterations in cell physiology: 1) self-sufficiency of growth signals, 2) evasion of apoptosis (programmed cell death), 3) insensitivity to antigrowth signals, 4) unlimited replicative potential, 5) sustained angiogenesis, and 6) invasion and metastasis (35). Two other “emerging” hallmarks functionally important for cancer are deregulation of cellular energy metabolism and immunoevasion (36). HIF-1α and HIF-2α were found to be overexpressed in most human cancers, also including those with more aggressive behavior displaying all the hallmarks of cancer (37–40). However, the precise link between tumor growth, aggressiveness, metastasis, and HIF overexpression has not yet been confirmed and may even be controversial (7,41,42).

Figure 2.

Genes targeted by hypoxia-inducible factor (HIF) upregulation. Pseudohypoxia- or hypoxia-related HIF upregulation targets many genes that are involved in cancer development, progression, and metastasis. These include genes stimulating erythropoiesis, angiogenesis, glucose metabolism, extracellular matrix formation, epithelial to mesenchymal transition, metastasis, chemotaxis, or cell proliferation/survival, which eventually protect hypoxic tissues (43,44–48). ADCY = adenylatecyclase; ALDOA = aldolase A; ALDOC = aldolase C; ANGPT2 = angiopoietin-2; ANGPTL4 =angiopoietin-related protein 4; BIRC5 = baculoviral inhibitor of apoptosis repeat-containing 5; CA9 = carbonic anhydrase 9; CA12 = carbonic anhydrase 12; C-MET = met protooncogene tyrosine kinase; CP = ceruloplasmin; CXCR4 = C-X-C chemokine receptor type 4 (fusin); DEC1 = differentiated embryo chondrocyte 1; EID2 = EP300 interacting inhibitor of differentiation 2; EPO = erythropoietin; ET1 = endothelin 1; FN1 = fibronectin 1; GAPDHS = glyceraldehyde-3-phosphate dehydrogenase, spermatogenic; GLUT1 = glucose transporter; GLUT3 = glucose transporter 3; GPI = glucose-6-phosphatase isomerase; HK1 = hexokinase 1; HK2 = hexokinase 2; HO-1 = hemeoxygenase 1; hTERT = human telomerase reverse transcriptase; IGF2 = insulin-like growth factor 2; IGFBP = insulin-like growth factor binding protein; KITLG = KIT-ligand; LDHA = lactate dehydrogenase A; LOX = lysyl oxidase; MMP2 = matrixmetallopeptidase 2; MMP14 = matrixmetallopeptidase 14; MSH2 = mitochondrial MutS protein MSH1; MSH6 = DNA mismatch repair protein MSH6; NBS1 = Nijmegen breakage syndrome protein 1; Nip3 = novel immunogenic protein 3; NOS2 = nitric oxide synthase 2; NOTCH1 = translocation-associated notch protein TAN-1; Oct-4 = octamer-binding transcription factor 4; P21 = cyclin-dependent kinase inhibitor 1A; P35srj = CBP/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2; PDGF = platelet-derived growth factor; PDGFB = platelet-derived growth factor beta polypeptide; PDK1 = pyruvate dehydrogenase kinase 1; PFKL = phosphofructokinase; PGF = placental growth factor; PGK1 = phosphoglycerate kinase 1; PIGF = phosphatidylinositol glycan anchor biosynthesis, class F; PKM2 = pyruvate kinase 2; PLAU = plasminogen activator, urokinase; SDF1 = stromal cell-derived factor 1; SDF4 = stromal cell-derived factor 4; SIP = SPB interaction protein; SNAIL = snail 1 zinc finger protein; TF = transferrin; TFR = transferrin receptor; TGFα = transforming growth factor alpha; TGFB3 = transforming growth factor beta 3; TWIST1 = class A basic helix-loop-helix protein 38; VEGF = vascular endothelial growth factor; VEGFR1 = vascular endothelial growth factor receptor 1; WSB1 = WD repeat and SOCS box containing 1; ZEB1 = zinc finger E-Box binding homeobox 1; ZEB2 = zinc finger E-Box binding homeobox 2.

