Genetics and molecular pathogenesis of pheochromocytoma and paraganglioma

Authors

  • S. R. Galan,

    Corresponding author
    • Division of Endocrinology & Diabetology, Faculty of Medicine, Philipp's University Marburg, University Hospital Giessen and Marburg, Marburg, Germany
    Search for more papers by this author
  • P. H. Kann

    1. Division of Endocrinology & Diabetology, Faculty of Medicine, Philipp's University Marburg, University Hospital Giessen and Marburg, Marburg, Germany
    Search for more papers by this author

Correspondence: S. R. Galan, Division of Endocrinology & Diabetology Faculty of Medicine, Philipp's University Marburg, University Hospital Giessen and Marburg, Marburg, Germany. Tel.: 0049 6421 586 3135; E-mail: galan@med.uni-marburg.de

Summary

Although most pheochromocytomas (PCCs) and paragangliomas (PGLs) are sporadic, molecular genetic medicine has revealed that a considerable number of patients with apparently sporadic PCC actually have a genetic predisposition to the development of these tumors. After decades of intensive research, several genes are now known to play an important role in the pathogenesis of PCC. At present, these are RET proto-oncogene, von Hippel–Lindau disease tumor suppressor gene (VHL), neurofibromatosis type 1 tumor suppressor gene (NF1), genes encoding the succinate dehydrogenase (SDH) complex subunits SDHB, SDHC, and SDHD, but also SDHA, the gene encoding the enzyme responsible for the flavination of SDHA (SDHAF2 or hSDH5), and the newly described TMEM127 and MAX tumor suppressor genes. In addition to these ten PCC susceptibility genes, two other genes, KIF1B and PHD2, have also been associated with PCC. Studying the pathogenesis and the molecular correlation of these mutations has revealed the existence of two main transcription signatures: a pseudohypoxic cluster (VHL and SDH mutations) and a cluster rich in kinase receptor signaling and their downstream pathways (RET, NF1, TMEM127, and MAX mutations). However, the general mechanism in the pathogenesis of a syndrome does not entirely apply in the particular pathogenesis of PCC as a manifestation of that syndrome. A better understanding of the complexity and high genetic diversity of PCC and PGL may lead to more efficient diagnosis and management of the disease.

Introduction

Pheochromocytomas (PCCs) are neuroendocrine tumors of neuroectodermal origin derived from chromaffin tissue of the adrenal medulla, which usually, but not always, secrete excessive catecholamines. PCC can cause a variety of clinical symptoms – the majority of patients, however, present with continuously or paroxysmally increased blood pressure due to catecholamine excess.[1] Although PCCs are relatively rare tumors, found in about 4% of adrenal incidentalomas and in about 0·2–0·4% of hypertensive patients,[2] autopsy series have revealed a much higher prevalence.[3] For this reason alone, it is highly advisable to better understand the pathology of PCC and to identify lesions at an early stage. Paragangliomas (PGLs) arise from paraganglia of the autonomic nervous system, which are distributed throughout the body and play important roles in homeostasis.[4] A topographical classification helps to classify the PGLs in the head and neck branchial PGLs and metameric PGLs. The head and neck PGLs, sometimes referred to as nonchromaffin PGLs, are located in the carotid body, in the vagal body, in the jugulotympanic region, in the superior and inferior laryngeal paraganglionic tissues, in the nasal cavity, or in the orbit. The metameric PGLs, or sometimes referred to as chromaffin PGLs or chromaffinomas, are located in thorax, abdomen, pelvis, and urinary bladder.[5, 6] The PGLs of the head and neck are associated with the parasympathetic nervous system and often lack endocrine activity. PCC and chromaffin PGLs are, on the contrary, more closely associated with the sympathetic nervous system and usually secrete higher levels of catecholamines.[4] Altogether, about 2–5% of the head and neck PGLs and more than 50% of abdominal PGLs produce hormones. While PCCs produce epinephrine predominantly, PGLs secrete norepinephrine. Up to 20% of PCCs and PGLs are malignant, with metastatic lesions or local relapse.

PCC and PGL can be divided into two types: familial and sporadic (not congenital). It has been shown that up to 24% of the patients with apparently sporadic PCC may have in fact familial PCC.[7] Some criteria such as younger age at the onset of symptoms, bilateral localization of the tumors, and a positive family history could indicate the presence of a familial syndrome. The molecular genetic diagnosis remains however decisive. Ten PCC susceptibility genes are currently known, yet the high genetic diversity has transcription correlates, which are reflected in the two main transcription signatures underlying these mutations: a pseudohypoxic cluster (VHL and SDH mutations) and a cluster rich in kinase receptor signaling and its downstream pathways (RET, NF1, TMEM127, and MAX mutations) (Fig. 1). In this paper, we will discuss the various syndromes that must be considered when investigating PCC/PGL and these are summarized in Table 1.

