Metastatic Pheochromocytomas and Paragangliomas- Proceedings of the MEN2019 Symposium

Abstract

Overview:

Pheochromocytomas and paragangliomas (PPGLs) are adrenal or extra-adrenal autonomous nervous system-derived tumors. Most PPGLs are benign, but approximately 15% progress with metastases (mPPGLs). mPPGLs are more likely to occur in patients with large pheochromocytomas, sympathetic paragangliomas and norepinephrine-secreting tumors. Older subjects, those with larger tumors and synchronous metastases progress more rapidly. Germline mutations of SDHB, FH, and possibly SLC25A11, or somatic MAML3 disruptions relate to a higher risk for metastatic disease. However, it is unclear whether these mutations predict outcome.

Meeting format:

Two two-hour multidisciplinary workshops on mPPGLs held in the setting of the 16th World MEN Conference

Identified gaps in knowledge:

Once diagnosed, there are no well-established predictors of outcome in mPPGLs, and some tumors progress rapidly with few therapeutic options and limited response. High-specific activity (HSA) metaiodine-benzyl-guanidine (MIBG) is the first FDA approved treatment and show clinical effectiveness for MIBG-avid mPPGLs.

Proposed next steps:

Ongoing and future investigations should involve validation of emerging candidate outcome biomarkers, including somatic ATRX, TERT and microRNA disruptions, and identification of novel prognostic indicators. Long-term effect of HAS-MIBG and the role of other radiopharmaceuticals should be investigated. Novel trials targeting molecular events prevalent in SDHB/FH mutant tumors, such as activated hypoxia inducible factor 2 (HIF2), angiogenesis or other mitochondrial defects that might confer unique vulnerability to these tumors should be designed and executed. As therapeutic options are anticipated to expand, multi-institutional collaborations and well-defined clinical and molecular endpoints will be critical to achieve higher success rates in improving care for patients with mPPGLs.

1. Introduction

The Multiple Endocrine Neoplasia 2019 (MEN 2019) symposium- the 16^th international workshop focused on new concepts and treatment of malignant manifestations of cancers that comprise hereditary endocrine cancer syndromes. This meeting took place between March 26–28, 2019 in Houston, Texas.

The meeting addressed the state of the field, discussed important outstanding research questions and challenges, in the form of symposium presentations and two-day workshops. Workshop attendees weighed in on key priority areas for future research and consensus points were collated. The next sections summarize presentations and workshop discussions on metastatic pheochromocytomas and paragangliomas (mPPGLs) and conclude with proposed plans for the coming years and research projects for bridging these gaps.

2. Diagnosis of ‘malignant’ pheochromocytoma and paraganglioma: state of the science.

Under the revised WHO classification (2018), pheochromocytomas and paragangliomas (PPGLs) are now referred to as ‘metastatic’ or ‘non-metastatic’ rather than ‘malignant’ or ‘benign’; metastases are still defined by deposits at sites where normal chromaffin tissue is not present(Lam, 2017). Approximately 10% of pheochromocytomas and ~40% of sympathetic paragangliomas will be associated with metastases(Ayala-Ramirez et al., 2011). Head and neck paragangliomas are more rarely metastatic (McCrary et al., 2019). Even though family history may be negative, a hereditary basis is present in 40–50% of mPPGLs. Synchronous metastases are present in 35–50% cases; metachronous metastases can develop within months or after more than a decade(Baudin et al., 2014, Roman-Gonzalez et al., 2017a). Median survival of mPPGL is ~6 years, although the rate of progression is highly variable. Death is usually from metastatic progression, although hypersecretion of catecholamines is often morbid and sometimes fatal.

Five large clinical series (Amar et al., 2007, Ayala-Ramirez et al., 2011, Turkova et al., 2016, Hamidi et al., 2017, Hescot et al., 2019, Fishbein et al., 2017a) of mPPGL are shown in Table 1. Despite the heterogeneous nature of mPPGL, group statistics for these series are strikingly similar.

Table 1.