Interestingly, the roles and functions of HIF-1α and HIF-2α in tumor growth vary among cell types (41,42,49,50). The basis for these discrepancies is not well understood but could reflect the consequences of preferential HIF-1α or HIF-2α activity/function in different cancer cell subtypes, different stages of tumor progression, and the tumor microenvironment (8,28,41,42,51–53). For example, HIF-1α antagonizes the c-Myc proto-oncogene and seems to be preferentially affected by mammalian target of rapamycin complex 1 (mTORC1) activity (51,54,55). By contrast, HIF-2α cooperates with c-Myc and is less responsive to mTORC1 modulation, so it has a more limited expression pattern than HIF-1α (55), and it has been thought to be more directly associated with later stages of tumor progression than HIF-1α. HIF-1α and HIF-2α can promote distinct, even opposing, effects in the same cell type, further supporting the existence of specific HIF-1α and HIF-2α target genes. Thus, although both HIF-1α and HIF-2α are overexpressed in many tumors, in advanced lesions it is mostly HIF-2α that is upregulated. HIF-2α also has been shown to regulate HIF-1α target gene expression in the absence of HIF-1α and vice versa [for a review, see (8,56–58)]. Furthermore, HIF-2α becomes stabilized at higher oxygen tensions than HIF-1α (11,59,60) and can escape degradation at near-normoxic conditions (61).

Multiple other oxygen-independent oncogenic pathways are involved in HIF-α regulation, including growth factor signaling pathways and a loss of tumor suppressor genes (Figure 3) (62). Some alterations in these pathways or genes induce HIF-α overexpression and increase transcription and/or translation of HIF-α target genes (27). For example, for tumor suppressor genes such as phosphatase and tensin homologue (PTEN) (63,64), protein p53 (p53) (65,66), or VHL (67), inactivation leads to an accumulation of HIF-α and activation of HIF-α target genes. Some oncogenes, such as H-RAS (Harvey rat sarcoma viral oncogene) (68,69), v-Src (viral sarcoma) (70), or c-Myc (71), have also been found to increase activation of the HIF pathway. In oxygen-independent pathways, HIF-α is activated by the phosphatidyl-inositol-3-kinase (PI3K) or mitogen-activated protein kinase (MAPK) pathways, and PI3K mediates its effect through its target Akt and the downstream mTORC kinase (43). Furthermore, hypoxia-associated factor (HAF, also known as SART1₈₀₀), an E3 ubiquitin ligase, binds to and ubiquitinates HIF-1α by an oxygen- and pVHL-independent mechanism and targets HIF-1α for proteasomal degradation (72). Thus, HAF overexpression lowers levels of HIF-1α and decreases its transactivating activity. By contrast, HAF also binds to HIF-2α and increases its transcriptional activity. This implies that HAF overexpression switches the hypoxia response of the cancer cell from HIF-1α to the HIF-2α–dependent transcription of genes, promoting tumor progression, invasion, and proliferation (73). Mutations in genes involved in the mitochondrial respiratory chain are often additional pseudohypoxic mechanisms involved in tumorigenesis and cancer progression, including in PHEO and PGL (see below). Mitochondrial complex 2 (succinate dehydrogenase complex [SDH]) subunit mutations result in HIF-α stabilization and are known to be involved in the development of several hereditary cancer syndromes (74–80).

Figure 3.

