SDHAF2 (PGL2-SDH5) and Hereditary Head and Neck Paraganglioma

This article is featured in Highlights of This Issue, p. 201

Paragangliomas arise from the paraganglia, small clusters of neuroendocrine tissue found throughout the body in the vascular and neuronal adventitia. Parasympathetic paraganglia of the head and neck generally lack endocrine activity, in contrast to the sympathetic paraganglia in the abdomen and thorax (1). Head and neck paraganglioma (HNPGL) are named according to their site of origin (2). The majority of these paraganglioma occur in the carotid body (78%), as (difficult to differentiate) tumors of the tympanic body and the jugular body (17.5%) and as neoplasms of the nodose ganglion associated with the vagus nerve (4.5%; ref. 3). Other localizations of HNPGL incidentally reported in literature are those in the larynx, the nose, the paranasal sinuses, and the orbit.

HNPGL is often present within families. Heredity in HNPGL was first recognized by Chase (4) and is currently identified in 9.5% to 50% of all cases; the remainder is defined as isolated cases or sporadic tumors (3, 5, 6). Germline mutations of genes encoding succinate dehydrogenase (SDH), SDHD (PGL1), SDHC (PGL3), and SDHB (PGL4) are a frequent cause of paraganglioma (7–9). We recently identified germline mutations in a previously uncharacterized gene at the PGL2 locus. This gene, named SDHAF2 (succinate dehydrogenase complex assembly factor 2), encodes an evolutionarily highly conserved flavin-adenine dinucleotide factor (10). SDHAF2 is involved in the flavination of SDHA. For a long period after the discovery of SDHD, no mutations of SDHA were reported in paraganglioma patients, despite the fact that SDHA is the largest gene and protein of the SDH complex and the major catalytic subunit. Very recently, the first SDHA gene mutation was identified. The patient had a catecholamine-secreting, extra-adrenal paraganglioma and no family history of disease (11).

The loss of SDHAF2 results in the loss of SDH function and a reduction in the stability of the SDH complex, leading to reduced levels of all subunits (10). We found a missense mutation of SDHAF2, c.232G>A (p.Gly78Arg), in a conserved region of the gene in our family and demonstrated that it is present in all affected members of this family and in several unaffected individuals who inherited the inactivated gene via maternal transmission.

The SDHAF2 mutation, similar to mutations in SDHD, shows a striking parent-of-origin inheritance of tumor susceptibility, with onset of tumor development only upon inheritance via the paternal line (6, 12). Both of these genes are located on chromosome 11. This phenotypic expression of the mutation only upon paternal inheritance suggests imprinting of the gene. In contrast, paraganglioma associated with mutation in the SDHB and SDHC genes (both located on chromosome 1) do not show this parent-of-origin inheritance. Van der Mey et al attempted to explain this phenomenon by invoking gene activation/silencing during female oogenesis and male spermatogenesis. This hypothesis became untenable once it was recognized that PGL1 encoded an essential gene involved in energy metabolism and was expressed by both parental alleles in all tissues tested (8). More recently, an alternative hypothesis has been proposed that takes note of the fact that the major imprinted locus of the human genome is also located on chromosome 11, at p15.5, and has been shown to contain genes with a tumor suppressor function (13, 14). This hypothesis, postulated by Hensen, takes into account the frequent loss of the entire maternal copy of chromosome 11 in paragangliomas and proposes a “three-hit” model in which the loss of the maternal chromosome results in the deletion of the remaining copy of SDHD and a maternally expressed gene at Chr11p15.5, in addition to the germline mutation present on the paternal chromosome (13)

In general, HNPGL exhibit a slow rate of growth, and most often present asymptomatically as a space occupying mass noted clinically or radiographically. Sporadic and hereditary HNPGL show a variety of differences in clinical manifestations. Hereditary HNPGL presents at a younger age and tumors at multiple sites are more commonly observed, compared with sporadic. Malignancy, present in approximately 10% of all paraganglioma (15, 16), is also more common in patients with a hereditary syndrome, particularly those with SDHB mutations (up to 54% of cases; ref 16). Treatment options for HNPGL include surgical removal, radiotherapy, and a wait-and-scan policy. Choosing a treatment modality is dependent on the localization and growth of tumor, surrounding tissues, age, condition, symptoms experienced, and the preferences of the patient.

