Summary. (Received 20 August 2012; returned for revision 6 September 2012; finally revised 10 October 2012; accepted 10 October 2012)

Transcription

1 Clinical Endocrinology (2013) 78, doi: /cen ORIGINAL ARTICLE Evaluation of SDHB, SDHD and VHL gene susceptibility testing in the assessment of individuals with non-syndromic phaeochromocytoma, paraganglioma and head and neck paraganglioma Mariam Jafri 1,2, James Whitworth 1,3, Eleanor Rattenberry 1,3, Lindsey Vialard 3, Gail Kilby 1,3, Ajith V. Kumar 4, Louise Izatt 5, Fiona Lalloo 6, Paul Brennan 7, Jackie Cook 8, Patrick J. Morrison 9, Natalie Canham 10, Ruth Armstrong 11, Carole Brewer 12, Susan Tomkins 13, Alan Donaldson 13, Julian Barwell 14, Trevor R. Cole 3, A. Brew Atkinson 15, Simon Aylwin 16, Steve G. Ball 17, Umasuthan Srirangalingam 18, Shern L. Chew 18, Dafydd Gareth R Evans 6, Shirley V. Hodgson 19, Richard Irving 20, Emma Woodward 1,3, Fiona Macdonald 3 and Eamonn R. Maher 1,3 1 Centre for Rare Diseases and Personalised Medicine, School of Clinical and Experimental Medicine, College of Medical and Dental Sciences, University of Birmingham, 2 Department of Oncology, University Hospital Birmingham Foundation Trust, 3 West Midlands Regional Genetics Service, Birmingham Women s Hospital NHS Trust, Birmingham, 4 Department of Clinical Genetics, Great Ormond Street Foundation Trust, Great Ormond Street Hospital, 5 Department of Clinical Genetics, Guy s and St Thomas Foundation Trust, Guy s Hospital, London, 6 Department of Genetic Medicine, Central Manchester Hospitals NHS Foundation Trust, Manchester Academic Health Sciences Centre, Manchester, 7 Northern Genetics Service, Teesside Genetics Unit, James Cook University Hospital, Middlesbrough, 8 Department of Clinical Genetics, Sheffield Children s Foundation Trust, Western Bank, Sheffield, 9 Department of Medical Genetics, Queen s University Belfast, Belfast HSC Trust, Belfast, 10 North West Thames Regional Genetics Service, Northwick Park Hospital, Middlesex, 11 Department of Clinical Genetics, Addenbrooke s Hospital, Cambridge University Hospital Foundation Trust, Cambridge, 12 Pennisula Genetics Service, Heavitree Hospital, Exeter, 13 Department of Clinical Genetics, St Michael s Hospital Bristol, Bristol, 14 Department of Cancer Genetics, University of Leicester, Leicester, 15 Department of Endocrinology, Royal Victoria Hospital Belfast, Belfast, 16 Department of Endocrinology, King s College London, London, 17 Department of Cardiology, University of Leeds, Leeds, 18 Department of Endocrinology, St Bartholomew s Hospital, London, 19 Department of Clinical Genetics, St George s Hospital, University of London, London and 20 Department of Otolaryngology, University Hospital Birmingham Foundation Trust, Birmingham, UK Summary Objectives Research studies have reported that about a third of individuals with phaeochromocytoma/paraganglioma (PPGL) have an inherited predisposition, although the frequency of specific mutations can vary between populations. We evaluated VHL, SDHB and SDHD mutation testing in cohorts of patients with non-syndromic PPGL and head and neck paraganglioma (HNPGL). Design Prospective, observational evaluation of NHS practice. Patients Individuals with PPGL/HNPGL referred to a supraregional genetics testing service over a 10-year period. Measurements Clinical (age, tumour site, malignancy, etc.), mutation frequencies and characteristics. Correspondence: Eamonn R. Maher, Centre for Rare Diseases and Personalised Medicine, School of Clinical and Experimental Medicine, College of Medical and Dental Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TH, UK. Tel.: ; Fax: ; e.r.maher@bham.ac.uk Results A total of 501 probands with PPGL (n = 413) or HNPGL (n = 88) were studied. Thirty-one percent of patients with PPGL presented had a pathogenic mutation in SDHB, SDHD or VHL. Mutation detection rates were highest in those with a positive family history (62%), malignancy (53%), multiple tumours (33%) or PGL (44%). Twenty-eight percent of individuals with a single sporadic phaeochromocytoma had a mutation. Overall, 63% of patients with HNPGL had a mutation (92% of those with a family history, 89% of those with multicentric tumours and 34% of those with a single sporadic HNPGL). Penetrance was calculated in 121 SDHB mutation-positive probands and 187 of their mutationpositive relatives. Most relatives were asymptomatic and lifetime penetrance in non-proband SDHB mutation carriers was <50%. Conclusions Practice-based evaluations of genetic testing in PPGL reveal high mutation detection rates. Although clinical criteria can be used to prioritize mutation testing, mutations were detected in low risk groups indicating a need for comprehensive and inexpensive genetic testing strategies for PPGL and HNPGL. (Received 20 August 2012; returned for revision 6 September 2012; finally revised 10 October 2012; accepted 10 October 2012) 898