Regulation of hypoxia-inducible factor (HIF)–α expression or stabilization by pathways other than hypoxia. Heat shock protein 90 (HSP90) interacts directly with HIF-1α and promotes its stabilization (a) (81–83). HSP90 inhibitors and histone acetylase inhibitors promote degradation of HIF-α in a von Hippel-Lindau protein (pVHL)–independent manner (83–86). Another pathway is sumoylation of HIF-α (b), but there is disagreement as to whether this leads to increased or decreased HIF-α stability (87–89). Another method of HIF regulation is reciprocal control of prolyl hydroxylase domain proteins (PHDs) and HIF-1α by the nuclear factor kappa B (NF-κB) signaling pathway linked to the activation of phosphoinositide 3-kinase (PI3K) and RAC-alpha serine/threonine-protein kinase (Akt) signaling (c) (90,91–93). Hypoxia-associated factor (HAF, known also as SART₁₈₀₀) is overexpressed in a variety of tumor types, and it binds to and ubiquitinates HIF-1α by an oxygen- and pVHL-independent mechanism and targets HIF-1α for proteasomal degradation (d) (72). The tumor suppressor protooncogene p53 (p53) seems to provide a route for HIF-1α degradation in hypoxic cells (94), and HIF-1α seems to promote p53 transcription [for a review, see (95)]. An important negative regulator of p53 is the murine double minute 2 (MDM2) ubiquitin protein ligase (96,97). It was suggested that it promotes HIF-1α degradation by binding to the HIF-1α/p53 complex (66), but the formation of trimeric p53/MDM2/HIF-1α complexes is controversial (e) [for a review, see (95)]. Conversely, the direct interaction of HIF-1α has been shown to increase HIF-1 activity and abundance in hypoxic cells (98,99). c-Myc = Myc proto-oncogene; CBP = cAMP-response element-binding protein; Cul2 = cullin 2; E2 = E2 ubiquitin-conjugating enzyme; FIH = factor inhibiting HIF; HRE = hypoxia-responsive elements; EGF = epidermal growth factor; EGF-R = epidermal growth factor receptor; FGF-R = fibroblast growth factor receptor; HSP90 = heat shock protein 90; IGF = insulin-like growth factor; IGF1-R = insulin-like growth factor 1 receptor; NGF = nerve growth factor; p300 = histone acetyltransferase p300; RET = Ret proto-oncogene; Rbx1 = ring box protein 1; SSAT2 = spermidine/spermine N1-acetyltransferase 2; SUMO1 = small ubiquitin-related modifier 1; TGFα = transforming growth factor alpha; UQ = ubiquitin; VEGF-R = vascular endothelial growth factor receptor.

HIF-α activity can also be regulated by sirtuins (Sirt), a family of redox-sensitive, NAD⁺-dependent deacetylases and/or ADP- ribosyltransferases. For example, Sirt1 forms a complex with HIF-2α and deacetylates conserved lysine residues, and this enhances HIF-2α transcriptional activity. With HIF-1α, Sirt1 has either a neutral, activating, or repressing role on HIF-1α transcription (100–103). Sirt1 has been also shown to increase HIF-2α regulation of EPO (102). Sirt6 increases the synthesis of HIF-1α as well as its stability (8,104). The mitochondrial deacetylase Sirt3 regulates HIF-α stabilization indirectly by repressing the formation of reactive oxygen species (ROS), which is followed by HIF-1α destabilization (105,107).

ROS signals are sufficient to trigger HIF-α stabilization, and antioxidant therapy was shown to substantially inhibit tumor progression in vivo by inhibiting HIF activation (108). HIF-α/HIF-β dimers potently enhance the glycolytic metabolism in cancer cells, with targets from glucose transporters through lactate dehydrogenase A (LDH-A) (109). HIF specifically blocks access of the glycolytic end products to the mitochondria, mediated by the HIF target pyruvate dehydrogenase kinase 1 (PDK1). PDK1 inhibits conversion of pyruvate to acetyl coenzyme A (acetyl-CoA) by phosphorylating pyruvate dehydrogenase (110,111). HIF also mediates a shift in the components of cytochrome c oxidase (COX). It substitutes COX4-2 for COX4-1 (112). This change results in enhanced electron transport chain efficiency under hypoxic conditions, with increased ATP production and decreased ROS production. Thus, ROS generation is required for HIF stabilization, but HIF stabilization prevents excessive ROS production in hypoxic cells and their damage. HIF dimer targets, such as superoxide dismutase 2 (SOD2), also protect cellular and mitochondrial components in the presence of oxidative stress (113). This indicates that limiting ROS is an important HIF metabolic adaptation to low oxygen levels. HIF-α can also be inhibited by nitric oxide, which is an endogenous inhibitor of respiration. Nitric oxide binds to COX and inhibits oxygen consumption in the mitochondria, increasing the availability of oxygen for PHDs and allowing their reactivation under hypoxia (114). Chronic HIF activation also decreases the overall mitochondrial mass by indirect modulation of c-Myc transcriptional activity (115,116).