Table 1. Overview of the 10 currently known pheochromocytoma/paraganglioma (PCC/PGL) susceptibility genes with their clinical features
GeneChromosomeFreq. (%)SyndromeFurther syndromic clinical manifestationsInh.Malignant PGL/PCCTumor localization
SinglePCCBilateral PCCTAPGLHNPGLMultiple PGL
  1. Freq. = Frequency of mutation reported to all patients with PCC or PGL. Inh. = Inheritance, AD = autosomal dominant, PI = paternal inheritance (maternal imprinting) TAPGL = thoracic or abdominal PGL, HNPGL = head and neck PGL.

RET 10q11·2Approximately 5MEN 2Amedullary thyroid cancer, hyperparathyroidism, cutaneous lichen amyloidosisAD++++
MEN 2Bmedullary thyroid cancer, mucosal neuroma, marfanoid habitus
VHL 3p25-26Approximately 9VHLhemangioblastomas, retinal angiomas, clear-cell renal cell carcinomas, endolymphatic sac tumors, serous cystadenomas and neuroendocrine tumors of the pancreas, papillary cystadenomas of the epididymis, and broad ligamentAD+++++++++
NF1 17q11·2Approximately 2NF1cafe au lait spots, neurofibromas, Lisch nodules of the iris, optic pathway and brainstem gliomas, astrocytomas, soft-tissue sarcomas, chronic myeloid leukemias of childhood, learning disabilities, seizures, macrocephaly, short stature, scoliosis and pseudoarthrosisAD++
SDHD 11q23Approximately 5PGL-1 ADPI++++++++++
SDHAF2 11q13·1<1PGL-2 ADPI+++++
SDHC 1q21Approximately 1PGL-3 AD++++++
SDHB 1p36·1Approximately 5PGL-4additional neoplasms, such as renal cell carcinoma, thyroid tumors, neuroblastoma, GISTAD+++++++++++++
SDHA 5p15<1  AD++
TMEM127 2q11Approximately 2  AD++++++++
MAX 14q23·3<1  ADPI++++++
Figure 1.

Schematic Please check and approve the edit made in the figure caption representation of the pheochromocytoma/paraganglioma classification in cluster 1 and cluster 2 based on the pathogenesis or signaling pathway; the mutated proteins are disrupting.

Multiple endocrine neoplasia type 2 (MEN 2)

About 5–10% of the patients with PCC develop the tumor in the context of a MEN 2 syndrome.[7] In this case, the PCC is usually bilateral, but may be asynchronous. The frequency of PCC in MEN 2 is 40–50%, while medullary thyroid cancer is the most common manifestation of MEN 2 syndrome with a frequency of over 90% and usually also represents the first manifestation, at ages between 5 and 25 years.[8] Only about 10–20% of MEN 2 patients develop multigland parathyroid hyperplasia with hyperparathyroidism. Rare cutaneous lichen amyloidosis has been described in some families with MEN 2A.[9] Three distinct clinical forms have been described depending on the phenotype: classic MEN 2A with the aforementioned manifestations, MEN 2B as an association of medullary thyroid cancer, PCC, mucosal neuroma, and marfanoid habitus, as well as FMTC (familial medullary thyroid cancer). The latter is usually associated with a low incidence of other endocrinopathies.[8]

The MEN 2 syndrome is an autosomal dominant syndrome and can be traced back to constitutional mutations of the RET proto-oncogene on chromosome 10, which is relevant to various growth- and differentiation processes. Its gene product, the RET protein, is a transmembrane receptor of the tyrosine kinase family. It is expressed in cell lineages derived from the neural crest and has a key role in regulating cell proliferation, migration, differentiation, and survival during embryogenesis.[10] Under normal conditions, the RET receptor can be activated through various factors such as glial-cell-line-derived neurotrophic factor (GDNF), neurturin, artemin, and persephin. This happens only in the presence of coreceptors from the GDNF-α receptor family (GFR-α-1, GFR-α-2, GFR-α-3, and GFR-α-4).[11] The interaction between these molecules induces dimerization of the RET protein followed by autophosphorylation of the RET receptor intracellular region and subsequently by intracellular signal transmission. Germ-line mutations of the RET proto-oncogene cause constitutive activation of the RET receptor and of intracellular signaling pathways (“gain of function”), ultimately resulting in the cellular transformation.[12] This particular feature of MEN 2 is actually rare, as most genetically determined tumors are the result of loss-of-function mutations, which inactivate specific tumor suppressor proteins. For example, mutations causing loss of function of the RET protein were found to be associated with Hirschsprung's disease, a developmental disorder characterized by the absence of enteric ganglia in the intestinal tract.[13] Extensive studies on large families reveal that the phenotype of MEN 2 correlates partly with specific mutations and their localization on the RET proto-oncogene.

