Genetics of pheochromocytoma and paraganglioma syndromes: new advances and future treatment options

Abstract

Purpose of review

To summarize the recent advances in the genetics of pheochromocytoma and paraganglioma (PHEO/PGL), focusing on the new susceptibility genes and dividing PHEOs/PGLs into two groups based on their transcription profile.

Recent findings

Recently, TMEM127, MYC-associated factor X, and hypoxia-inducible factor (HIF) 2α have been described in the pathogenesis of PHEOs/PGLs. Thus, now about 30–40% of these tumors are linked to the germline mutations, which also include mutations in the VHL, RET, NF1, SDHx, and SDHAF2 genes. Furthermore, PHEOs/PGLs have been divided into two groups, cluster 1 (SDHx/VHL) and cluster 2 (RET/NF1), based on the transcription profile revealed by genome-wide expression microarray analysis.

Summary

PHEOs/PGLs are the most inherited tumors among (neuro)endocrine tumors. Future approaches in genetics, including whole-genome sequencing, will allow the discovery of additional PHEO/PGL susceptibility genes. The current division of PHEOs/PGLs into cluster 1 and 2 provides us with additional knowledge related to the pathogenesis of these tumors, including the introduction of new treatment options for patients with metastatic PHEOs/PGLs. New discoveries related to the role of the HIF-1/HIF-2α genes in the pathogenesis of almost all inherited PHEOs/PGLs may call for a new regrouping of these tumors and discoveries of new treatment targets.

INTRODUCTION

Pheochromocytomas and paragangliomas (PHEOs/PGLs) are rare neuroendocrine tumors that produce catecholamines [1]. PHEOs/PGLs arise from three distinct parts of the neural crest: the adrenal medulla (PHEOs) and the sympathetic and parasympathetic paraganglia (PGLs) [2]. Initially, the pioneering work of Neumann et al. [3] showed that about one-quarter of these tumors were hereditary. According to the recent publications, up to 30–40% of these tumors are genetically inherited [3,4,5■■,6]. Now, the group of susceptibility genes includes the following genes: the von Hippel–Lindau (VHL) tumor suppressor gene, the rearranged during transfection (RET) protooncogene, the neurofibromatosis type 1 (NF1) tumor suppressor gene, genes encoding the four subunits (A, B, C, and D) of the succinate dehydrogenase (SDH) complex, and a gene encoding the enzyme responsible for flavination of the SDHA subunit (SDHAF2). In addition to these PHEO/PGL susceptibility genes, two other genes, KIF1Bβ and PHD2, have also been associated with PHEO/PGL development, although very rarely. Recently, three other genes, TMEM127, MYC-associated factor X (MAX), and hypoxia-inducible factor 2α (HIF2A) (both somatic and germline mutations found), have been described [7■,8,9■■,10,11,12■]. Somatic mutations in the known PHEO/PGL susceptibility genes have been reported as an extremely rare event, but recently, Burnichon et al. [13] detected somatic mutations of the RET and VHL genes in about 14% of sporadic PHEOs/PGLs. Also, Welander et al. [14] suggest that the NF1 gene constitutes the most frequent (24%) target of somatic mutations so far known in sporadic PHEOs. Furthermore, somatic mutations in MAX were found in 1.65% of tumors [15]. Thus, the proportion of all patients with PHEOs/PGLs because of the gene mutations described above is estimated to be approximately 50% at present (Table 1). In the present article, we focused on the role of new genes linked to the pathogenesis of PHEOs/PGLs and on clustering hereditary PHEOs/PGLs based on their transcriptional profile as described previously [13]. Discoveries of new genes and specific transcriptional profiles for PHEOs/PGLs will further provide new treatment strategies for these tumors.

Table 1.

