Update on the genetics of paragangliomas

Paragangliomas (PGL) of the adrenal (also known as pheochromocytomas) or extra-adrenal neural crest-derived cells are highly heritable tumors, usually driven by single pathogenic variants that occur mutually exclusively in genes involved in multiple cellular processes, including the response to hypoxia, MAPK/ERK signaling, and WNT signaling. The discovery of driver mutations has led to active clinical surveillance with outcome implications in familial PGL. The spectrum of mutations continues to grow and reveal unique mechanisms of tumorigenesis that inform tumor biology and provide the rationale for targeted therapy. Here we review recent progress in the genetics and molecular pathogenesis of PGLs and discuss new prospects for advancing research with new disease models and ongoing clinical trials presented at the recent International Symposium of Pheochromocytomas and Paragangliomas (ISP2022) held in October 2022 in Prague.


Introduction
The sixth International Symposium on Pheochromocytoma (ISP2022) took place in Prague, Czech Republic, from 19 to 22 October 2022, 5 (and not the usual 3) years after the previous edition held in Sydney, Australia, in 2017, due to delays caused by the COVID-19 pandemic. In the intervening years, two international meetings related to pheochromocytomas (PCC) and paragangliomas (PGL) took place, Multiple Endocrine Neoplasia Workshop 2019 (in Houston) and 2021 (virtual). In 2019, the session dedicated to the metastatic forms of PCC and PGL was reported in a special issue of Endocrine-Related Cancer (Dahia et al. 2020). Here we summarize some key insights on the genetics of these tumors published or presented over the past years, along with our predictions for future progress in this active field of research.

New WHO classification
The World Health Organization (WHO)/The International Agency for Research on Cancer (IARC) has historically defined PCC and PGL as distinct entities. The fifth series of the WHO classification of endocrine and neuroendocrine tumors published in 2022 describes clearly PCC as part of the PGL family of tumors by defining PCC as a neuroendocrine neoplasm that originates from chromaffin cells of the adrenal medulla and as an intraadrenal PGL (Mete et al. 2022). A sympathetic PGL is defined as a neuroendocrine neoplasm that develops from neural crest-derived progenitors in paraganglia associated with the prevertebral and paravertebral sympathetic chains, sympathetic nerve plexuses, and sympathetic nerve fibers. Parasympathetic PGLs are well differentiated non-epithelial neoplasms derived from paraganglion cells of the autonomic nervous system. In this review, we will adopt this new nomenclature.
In the absence of clear features that can predict metastatic behavior, all patients with PGL are currently considered to have a lifelong risk of metastases; therefore, conceptually all PGLs of any location are considered potentially metastatic. On the other hand, it is worth noting that the identification of metastases is especially complex in patients with germline predisposition syndromes, as multiple lesions may represent multifocal primary tumors rather than metastatic spread. The review also notes that even the identification of PGL tissue in unusual locations such as the lung or liver does not necessarily imply the diagnosis of metastases, as these may be primary paraganglial locations. The value of sustentacular cells and Ki67 labeling as prognostic features is also discussed in this new classification. A TNM staging system for PGLs (adrenal and extra-adrenal sympathetic PGLs), introduced in the eighth edition of the AJCC Cancer Staging Manual, has now been included in the WHO classification. From a molecular perspective, WHO experts report that germline mutations in SDHB confer the highest risk of metastasis and that mutations in ATRX, or SETD2, high total somatic mutation burden, MAML3 fusion genes, altered WNT pathway, and TERT activation have also been associated with increased metastatic risk (Mete et al. 2022).