The majority of mutations in MEN 2A patients and most in FMTC patients involve one of the five cysteine residues in the cysteine-rich region of the RET protein's extracellular domain encoded in exon 10 (codons 609, 611, 618, and 620) or exon 11 (codons 630 or 634) of the RET proto-oncogene.[14, 15] The most common genetic defects in MEN 2A correspond particularly to mutations in codon 634 in exon 11 and are associated with a high probability of PCC. Rarely, mutations involve the intracellular tyrosine kinase (TK) domain of RET receptor (exon 13: codon 768, exon 14: codon 804, and exon 15: codon 891), which are usually associated only with mild cases of FMTC with almost no presence of PCC. Less common mutations in MEN 2A and FMTC occur in exon 13 in codon 790 and were also thought to be specific for FMTC.[7] The RET p.Tyr791Phe and p.Ser649Leu mutations were also thought of being pathogenic until Erlic et al. proved the contrary and cataloged them as polymorphisms.[16] MEN 2B-associated tumors are caused by mutations in the intracellular tyrosine kinase TK2 domain of the RET receptor. A single amino acid exchange (methionine to threonine) at codon 918 in exon 16 is responsible for over 95% of cases of MEN 2B and is found only in this disorder. Mutations at codon 883 in exon 15 or at codon 922 in exon 16 have been found in about 3–5% of unrelated MEN 2B patients.[17]

Recent reports have described further mutations of the RET proto-oncogene localized in other exons than expected. For example, Castellone et al. reported a novel heterozygous germ-line RET mutation in exon 5 (p.Val292Met) in a 44-year-old patient with occult medullary thyroid carcinoma and unilateral PCC. The mutation mapped to the third cadherin-like domain of the RET protein.[18] Furthermore, a new missense point mutation in exon 8 of the RET proto-oncogene (p.Gly533Cys) corresponding to the cysteine-rich domain of RET protein has been reported in 76 patients from a sixth-generation Brazilian family with 229 subjects, with ascendants from Spain.[19] The same mutation has also been identified in two Greek families and seems to cause FMTC.[20] Nevertheless, there is still a small group of MEN 2 patients (approximately 2%) for which no mutation has been found despite extensive molecular screening.

Von hippel–lindau syndrome (VHL syndrome)

VHL syndrome is an inherited, autosomal dominant syndrome, manifested by a variety of benign and malignant tumors, which affect one in 36 000 individuals.[21] The spectrum of VHL-associated tumors includes hemangioblastomas of the brain and spine, retinal angiomas, clear-cell renal cell carcinomas, endolymphatic sac tumors of the middle ear, serous cystadenomas and neuroendocrine tumors of the pancreas, papillary cystadenomas of the epididymis and broad ligament, as well as PCCs. A clinical classification of VHL syndrome can be based on the absence (type 1) or presence (type 2) of PCC. Type 2 patients with low risk of developing renal cell carcinoma are further subclassified as VHL syndrome type 2a, whereas those with high risk are included in VHL syndrome type 2b. Type 2c patients are those patients who only develop PCC without the other classical manifestations of VHL syndrome.[22] PCCs arise in about 10–20% of patients with VHL disease. The mean age at presentation is about 30 years. The onset can sometimes be before the age of 10. PCC in VHL disease can be multiple and bilateral, or even arise as extra-adrenal PGLs. Five percent of all PCCs are malignant.[23]

The VHL gene has been mapped to chromosome 3p25–26 and encodes the VHL tumor suppressor protein. Currently, more than 500 different germ-line VHL mutations linked to VHL disease have been reported. The VHL mutations are extremely heterogeneous and are distributed throughout the coding sequence, except that intragenic missense mutations are rarely seen within the first 50 codons.[24] Missense, nonsense, splice site mutations, microdeletions, and insertions are detected in about two-thirds of the patients, whereas large deletions (4–380 kb) are found in the remaining one-third of the VHL families. Deletions in VHL and nonsense and frameshift mutations appear to be more common in type 1 disease, while missense mutations may be more common in type 2 disease.[25] Furthermore, 90% of families with PCC have missense mutations, whereas families with renal cell carcinoma show the entire spectrum. Missense mutations at codon 167 are associated with a particularly high risk of PCC (over 80% by age 50).[26]

