Pheochromocytoma (PCC) and paraganglioma (PGL) are tumors originating from chromaffin cells in the adrenal medulla and extra-adrenal paraganglia of the sympathetic and parasympathetic nervous chains, respectively [1]. These neuroendocrine tumors are collectively referred to as pheochromocytoma and paraganglioma (PPGL). PCCs and sympathetic PGLs can synthesize and secrete catecholamines, mainly adrenaline and noradrenaline, and even dopamine, as they originate from neuroendocrine cells derived from the embryonic neural crest, while parasympathetic PGLs rarely produce catecholamines [2,3].

PPGLs are rare tumors with a incidence of 0.95 per 100,000 person-years in the United States. They account for 0.02% to 0.5% of patients with hypertension [4]. Classically, PCC was called “10% tumor,” as this tumor had a 10% rate of malignancy, bilateral occurrence, pediatric-onset, being asymptomatic, familial inheritance, and having an extra-adrenal location [5,6]. Regarding the pathogenesis, 90% of PCCs were considered sporadic, with only 10% being hereditary with germline mutations [7]. However, this rule was disproved as techniques for genetic testing evolved, and the prevalence of PPGL with germline mutations was reported to be approximately 30% to 40% [8–10].

A traditional familial cause of PPGL was noted in syndromes such as multiple endocrine neoplasia (MEN) type IIA and IIB, neurofibromatosis (NF) type 1, and von Hippel-Lindau (VHL) disease [5]. However, with the advancement of genetic analysis techniques, more susceptibility genes for hereditary PPGL syndromes have been discovered. Mutations in subunits of the succinate dehydrogenase (SDH) complex, a mitochondrial enzyme for oxidative phosphorylation [11], is now a well-known cause of hereditary PPGL [12,13]. An astonishing feature of hereditary PPGLs is the associated high degree of genetic heterogeneity. Some of the susceptibility or driver genes, mutations known to cause hereditary PPGLs [9], are involved in the hypoxic response during cellular metabolism. Others are involved in kinase signaling pathways [14]. Due to the complexity of analyzing the genetics of PPGLs, there has been an attempt to approach this issue by classifying the genes into molecular clusters: pseudohypoxia, Wnt signaling, and kinase signaling group [15].

The high prevalence of PPGLs is known to be caused by a certain germline mutation of susceptibility genes, making them the most heritable tumor in humans [9]. Therefore, genetic testing is recommended in all patients regardless of family history [16]. Genetic testing has become more accessible with the utilization of next-generation sequencing (NGS), an accurate, fast, and inexpensive way of genetic testing and has shifted the paradigm of genetic research. The NGS in PPGL (NGSnPPGL) study group suggested that targeted NGS is the most preferred method for genetic testing of PPGL patients [17]. Targeted NGS analyzes only the genes in the panel previously reported to have an association with tumorigenesis of PPGLs.

Regarding the development of valuable NGS technology, more susceptibility genes for hereditary PPGLs are being discovered, but the relative frequencies of the mutations found to vary between studies and geographic locations [17–19]. Moreover, since PPGL is a very rare type of tumor, large-scale studies on clinical relationships are limited. The genotype-phenotype relationship has not been thoroughly examined, especially in Korea. Regarding the clinical features of these tumors, a study by Kim et al. [20] on 119 patients with PCCs from 1986 to 2009 is so far the largest in Korea; however, it only included PCCs and not PGLs. Further, this study lacked information on the genetic aspect of this issue. Another study by Kim et al. [21] demonstrated a correlation between germline mutations and the clinical characteristics of PPGLs by examining 53 patients with non-syndromic PPGLs. In this study, seven patients (13.2%) had germline mutations: rearranged during transfection (RET; n=2), VHL (n=1), SDHB (n=2), and SDHD (n=2). The prevalence rate of germline mutations is much lower than the recently reported prevalence worldwide. This is due the fact that the information on germline mutations in this study was limited to only four genes screened by direct sequencing. Other publications on this topic are limited to case reports. Since NGS has been applied to the clinical hospital setting only recently in Korea, there is no report examining germline mutations of susceptibility genes related to PPGLs in Korea.

We aimed to investigate the prevalence of gene mutations found in PPGL patients in the Korean population and to examine its associated clinical characteristics using genetic analyses, including targeted NGS. To this end, we reviewed the clinical and genetic features of patients with PPGLs in Severance Hospital.