Clinical series of metastatic pheochromocytomas and paragangliomas

	Amar et al (Amar et al., 2007)	Ayala-Ramirez et al (Ayala-Ramirez et al., 2011)	Turkova et al (Turkova et al., 2016)	Hamidi et al (Hamidi et al., 2017)	Hescot et al (Hescot et al., 2019)	Fishbein et al (Fishbein et al., 2017)
N	54	131	132	272	169	71
Timeframe	n/r	1960–2009	2000–2014	1960–2016	1998–2010	200–2016
PC/sPGL/HNPGL (%)	54/46/0	52/48/0	29/71/0	40/47/11	53/37/10	28/45/211
Synchronous metastasis	24 (44%)	67 (51%)	26 (20%)	96 (35%)	79 (47%)2	18 (25%)3
Age (Dx primary), y	37.9	n/r		39	n/r	41
Age (Dx metastases), y	42	n/r		44	48	6.2 (0–44.77)4
SDHB	23/54 (43%)	9/215 (43%)	73/132 (55%)	42/272 (15%)5	63/1515 (42%)	37/605 (62%)
Died of disease	26/54 (48%)	87/131 (66%)	39/132 (30%)	73/272 (27%)	72/169 (43%)	19/71 (27%)
Age at death (range), y	45	n/r		54 (7–91)	n/r	n/r

¹an additional 5.6% had more than one tumor location;

²metastases within 1 year;

³metastases within 3 months;

⁴years (range) after primary diagnosis;

⁵not all cases had genetic testing; Dx primary: at first diagnosis, Dx metastases: at the time of metastases identification, n/r: not reported

Risk factors for mPPGL include clinical, pathological and genetic factors. The strongest clinical risk factors are tumor size and thoraco-abdominal (sympathetic) PGL. Tumor size is incorporated in the newly published AJCC Cancer Staging system for PPGL(Amin et al., 2017): T1 tumors are <5cm in greatest dimension, T2 tumors are ≥ 5 cm or sympathetic PGLs of any size, and T3 tumors are of any size with invasion into surrounding tissues. Metastasis is rare for pheochromocytomas below 4 cm; conversely, metastatic paraganglioma can occur from smaller primaries(Ayala-Ramirez et al., 2011). A retrospective study by Hamidi et al (Hamidi et al., 2017) examined factors associated with rapid progression (death within 5y of initial presentation) and found the strongest predictors to be older age at primary diagnosis (OR 1.054 [95% confidence interval 1.01–1.08]/y), primary tumor size (OR 1.12 [1.009–1.25]/cm) and the presence of synchronous metastases (OR 10.24 [3.76–31.18]). Another clinical risk factor for the rate of progression is the site of metastatic disease: skeletal only metastases tend to be more indolent (median survival 12y), compared with non-osseous metastases (median survival 7.5y)(Ayala-Ramirez et al., 2013). Biochemically, mPPGL is less likely in epinephrine-secreting tumors(Ayala-Ramirez et al., 2013); conversely, elevated plasma methoxytyramine, a byproduct of dopamine, is associated with mPPGLs(Eisenhofer et al., 2012). The ASES (age, size, extra-adrenal, secretory) score attempts to bring these clinical risk factors together, by giving one point each for age ≤35 y, tumor size ≥6 cm, extra-adrenal location and norepinephrine secretion(Cho et al., 2018). As the ASES score does not include genetic data, the younger age likely acts as a surrogate for genetic risk. An ASES score of ≥2 had a sensitivity of 61%, specificity of 80%, negative predictive value of 96.5% but a positive predictive value of only 18.4% for mPPGL. Ten-year survival rate was 30% for ASES ≥2 compared with 86% for ASES <2(Cho et al., 2018).

PET imaging has revolutionized the detection of mPPGL: ⁶⁸Ga-DOTATATE has the highest sensitivity in most cases, in particular for SDHB-associated disease (98.6% overall detection rate)(Janssen et al., 2015, Janssen et al., 2016a, Janssen et al., 2016b, Janssen et al., 2017, Taieb et al., 2018). Unfortunately, PET imaging has not been shown to predict the tempo of mPPGL, thus prognostic and disease modeling are not achievable with this imaging modality alone. Diagnosis of PPGL is usually considered an indication for surgery. An important albeit retrospective study by Roman-Gonzalez et al showed that surgical resection of the primary lesion was associated with improved survival in synchronous mPPGL (Roman-Gonzalez et al., 2017b).

There is no consensus on the utility of tumor grading systems for predicting mPPGL; both the PASS and GAPP systems are limited by inter-observer variability(Baudin et al., 2014). Absent immunostaining for SDHB is strongly associated with germline SDHB pathogenic variants(Papathomas et al., 2015). Tumoral succinate, measured by LC/MS-MS, is higher in PPGLs associated with metastases compared with non-metastatic cases, likely due to the presence of SDHB mutations in the former(Richter et al., 2014).