A large SDHAF2 (PGL2-SDH5) family showing frequent occurrence of HNPGL was collected by the Department of Otorhinolaryngology at the Radboud University Nijmegen Medical Centre over a 30-year period and was originally described in 1981 by van Baars et al. (3). Here, we present a phenotypic and genetic description of a large novel branch of this unique paraganglioma family (3, 17), presently the only large SDHAF2 family known. We present the results of linkage analysis, mutation analysis, and the observed inheritance characteristics. Phenotypic analysis includes the description of observed tumor localization(s), tumor size(s), multifocality, malignancy, gender distribution, age at onset of symptoms/detection, and presenting symptoms. In addition, we place these findings in the context of recent literature of other SDH-associated paraganglioma genes and the previous SDHAF2 (PGL2) studies.

A hereditary paraganglioma syndrome was suspected in a family that was known with paraganglioma at the Radboud University Nijmegen Medical Centre. Linkage analysis was performed as described (18) that indicated linkage to the PGL2 locus on Chr11q13.1, previously detected in another Dutch PGL2 family. Further genetic testing confirmed the suspicion of a relationship between these 2 families. Additional at-risk family members were identified using telephone questionnaires and invited to participate in a clinical genetic study. Patients treated elsewhere were also invited to participate and their clinical data were gathered. Linkage analysis was used to identify mutation carriers, and all carriers underwent a thorough clinical screening including otolaryngological examination and magnetic resonance imaging (MRI) of the head and neck region. MRI of thorax/abdomen and catecholamine secretion screening have not been performed. Treatment options, including surgery, radiotherapy, or a wait-and-scan policy were advised, dependent on size and localization of detected tumor(s). Upon the identification of the SDHAF2 mutation, c.232G>A, p.Gly78Arg, all family members were analyzed by sequencing for the presence of the mutation. The following primers were used for the amplification of exon 2 of the C11orf79/SDHAF2 gene: (SDH5–2F: 5′-GTTGACCTTCCCAGGCTC-3′ and SDH5–2R: 5′-GAGGTTCAGCTGCTTTTCTG-3′). Thirty nanogram of genomic DNA from each patient was amplified, and primer annealing was performed at 60°C. PCR fragments were purified using the Nucleofast 96-Well kit (Macherey-Nagel). Sequencing was performed using standard protocols. Sequences were analyzed using the Mutation Surveyor software package (Softgenetics).

In 3 generations, 72 living family members were traced. Linkage analysis was carried out in 57 members willing to participate. Of the cases in which reliable conclusions could be drawn, 31 did not carry the disease halotype and 23 did (13 male and 10 female) including 7 risk-free carriers showing inheritance via maternal transmission. Our experience with this family indicated an “imprinting” phenomenon with no observed cases of maternal transmission (17). These family members would not develop paraganglioma and hence are called risk-free carriers. We examined 4 of these cases by MRI scanning of the head and neck region and confirmed the absence of tumors. The remaining 16 family members were at risk for developing paraganglioma due to paternal inheritance. MRI investigation was offered to all, and 13 individuals were investigated, showing tumors in 11. Five of these individuals were already patients under treatment. In total, 12 of the 16 investigated at-risk family members (paternal transmission) were considered to be affected (75%). All clinical data are summarized in Table 1.

The pedigree of the novel family branch is shown in Figure 1. The observed inheritance pattern is autosomal dominant, with the characteristics of maternal genomic imprinting. This genealogic tree also depicts the relation with the family as previously described families (3). The relationship was confirmed genetically. On pedigree examination, the families appeared to be linked to the same ancestor dating back to the year 1771.