Genotype-phenotype correlations in pheochromocytoma/paragangliomas because of mutations in susceptibility genes

Gene	Locus	Syndrome	Inh.	Malignant PHEO/PGL	Single PHEO	Bilateral PHEO	TAPGL	HNPGL	Multiple PGLs	Biochemical phenotype
SDHA	5p15		AD	?	−	−	+	+	−	?
SDHB	1p 36.13	PGL4	AD	+++	++	+	+++	++	++	MT, NMN, MTY or NS
SDHC	1q 21	PGL3	AD	±	±	−	+	++	+	MT, NMN or NS
SDHD	11q 23	PGL1	AD PI	+	+	+	++	+++	+++	MT, NMN, MTY or NS
SDHAF2	11q 13.1	PGL2	AD PI	?	−	−	−	+++	++	?
VHL	3p25–p26	VHL	AD	+	++	+++	+	±	+	NMN
NF1	17q11.2	NF1	AD	+	+	±	−	−	−	MN, NMM
RET	10q11.2	MEN	AD	±	++	++	−	−	−	MN, NMM
MAX	14q23.3		AD PI	+	++	++	−	−	−	Mixed: NMN + MN
TMEM127	2q11.2		AD	±	+++	++	±	±	±	MN
HIF2A	2p21–p16		Somatica	?	±	+b	++	−	++	NMN

AD, autosomal dominant; HNPGL, head and neck paraganglioma; Inh, inheritance; MN, metanephrine; MTY, methoxytyramine; NMN, normetanephrine; NS, nonsecreting; PGL, paraganglioma; PHEO, pheochromocytoma; PI, paternal inheritance; TAPGL, thoracic or abdominal paraganglioma; ?, unknown. SDHx-related paragangliomas can be also associated with SDH-deficient gastrointestinal stromal tumors (GISTs). This autosomal-dominant familial paraganglioma and GIST syndrome is known as Carney–Stratakis syndrome [11]. Recently, Pacak–Zhuang syndrome including paraganglioma, somatostatinoma, and polycythemia in women has been introduced. This syndrome is associated with somatic HIF2A gain-of-function mutation [16■].

^aRarely germline.

^bTaieb, Pacak, unpublished observation.

TMEM127 MUTATION

Recently, Qin et al. [8] reported heterozygous germline mutations in the TMEM127 gene in seven patients affected by PHEO. The TMEM127 gene, located on chromosome 2q11, encodes a transmembrane protein of 238 amino acids. TMEM127 is a highly conserved and broadly expressed protein with three transmembrane regions, but has no known functional domains. The TMEM127 protein associates dynamically with the endosomes and may participate in the protein trafficking between the plasma membrane, the Golgi, and lysosomes [8]. Tumors with TMEM127 mutations have a transcription signature comparable to that of RET-mutated and NF1-mutated PHEOs. However, neither RAS activation nor AKT phosphorylation was seen, indicating that TMEM127 loss is not identical to either NF1 or RET [8]. In vitro and in primary tumors, it was found that TMEM127 functions as a negative regulator of the mammalian target of rapamycin (mTOR), or more specifically of mTORC1, but the mechanisms underlying this interaction have not yet been established. However, TMEM127 knockdown, leading to an increase in the phosphorylation of mTORC1 targets, results in larger cells with higher rates of proliferation and hyperphosphorylation of mTOR effector proteins [8]. Thus, TMEM127 is a new tumor-suppressor gene involved in hereditary PHEOs/PGLs. TMEM127 missense, frame shift, and nonsense mutations were detected in all three coding exons of the gene, but no large TMEM127 deletions or duplications were found. Genetic studies of PHEO/PGL patients indicate a low prevalence of TMEM127 mutations (~2% of all cases negative for other PHEO/PGL susceptibility mutations) [8,17■,18]. A unique finding in these patients was the older average age (42 years) at presentation, which is similar to sporadic cases, but older than that of carriers of mutations in other susceptibility genes [8,17■,18]. In most cases, TMEM127 mutation carriers suffered from PHEOs only (unilateral as well as bilateral tumors) and secreted a high level of metanephrines. Only two mutations have been reported in a few patients with PGLs [19], one associated with multiple PHEO and retroperitoneal PGL and the other with bilateral carotid PGLs [8,17■,18,20]. Malignancy has been reported rarely [8,17■,18].