Genetic risk factors for mPPGL include germline pathogenic variants in SDHB, SDHA, SDHD, FH, MAX and SLC25A11 (Buffet et al., 2018, Castro-Vega et al., 2014) and somatic MAML3 fusions(Fishbein et al., 2017b). The strongest genetic risk factor for mPPGL is a pathogenic SDHB variant, present in 40–50% cases (Amar et al., 2007, Ayala-Ramirez et al., 2011, Turkova et al., 2016, Hamidi et al., 2017, Hescot et al., 2019); nevertheless, accumulating evidence suggests that SDHB mutations are not associated per se with rapid progression of disease(Crona et al., 2019, Hamidi et al., 2017, Hescot et al., 2019). Recently, aberrant telomere maintenance mechanisms (TMMs) have been associated with mPPGL: somatic ATRX mutations associated with alternate lengthening of telomeres were first reported in mPPGL by Fishbein et al(Fishbein et al., 2015); high TERT expression was then reported by Liu et al in many mPPGLs(Liu et al., 2014); Dwight et al reported structural variants in TERT associated with telomerase re-expression(Dwight et al., 2018); and most recently Job et al (Job et al., 2018)have shown that aberrant TMM is associated with worse prognosis, and that a somatic hot-spot TERT promoter mutation (C228T) is specifically associated with poor outcomes in SDHB-associated disease, as discussed later in this article.

All patients with mPPGL should be referred for genetic testing after appropriate counseling; multi-gene panels facilitate comprehensive evaluation of 12 well-defined hereditary loci for PPGL(Group et al., 2017), as well as additional genes recently reported as related to the disease (Buffet et al., 2018, Remacha et al., 2017, Remacha et al., 2018, Remacha et al., 2019). Discovery of a germline variant has implications for the patient and their first-degree relatives. A recent retrospective study found that delayed diagnosis of an underlying germline variant in PPGL was associated with increased recurrence risk and reduced survival (Buffet et al., 2019).

In summary, mPPGL is more common in pheochromocytomas > 5cm or in sympathetic PGLs, associated with norepinephrine and/or dopamine secretion, and pathogenic germline SDHB, FH, or possibly also SLC25A11 variants. Rapid progression of metastatic disease is more likely in older subjects, with larger tumors and synchronous metastases. The presence of somatic mutations in ATRX or TERT is also associated with worse prognosis. Conversely, pheochromocytomas <4 cm secreting epinephrine, or head and neck PGLs, are associated with a low likelihood of metastatic disease (Table 2).

Table 2.

Factors associated with metastatic risk in PPGLs

High risk	Low risk
Tumor >5cm	Tumor <5cm
Norepinephrine secretion	Epinephrine secretion
Paraganglioma*	Pheochromocytoma
Older age (sporadic)	Younger age (sporadic and certain genetic groups)
SDHB, FH, SLC25A11 germline mutation	VHL, RET mutation
ATRX, TERT, MAML3 somatic disruptions	-

^*excluding head and neck paragangliomas, which have intrinsically low risk of metastasis

3. Advances in Molecular Aspects of PPGLs

a. Enabling Replicative Immortality in PPGLs

At the beginning of the third millennium, pioneer retrospective studies carried out on the large PPGL collection of the French COMETE network, revealed that SDHB mutation is a high risk factor of malignancy and of poor prognosis (Gimenez-Roqueplo et al., 2003, Amar et al., 2007). Independent studies, utilizing different genomics technologies, including multi-omics integrative studies, classified PPGL into two main different clusters driven mainly by germline or somatic mutation in a PPGL susceptibility gene (Dahia et al., 2005, Eisenhofer et al., 2004, Lopez-Jimenez et al., 2010, Castro-Vega et al., 2015), one containing genes enriched for hypoxia-related response (named ‘Cluster 1), and another expressing predominantly tyrosine kinase signaling genes (Cluster 2). Further analysis revealed that within the hypoxia-related cluster, the subcluster C1A contains PPGL with higher metastatic potential, i.e. tumors related to germline mutations in a gene encoding for a tricarboxylic acid cycle (also known as the Krebs cycle) protein such as SDHx (SDHA,-B,-C,-D,-F2), FH, MDH2, SLC25A11, GOT2, etc(Castro-Vega et al., 2014, Cascon et al., 2015, Remacha et al., 2017, Buffet et al., 2018). The subcluster C1A presents global DNA hypermethylation, transcriptional signatures of reactivation of epithelial to mesenchymal transition and of activation of angiogenesis/hypoxia signaling and overexpression of miRNA cluster 182/96/183, miR-210, miR-483. A second group, subcluster 1B (germline or somatic VHL mutations) tumors have intermediate methylation levels, also share a hypoxic-like transcriptional profile, and display an overexpression of miR-210. Cluster 2 tumours (NF1-, RET-, TMEM127-, MAX-, HRAS-, MET- or FGFR1-related tumors) are characterized by global hypomethylation, activation of RAS/MAPK signaling and down-regulation of DLK-MEG3 miRNA cluster(Castro-Vega et al., 2014). More recently, an additional molecular group (MAML3 gene fusions, CSDE1 somatic mutations) was described by The Cancer Genome Atlas program that exhibits activation of the Wnt signaling pathway(Fishbein et al., 2017b).