HIF and PHEO/PGL

PHEOs and PGLs are neuroendocrine tumors derived from neural crest cells. PHEOs are tumors of the adrenal gland; PGLs arise from extra-adrenal sympathetic- or parasympathetic-derived chromaffin tissue. PHEOs and PGLs can occur as sporadic tumors or within the spectrum of hereditary tumor syndromes, including most commonly multiple endocrine neoplasia 2 (MEN2), caused by mutations of the RET protooncogene (RET); von Hippel-Lindau (VHL) disease, which results from mutations of the VHL tumor suppressor gene; and neurofibromatosis type 1 (NF1), caused by mutations of the NF1 tumor suppressor gene (117,118). The other known PHEO and PGL susceptibility genes include the SDH subunits—SDHA, SDHB, SDHC, and SDHD (74–77)—and succinate dehydrogenase complex assembly factor 2 (SDHAF2) (119). Recent studies have identified new genetic candidates that predispose patients to the development of PHEO and PGL: transmembrane protein 127 (TMEM127) (120,121), Myc-associated factor X (MAX) (122), the newly described H-RAS (123) and, by anecdotal reports, kinesin family member 1B, transcript variant beta (KIF1Bβ) (124), isocitrate dehydrogenase (IDH) (125) and PHD2 (126). Currently, germline mutations can be identified in up to 35% of PHEOs and PGLs (127). Recently, new somatic and germline mutations in HIF2A have been found to be the cause of a syndrome consisting of multiple and recurrent PHEOs and PGLs and polycythemia, also associated with multiple duodenal somatostatinomas in some patients (128–131). The HIF pathway has been found to be dysfunctional in some PHEOs and PGLs, depending on their genetic background, resulting in or contributing to their development, as described in detail below (7,74,132–136). Furthermore, recently somatic mutations in HIF2A have been found in sporadic PHEOs and PGLs, supporting the relevance of studying hypoxia-related mechanisms in the pathogenesis of apparently sporadic PHEOs and PGLs (137,138).

(Pseudo)hypoxia was anticipated to be an underlying mechanism of PHEO and PGL development for a long time, but a direct link between the hypoxic pathway and hereditary PHEO and PGL development was not discovered until 1999 by the elucidation of HIF regulation and the association of VHL inactivation with the hypoxic pathway [for a review, see (23,132)]. VHL and SDH inactivation have been linked to HIF-1α or HIF-2α dysregulation, respectively (139,140). RET- and NF1-associated PHEOs and PGLs exhibit activation of the Ras (rat sarcoma)–mediated MAPK pathway (141–143), which leads to the stimulation of mTORC activity and mTORC-mediated cap-dependent translation of HIF-1α (144,145).

However, some VHL mutations associated with a high risk of PHEO and PGL were found not to affect pVHL-related regulation of HIF (146,147), resulting in the description of a HIF-independent pathway linked to inherited causes of VHL PHEO and PGL, but also including SDH mutations (148). The authors suggested that germline VHL or SDHB mutations disrupt the normal developmental apoptosis pathway (c-Jun–dependent apoptosis) of sympathetic neuronal precursor cells, which leads to the abnormal survival of neuronal cells that give rise to PHEO and PGL. This suggests that mitochondrial dysfunction and the activation of the pseudohypoxic pathway are most likely interconnected with other mechanisms of tumorigenesis, including apoptosis resistance (149).

Currently, based on the main signaling pathway signatures resulting from specific PHEO and PGL gene mutations, PHEOs and PGLs are divided into two clusters: 1) cluster 1, a pseudohypoxic cluster (SDH [cluster 1A] and VHL [cluster 1B] mutations), and 2) cluster 2, a cluster rich in kinase receptor signaling and protein translation pathways (RET, NF1, TMEM127, MAX, and KIF1Bβ mutations) (127,142). Although these two clusters seem to show distinct cell signaling, recent data suggest that both clusters are indeed interconnected, turning the rudder to HIF signaling as the major driver in their tumorigenesis.

HIF in Cluster 1 PHEOs and PGLs

Cluster 1 tumors are mostly extra-adrenal, except for VHL tumors, which are usually adrenal (150–158). VHL PHEOs and PGLs show a low risk of malignancy; SDH-related tumors, especially those associated with SDHB, are more aggressive and often metastatic (78,154,156,157,159–161). Cluster 1 PHEOs and PGLs commonly present with a dopaminergic and/or noradrenergic biochemical phenotype (162–164).