MAX MUTATION

In 2011, Comino-Mendez et al. [7■] identified MAX as a new PHEO tumor-suppressor gene in three independent patients with familial antecedents of the disease. The protein encoded by the MAX gene is a member of the basic helix–loop–helix leucine zipper (bHLHZ) family of transcription factors. The MAX protein is a ubiquitous, constitutively expressed protein that plays a central role in controlling the MYC/MAX/MXD1 (MAX dimerization protein 1) axis. MYC activates transcription binding to E-box DNA recognition sequences in target gene promoters through heterodimerization with MAX; heterodimers of MAX with MXD1 antagonize MYC-dependent cell transformation by transcriptional repression of the same E-box target DNA sequences [21,22]. MAX mutations are associated with bilateral PHEOs and an apparent paternal transmission of the disease [7■]. In a large international study, Burnichon et al. [15] confirmed that MAX germline mutations are responsible for PHEOs and PGLs in 1.12% of cases. In this study, patients with pathogenic MAX mutations had adrenal tumors, including 13 with bilateral or multiple PHEOs within the same adrenal gland. MAX-related thoracoabdominal PGLs were also found in the same study [15]. Thirty-seven percent of these patients had familial antecedents. Malignant disease developed in about 10.5% of patients [15]. However, Comino-Mendez et al. [7■] found metastasis at diagnosis in three out of eight probands. Therefore, mutations of MAX can be associated with a high risk of malignancy. Furthermore, somatic mutations in MAX were found in 1.65% of tumors [15]. The mutations were found in all five exons of the MAX gene, but were especially frequent in exons 3 and 4, corresponding to some of the most important residues within the conserved bHLH–Zip domain of MAX. The majority of the mutations led to the presence of truncated proteins, which resulted in the absence of the protein, as determined by immunohistochemistry [15]. How MAX mutations contribute to the pathogenesis of PHEO/PGL remains unclear. However, the ability of MYC to function independently of MAX has been demonstrated [23]. Therefore, Cascón and Robledo [21] suggested that the pivotal role of MAX in the MYC/MAX/MXD1 network is related more to the repression rather than to the activation of the MYC network.

HIF2A MUTATION

Zhuang et al. [9■■] identified novel somatic mutations in the gene encoding HIF-2α in multiple PGLs and duodenal somatostatinomas associated with polycythemia, suggesting the existence of a new syndrome. Also, Favier et al. [24] found somatic HIF2A mutations that may lead to the pathogenesis of PHEO in a female patient. The existence of a new syndrome (potentially to be named Pacak–Zhuang syndrome) was then confirmed in a larger series of female patients [16■]. These patients were found to have polycythemia either at birth or in childhood together with multiple PGLs and somatostatinomas. Subsequently, Lorenzo et al. [12■] reported a novel HIF2A germline mutation in a patient with congenital polycythemia with multiple PGLs.

HIFs are transcription factors controlling energy, iron metabolism, erythropoiesis, development, glycolysis, and other cell functions [25,26■]. The HIF-β subunit is constitutively expressed, whereas the α subunits are inducible by hypoxia [27]. When these proteins are dysregulated, they often contribute to tumorigenesis and cancer progression [28,29]. However, mutations in the genes encoding HIFs have not previously been identified in any cancer. Dominantly inherited gain-of-function mutations of HIF2A were previously found to be associated with an increase of erythropoietin and congenital polycythemia [30,31]. In HIF-2α-related polycythemia with or without PGLs, these mutations have been found mainly at hot spots in exon 12 [9■■,30,31]. HIF-2α gene mutations were found to disrupt prolyl hydroxylation of HIF-2α and, in turn, recognition by the VHL protein, resulting in a failure of HIF-2α ubiquitylation and its degradation. Thus, the longer half-life of the mutant HIF-2α protein resulted in the upregulation of downstream HIF-2α targets (EDN1, EPO, GLUT1, and VEGF), which is currently believed to be the pathogenic mechanism that leads to tumor development [9■■]. Very recently, we have also identified a somatic HIF2A gene mutation in a patient with multiple PHEOs associated with polycythemia (unpublished observation).