More than ten years after the introduction of the PPGL genetic testing in routine practice, it was demonstrated that knowledge of the genetic status in the first year after PPGL diagnosis has improved the patients’ outcome, even for patients with an SDHB mutation(Buffet et al., 2019). This finding provides strong support for widespread genetic testing in all patients with PPGL and not just patients with high-risk tumors.

Mechanistically, there has been important progress in understanding aberrant pathways in PPGL. Letouzé et al demonstrated that the inactivation of succinate dehydrogenase in SDHx-mutated PPGL led to marked accumulation of succinate, and shed light on its role as an oncometabolite(Letouze et al., 2013). Excess succinate is able to inhibit multiple 2-oxoglutarate-dependent dioxygenases, as prolyl-hydoxylases or DNA/histones demethylases(Xu et al., 2011), and to promote angiogenesis and global DNA hypermethylation that contribute to tumorigenesis(Letouze et al., 2013). Nevertheless, only 50% of SDHx-related PPGLs go on develop a metastatic phenotype suggesting that additional molecular(s) mechanism(s) promote malignant properties. Furthermore, patients without SDHB mutations can also develop metastatic disease. Job et al recently addressed the question of the relative contribution of immortalization mechanisms to metastatic progression in PPGL and, for that purpose, carried out a comprehensive analysis of two immortalization mechanisms (telomerase reactivation and alternative lengthening of telomeres) in the large well-annotated series of PPGL collected by the French COMETE network, which had previously undergone integrative genomic analyses(Job et al., 2018). They found that 70% of mPPGLs become immortalized, including every metastatic case classified into cluster C1A. Molecularly, these tumors presented either transcriptional activation of TERT (via a TERT promoter mutation, promoter hypermethylation or copy number variation at the TERT locus) or somatic ATRX mutation. Importantly, telomerase activation and ATRX mutation are independent risk factors for malignancy strongly associated with both metastasis and overall survival. These two indicators appear to more accurately discriminate metastatic from non-mPPGL compared to an SDHB mutation and are able to predict metastatic behavior irrespective of the SDHB status(Job et al., 2018). These new PPGL biomarkers are promising candidates for prognostication of SDHB-related PPGL, for improving risk stratification and for tailoring patients’ monitoring.

b. Epigenetic characteristics of SDHB-related paragangliomas

If we consider cancers in general, 5–10% of cases are classified as “hereditary” due to inherited genetic pathogenic variants (mutations), which can be transmitted to the next generation(Rahman, 2014). There is an additional 10 to 15% of patients that show familial cancer aggregation, which may be due to the combined effect of genes and other shared factors, such as environment and lifestyle. However, the vast majority of cancers are sporadic and associated with non-inherited gene mutations. However, when we consider PPGLs, this scenario markedly changes. Up to 40% of cases are hereditary and related to at least 12 well-recognized susceptibility PPGL genes. An additional 30% of PPGLs have somatic mutations, and the remainder are included in a heterogeneous group without a clear identifiable driver event (Favier et al., 2015, Dahia, 2017).