Of the 78 Korean patients with PPGLs, we exclusively evaluated 57 patients and examined the clinical and genetic features of the disease. Among those 57 patients who underwent either targeted NGS or Sanger sequencing, 28 patients were found to have pathogenic germline mutations of PPGL-related genes. Considering the fact that targeted NGS was implemented recently as a routine modality, prevalence rate of germline mutation is 32.6% (14 out of 43). There was a higher likelihood of family history of PPGL in patients with germline mutations than those with no mutations. There were significant differences in the type of PPGL, percentage of family history, metastasis rate, the presence of other tumors, and biochemical profile between three molecular clusters. Nineteen different germline mutations from seven genes (SDHB, RET, VHL, NF1, MAX, SDHA, and SDHD) were found, most of which were missense mutations. Seven patients had VOUS mutations of PPGL-related genes (SDHB, EPAS1, KIF1B, and SDHC). EPAS1, KIF1B, MAX, SDHA, SDHC, and TMEM127 mutations were the first to be reported among Korean patients with PPGL.

In this study, 70.2% of patients were diagnosed with PCCs, while 29.8% were diagnosed with PGLs, which was in line with the previously reported percentage of PCC (69% to 89%) [18,22,31]. The mean age of onset in this study was 42.8 years, and 54.4% of the patients were women. This information is concordant with that reported in a previous study (the peak age of onset, 40 to 50 years). Furthermore, PPGLs were reported to have an approximately equal distribution among men and women [31]. Family history was found only in nine patients (15.5%), all of whom were found with germline mutations. However, most of the patients with germline mutations did not have family histories. A previous study also found that a high incidence of PCCs positive for mutations occurred without a family history [32]. They suggested a potential explanation for this phenomenon: a high rate of spontaneous mutations, low penetrance of the gene mutations, or maternal imprinting.

Malignant PPGL, defined as the presence of distant metastasis, is reported to have a prevalence of approximately 10% [16, 22,24]. However, in our data, the metastatic rate was as high as 26.3% because our center is a tertiary referral hospital. Frequent sites of metastases were the bone, lymph nodes, liver, and lung, which were concordant with a previous study [33]. It is known that the risk of metastasis is higher in extra-adrenal location, with a higher level of plasma methoxytyramine (metabolites of dopamine), and in SDHB mutations [34–36]. Our study only showed a significant difference in metastatic rate depending on SDHB mutations. Among 15 patients with metastatic PPGL, six patients had SDHB mutations, three patients had VHL mutations, one patient had SDHA mutation, and the other five patients were sporadic.

A previous systematic review had shown that the incidence of recurrence, including local and systemic recurrence, and appearance of new tumors, is 0.95 per 100 person-years, which is 4.7% of occurrence in 5 years [22]. One-fifth of the patients were found to have the recurrent disease in this study, similar to the previous report [20]. However, some studies have reported that the presence of syndromic disease, larger tumor size, and PGL relative to PCC being related to an increased risk of recurrence [22]; our study found no association between them. Owing to varying results between studies and the absence of precise features that predict recurrence, all patients should undergo follow-up, including clinical history-taking, imaging, and biochemical evaluation.

With the development of imaging techniques, the rate of incidental diagnosis of PCC dramatically increased to nearly 30% [37]. In Korea, this rate was even higher, with 36.1% reported in a previous study [20]. CT was the frequently used diagnostic tool because the cost is lower than that in other countries due to the National Health Insurance System. In our study, PPGLs were discovered incidentally in 41.2% of the cases, showing that the rate of incidental diagnosis is steadily increasing. This phenomenon could be attributed to the growing accessibility of imaging procedures, including abdominal CT and sonography. Meanwhile, in 54.9% of the cases, the tumor was discovered because of the symptoms and signs. Similar to that in a previous study, hypertension was the most common clinical sign, whereas palpitation and headache were the most common symptoms [7]. About half had either paroxysmal or sustained hypertension, and three-fourth of the patients had symptoms and signs associated with PPGLs. Among functional PPGLs, the noradrenergic type was the most common. There was no significant difference in the biochemical profile between PPGLs with and without germline mutations.

To our knowledge, the current result on the frequency of germline mutations in Korean PPGL patients, 32.6%, is relatively higher than the previously reported prevalence in Korean so far (13.2%). However, this is concordant with the worldwide prevalence of germline mutations in PPGL known to be around 30% to 40% [8–10]. Similar to previous studies, SDHB mutation was the most common [17,38]. VOUS mutations were excluded from significant germline mutations due to lack of data to support their pathogenicity. Indeed, VOUS mutations found in our study were from SDHB, EPAS1, SDHC, and TMEM127 genes, and except for SDHB, other genes are known to be extremely rare in frequencies [17,38]. Nevertheless, as VOUS indicates “missing pathogenicity,” the possibility of reclassification to other ACMG classes exists with the additional accumulation of clinical data [39].