THE CLUSTERS OF HEREDITARY PHEOCHROMOCYTOMAS AND PARAGANGLIOMAS BASED ON TRANSCRIPTIONAL PROFILE

Hereditary and sporadic PHEOs/PGLs can be divided into two groups based on the transcription profile revealed by the genome-wide expression microarray analysis. The first group (cluster 1) includes tumors carrying VHL and SDHx (SDHD, SDHB, SDHC, SDHA, and SDHAF2) mutations and also accounts for about 30% of sporadic tumors [13,32,33]. The second group (cluster 2) represents the tumors carrying NF1, RET, and KIF1Bβ mutations, and also includes about 70% of sporadic tumors [5■■,13,34■,35■]. The new TMEM127 and MAX genes are most likely associated with cluster 2 and HIF2A with cluster 1 [9■■,12■,13,15].

Cluster 1

VHL/SDHx mutations lead to impaired degradation and accumulation of HIF-1/HIF-2α and display signatures of pseudohypoxia, angiogenesis, increased reactive oxygen species (ROS), and reduced oxidative response, resulting in changes in the cell metabolism (energy metabolism regulation). HIF-α heterodimerizes with HIF-β and acts as an active transcription factor. Succinate that accumulates in the mitochondrial matrix owing to SDH dysfunction leaks out into the cytosol, where it inhibits the activity of HIF-1/HIF-2α prolyl hydroxylase enzymes (PHDs – PHD1, PHD2, and PHD3, also known as EglN2, EglN1, and EglN3, respectively) that hydroxylate two prolyl residues [36]. Hydroxylated HIF-α is recognized by the VHL protein and marked for degradation in the proteasome. If the VHL gene is mutated, the protein is not formed, so HIF-α cannot be degraded and therefore accumulates. Burnichon et al. [13] showed that SDHx-related and VHL-related PHEOs/PGLs shared overexpression of several genes involved in angiogenesis and the hypoxic pathway. Some genes were specifically overexpressed in SDHx-related tumors. These genes are involved in transcription regulation, protein transport, proliferation, energy metabolism, and cell adhesion. Also, several specifically overexpressed genes have been found in VHL-related tumors. These genes are EGLN3 and KISS1R, as well as genes involved in glycolysis [13]. Burnichon et al. [13] found, which was also confirmed by Lopez-Jimenez et al. [33], that although VHL/SDHx-related tumors are associated with pseudo-hypoxia, some HIF target genes were differentially expressed between SDHx and VHL-related tumors. Most of these genes, such as ENO1, BNIP3, or CA9, are considered to be HIF-1α-specific targets and were specifically induced in VHL-related tumors [13]. Therefore, these tumors were further divided into cluster 1A (SDHx) and cluster 1B (VHL).

The SDHB mutation is associated with malignancy and poor prognosis [37]. Burnichon et al. [13] found genes specifically overexpressed in SDHB-related PHEOs/PGLs, including MMP24, DSP, SIX1, LGR5, LAPTM4B, and β-catenin, most of which are important in the development of metastasis.

Cluster 2

Cluster-2-related PHEOs/PGLs are linked together by the activation of kinase signaling pathways driven by the oncogenes that involve translation, initiation, and protein synthesis, and genes involved in neural/neuroendocrine identity [13,32, 35■,38–40]. The proto-oncogene RET is a tyrosine kinase receptor primarily expressed in the neural crest cells. RET mutations have been associated with increased activation of PI3K/v-Akt signals [41]. NF1 encodes for the protein neurofibromin, a GTPase-activating protein in the RAS signaling cascade and mTOR signaling pathway [42,43]. Thus, RET and NF1 mutations lead to activation of the PI3K/AKT/ mTOR and RAS/RAF/MAPK signaling pathways. TMEM127 mutations enhance mTOR activity independent of the RET and NF1 pathways. Finally, MAX gene mutations result in the dysregulation of the MYC–MAX–MXD1 network connected with mTOR pathway function [15].