Due to the large number of genes responsible for PPGL, patients are required to undergo comprehensive genetic screening not only to assist with adequate genetic counseling but also as an attempt to estimate the risk of metastatic disease. In this regard, the association between higher metastatic risk and the presence of mutations affecting the SDHB gene has been well-established (Hescot et al., 2019), as described in the preceding sections. Over the past few years, more genes associated with increased risk of metastatic disease have also been reported(Cascon et al., 2019). These additional genes have received little attention because their mutations are rare and individually account for only a small percentage of patients. However, if one considers this set of genes in aggregate, a common theme can be recognized: they belong to the energy metabolism, in particular the tricarboxylic acid cycle. The use of high-throughput platforms in PPGLs has made it possible to identify genomic characteristics associated with clinical variables, grade of differentiation or with specific mutations, allowing classification of these tumors beyond the driver mutation(Dahia, 2017, Fishbein et al., 2017a). One of the genomic features shared by PPGLs with mutations in Krebs cycle genes is a global hypermethylation phenotype, also known as CIMP (CpG island methylator phenotype)(Letouze et al., 2013). The mechanism underlying this epigenetic phenotype relates to an effect of the actual driver genetic mutation. Defective Krebs cycle enzymes lead to the accumulation of their corresponding substrate (for instance, succinate or fumarate in mutant SDH or FH, respectively), which in turn inhibits the activity of enzymes dependent on 2-oxoglutarate (also known as alpha-ketoglutarate). Succinate and fumarate are structurally similar to 2-oxoglutarate, and thus abundance of either substrate outcompetes the latter, leading to reduced activity of 2-oxoglutarate dependent enzymes. This broad group of enzymes includes the TET family of DNA modifying enzymes and the JmjC domain-containing histone lysine demethylases (KDMs), leading to DNA and histone hypermethylation seen in these tumors. The resulting epigenetic alterations lead to cell differentiation arrest and promote malignant transformation(Frezza et al., 2011, Xu et al., 2011).

As SDHB is part of mitochondrial complex II, along with SDHA, SDHC and SDHD, it is difficult to explain why the former is uniquely associated with a higher risk of metastasis. In this regard, it has been suggested that the mean methylation levels across all CpG sites is higher in SDHB-mutated PPGLs than in other SDH mutant PPGLs(Yang et al., 2013). It is possible that loss of SDHB function results in complete inactivation of the SDH complex, whereas enzyme activity through mutations in other subunits may not lead to full loss of SDH function. This might explain higher succinate accumulation and stronger inhibition of demethylation in SDHB mutant tumors, relative to mutations in the other SDH subunits (Yang et al., 2013). Additional studies are required to fully demonstrate this concept and its attendant consequences on clinical behavior of PPGLs.

Having demonstrated the usefulness of high-throughput platforms to define genomic characteristics in PPGLs (Dahia et al., 2005, de Cubas et al., 2015, Castro-Vega et al., 2014, Fishbein et al., 2017b)a potential application of these findings is to define therapeutic targets, or identify drug resistance mechanisms. In this regard, it has been recently reported that tumors with SDHB mutations have increased activity of mitochondrial complex I of the electron transport chain (Pang et al., 2018). This augmented complex I activity could lead to higher NAD+ (nicotinamide adenine dinucleotide) availability, and therefore, to a more efficient DNA repair process(Tateishi et al., 2015). It has been proposed that targeting poly(ADP-ribose) polymerase (PARP), a highly conserved enzyme involved in DNA break repair and stabilization of DNA replication, could potentiate the therapeutic effect of genotoxic agents, such as temozolomide (Pang et al., 2018). These preliminary findings need further validation.

4. Translational Progress and Treatment Opportunities

a. Pseudohypoxia as a Therapeutic Target: Lessons from PPGL Clusters

As discussed in the previous sections, there are currently no biomarkers that can reliably and prospectively distinguish metastatic from non-metastatic PPGLs. As explained in above, the establishment of ‘malignancy’ in PPGL requires the documentation of metastasis in non-paraganglial tissue, which by definition constitutes either an advanced(late) or a retrospective diagnosis, two undesirable attributes in contemporary oncology(Baudin et al., 2014). The realization that the natural evolution of PPGLs with metastatic potential is poorly known has practical implications beyond academic knowledge, as it impedes therapeutic progress for these tumors. Current standard therapies for mPPGL result at best in partial responses, do not stratify patients based on their molecular or risk group, and trials adopting emerging therapies are still in early investigational stages (Baudin et al., 2014, Roman-Gonzalez et al., 2017a, Pryma et al., 2019). Thus, identification and testing of effective drugs to treat mPPGL remain an unmet clinical need.