The VHL tumor suppressor protein targets especially hypoxia-inducible factor-1 (HIF-1), but also several other proteins potentially involved in tumorigenesis, such as matrix metalloproteinases (MMP), MMP inhibitors, and atypical protein kinase C. Through proteasomal degradation, the VHL protein regulates the levels of the above-mentioned proteins within the cell.[27] HIF-1 is involved in erythropoiesis through its ability to induce transcription of mRNA coding for erythropoietin. Furthermore, it regulates several growth factors, such as vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF)-beta, and transforming growth factor (TGF)-alpha.[28] In accordance with Knudson's two-hit tumor suppressor model, tumors from patients with VHL who have a germ-line VHL mutation display somatic inactivation of the remaining wild-type allele. The result is abnormal or absent VHL protein function and therefore decreased proteasomal degradation of many growth factors, in particular HIF-1, followed by increased mRNA transcription of erythropoietin (as during hypoxia). The combined effect of the increased angiogenic and growth factors may be responsible for an autocrine loop that provides an uncontrolled growth stimulus and, thus, the development of VHL tumors.[27]

However, chromaffin cells and their precursors seem to require residual VHL function and cannot tolerate complete loss of VHL protein function in contrast to renal cells.[29] Complete deletion or truncating mutations in the VHL gene are extremely rare in PCC cases, while they are often found in renal carcinoma-related disease. Furthermore, VHL type 2c patients, manifesting only PCC, are still able to maintain the ability to down-regulate HIF-1,[29, 30] whereas the mutations associated with type 2a and 2b affect the proteasomal degradation of HIF-1 to different degrees. This might suggest that HIF-1 is not relevant to PCC development and that there must be other molecular mechanisms and VHL substrates involved in the PCC formation in VHL patients. On the other hand, transcription of the tyrosine hydroxylase (TH) gene, which encodes the final and rate-limiting enzyme in catecholamine synthesis, is hypoxia inducible and seems to be regulated by VHL protein.[31, 32] Moreover, VHL type 2c patients seem to be incapable of binding and regulating the assembly of fibronectin.[29, 30] This indicates that abnormal extracellular matrix formation may play a role in the pathogenesis of PCC.

Neurofibromatosis type 1 (NF1)

NF1, the classical von Recklinghausen's disease, is an autosomal dominant genetic disorder with an incidence of about one in 3 000 individuals. The condition is characterized by multiple cafe au lait spots, neurofibromas, and Lisch nodules of the iris. Although each of the three characteristics occurs in over 90% of all NF1 patients by puberty, the number of lesions is extremely variable. Other features, such as learning disabilities, seizures, macrocephaly, short stature, scoliosis, and pseudoarthrosis, are present in only a minority of NF1 cases. Patients with NF1 develop both benign and malignant tumors at increased frequency throughout life, such as optic pathway gliomas, astrocytomas, and brainstem gliomas, but also soft-tissue sarcomas, chronic myeloid leukemias of childhood, and PCC.[33] Historical review of families with NF1 has identified a 0·1–5·7% incidence of PCC. However, PCCs have been found in 3·3–13% of patients with NF1 disease at autopsy.[34] The mean age at diagnosis of PCC in patients with NF1 is 42 years. Most of the patients have solitary PCC (approximately 84%), while only a few develop bilateral adrenal or extra-adrenal tumors.[35] Malignant behavior is slightly more frequent than in VHL or in MEN 2 syndromes.

The NF1 gene has been localized on chromosome 17qll.2 and encodes its protein product, neurofibromin, which is found in neurons, oligodendrocytes, and Schwann cells, but also in keratinocytes, adrenal medulla, and leukocytes.[36] The GAP (Ras-GTPase-activating protein)-related domain (GRD) is a small part of neurofibromin (only 360 amino acids) and has the important role of stimulating the intrinsic GTPase of p21-Ras-GTP to hydrolyze GTP to GDP, thus inactivating p21-Ras. P21-Ras is a key component of many growth factor signaling pathways, and neurofibromin acts therefore as a tumor suppressor protein. Most mutations in the NF1 gene result in the truncation of neurofibromin and thus in the loss of functional protein causing the wide spectrum of clinical findings. In the absence or at decreased levels of neurofibromin in NF1, signaling is increased through various pathways resulting in the cell proliferation and inhibited apoptosis.[36] The NF1 mutations include translocations, splicing, deletions, duplications, insertions, point mutations, and substitutions. NF1 also has a high rate of new mutations of about 50%. Several studies have revealed that homozygosity is lethal to embryos. All affected living individuals are therefore heterozygous for a NF1 mutation.[37] The clinical expression of NF1 mutations varies greatly, even within a given family carrying the same mutation.