The PGL genetic testing in routine practice changes the patients' outcome
In 2020, the working group on endocrine hypertension of the European Society of Hypertension clearly stated that genetic studies should be considered for any patient diagnosed with PGL because these tumors carry the highest known heritability rate of any human neoplasm and because genetic alterations currently explain almost 80% of all cases. Regarding the latter, approximately 40-50% of the cases are explained by germline mutations in one PGL susceptibility gene identified in patients affected by a hereditary form of the disease, and 40-60% of them by somatic mutations identified in the tumor(s) of patients with the sporadic disease (Lenders et al. 2020).
The impact of the knowledge of the genetic status at the time of the first PGL diagnosis was evaluated by Buffet and collaborators in a retrospective multicentric study (Buffet et al. 2019). The study compared the management and outcome of 221 patients diagnosed with PGL carrying mutations in SDHx or VHL who were informed of their positive genetic status either within the first year or more than 7 years after the initial diagnosis of PGL. Fewer patients were lost to follow-up in the group who received their genetic result in the first year after diagnosis compared to those who were tested later (9.6 vs 72%, respectively), and during follow-up, the former group developed smaller new PGLs with lower metastatic spread. In addition, patients of this group who developed metastases had a better 5-year survival rate than patients who did not undergo genetic testing at diagnosis (Buffet et al. 2019). That study showed for the first time the positive impact of early knowledge of genetic status, especially for those with SDHB variants, as this diagnosis triggers a specialized active surveillance screening program for the management of mutation carriers. This impact on the prognostic value of SDHBrelated PGL was illustrated in the MAPP-Prono study (169 patients with metastatic PGL) whereby the identification of an SDHB mutation lost its significance as a prognostic factor of worse overall survival in these closely monitored individuals, supporting the value of active surveillance in these patients (Hescot et al. 2019).
The impact of genetic screening was recognized, among other recommendations, as relevant for the management of metastatic and unresectable PGLs by a multidisciplinary panel convened by The North American Neuroendocrine Tumor Society (Fishbein et al. 2021).

New PPGL susceptibility genes
Over the years, new genes related to the susceptibility to develop PGL have been identified. In this regard, it is worth highlighting the power of massive sequencing platforms, which, together with the genomic characterization of these tumors, have made it possible to recognize the genetic drivers of an additional percentage of patients, whose families will now be able to benefit from appropriate genetic counseling and clinical surveillance. Endocrine-Related Cancer e220373 responsible for establishing DNA methylation patterns during embryonic development and gametogenesis in mammals (Bestor 2000). The implication of this epigenetic modifier in the development of PGL was first demonstrated by Remacha and colleagues (Remacha et al. 2018), who described a de novo germline mutation in DNMT3A in a 22-year-old woman diagnosed with nine head and neck (H&N) PGLs. Functional characterization of the variant detected in this patient, c.896A>T; p.Lys299Ile and detailed analysis of the genomic characteristics of the available PGLs supported an impact of this mutation in DNMT3A function. Overall, these experiments suggested a gain-of-function effect, which was consistent with the hypermethylated profile observed in the tumors with this variant. A second study confirmed the role of DNMT3A in the susceptibility to develop H&N PGL ( Table 1). The patient presented with multiple clinical features, a finding that suggested a heterogeneous phenotypic spectrum related to DNMT3A germline variants (Mellid et al. 2020).

Dihydrolipoamide S-succinyltransferase (DLST) gene
DLST encodes the E2 subunit of the mitochondrial α-ketoglutarate (αKG) dehydrogenase complex, which catalyzes the overall conversion of αKG to succinyl-CoA and CO 2 in the tricarboxylic acid (TCA) cycle. Mutations in DLST cause the PGL7 tumor predisposition syndrome (OMIM 618475) and have been found in 0.6-3% of PPGL patients. All DLSTrelated patients described so far were diagnosed with multiple tumors in the thoracoabdominal region, without mutations in other PPG-related genes (Remacha et al. 2019 (Table 1). PGLs harboring these DLST mutations display altered methylation and transcriptional profiles similar to those observed in EPAS1-mutated tumors, suggesting a connection between DLST functional abrogation and pseudohypoxia.

Succinate-CoA ligase GDP-forming subunit beta (SUCLG2) gene
Succinyl-CoA ligase is a TCA cycle enzyme which catalyzes the reversible conversion of succinyl CoA and adenosine diphosphate (ADP). It is composed of a heterodimer comprising a subunit encoded by SUCLG1 and an ATPforming encoded by SUCLA2 or a GTP-forming subunit encoded by SUCLG2. Among 352 patients with PGL, 1 frameshift and 7 missense variants were recently identified in the SUCLG2 gene (Hadrava Vanova et al. 2022), but this first report suffered limitations (no familial aggregation, two SUCLG2 variants classified as benign or likely-benign variants due to their frequency in gnomAD, incomplete functional studies and lack of evidence for loss of heterozygosity, etc), which were described in the editorial accompanying the paper (Ney & Stewart 2022). Thus, additional studies are still needed before considering SUCLG2 as a new PGL susceptibility gene (Table 1).