As discussed above, about 40–50% of the patients with mPPGL carry a germline mutation in the succinate dehydrogenase subunit B (SDHB) gene(Amar et al., 2007, Ayala-Ramirez et al., 2011, Astuti et al., 2001). These mutations inhibit degradation of hypoxia inducible factor (HIF) transcription factor, ultimately leading to its stabilization and constitutive activation, a phenomenon known as pseudohypoxia(Dahia et al., 2005, Fishbein et al., 2017b). As described above, mPPGLs associated with other mutations, including prominently other components of the energy cycle, often share this pseudohypoxic profile(Favier et al., 2015, Cascon et al., 2019). HIF target genes are involved in angiogenesis, cell proliferation, metastases, metabolic reprogramming, etc(Kaelin et al., 2008). Thus, pseudohypoxia and HIF represent early (or truncal) and critically important events for PPGL initiation, and possibly also for tumor maintenance, rendering them relevant targets for mPPGL treatment(Toledo, 2017).

Transcriptionally active HIFs are heterodimers that comprise an inducible, short-lived alpha subunit and a constitutive beta-subunit (ARNT). In particular, multiple lines of evidence suggest that the HIF2α subunit is the most biologically relevant of the HIF subunits for oncogenesis(Yan et al., 2007) in general, and, for chromaffin cell and sympathetic nervous system development and transformation in particular(Tian et al., 1998, Comino-Mendez et al., 2013, Toledo et al., 2013).

Transcription factors have historically been considered ‘undruggable’ targets(Kaelin, 2018). However, the identification of a large protein cavity in the HIF2α PAS-B domain opened the path to development of inhibitors of this subunit (Scheuermann et al., 2013, Scheuermann et al., 2009). Exploiting this structural feature, clinical-grade, potent small molecules (PT2385 and PT2977) that efficiently and specifically prevent HIF2α/ARNT dimerization, thus blocking HIF2 transactivation, were identified and developed. These HIF2a- antagonists showed tumor inhibition capacity in clear cell renal cell carcinomas in vitro and in vivo(Busch et al., 2016, Cho et al., 2016). Like mPPGLs, clear cell renal carcinomas are pseudohypoxic(TCGA, 2013). HIF2α inhibitors have since been tested in advanced renal carcinomas. A phase 1 clinical trial with PT2385 in heavily pre-treated patients with renal carcinoma showed an overall response rate (ORR) of 66%, with favorable safety and tolerability profile(Courtney et al., 2018). Recently, a Phase 1/2 dose-escalation trial with PT2977 (more potent and with superior pharmacokinetics than PT2385) was performed in patients with advanced solid tumors. The Phase 2 portion included 52 patients with advanced, previously treated renal carcinoma. Interim results of this trial showed 22% patients with a confirmed partial response. Based on these results, a Phase 3 trial for renal carcinoma is scheduled to start in the Fall of 2019. Importantly, one patient from the UTHSCSA cohort with metastatic paraganglioma, carrier of an SDHB mutation, enrolled in the Phase 1 portion of the PT2977 trial(Papadopoulos et al., 2018). This patient, heavily pre-treated, remained stable with sustained reduction of plasma normetanephrines and good tolerability to PT2977 for 30 weeks, suggesting that HIF2 inhibition may have a place in the therapeutic arsenal for mPPGL. A proposal for a pilot trial with this drug is currently undergoing evaluation.

b. Advances in Therapeutics: High Specific Activity MIBG, Tyrosine Kinase Inhibitors, Immunotherapy and other treatments

Metastases happen in approximately 15–20% of patients with PPGs(Ayala-Ramirez et al., 2011). Tumors more commonly spread to the lymph nodes, skeletal tissue, lungs, and liver(Ayala-Ramirez et al., 2013). A metastatic phenotype is more commonly observed in tumors associated with pseudo-hypoxia(Dahia, 2014). As mentioned before, 40–50% of patients with mPPGLs carry germline mutations of the SDHB gene. In addition, many other hereditary and apparently sporadic mPPGLs share a similar molecular profile with the SDHB mutant tumors(Dahia et al., 2005). However, survival curves indicate that the outcomes of patients with mPPGLs are quite heterogeneous. Some patients may have very aggressive tumors with poor clinical outcomes and die shortly after diagnosis, while others have very slow growing or ‘static’ tumors that may not even require systemic intervention. Most patients have tumors that exhibit progression over time and therefore, they need systemic therapy(Jimenez et al., 2013).