However, it seems that in the pathogenesis of NF1-associated PCC, the cysteine-serine-rich domain (CSR) of neurofibromin plays a more important role than the GRD. This consideration is based on the observation in a recent study that only 13% of the germ-line mutations in a series of PCC patients with NF1 were within the RGD, in contrast to 35% in the CSR.[38] These data suggest that the CSR could be an equally important functional domain in neurofibromin and a considerable mutational target, at least in NF1-associated PCC. The spectrum of NF1 mutations in patients with PCC has not been fully established, as diagnosis is based on the clinical parameters rather than routine genetic screening. Unlike mutations in VHL disease, NF1 mutations which put a patient at risk of PCC have not been identified.[35] However, it seems also that in the pathogenesis of PCC in NF1 patients, Knudson's two-hit tumor suppressor model could be applied, resulting in a loss of heterozygosity at tumor level. This process leads to the lack of expression of neurofibromin in PCCs.[39]

Familial paraganglioma (familial PGL)

Familial PGL is an autosomal dominant disorder. Currently, four hereditary PGL syndromes have been described to be associated with distinct genetic susceptibility loci: PGL-1 with mutations in the SDHD gene on chromosome 11q23,[40, 41] PGL-2 with mutations in the SDH5/SDHAF2 gene on chromosome 11q13,[42] PGL-3 with mutations in the SDHC gene on 1q21,[43] and PGL-4 with mutations in the SDHB gene on 1p21.[44] SDHA, SDHB, SDHC, and SDHD are subunit genes of succinate dehydrogenase (SDH), which compose portions of mitochondrial complex II and are involved in the electron transport chain and Krebs cycle.[45] It has been suggested that SDH mutations cause an accumulation of succinate and reactive oxygen species (ROS), which, among other things, results in stabilization of HIF-1, thus activating hypoxia signaling pathways.[46] Here, also the Knudson's two-hit tumor suppressor model could be applied.[47]

Mutations in SDHD and SDHC are more probable in the silent head and neck PGLs. SDHC mutations are usually less common than SDHD mutations, but have been identified nonetheless in up to 4% of patients with the head and neck PGLs.[48] Some of these patients present with multiple PGLs, a few being located in the thorax. However, no PCCs have been associated with SDHC mutations. The tumors are typically benign, but malignancy has been reported in one case.[49] Patients with SDHD mutations usually develop head and neck PGLs, but they can also present with thoracic or abdominal PGL and PCC, as one of the multiple PCC/PGL tumors throughout the body. In addition to multifocal tumors, also typical for SDHD mutations is maternal imprinting, these mutations being penetrant only when inherited from the father.[50] The associated risk of malignancy for SDHD mutations carriers is altogether low.

SDHB mutations have been mostly associated with abdominal, pelvic, and thoracic catecholamine-secreting familial PGL. Furthermore, the carriers are more likely to develop malignant PGLs and additional neoplasms, such as renal cell carcinoma, papillary thyroid tumors, neuroblastoma, or gastrointestinal stromal tumors (GIST).[51, 52] About 20% of SDHB mutation carriers will develop malignant PGLs, and up to 50% of patients with a malignant PGL harbor a SDHB mutation.[51] The associated risk of malignancy varies significantly from one study to another, depending on the number of analyzed subjects and on the applied definitions of metastatic disease, including or not the sites where chromaffin tissue is normally present. In addition, more lifelong follow-up studies are needed, to correctly estimate the lifelong malignant risk in SDHB mutation carriers. Germ-line mutations of the SDHB, SDHC, and SDHD genes were also found in the Carney–Stratakis syndrome, an autosomal dominant disorder characterized by the dyad of PGLs and GIST.[53]