VHL disease is caused by mutations in the tumor suppressor gene VHL (165). pVHL, the VHL gene product, is the substrate recognition unit of the VBC E3 ubiquitin ligase complex that targets HIF for proteasomal degradation in the presence of oxygen (67) (Figure 1). HIF-α is hydroxylated by PHD for its recognition by pVHL; therefore this interaction is exclusively oxygen dependent (166–169). Mutations in VHL cause abnormalities in oxygen sensing and therefore affect HIF-associated transcription (1). For example, type 1 VHL mutations, commonly nonsense or deletion mutations, result in catastrophic dysfunction of this tumor suppressor, which leads to full activation of HIF; PHEOs are more rare in these patients. Type 2 VHL, more commonly found as missense mutations, results in a higher risk of PHEOs. They establish mild changes in pVHL structure, allowing normal assembly of the VBC complex. However, the activity of the degradation complex is compromised, depending on mutations that eventually lead to HIF-α stabilization with different ranges (170–173). Residual HIF-α escapes recognition by pVHL and activates the transcription of many downstream genes (3). Thus, under normal conditions, pVHL has been found to be a negative regulator of HIF-inducible genes, including glucose transporter 1 (GLUT1), vascular endothelial growth factor (VEGF), EPO, and many others (2,132,174,175). Under conditions associated with defective (mutated) pVHL, HIF activity is increased, resulting in the activation of HREs in downstream genes (Figures 1 and 2), tightly linked to tumorigenesis (2,139,142,149,175).

SDH-related tumorigenesis is also believed to be associated with HIF stabilization and activation and stability of its pathway (132,142,176). SDH mutations disrupt the activity and stability of the SDH enzyme, leading to its accelerated proteasomal degradation (177–179). SDH is a mitochondrial enzyme (mitochondrial complex 2) and is a part of the mitochondrial respiratory chain. SDH consists of four functionally different subunits: A, B, C, and D [for a review, see (180–182)]. Its function is to couple the oxidation of succinate to fumarate in the mitochondrial matrix (as part of the Krebs cycle) with the reduction of ubiquitin in the membrane (as part of the electron transport chain) (74,135,136). Mutations in SDH subunits lead to disrupted oxidation of succinate to fumarate and therefore succinate accumulation and to the inhibition of PHD-catalyzed HIF-α hydroxylation, which is required for HIF-α proteasomal degradation (Figure 4) (135,139,177,183).

Figure 4.

Hypoxia inducible factor (HIF) signaling in cluster 1 and cluster 2 pheochromocytoma (PHEO) and paraganglioma (PGL). Akt = RAC-alpha serine/threonine-protein kinase; CBP = cAMP-response element-binding protein; c-Myc = Myc proto oncogene; Cul2 = cullin 2; E2 = E2 ubiquitin-conjugating enzyme; EGF = epidermal growth factor; EGF-R = epidermal growth factor receptor; eIF-4E = eukaryotic translation initiation factor 4E; ERK = extracellular signal-regulated kinase; FGF-R = fibroblast growth factor receptor; H-RAS = Harvey rat sarcoma viral oncogene; HRE = hypoxia-responsive elements; HSP90 = heat shock protein 90; IGF = insulin-like growth factor; IGF1-R = insulin-like growth factor-1 receptor; MAX = myc-associated factor X; mTORC1 = mammalian target of rapamycin complex 1; mTORC2 = mammalian target of rapamycin complex 2; NF1 = neurofibromin 1; NF-κB = nuclear factor kappa B; NGF = nerve growth factor; p300 = histone acetyltransferase p300; PHD = prolyl hydroxylase domain protein; PI3K =phosphoinositide 3-kinase; pVHL = von Hippel-Lindau protein; Raptor = regulatory associated protein of mTOR; Ras = rat sarcoma oncogene; RET = Ret proto-oncogene; Rbx1 = ring box protein 1; Rheb = Ras homolog enriched in brain; S6K = S6 kinase; SDH = succinate dehydrogenase; SSAT2 =spermidine/spermine N1-acetyltransferase 2; TGFα = transforming growth factor alpha; TMEM127 = transmembrane protein 127; TSC1/2 = tuberous sclerosis complex 1/2; UQ = ubiquitin; VEGF-R = vascular endothelial growth factor receptor.

Hydroxylation of HIF-α by PHDs generates a binding site for pVHL and thus promotes HIF-α degradation. PHD2 is one of the PHD isoforms that is crucial for oxygen-dependent HIF-α stabilization and degradation (6,7). A germline loss-of-function PHD2 mutation was described in one family with polycythemia and PGL (184). Because PHD2 hydroxylates HIF-α, its loss-of-function mutation leads to HIF-α stabilization and its activation (Figure 1).