Chemotherapy, mainly with a combination of cyclophosphamide, vincristine, and dacarbazine (CVD), was the first therapy introduced for patients with mPPGLs. Approximately, 30–40% of these patients respond to chemotherapy (Asai et al., 2017, Roman-Gonzalez et al., 2017a). Responses mainly include disease stabilization and tumor size reduction with improvement of symptoms of catecholamine excess and a reduction of the risk for tumor burden-related complications. In contrast, complete responses are exceptional (Niemeijer et al., 2014). It is difficult to predict which patients will benefit from chemotherapy and responsive patients may remain on treatment for a long period. Nevertheless, chemotherapy may cause severe toxicity (i.e. bone marrow suppression, neuropathy); toxicity becomes more obvious over time and limits its long-term use. Chemotherapy is, however, the only treatment available worldwide. Interestingly, SDHB mutation carriers were noted to display better response to CVD than non-carriers (Fishbein et al., 2017a). Taken together with other observations of improved outcome after sunitinib (Ayala-Ramirez et al., 2012) or temozolomide (Hadoux et al., 2014) in SDHB mutation carriers, these studies may support the notion that an SDHB mutation, although increasing the risk of metastatic disease, may in fact be associated with better outcome and greater therapeutic response. Additional, prospective studies evaluating the progression-free survival and overall survival of SDHB-mutation carriers are warranted.

In the 1980s, radiopharmaceutical medications became a therapeutic option to consider as treatments for patients with mPPGLs. The first medication was meta-iodine-benzyl-guanidine (MIBG) labeled with Iodine-131, prepared through a simple isotope exchange methodology. I-131-MIBG has been offered since then to patients with mPPGLs that express the noradrenaline transporter (MIBG-avid tumors). Approximately, 60% of mPPGLs are MIBG-avid(Ilias et al., 2008, Jimenez et al., 2019). Close to one third of patients treated I-131-MIBG exhibit tumor size reduction and stabilization with improvement of symptoms of catecholamine excess(Jimenez et al., 2019). Similar to chemotherapy, only rare patients achieve a complete response(Jimenez et al., 2019). Over the last two decades, the manufacturing of I-131-MIBG has been optimized through a selective resin (Ultratrace)(Vallabhajosula et al., 2011). This resin prevents unlabeled MIBG from being carried from the production reaction to the final solution. Thus, the specific radioactivity of Ultratrace I-131-MIBG is much higher than then simple isotope exchange I-131-MIBG (92.5 vs. 1.59 MBq/kg), implying a much higher delivery of radioactivity to the tumor per dose(Jimenez et al., 2019). The results of a phase 2 clinical trial with Ultratrace I-131-MIBG showed that more than 90% of patients achieved tumor stabilization and reduction one year after the first infusion; these patients had improvement of symptoms of catecholamine excess(Pryma et al., 2019). Ultratrace I-131-MIBG did not cause cardiovascular toxicity. Less than a quarter of patients had severe bone marrow insufficiency that required supportive treatment (i.e. platelet and red blood cell transfusion, granulocyte colony stimulating factors). All patients recovered bone marrow function and no patients required bone marrow transplant(Pryma et al., 2019). The United States Food and Drug Administration (FDA) approved Ultratrace I-131-MIBG for the treatment of patients with MIBG-avid mPPGLs in 2018. Ultratrace I-131-MIBG is currently, the only FDA approved therapy in the US.

mPPGLs frequently express somatostatin receptors in the cell membranes; in fact, Gallium-68-DOTATATE scintigraphy is the most sensitivity study to anatomically characterize these tumors(Janssen et al., 2016a). Therefore, somatostatin analogues labeled with radioactivity are attractive medications to study in prospective clinical trials. Lutetium-177 labeled DOTATATE is FDA-approved for the treatment of gastro-enteropancreatic neuroendocrine tumors. Retrospective studies and meta analyses suggest that this radionuclide may be an effective option in mPPGLs (Satapathy et al., 2019, Taieb et al., 2019).This drug is currently evaluated in recently activated clinical trials for patients with mPPGLs (NCT04106843 and NCT03206060).