Mutations in SDHA have been shown to be associated with juvenile encephalopathy, also known as Leigh syndrome,[54] and were initially thought not to cause hereditary PGL or PCC. However, Burnichon et al. recently identified a patient with abdominal PGL resulting from a germ-line mutation in SDHA (p.Arg589Trp) associated with the loss of heterozygosity in the tumor, providing the first evidence that mutations in SDHA can cause PGL and that SDHA, like other SDH genes, can act as a tumor suppressor gene in accordance with the Knudson's two-hit model.[55] Recently, a SDHA immunohistochemistry analysis allowed the identification of further SDHA mutations in at least 3% of a total of 316 patients affected by apparently sporadic PGL and PCCs.[56] The gene for PGL-2 had initially not been identified, but mapped to chromosome 11q13·1.[42] In 2009, Hao and Rutter showed that 11q13·1 is actually the genomic locus of the so-called hSDH5 protein[57] and studied in detail its yeast ortholog, the SDH5 protein. It has been found that hSDH5 binds to SDHA (SDH1 in yeast) and promotes its flavination, which is required for SDH complex assembly and function. The loss of SDH5 in yeast decreases the abundance of the other SDH components, probably caused by enhanced degradation as a result of altered SDH complex formation. Hao and Rutter also identified a nonsynonymous hSDH5 variant in one Dutch family with hereditary PGLs of the head and neck and confirmed that it was defective with respect to SDHA flavination and restoration of SDH activity. Similarly to SDHD, families with PGL-2 syndrome also seem to exhibit maternal imprinting.[57] Because of the discovery of SDH5 gene by Hao and Rutter, other families with SDH5 mutation, now called SDHAF2 mutation, have been identified. However, no germ-line or somatic SDHAF2 mutations have been found in patients with adrenal PCC or tumors of sympathetic paraganglia of the abdomen and thorax. Mutations of SDHAF2 seem to be only a rare cause of the head and neck PGL.[58]

TMEM127

Recently, Qin et al. reported heterozygous germ-line mutations in a new PCC susceptibility gene, TMEM127, on chromosome 2q11, in seven patients affected by PCC.[59] The TMEM127 gene seems to encode a transmembrane protein of 238 amino acids, the TMEM127 protein, which associates dynamically with endosomes and may participate in protein trafficking between the plasma membrane, the Golgi and lysosomes. in vitro and primary tumor analyses have indicated that TMEM127 is associated with kinase receptor signals and is a negative regulator of mTOR, or more specifically of mTORC1, which promotes cell growth and protein translation. TMEM127 acts therefore as a tumor suppressor gene.[59] The wild-type allele was consistently deleted in tumor DNA, in agreement with the Knudson's two-hit tumor suppressor model. Furthermore, TMEM127 missense, frameshift, and nonsense mutations were detected in all three coding exons of the gene, but no large TMEM127 deletions or duplications. Surprisingly, the patients with TMEM127 mutations developed PCCs on average at 45·3 years of age, which is older than the average age at presentation of the other familial PCCs. The fact that the average age at onset of tumors in patients with TMEM127 mutations was similar to that of patients with sporadic disease has been confirmed by larger studies.[60, 61] In most cases, the tumors arose from the adrenal medulla, while only two mutations have been reported in patients with PGL: one associated with multiple head and neck PGL and the other with retroperitoneal PGL and PCC.[62] The TMEM127 mutation carriers can present with unilateral as well as bilateral tumors. Malignancy has been rarely reported.

MAX

The MYC-associated factor X (MAX) gene is the most recently reported PCC susceptibility gene. Comino-Mendez et al. sequenced the exomes of three unrelated individuals with hereditary PCC and identified mutations in MAX.[63] MAX is a gene of five exons, located on chromosome 14q23·3. It encodes a transcription factor, MAX protein, which belongs to the basic helix-loop-helix leucine zipper (bHLHZ) family and plays an important role in regulation of cell proliferation, differentiation, and apoptosis as a part of the MYC-MAX-MXD1 network. In this network, MAX is the common interaction partner for both MYC and MXD1 proteins. While heterodimerization of MYC with MAX mediates their function as transcription factors, heterodimers of MAX with MXD1 antagonize MYC-dependent cell transformation by transcriptional repression.[64] Already in 1995, it was known that PC12A, a single-cell clonal line established from a transplantable rat adrenal PCC, expressed a mutant form of MAX incapable of dimerization with MYC. Reintroduction of normal MAX in these cells resulted in transcriptional repression and growth reduction. This implied that MAX can function as a tumor suppressor. Furthermore, the ability of the tumor cells to divide, differentiate, and apoptose despite the mutated MAX suggested for the first time that these processes can occur via MAX-independent mechanisms.[65] The recent reported absence of MAX protein in the tumors and the loss of heterozygosity (LOH) of the wild-type allele in tumor DNA in the analysis of Comino-Mendez et al. of three unrelated individuals with hereditary PCC confirmed the tumor suppressor role of MAX in humans. A study of an additional 59 selected patients with PCC identified five further MAX mutations and suggested an association of tumor bilaterality with malignant outcome as well as preferential paternal transmission.[63] An even more recent study sequenced MAX in 1694 patients with PCC or PGL but no mutations in the major susceptibility genes and established that MAX germ-line mutations are responsible for PCC or PGL in 1·12% of cases.[66] Furthermore, the study supported previous suggestions of a paternal mode of transmission, extended the spectrum of MAX-related tumors to PGLs, determined no special predisposition to malignancy, and demonstrated that MAX tumors produce predominantly norepinephrine, but with some capacity to also produce epinephrine. The age at diagnosis of PCCs in MAX mutation carriers was lower, compared with the nonmutated cases (34 vs 48 years). The identified mutations were distributed along the gene, but were especially frequent in exons 3 and 4. The majority of mutations lead to truncated proteins. The most frequently found mutation was the c.97C>T variant, representing the first hot spot mutation of MAX.[66]