News from 'old' PGL susceptibility genes Kinesin family member 1B (KIF1B) gene
The KIF1B gene has been suspected to be a PGL susceptibility gene in few reports, but its involvement is still debated. Cardot-Bauters and collaborators published in 2008 a family carrying a KIF1B missense variant but without a second mutation in the other allele or loss of heterozygosity at the somatic level (Yeh et al. 2008). They recently extended their study because one brother of the proband, who did not carry the KIF1B variant, developed a bilateral PCC at 31 years. A MAX variant was identified in the germline DNA of that patient but also in all his relatives affected by PCC suggesting that the genetic susceptibility to PCC is linked to the MAX variant rather than to the KIF1B's one in this family . While evidence for the role of KIF1B in neural crest-related tumorigenesis and in neuroblastomas seems to be supported in independent studies (Fell et al. 2017), the Cardot-Bauters et al. report further strengthens the view that KIF1B is probably not a PGL susceptibility gene.

Endothelial PAS domain protein 1 (EPAS1) gene
Somatic mutations of the EPAS1 gene, encoding for the hypoxia-inducible factor 2α (HIF2α) transcription factor, are highly prevalent in PGLs. Notably, these mutations are detected in patients with congenital cyanotic heart disease at a higher than 10-fold rate compared with the frequency of EPAS1 mutations in sporadic PGL (~90% vs ~6-7%), suggesting that these mutations are under selective pressure in the specific clinical/environmental conditions experienced by these patients (Vaidya et al. 2018, Ogasawara et al. 2022. Of high translational relevance, the first description of a sustained therapeutic response was recently reported in a patient with polycythemia and multiple inoperable PGL caused by a mosaic germline EPAS1 variant who was treated with belzutifan, a selective small-molecule inhibitor of HIF2α (Kamihara et al. 2021). This experience should encourage molecular geneticists to investigate such variants in tumoral DNA and also in the germline DNA (in search for mosaicism) by deep sequencing or digital droplet PCR and should prompt the design of new trials to determine the efficacy of this drug for EPAS1-mutated states (Toledo et al. 2022).

Refined methodology to improve genetic variant identification
Over the past few years, technical advances in nextgeneration sequencing methods applied to germline and tumoral DNA associated with the decreased cost of sequencing have allowed the introduction of whole exome or genome sequencing in research and also in routine practice. Non-classical pathogenic variants and rare genetic variants became more easily accessible and were reported in patients for whom the causative germline or tumoral variant was still unknown (Ben Aim et al. 2022). For instance, the diagnosis of germline mosaicism in PGL susceptibility genes became possible in the same assay, by identification of a variant in a minority of reads in germline DNA and in the majority of reads in its matched tumoral DNA. Whereas a minor peak on electropherogram of Sanger sequencing can be misinterpreted as an artifact, droplet digital PCR allows to precisely quantify a low level of mosaicism in DNA extracted from leukocytes or other tissues. In this way, the first case of constitutional mosaicism of SDHB mutation was reported in a young patient with a norepinephrine-producing extra-adrenal PGL (Cardot-Bauters et al. 2019).
Intronic deep mutations were reported in PGL susceptibility genes. An international effort from the US and Europe brought new evidence for the pathogenicity of deep intronic variants in the VHL gene in a cohort reassembling 1167 patients with previous negative genetic testing. Six different genetic variants were discovered in a cryptic exon of VHL, named E1', which was previously identified in deep intronic sequence but not usually included in the target gene panels (Buffet et al. 2020). A study from Australia demonstrated the activation of an exonic splicing enhancer by a genetic variant located in the +74 position of the SDHC gene (De Sousa et al. 2020). Other deep intronic pathogenic variants would likely be more detected and involved in PGL pathogenesis in the upcoming years. Synonymous variants at non-canonical splice sites but which may nonetheless impair splice, such as in the case of the VHL gene, were demonstrated as being pathogenic in patients affected by a von Hippel Lindau disease or familial erythrocytosis (Lenglet et al. 2018, Flores et al. 2019, Liu et al. 2020, Buffet et al. 2020. Incidental genetic findings were also discovered due to these highly sensitive technologies. This was the case in two recently published case reports: the first described the simultaneous identification of germline mutations both in SDHB and in TP53 in a patient with metastatic PCC (Gniado et al. 2020), and the second reported germline mutations in the FLCN and SDHB genes in a patient with metastatic renal cell carcinoma (Boland et al. 2020). Noteworthy, these discoveries raise new challenges for the practice of genetic counseling and recommendations on the surveillance of patients and relatives.