We also investigated the clinical characteristics based on the molecular clusters. Genes were classified into three molecular clusters, pseudohypoxic TCA cycle-related (known to be 10% to 15% of PPGLs), pseudohypoxic VHL/EPAS1-related (15% to 20%), and the kinase signaling (50% to 60%) group. In our study, there were significant differences in the type of PPGLs, percentage of family history, metastasis rate, the presence of other tumors, and biochemical profile. While PPGLs related to the TCA cycle can occur in any location, PPGLs related to VHL/EPAS1 are not likely to be located on the head and neck areas. Kinase signaling-related PPGLs are most likely to be PCCs with an uncommon incidence of parasympathetic PGLs, and sympathetic PGLs are not likely to be present [15]. These findings are concordant with our data that the pseudohypoxic TCA cycle-related group included all of PCCs, sympathetic and parasympathetic PGLs, but the pseudohypoxic VHL/EPAS1-related group only included PCCs and parasympathetic PGLs. Most of the patients in the kinase signaling group had PCCs except one with parasympathetic PGL. Our result showed that the kinase signaling group had 50% of family history but none in the pseudohypoxic TCA cycle-related group. The discrepancies between the groups must be attributed to the rate of spontaneous mutation and penetrance. A previous review stated that the risk of metastasis is highest in the pseudohypoxic TCA cycle-related group and lowest in the kinase signaling group, and the result was largely the same in our data. Pseudohypoxic TCA cycle-related PPGLs are reported to have the lowest paraganglia cell differentiation with a decreased expression of catecholamine-synthesizing enzymes leading to depressed catecholamine levels. Pseudohypoxic VHL/EPAS1-related PPGLs have an intermediate level of catecholamine production, and kinase-signaling PPGLs, being the most differentiated, are likely to be most functional [15]. Our study results are in line with those of a previous study, which reported that pseudohypoxic PPGLs are purely noradrenergic and never adrenergic because they always lack phenylethanolamine N-methyltransferase, an enzyme that catalyzes the synthesis of epinephrine from norepinephrine [15].

There are several limitations to this study. First, our study population has a small sample size in a single center, so our results might have restricted value in terms of generalization. However, considering the rare prevalence of PPGL, an analysis of 57 patients of clinical and genetic characteristics in Korea is noteworthy. Second, this study has a potential for biases. We included PPGL patients managed by a single endocrinologist. Although vast majority of PPGL patients in our hospital were treated by this physician, few patients might have been missed by being registered to other endocrinologists or endocrine surgeons. Also, we excluded two patients with no mutations and one patient with silent mutation analyzed by only Sanger sequencing but included 17 patients in whom germline mutations were found. This exclusion was inevitable due to the characteristic of Sanger sequencing, by which we can only seek for mutation at a gene of interest, but this could have created selection bias. To minimize the bias, the frequency of germline mutation was calculated among the patients who underwent genetic analyses after the targeted NGS technique was adopted as a routine test. Consecutive use of targeted NGS by recommending it routinely to all the PPGL patients will be necessary for future studies. Difficult interpretation of VOUS resulted from targeted NGS was one of the greatest problems in our study. We approached this issue in rather conservative manner, as we considered VOUS as negative. This might have led to underestimation of the frequency as we acknowledge the possibility that some VOUS mutations might be reclassified to pathogenic or likely pathogenic over time with additional reports of data. Meanwhile, patients who had no mutations of susceptibility genes by targeted NGS cannot be considered as patients with absolutely negative results, as we only analyzed the known susceptibility genes of PPGLs. For those patients with negative results with NGS, we might recommend whole-genome sequencing [17]. Whole-exome sequencing of these patients might uncover new susceptibility genes that are still unknown, but the validation of the causative relationship between the mutation and the PPGL will be a difficult process. Moreover, epigenetic changes were not taken into consideration in this study. Previous studies have shown that epigenetic gene silencing of tumor suppressor genes due to promotor methylation is present and considered pathogenic in some patients with PPGLs [40,41]. This finding suggests that the percentage of PPGL patients with genetic background might be higher than the value presented.

In conclusion, our data present the frequency of germline mutations in Korean patients with PPGL. The expected prevalence of germline mutations of susceptibility genes in PPGL patients is estimated to be approximately 32.6%. With the utilization of NGS, genetic testing has become more accessible, simultaneously providing us with abundant information. Further research into the molecular characterization of PPGL can lead us to pursue and take a step toward the concept of precision medicine.