In total, 24 tumors were detected in 11 of the 13 radiographically investigated individuals at-risk for developing HNPGL. The most commonly detected HNPGL was the carotid body tumor (CBT), of which 17 were detected in a total of 10 patients, representing 71% of all tumors. Vagal paraganglioma accounted for 17% (n = 4) of all tumors and jugulotympanic paraganglia for the remaining 12% (n = 3). Tumors at multiple sites (multifocality) were detected in all affected patients, with the exception of patient 12 (91%). HNPGL at other less common sites in the head and neck region were not detected.

The average age at onset of symptoms/detection was 33 years (range: 22–47 years). A symptomatic presentation was reported in 58% of all proven affected individuals. Their symptoms at presentation are summarized in Table 2. The average age at detection in the asymptomatic group was at age 29 years (range: 22–35 years; n = 5). In the group presenting with symptoms, the average age of presentation/onset was 38 years (n = 4; range: 32–47; in 2, the age was unknown).

Of the 24 detected tumors, 17 were treated surgically, 1 using radiotherapy and the remaining 6 were treated followed (imaging) closely but surgically untreated. Surgical excision of 11 of these tumors was carried out at the Radboud University Nijmegen Medical Centre (1× Vagal tumor, VT; 1× Jugulotympanicum tumor, JTT; 9× CBT), the 6 remaining were operated elsewhere and will not be subject of this article due to lack of information. No major technical problems were encountered during surgery. Malignancy was not detected. Postoperative complications were encountered with 4 of the 11 operated tumors (36%). In 3 patients, postoperative hypertension with related symptoms was reported after unilateral CBT resection. The complaints resolved after several months and spontaneously normotension was eventually achieved in 2 patients and after drug treatment in 1.

Vagal nerve damage was reported in 1 patient after removal of a carotid tumor. Speech therapy improved the voice within 1 year after resection and laryngoscopy showed recurrence of mobility of the right vocal cord.

SDHAF2 mutation analysis confirmed the findings of the previously performed linkage analysis. No discrepancies were detected. Of the 3 patients in whom linkage was inconclusive, 2 were tested negative. The third patient at risk for developing paraganglioma due to possible paternal inheritance, tested positive on PGL2 mutation analysis and will be further investigated and treated when necessary

All the genotypic and phenotypic characteristics of this PGL2 family are summarized in Tables 3 and 4.

A large related Dutch PLG2 family was originally described in 1981, and a pattern of autosomal dominant inheritance and generation skipping was noted (19). In 1989, a genomic imprinting hypothesis was proposed to explain the striking lack of disease expression upon maternal inheritance in SDHD (PGL1) families (6) and we later demonstrated that the PLG2 families show the same inheritance pattern (17, 18). In no case did a child of a female carrier develop disease, with all affected patients inheriting the trait exclusively from their father.

The overall gender distribution of HNPGL patients shows a slight female predominance (female:male, 60% vs. 40%; ref. 20). A PLG2 family was reported with obvious male predominance (female:male, 33% vs. 77%), but only included clinically affected individuals (3). In the novel family branch described here, there is an equal gender contribution between all individuals at-risk for developing HNPGL (n = 16). If we exclude the at-risk individuals without tumors (carriers, at risk), the male:female ratio is calculated to be also an equal distribution (n = 12). Struycken et al. reported that the sex ratio among affected offspring appears to be influenced by the paternal or maternal origin of the gene of the transmitting father (17). In our study, no such conclusion could be drawn.

A symptomatic presentation is reported in only 58% of all affected individuals. The symptoms are summarized in Table 2. The average age at detection of tumors in the asymptomatic group is 29 years (range: 22–35), which is below the average age at onset of symptoms/detection of the whole investigated family (33 years). In the group presenting with symptoms, the average age at presentation is 38 years (n = 4, range: 32–47). The growth rate of paragangliomas is low, with a median increase of 1 mm per year and a doubling time of 4.2 years (21). This is reflected in the time lag of 9 years to the development of clinical symptoms in this family.