There is no consensus whether HIF-1α or HIF-2α promotes tumorigenesis in cluster 1 PHEOs and PGLs. Gimenez-Roqueplo et al. (136,140) described overexpression of HIF-2α and VEGF in patients with PHEOs and PGLs carrying SDHB and SDHD mutations compared with sporadic PHEOs and PGLs. However, in another study, Pollard et al. (134,139) found relatively more common HIF-2α overexpression in VHL PHEOs and PGLs, whereas in SDH tumors nuclear HIF-1α staining was more prominent. In the experiments of Favier et al. (185), HIF-2α mRNA was found to be overexpressed in both VHL and SDH PHEOs and PGLs. The findings by Koh et al. (73) support the leading role of HIF-2α in tumor development and progression in cluster 1 tumors.

Recently novel somatic and germline gain-of-function mutations have been identified in the HIF2A gene in patients with multiple or recurrent PHEOs and PGLs and polycythemia, as well as in some patients with multiple duodenal somatostatinomas (128–131,137,138). Mutations were shown to substantially disrupt HIF-2α prolyl hydroxylation, as well as the binding of mutated HIF-2α to pVHL, with upregulation of various HIF-2α downstream genes (Figure 1).

HIF in Cluster 2 PHEOs and PGLs

PHEOs and PGLs in this cluster are most often adrenal with a very low risk of malignancy, except MAX-related tumors (156,186–193). They are most commonly characterized by an adrenergic biochemical phenotype (163,164,194–196).

In MEN2 and NF1, HIF upregulation is driven through activation of the Ras/MAPK pathway (Figure 4). Ras activation results in the activation of downstream pathways: MAPK, PI3K, and mTORC. These pathways are involved in proliferation, differentiation, survival, migration, and angiogenesis (197–199), and they increase HIF-α signaling in many cancers (144,145,200–206). MAPKs also promote nuclear accumulation of HIF-1α (207). In NF1, the inactivating mutation in the tumor suppressor gene NF1, which encodes a large GTPase activating protein (GAP) neurofibromin (208), causes the activation of Ras; in MEN2, activation of the Ras/MAPK and PI3K/AKT pathways is caused by gain-of-function mutations in RET (Figure 4) (209–215). Extracellular signal-regulated kinase (ERK), a component of the Ras/MAPK pathway, directly phosphorylates HIF-1α and the transcription factor specificity protein 1 (Sp1) and, therefore, induces transcription of VEGF, a key regulator of angiogenesis (216,217). Moreover, another downstream target of PI3K is nuclear factor kappa B (NF-κB), and crosstalk between the HIF-α pathways and NF-κB signaling has been proposed. Overexpression of NF-κB subunits induces HIF-1α mRNA expression and transcription (Figure 4) (2,90,218,219).

In tumors caused by TMEM127 and MAX mutations, HIF-α levels seem to be increased due to mTORC activation (220,221). TMEM127 is a negative regulator of the mTORC signaling pathway, and mutations in the TMEM127 gene lead to its inactivation and to enhancement of mTORC signaling (120) in a Ras- and PI3K/AKT-independent manner (Figure 4) (121,222). The TMEM127 protein is dynamically associated with various endosomal pools (120). This suggests it participates in protein trafficking and/or the protein recycling endocytic machinery (127), and endomembranes of the protein secretory pathway appear to be important in mTORC signaling (223,224).

MAX acts as a tumor suppressor, and mutations in MAX deregulate c-Myc signaling (Figure 4) (122,127). c-Myc controls genes that encode translation intitiation factors, including the eukaryotic initiation factor 4E (eIF4E) (225–229), which is also a downstream effector of the PI3K/mTORC pathway (230,231,232). HIF-1α is also one of the transcription targets of c-Myc (by eIF4E) (203). Thus, MAX mutations cause a Myc neoplastic switch (233). Moreover, crosstalk between the Myc/MAX/MXD1 (MAX-interacting protein 1) and Ras/PIK3CA/AKT1/mTORC pathways has been proposed (234,235), and this suggests that MAX affects the mTORC pathway as well.

Recently, somatic H-RAS mutations were found in PHEOs and PGLs in male patients, causing the activation of the Ras/RAF/ERK signaling pathway, a part of the Ras/MAPK pathway (236). Ras/MAPK pathway acitivation leads to an increase in HIF-α signaling and to the transcription of HIF target genes.

KIF1Bβ is a tumor suppressor gene that acts downstream from PHD3 and is necessary for neuronal apoptosis (124). Because KIF1Bβ was described as a downstream target of PHD3, there is some possibility that it is also connected with the HIF signaling pathway, but currently there is no data to support this.