Angiogenesis is an important hallmark for mPPGL development. Several tyrosine kinase inhibitors that block the vascular endothelial growth factor receptors and other receptors involved in angiogenesis are currently under evaluation in clinical trials. These drugs include axitinib, cabozantinib, lenvatinib, pazopanib, and sunitinib. These drugs can lead to tumor stabilization and size reduction with improvement of symptoms of catecholamine excess(Jimenez, 2018). Nevertheless, the positive tumor responses reported by the clinical trials with axitinib, pazopanib, and sunitinib were blunted by toxicity associated with their increasing doses(Jasim et al., 2017, O’Kane et al., 2019). The preliminary results of a phase 2 clinical trial with cabozantinib show a high objective response rate with acceptable toxicity. Different from the trials with axitinib and pazopanib, the trial with cabozantinib was designed to allow for dose adjustment based on patient’s tolerability(Jimenez, 2018). Cabozantinib inhibits the c-met pathway implicated in the development of metastases and tumor resistance and is arguably the most potent antiangiogenic drug currently available in clinical practice(Salgia et al., 2019). Therefore, cabozantinib may become the most effective tyrosine kinase inhibitor to treat patients with mPPGLs. Evaluation of current trials is ongoing.

The recognition of the hallmarks involved in the origin of mPPGLs is leading to the discovery of medications that benefit these patients. In addition to radiopharmaceuticals and tyrosine kinase inhibitors, immunotherapy is currently evaluated in clinical trials(Jimenez, 2018). None of these therapies is expected to be curative. However, the results of these clinical trials are helping to identify pathogenic complementary pathways. The success rate of systemic therapy for patients with mPPGLs is expected to increase in the years to come, as clinical trials will likely explore combinations of therapies.

5. Expanding opportunities for diagnosis, surveillance and therapy of mPPGLs: a global perspective and international collaborations

There has been enormous progress in our understanding of the genetic basis and biological stratification of PPGLs. Nevertheless, several important gaps remain in this disease. The symposium participants and, in particular, the well-attended (see list of participants in Supplementary Information) workshop sessions, identified the most critical limitations currently perceived in the field of mPPGLs and attempted to define research areas that should be undertaken to help narrow these gaps (Table 3).

Table 3.

Current gaps recognizing in research and clinical practice of metastatic PPGLs

Risk assessment models

Guidelines for care of metastatic PPGL

Surveillance programs for mutation carriers

Access to treatment opportunities

Genetic testing: not widely available

Opportunities for communication and discussion of care/surveillance of specific cases

Goals for the future

At the two-day workshop on metastatic paragangliomas, extensive and productive discussions reached consensus on several relevant action items to be developed over the coming five years. They are summarized below:

• The group recognizes the need for a large, publicly accessible database of pheochromocytoma/paraganglioma as a critical resource to better understand the natural history of the disease, to help refine and improve risk assessment models.

• A means to achieve this level of information will likely be operationalized by the creation of multi-institutional, international networks with well-defined data entry. This will facilitate the exchange of uniform data points.

• Existing infrastructure such as the European ENS@T (http://www.ensat.org/) for the study of adrenal tumors is a successful model that can be leveraged for new networks. The A5 Alliance (www.a5adrenalalliance.com) has been created in 2015 utilizing a model similar to the ENS@T, and can spearhead network studies outside of Europe.

• An important point of discussion was the widely perceived need for developing guidelines specific for management of metastatic pheochromocytomas and paragangliomas.

• Additional proposals involved more specific studies for long-term surveillance to investigate the role of SDHB in patient outcome. Such studies would require that genetic testing be performed routinely, a practice which has not yet been adopted outside of US, parts of Europe, and Australia; the widespread standardization of SDHB immunohistochemistry as a first-level screen for patients carrying an SDHB mutation was discussed as a potential short-term measure to circumvent genetic testing limitations.

• The need for development of multi-institutional infrastructure for the design and implementation of pragmatic trials was recognized; for example, the efficacy and potential genotype impact on clinical response of existing cytotoxic systemic therapies (e.g CVD, temozolomide) has not yet been evaluated prospectively in large cohorts.

• It has been acknowledged that the design of clinical trials of mPPGLs should take into account the unique challenges posed by rare diseases/cancers, and should involve innovative design and operational approaches, which allow for the participation of multiple sites across national barriers.

• In times of limited funding, research on rare diseases is particularly vulnerable. Future multi-institutional research should develop innovative funding models to carry out studies that can have wide impact in the field.

Due to space constraints, the authors acknowledge that there are other areas relevant to mPPGLs that have not been included in this article. For example, studies to validate novel genetic risk mutations, including MAML3 fusions, CSDE1 mutations, and SLC25A1 variants; to investigate the role of co-existing genetic events, e.g. germline SDHB associated with somatic ATRX or TERT variants, or to characterize the contribution of co-existing genetic and epigenetic alterations. These areas represent emerging fields revealed by multiple ‘omics’ studies, and their exploration may hold promise for future risk assessment and development of novel therapeutic opportunities.