KIF1B and PHD2/EGLN1

In addition to the ten PCC susceptibility genes already discussed, two novel genes, KIF1B and PHD2, have also been associated with PCC. Schlisio et al. sequenced 52 PCCs among other tumors and identified KIF1B missense variants in two PCCs (Ser1481Asn and Glu628Lys).[67] The Ser1481Asn variant was reported in a 28-year-old woman who, at 17 months of age, presented with a neuroblastoma and, in adulthood, developed a mature ganglioneuroma and bilateral PCC. However, the Ser1481Asn variant was not associated with the loss of the remaining wild-type allele. Her paternal grandfather also developed bilateral PCC, while the girl's father did not show any signs of the disease. KIF1B is a large gene of about 50 exons mapping to chromosome 1p36·22. The gene has two splice variants, KIF1Bα and KIF1Bß, which are motor proteins implicated in anterograde transport of mitochondria and synaptic vesicle precursors. KIF1Bß functions as a tumor suppressor that is necessary for neuronal apoptosis. Germ-line KIF1Bß mutations were also identified in patients with neuroblastoma, medulloblastoma, ganglioneuroma, leiomyosarcoma, and lung adenocarcinoma.[67, 68]

In 2008, a germ-line mutation (His374Arg) in PHD2, also called EGLN1, was reported in a 43-year-old patient with erythrocytosis and recurrent para-aortic PGL.[69] PHD2 is a gene of five exons located on chromosome 1q42·1, encoding the prolyl hydroxylase domain protein 2 (PHD2). PHD proteins (PHD1, PHD2, PHD3) play a major role in regulating the hypoxia-inducible factor (HIF) that is involved in angiogenesis, erythropoiesis, cell metabolism, and proliferation. Germ-line mutations in PHD2 gene have previously been reported in patients with familial erythrocytosis but not in association with tumors.[70] Furthermore, a recent mutation analysis in 82 patients with features of inherited PCC, but no evidence of germ-line mutations in known susceptibility genes, detected no PHD1, PHD2, or PHD3 mutations. This suggests that mutations in PHD genes are not a frequent cause of inherited PCC.[71]

Genetic tests in patients with PCC and PGL

Molecular genetic medicine has revealed that a considerable number of patients with apparently sporadic PCC actually have a genetic predisposition to the development of these tumors, a lot more than was initially suggested. After decades of intensive research, several genes are now known to play an important role in the pathogenesis of PCC. At present, these are RET proto-oncogene, von Hippel–Lindau disease tumor suppressor gene (VHL), neurofibromatosis type 1 tumor suppressor gene (NF1), genes encoding the succinate dehydrogenase (SDH) complex subunits SDHB, SDHC, and SDHD, but also SDHA, the gene encoding the enzyme responsible for the flavination of SDHA (SDHAF2 or hSDH5), and the newly described TMEM127 and MAX tumor suppressor genes (Fig. 2).

Figure 2.

Frequency of susceptibility genes in patients with pheochromocytoma (PCC) or paraganglioma (PGL). Among the sporadic tumors, there is a considerable proportion (up to 10%) of patients with PCC/PGL that are good candidates to carry a germ-line mutation in a susceptibility gene but test negative on mutation screening.