2 SDHB, SDHD and VHL mutations in phaeochromocytoma and HNPGL 899 Introduction Phaeochromocytoma and paraganglioma (PPGL) are rare tumours with an incidence of 2 8 cases per million per year. 1,2 The WHO defines phaeochromocytoma as a tumour arising from the chromaffin cells of the adrenal medulla, whereas all other extra-adrenal tumours are defined as paraganglioma (PGL). PGL can be functional, that is, secreting catecholamines, or nonfunctional. In the literature and in this study, the term phaeochromocytoma refers to adrenal tumours and PGL to functional extra-adrenal, nonhead and neck tumours originating from the sympathetic nervous system. These are analysed together and abbreviated to PPGL. Head and neck paraganglioma (HNPGL) refers to, generally, nonfunctioning tumours, derived from parasympathetic tissues. 3 Phaeochromocytoma/paraganglioma and HNPGL may present in individuals with well-characterized clinical syndromes such as von Hippel Lindau disease (VHL), multiple endocrine hyperplasia type 2A (MEN2A) or type I neurofibromatosis (so-called syndromic phaeochromocytoma). These conditions are associated with inherited mutations in the VHL, RET and NF1 genes, respectively. Historically, around 10% of phaeochromocytoma were believed to be inherited in an autosomal dominant manner. However, in the past two decades, developments in genetics research have identified further genetic causes for phaeochromocytoma leading to the recognition that around a third of non-syndromic cases have a genetic cause. 4 Genes forming part of the mitochondrial succinate dehydrogenase complex, notably SDHB and SDHD are associated with phaeochromocytoma and HNPGL. 5,6 In a population-based research cohort, the frequency of VHL, SDHB and SDHD mutations among patients with non-syndromic cases PPGL was 196% 7 leading to suggestions that all patients with PPGL should be offered genetic testing. 7 However, such a testing strategy is expensive and there has been a move towards targeted testing that prioritizes certain subgroups of patients for testing. Thus, it is noteworthy that in the population-based study reported by Neumann et al., 7 only 04% (1/ 271) of patients aged >50 years had a detectable mutation in SDHB, SDHD, VHL or RET. Within certain populations, founder mutations may account for significant numbers of patients with PPGL or HNPGL e.g. c. 505T>C VHL mutation in Germany 8 and up to 90% of all Dutch patients with SDH X mutations can be accounted for by just six founder mutations including the c274g>t SDHD mutation which is found in 46% of all patients with HNPGL in the Netherlands. 9 Hence, the optimum screening strategy for detecting inherited PPGL or HNPGL may vary from country to country. In order to evaluate our UK-based experience of genetic testing of VHL, SDHB and SDHD for patients with PPGL or HNPGL, we reviewed data from a 10-year period in a supraregional genetics testing service. Midlands PPGL/HNPGL database between 1 September 2001 and 1 September This database included all individuals referred to the West Midlands Regional Genetics Service (WMRGS) for genetic testing for SDHB, SDHD or VHL. If more than one member of a family was referred, only the proband case was included in this study (although data on non-proband symptomatic and asymptomatic SDHB mutation carriers were included in penetrance calculations (see below)). Patients were referred from clinical genetics departments and endocrinologists from across the United Kingdom. Data collated Referring clinicians were contacted regarding demographic and clinical features of cases studied. Data were collated on a standardised proforma and entered on a specified database. We collected demographic and clinical data including the following: gender, age at presentation, method of presentation (sporadic vs familial vs multicentric), location of tumour, presence of bilateral disease and evidence of malignancy. Malignancy was defined as the presence of distant or local regional metastasis. Information on the clinical and genetic features of tested relatives was collated in order to determine the penetrance of SDHB mutations. Molecular genetic analysis The WMRGS received 10 ml of EDTA or ACD anticoagulant blood for each individual included in the study. Genomic DNA was extracted from peripheral blood leucocytes according to standard protocols described previously by our group. 10 Briefly, all DNA samples were analysed in the same genetic laboratory (WMRGS). Mutation analysis was performed on the exonic and intronic flanking regions including splice sites for all exons of VHL (NM_ ), SDHB (NM_ ) and SDHD (NM_ ). In addition to PCR-based mutation scanning, multiplex ligation-dependent probe amplification (MLPA) was performed for SDHB, SDHD and VHL (MRC Holland). All genetic alterations were reviewed by consultant clinical geneticists and compared with available mutation data Statistical considerations Data were analysed using the IBM SPSS Statistical Package (v18, Somers, NY, USA). Logistic regression was used to determine the predictive power of models explored and ROC curves drawn to determine the accuracy of models using different age cut-offs. Student s t-tests were used to compare the ages in different subgroups. Methods Patients We included probands with non-syndromic presentation of PPGL and HNPGL entered prospectively into the West Ethical considerations All patients gave written informed consent for genetic testing, and evaluation of the genetic testing service was approved by Birmingham Women s Hospital NHS Trust Research and Development Department.