Future Perspectives

Over the past few years, increasing and solid evidence has emerged that HIF plays an important role in cancer as well as in PHEO and PGL pathogenesis (8,132,142,182,237–239). HIF is a gatekeeper of many, often overlapping, signaling cascades, joining oncogenic or abnormal cellular signaling in response to (pseudo)hypoxia. This designates HIF-α as a promising therapeutic target in cancer. Although the expression of HIF-α subunits differs in distinct cancers, it seems that in most cancers, including PHEO and PGL, HIF is upregulated, resulting in the downstream activation of hypoxia-related genes that promote rather than inhibit tumorigenesis (Figure 2). However, genetic intervention of HIF has not always been associated with reduced tumor growth. For example, homozygous deletions of the HIF-1α locus on 14q in clear cell renal carcinoma cell lines promote tumor growth, supporting the concept that HIF-1α is a kidney cancer suppressor gene (42). It is also very likely that HIF-associated tumor formation requires other important signaling pathways, such as fumarate hydratase, a Krebs cycle enzyme (240,241). Additionally, a very delicate balance between HIF-1α and HIF-2α may be the key to initial tumor development, progression, invasiveness, and finally metastasis (242,243). Recently discovered somatic mutations in HIF2A in PHEO and PGL with the activation of various HIF-related downstream genes may further support this concept (128,130,131,137,138). Finally, HIF activation may not only directly affect tumor cell growth but also affect the microenvironment adjacent to the tumor stroma (41). However, to further elucidate the exact biology of HIF in PHEO and PGL tumorigenesis, research performed on a suitable animal model is crucial and preferable. However, there is currently no human PHEO and PGL cell line or relevant mouse model derived from such a cell line, hampering any further progress and the introduction of new therapeutic approaches for these tumors based on HIF regulation.

Nevertheless, new and rapidly growing information about HIF function and regulation in cancer has led to the rapid discovery, development, and proposal of novel therapeutic agents targeting the hypoxia signaling pathway (244,245). Several possible methods to affect HIF signaling directly or indirectly include inhibition of 1) oncogenic signaling pathways (mTORC and EGFR inhibitors); 2) HIF-1α protein accumulation (topotecan, 2-methoxyestradiol; Hsp90 inhibitors; histone deacetylase inhibitors; thioredoxin inhibitors; nonsteroidal anti-inflamatory drugs [nonselective and cyclooxygenase-2 selective]; ascorbate; 9-beta-D-arabinofuranosyl-2-fluoroadenine [Fara-A], flavopiridol, isoflavonoidgenistein); 3) HIF-1α DNA binding activity (polyamides, echinomycin); and 4) HIF-1α transcriptional activity (chetomin, proteasome inhibitors, amphotericin B). Some natural products have also been used to target the HIF signaling pathway, such as some alkaloids or herbs [for a review, see (246)]. Although HIF-2α seems to be crucial in cancer progression, invasiveness, and metastasis (which may also apply to PHEO and PGL aggressive behavior) (182,243), drugs selectively targeting HIF-2α signaling have not yet been described but are under development (245,247). Moreover, because HIF-1α and HIF-2α can alternatively activate target genes depending on the stage of tumor progression/development, there is a huge interest in introducing drugs that would target both HIF-1α and HIF-2α [for a review, see (56)]. HIF-2α is also associated with stem cell maintenance and promotes tumor progression through the transcription of genes promoting invasion and proliferation of tumor stem cells, especially during prolonged hypoxia. Koh et al. (73) showed that HIF-2α transactivation and HIF-1α degradation is driven by HAF. HAF plays the critical role in hypoxia signaling and switches cells from HIF-1α to HIF-2α signaling. This suggests another therapeutic approach to PHEO and PGL, changing the balance of HIF-1α and HIF-2α by modulating HAF signaling in tumors. It remains to be investigated whether HIF signaling pathway inhibition may result in the activation or inhibition of other signaling pathways participating in tumorigenesis and if so, whether combined therapies are necessary for more optimal tumor response balance, keeping in mind any therapy-related toxicity and resistance. Finally, studies uncovering the unique participation of specific (pseudo)hypoxia-independent pathways regulating HIF in various PHEOs and PGLs, as outlined in Figure 3, should be carried out.