Differences in the cost and availability of genetic tests throughout Europe and North America have raised the issue of optimum strategy for molecular investigation in PCC and PGL patients. To choose an appropriate genetic test, the biochemical profile of catecholamine secretion, age of the patient, localization of the primary tumor, and previous family history should be carefully evaluated and included in the genetic algorithm.[72] For example, epinephrine-secreting PCC suggests the presence of MEN 2 syndrome or NF1, while norepinephrine-secreting PCC indicates VHL. Elevation of dopamine together with norepinephrine has been seen in some SDHB-related PGLs. Bilateral PCC is usually associated with MEN 2 or VHL disease; mutations should therefore be sought initially in the RET and VHL genes and then in the SDHB and SDHD genes. In patients with catecholamine-secreting abdominal PGLs, mutations of SDHB, SDHD, and VHL should be screened, especially if the tumors are multifocal. Mutations in the SDHC gene have so far been reported only in parasympathetic PGLs, and testing for SDHC is therefore only indicated in these cases. If a patient has a skull-base or neck PGL, tests for mutations in the SDHD, SDHC, SDHAF2, and SDHB genes should be ordered sequentially. Genetic testing of the SDHB gene is also indicated in all patients with malignant PGL or PCC, as several studies revealed that in patients with malignant extra-adrenal pheochromocytoma, up to 50% have SDHB mutations (Fig. 3). Family history is often helpful in MEN 2, VHL, and NF1 tumors, but only 10% of the currently investigated patients with SDHB mutations have a positive family history of pheochromocytoma or PGL. For inherited forms of the disease, patients tend to be younger at presentation, especially for VHL disease, but the age range at presentation is quite wide, being 5–69 years for the mutation carriers and 4–81 years in the group of sporadic pheochromocytomas.[73] Genetic testing should therefore be preferred for patients with early onset disease, but not reserved exclusively for them. The diagnosis of NF1 is based on the clinical features and genetic testing is usually not required, as most people with NF1 will have enough signs of the condition by age 5 years for a specialist to diagnose them with confidence. Genetic testing for changes in the NF1 gene is not widely available and is currently expensive but can be helpful in some situations, such as screening for a pheochromocytoma-inducing mutation, when no other mutations in RET, VHL and SDH have been found. Most genetic screening algorithms also recommend that if a mutation is identified at any point in the testing algorithm, no further testing should be performed. With each new identified susceptibility gene, initial recommendations must be revised and rectifications of the current genetic algorithm must be undertaken. As there is still a considerable proportion (up to 10%) of patients with PCC/PGL that are good candidates to carry a germ-line mutation in a susceptibility gene but test negative on mutation screening, the discovery of additional genes is to be expected.

Figure 3.

Algorithm for genetic testing for genes associated with pheochromocytoma and paraganglioma. Genetic testing should be performed in patients with early onset disease (age < 45), but not reserved entirely for them. The diagnosis of NF1 is usually based on the clinical features, and genetic testing is required only in special cases. The newly identified mutations, such as those of SDHA, TMEM127, and MAX genes, are still considered to be extremely rare. They should therefore be analyzed only when no other gene defect has been identified.

Identification of a gene mutation in a patient with PCC or PGL may be crucial, especially when it leads to an early diagnosis and treatment, regular surveillance, and a better prognosis for patients and their perhaps not-yet-diagnosed relatives. Children can profit especially from early identification of RET mutations as prophylactic thyroidectomy has proved to be ideal in preventing the development of medullary thyroid carcinoma, the life-limiting manifestation of MEN 2 disease. Currently, genetic analysis also offers the possibility to define genotype–phenotype correlations and to adjust the timing of prophylactic surgery.[72] As, in some cases of aggressive mutations (i.e., RET mutations 883, 918, 922, 804–805, 804–806, 804–904), early therapy is recommended as soon as in the first year of one's life, genetic analysis and even biochemical and imaging screening in family members should not be postponed too long. There is however no current medical consensus regarding when, how, and how often biochemical screening and imaging should be performed in at-risk individuals depending on the underlying mutation. Benn et al. estimated that if lifelong screening was to begin at ten years of age, PCC/PGL would be detected in all SDHD mutation carriers and in 96% of persons with SDHB mutations.[74] The detection rate could however be improved by screening at an even younger age. Recommended are lifelong annual history and physical examination (including blood pressure measurements), annual biochemical testing (i.e., blood tests to assess plasma-fractionated metanephrines), and annual full-body MRI. The advantage of gaining further data by performing CT, 123MIBG scintigraphy, or FDG-PET must be weighed up against the cumulative exposure to radiation, especially in the young children and adults. Other manifestations such as medullary thyroid cancer and hyperparathyroidism in the case of RET mutations or renal cell carcinoma and GIST in the case of SDHB mutations should also be considered.

To better guide surveillance strategies and optimize therapy choices in an evidence-based way, further clinical studies with proper genotype–phenotype analysis of all the known syndromes which present with PCC or PGL are essential. With the help of next-generation sequencing methods, which will lead to more available and cost-effective molecular genetic examinations, a further step may be taken toward conducting the necessary clinical trials.

Ancillary