3 900 M. Jafri et al. Results General demographic data for the cohort Five hundred and one patients were identified as having either a PPGL or HNPGL. Four hundred and thirteen individuals had a PPGL and 88 individuals had a HNPGL. The median age of patients with PPGL was 36 years (range 5 94 years), the median age of patients with HNPGL was 39 (range 8 74). The maleto-female ratio was 084 for PPGL and 102 for patients with HNPGL. Clinical and mutational data for individuals undergoing susceptibility testing with PPGL Of the PPGL cohort (n = 413), 99 (24%) patients had a family history of either phaeochromocytoma or HNPGL and 314 had a sporadic presentation. Two hundred and eighty-six of the 314 patients (69% of total PPGL cases) with a sporadic presentation had a single tumour. Forty had bilateral phaeochromocytoma and 30 patients had multiple PPGL (at least one PGL) at the time of referral. Fifty-seven patients (14% of the total PPGL referrals) had evidence of malignancy (i.e. metastatic spread). One hundred and thirty-one of the PPGL probands had a PGL (36 of 99 familial cases, 86 of 286 sporadic cases, nine of 30 multicentric cases and 29 of 35 malignant cases). Comparison of the age distribution of UK sporadic cases referred for genetic testing with those in the population-based cohort described by Neumann et al. 7 revealed a slight enrichment of younger cases in our series (see Fig. 1). SDHB, SDHD and VHL mutation data Overall, 129 of 413 (31%) probands with PPGL had a detectable mutation in either SDHB, SDHD or VHL. Mutations previously not described by our group 10 are listed in (Table 1). Overall, SDHB mutations were detected in 97 probands (accounting for % of cohort UK EAPR Age (years) Fig. 1 The age distribution of our UK cohort (labelled United Kingdom) compared to those presented in the population-based series described by Neumann et al. 7 (labelled EAPR). Table 1. Mutations in SDHB and SDHD described in our cohort which have not previously noted in the UK cohort 10 Gene Exon DNA mutation Predicted result SDHB 1 c. 17_42dup26 Frameshift p. Ala15Hisfs*8 2 c. 88delC Frameshift p. Gln30Argfs*47 2 c. 158G>A p. Gly53Glu 4 c G>C Splice site 6 c. 591delC Frameshift p. Ser198Alafs*22 6 c. 587G>A p. Cys196Tyr 7 c. 685_6ins13 p. Glu229Alafs*31 7 c. 745_748dupTGCA p. Thr250Metfs*7 7 c. 724C>T p. Arg242Cys SDHD 3 c. 296delT Frameshift p. Leu99Profs*36 3 c. 191_192delTC Frameshift p. Leu64Profs*4 75% of mutations detected), SDHD in 13 probands (10%) and VHL in 19 probands (15%). Eleven of 97 probands with a pathogenic SDHB mutation had a deletion of one or more exons, 10 had a nonsense mutation, 13 had a frameshift mutation, 19 had a splice site mutation and 43 had a missense mutation. There was one insertion. SDHD mutations were detected in 13 probands (missense = 5, nonsense = 4, frameshift = 2, splice site mutation = 1 and a deletion was each seen in one case). Pathogenic VHL mutations in 19 cases consisted of missense mutations (n = 17), deletions of one or more exons (n = 1) and frameshift mutations (n = 1). Characteristics of mutation-positive and mutationnegative probands Family history. Sixty-one of 99 (62%) probands with PPGL and a positive family history of PPGL or HNPGL had a detectable mutation in SDHB (n = 47), SDHD (n = 6) or VHL (n = 8). The median age of probands with a positive family history was 26 years (range 7 69). Twenty-two of 99 had multicentric tumours, 31 of 99 had a PGL and 18 of 99 had a malignant PPGL. The age of those with a family history and mutation was significantly less (t = 29, P = 0004) than those without a mutation (mean 26 vs 36 years). Multiple tumours. Ten of 30 probands with multiple tumours and no family history of PPGL or HNPGL had a detectable mutation in SDHB (n = 7), SDHD (n = 2) or VHL (n = 1). The median age of these probands with multiple tumours was 44 years (range 12 83), three had malignant tumours and 10 had one or more extra-adrenal tumours. The mean age at diagnosis for a proband with multicentric tumours without a positive family history with a mutation and those without a mutation was 395 and 51 years, respectively (t = 309, P = 0003). Malignancy. Thirty of 57 (53%) probands with a malignant tumour had a detectable mutation in SDHB (n = 26), SDHD (n = 3) or VHL (n = 1). Excluding those with a positive family history and/or multiple tumours, 16 of 35 (46%) probands with