FUTURE TREATMENT

Whereas the most optimal treatment for benign PHEOs/PGLs is surgical resection, therapy for malignant/metastatic disease is unsatisfying at best. Understanding the specific genetic alterations of various PHEOs/PGLs will undoubtedly open promising new options for targeted therapies in the near future. Also, clustering PHEOs/PGLs by expression alterations can lead us to choose certain treatment targets. Thus, expression microarrays can be a more powerful tool for the detection of target genes and associated pathways involved in PHEOs/PGLs in the future [13,15,24,28,42,43].

Yang et al. [44■] have recently demonstrated that the loss of SDHB function was because of a reduction in the mutant protein half-life by rapid proteasome degradation. The authors found that histone deacetylase inhibitors (HDACi) stabilized the half-life of mutated SDHB proteins and their activity. Direct activation of PHD by the activator KRH102053 increases HIF-1/HIF-2α hydroxylation and promotes its degradation [45,46].

Metastatic SDHx-related PHEO/PGL overexpresses heat shock protein 90 (HSP90), a molecular chaperone that assists in binding to HIF-1/HIF-2α and promotes its stability by preventing ubiquitination and proteasomal degradation of HIF-1/HIF-2α [25,47,48]. Thus, the inhibitors of HSP90, such as geldanamycin and analogs 17-allylaminogeldanamycin (17-AAG; tanespimycin) and 17-dimethylaminoethylamino-17-demethoxygeldanamycin (17-DMAG; alvespimycin) or other new ones, are promising novel anticancer therapeutic agents [49,50]. Small-molecule inhibitors of glucose transporter 1 (GLUT1), such as WZB117 or STF-31, downregulate glycolysis and inhibit cancer cell growth in vitro and in vivo [51,52]. Favier et al. [53] showed that cluster 1 tumors display angiogenic markers such as vascular epithelial growth factor (VEGF), its receptors, HIF-2α, angiopoietin-2, and the endothelin receptors ETA and ETB. Thus, these results suggest that there is a rationale for antiangiogenic therapy, including targeting the VEGF pathway using either humanized monoclonal anti-VEGF antibodies (Bevacizumab) or small tyrosine kinase inhibitors such as sunitinib or sorafenib [53]. Also, targeting HIF-1/HIF-2α by HIF-1α inhibitors (direct inhibitor PX-478 and indirect inhibitor PX-12) has shown antitumoral activity inhumantumor xenografts in mice and also seems to be promising for malignant PHEO/PGL [54–56].

Favier et al. [53] showed that the mTOR pathway was potentially activated in half of PHEOs/PGLs. Nolting et al. [57] showed that the combination treatment with dual NVP-BEZ235 (PI3K/mTORC1 inhibitor) and lovastatin (inhibitor of ERK signaling) had a significant additive effect in mouse PHEO cells and resulted in inhibition of both AKT and mTORC1/p70S6K signaling without ERK upregulation.

The new therapeutic options outlined above are expected to be introduced in the near future.

CONCLUSION

In the near future, next-generation sequencing technology is predicted to replace conventional sequencing methods and a stepwise procedure for genetic screening will likely no longer be required. Similarly, whole-genome sequencing will allow the discovery of newer PHEO/PGL susceptibility genes. Molecular targeted therapies will appear as the most promising strategies for the management of patients with metastatic PHEO/PGL.

KEY POINTS.

• TMEM127, MAX, and HIF2A are the new susceptibility genes for PHEOs/PGLs.

• New syndromes (such as Carney–Stratakis syndrome, comprised PGL and GIST, and a new syndrome potentially to be named Pacak–Zhuang syndrome, comprised multiple PGLs and duodenal somatostatinomas associated with polycythemia in women) have been introduced.

• Genome-wide studies led to the identification of somatic mutations in about 14% of sporadic PHEOs/PGLs.

• Hereditary and sporadic PHEOs/PGLs are currently divided into two groups (SDHx/VHL and RET/NF1) based on the transcription profile revealed by genome-wide expression microarray analysis or by other approaches.

• Deep knowledge of the genetic changes and expression profiles will introduce new targeted therapy for malignant PHEOs/PGLs.