Fusion genes
Recombinant fusions have only recently begun to be recognized and evaluated in PGL. Rare fusion genes, involving MAML3, BRAF, NGFR, and NF1, were first discovered by RNA sequencing in The Cancer Genome Atlas PPGL study (Fishbein et al. 2017). Fusions recurrently involving the MAML3 transcription factor, especially the UBTF::MAML3 fusions, were associated with a novel molecular cluster that had not been previously recognized in PGL oncogenesis. The UBTF::MAML3 fusion leads to the expression of Wnt targets and appears to be associated with an aggressive phenotype (Alzofon et al. 2021). UBTF::MAML3, but also EWSR1::CREM fusion gene, which was recently reported in a patient with a metastatic PGL should be further investigated as prognostic biomarkers (Javaid et al. 2023).
The RET gene is a well-established PGL susceptibility gene as part of Multiple Endocrine Neoplasia type 2A and type 2B syndromes, and less commonly as a somatically mutated oncogene, but only recently was it detected as part of recombinant fusion in PGLs (Mweempwa et al. 2021, Estrada-Zuniga et al. 2022. These fusions are distinct from conventional RET rearrangements detected in epithelial cancers such as lung and thyroid (Grieco et al. 1990, Kohno et al. 2012, Santoro et al. 2020 in the positioning of the fusion partners. In PGLs, RET is the 5′ partner of the fusion, while in epithelial tumors, it is invariably positioned as the downstream partner of the fusion. Despite this distinct recombination architecture, RET fusions in PGLs share other features of the more typical RET recombinant proteins: they lead to constitutive activation of RET and its downstream effectors, endow target cells with oncogenic phenotypes (Ou & Zhu 2020, Santoro et al. 2020, Estrada-Zuniga et al. 2022, and are responsive to highly selective, clinical grade RET inhibitors selpercatinib and pralsetinib (Subbiah et al. 2018, Wirth et al. 2020, Mweempwa et al. 2021, Thein et al. 2021, Estrada-Zuniga et al. 2022. These findings suggest that the identification of gene fusion involving the RET gene should open the way to treatment with RET inhibitors in patients with metastatic or inoperable PGL. Due to its prognostic and therapeutic value, the search for fusion genes in PGL at the tumor level by transcriptome or whole genome sequencing should be added to the genetic testing arsenal of PGLs that remain without a recognizable driver event, especially those belonging to the kinase cluster.

New tools for accurate variant's classification
Nowadays, the major challenge for the molecular geneticists in charge of PGL diagnosis remains to correctly classify molecular variants identified by NGS. In 2017, the NGS in PPGL (NGSnPPGL) Study Group (ENS@T/ PRESSOR) published standardized recommendations and initiated an international effort to collect, annotate, and classify variants in order to develop gene-centric curated database of PGLs (The NGS in PPGL (NGSnPPGL) Study Group et al. 2017). Following this initiative, 223 SDHB variants from 737 patients were collected worldwide and manually curated by a panel of experts from the NGSnPPGL Study Group who established a consensus classification. This curation reduced by half the variants initially classified as variants of unknown significance (Ben Aim et al. 2022). The SDHB variants classified by these experts are now freely available and publicly accessible via the Leiden Open Variation Database (LOVD) system (https://databases.lovd.nl/shared/genes/SDHB).
There have also been advances in our annotation of TMEM127 variants. Recent structure-functional studies revealed additional features of TMEM127, including a fourth transmembrane domain and an endocytic domain . These findings provided the basis for preliminary evaluation of a new classification of variant pathogenicity in 111 carriers and support the location of nonconserved missense mutations in transmembrane domains as a likely feature of pathogenic variants (Armaiz-Pena et al. 2021) Most of these variants have been deposited in LOVD (https://databases.lovd.nl/shared/ genes/TMEM127).