Patients with hereditary paraganglioma tend to develop the disease at a younger age than those with sporadic paraganglioma. In a PGL2 family, an average age at onset of 37 years has been described (range 16 to 80 years; ref. 22). The average age at onset of symptoms/detection in our family branch is 33 years (n = 9). Only 2 genetically at-risk individuals, aged 24 and 39, were negative for tumors on MRI scanning. The term age at onset of symptoms/detection is used, because 42% of the tumors are detected in asymptomatic individuals after screening. Clearly, active genetic screening and advanced high-resolution imaging have led to the earlier detection of otherwise occult or symptomless tumors, consequently decreasing the age at onset of symptoms/detection. We anticipate that with advances in imaging techniques and wider genetic screening for the SDHAF2 mutation, the age at onset of symptoms/detection will further decrease.

In the present family branch, CBTs are most common representing 71% of all tumors. This is in accordance with the literature (78%; ref. 3). JTTs were detected in 12% (n = 3) and are relatively underrepresented compared with values reported previously (17.5%; ref. 3). In contrast, VTs are overrepresented at 17% (n = 4) compared with Van earlier studies (4.5%). Only a larger study of this pedigree would allow definite conclusions to be drawn on the relationship of SDHAF2 and tumor incidence. As expected, tumors at other less common sites in the head and neck region were not detected because they are very rare. Screening for thoracic and abdominal paraganglioma and catecholamines secretion has not been performed, but is now part of the routine clinical care.

Another interesting aspect of hereditary paraganglioma is the tendency to be multifocal. Only 10% of sporadic paragangliomas are multifocal, compared with incidences of 70% to 87% in hereditary cases (23). van Baars et al. reported multifocality of tumors in only 50% in the first described PGL2 family (3, 19), in contrast to the present family branch, with 10 of the 11 proven affected subjects showing tumors at multiple sites (91%). This discrepancy may be a result of improved imaging techniques leading to detection of otherwise occult or symptomless tumors and hence higher multifocality of tumors. We anticipate that multifocality increases during follow-up. These patients will be followed clinically and radiographically, which may verify this assumption later on.

Of the patients treated operatively, postoperative complications, including postoperative hypertension, were reported in 4 of the 11 (36%) tumors (all CBT) operated at the Radboud University Nijmegen Medical Centre that resolved without sequelae. Hypertension is a known complication bilateral carotid body resection and is attributed to a failure of the baroreflex. In the long term, the blood pressure may remain slightly elevated with markedly increased variability (24). The rates of postoperative morbidity and serious sequelae reported in literature are in general higher than in our series (25). In our opinion, the low rates in our series justify the “wait-and-scan” policy often used at our medical centre, leading to less operation of extensive tumors that would result in a higher complication and mortality percentages. Paraganglioma of the head and neck region are generally asymptomatic and slow-growing benign tumors. With more sophisticated and precise imaging techniques at hand, the growth of these tumors and the association with surrounding tissues could be better assessed, giving us the opportunity to postpone or even avoid an operation. Prior to possible surgical intervention, a variety of issues should be taken into account including tumor growth, tumor localization, surrounding tissues, possible symptoms, age, condition, and the wishes of the patient. Follow-up on the natural course of tumor development in these patients is ongoing to test this policy and to be published in due time.

After establishing the genetic and clinical aspects of this novel branch of the SDHAF2 family, a comparison with the other hereditary head and neck paraganglioma syndromes could be made, as presented in Table 5.

This comparison shows a young age at onset and early full penetrance. Mutifocality levels are high compared with other HNPGL.

We established the SDHAF2 mutation status of PGL2 family members. Phenotypic characterization of this family confirms the association of SDHAF2 mutations with HNPGL. This family shows a young age at onset and a high level of multifocality. A high percentage of patients were asymptomatic at time of detection.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.