New insights for asymptomatic mutation carriers
An international panel of experts has established a consensus statement in following the Delphi method focused on the management of SDHx asymptomatic mutation carriers detected by familial genetic testing. An algorithm for screening and follow-up was proposed during adulthood and childhood. The experts proposed first screening at an earlier age (6-10 years old) for asymptomatic SDHB mutation carriers than for carriers of mutations in the other SDHx genes (10-15 years old) and recommended using magnetic resonance imaging as firstline imaging in children. If a SDHx mutation carrier never developed any tumor related to SDH deficiency, screening tests could be delayed to every 5 years after 70 years of age and follow-up could be stopped at 80 years of age (Amar et al. 2021). A retrospective multicentric study reported 249 asymptomatic SDHx mutation carriers who benefited from at least 1 imaging work-up. Imaging screening detected tumors in 20% of asymptomatic SDHx mutation carriers with a median age of 41 years old (11-86) (Saie et al. 2021). Similar proportions have been reported in other two independent multicenter studies (Greenberg et al. 2020, Davidoff et al. 2022, demonstrating that current SDHx screening protocols are effective at identifying SDHx-related tumors. The benefits of surveillance of asymptomatic mutation carriers starting in childhood have been reported by Vibert and collaborators in the context of the von Hippel-Lindau disease, for which genetic testing is recommended starting at 5 years of age. In a small series of 16 children diagnosed as VHL mutation carriers, follow-up examinations performed in a specialized expert network detected 11 tumors in 6 children but all had a favorable outcome (Vibert et al. 2022).

ISP2022 highlights in approaches and models for PGL research
A recent multi-institutional research effort led by Richard Tothill examined a group of 30 PGLs carrying mutations in various susceptibility genes at single nuclei transcriptome resolution (Zethoven et al. 2022). Different expression subgroups were identified corresponding to former recognized transcriptomic clusters driven by genotype: C1A (SDHx), C1B1 (VHL), C1B2 (EPAS1), C2A (Kinase), C2B1 (MAX), and C2B2 (MAML3). In addition, VHL, SDH, and EPAS1-mutated groups were enriched in stromal cells and tumor nuclei and showed enhanced hypoxia-related signaling. Intriguingly, tumors with MAML3 fusions, often linked to aggressive outcomes, showed a high expression of VEGFA and EPAS1 suggesting a HIF-pathway activation by still unknown mechanisms. Metastatic SDHx-related PGLs displayed an increase in proliferation markers and a reduced number of Schwann-cell-like cells. The orphan receptor GPR139 emerged as one of the overexpressed genes in metastatic tumors and should be further investigated as a potential target for treatment.
The scarcity of experimental models in PGLs has been an important unmet need in research. New and promising research models were presented at ISP2022, including two new Sdhb-deficient mouse strains (Armstrong et al. 2022, Gupta et al. 2022). Another emerging model of patient tumor-derived organoid cultures that may be amenable to drug screens may provide interesting information on the therapeutic profile of PGLs (Dahia, Soragni et al., unpublished observations). Follow-up work in these models is highly anticipated.

Future of clinical trials in PGL
Lastly, the next big frontier in PGL research will come from new therapies based on strong biological rationale and which are being tested in new clinical trials. The results of the first international randomized study in metastatic progressive PGL (FIRSTMAPPP, NCT01371201) investigating sunitinib (37.5 mg/day) or placebo were first presented during European Society of Medical Oncology Congress 2021 and at ISP2022. The median PFS was 8.9 months in the sunitinib arm vs 3.6 months in the placebo. This academic double-blind trial, in which 78 patients were enrolled, provided the highest level of evidence available thus far and established sunitinib as the first-line option for affected patients with progressive metastatic PGL (Baudin et al. 2021). On the heels of its approval by the FDA for the treatment of VHL-related tumors (Jonasch et al. 2021), an international, multi-institutional phase 2 clinical trial of belzutifan was launched for patients with locally advanced or metastatic PPGLs (NCT04924075). This trial recently completed its recruitment, and the PGL field expects with great anticipation the results of this and other trials grounded by preclinical research to guide the future of PGL treatment.

Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this review.