4 SDHB, SDHD and VHL mutations in phaeochromocytoma and HNPGL 901 a sporadic malignant tumour had a detectable mutation in SDHB (n = 14), SDHD (n = 1) or VHL (n = 1). The mean age at diagnosis in this latter group for those with a mutation and those without a mutation was 29 and 50 years, respectively, and there were no differences in the frequency of extra-adrenal phaeochromocytoma between the two groups. Adrenal or extra-adrenal location. Overall, 58 of 131 (44%) probands with an extra-adrenal PGL had a detectable mutation in SDHB (n = 40, respectively), SDHD (n = 6) or VHL (n = 12). For patients with a PGL, the mean age of those with a mutation was 28 years compared to 41 years in those without a detectable mutation (t = 413, P < 0001). For individuals with a solitary, benign, adrenal phaeochromocytoma, 29% (71/242) had a detectable mutation in SDHB (n = 55), SDHD (n = 6) or VHL (n = 10). Sporadic phaeochromocytoma patients with a mutation were significantly younger than those without a mutation (mean age 29 vs 40 years, respectively (t = 44, P < 0001). The frequency of mutations in patients with a single sporadic phaeochromocytoma in different age groups is shown in Fig. 2a. Mutations were detected in all age groups. The proportions of mutations detected in patients with a single sporadic phaeochromocytoma at different age-at-diagnosis cutoffs (<40, <50, <60 and >60 years) are shown in Fig. 2b. As the (a) Percentage (b) 100 Percentage < >60 Age group % Mutation positive in age group % of all mutations All Age group tested % Mutation positive in age group % of all mutations Fig. 2 (a) Proportion of individuals with a solitary adrenal phaeochromocytoma with a detectable SDHB, SDHD or VHL mutation according to age at diagnosis. (b) Comparison of proportion of mutationpositive patients detected and proportion of cases tested with different age cut-offs for offering testing. age cut-off is increased, the proportion of mutations detected increases at the expense of less specificity. Presence of additional tumour types. In addition to PPGL and HNPGL, a variety of other tumour types have been reported to be associated, or potentially associated, with SDHB or SDHD mutations. We therefore reviewed whether the presence of a different tumour increased the mutation detection rate in individuals with PPGL. In our cohort, there were three individuals with PPGL and breast cancer and none had a detectable mutation. Two individuals with PPGL also had thyroid carcinoma and neither had a mutation. Three individuals with PPGL and renal tumours were tested and one individual (reported previously 13 Henderson et al.) had a mutation. Neither of two patients with PPGL and a second endocrine tumour (prolactinoma or adrenal adenoma) had a mutation. Characteristics of mutation-positive cases by specific genes SDHB mutation-positive cases. The median age of SDHB mutation-positive probands (n = 97) was 27 years. Forty-nine percent of SDHB mutation-positive probands had a family history and 7% had multicentric tumours at presentation. Fortythree percent of individuals with an SDHB mutation had a sporadic presentation. Twenty-two percent of those with a bilateral phaeochromocytoma had an SDHB mutation. Twentyseven individuals with SDHB mutations had a malignant tumour (26%) and 31 of those with SDHB disease had extraadrenal tumours (32%). This indicates that SDHB mutations are more likely to be associated with malignancy and extra-adrenal presentation. SDHD mutation-positive probands (n = 13) had a median age of 23 years and only one patient aged over 50 had an SDHD mutation. Forty-six percent (6/13) had a positive family history, 15% (2/13) had multicentric disease and 38% (5/13) had a sporadic presentation. The proportion of those with multicentric disease included those with HNPGL which are strongly correlated with SDHD mutations. Twenty-three percent of individuals with SDHD mutations had bilateral tumours, 15% had malignant tumours and 31% of individuals had extra-adrenal tumours. VHL mutation-positive probands (n = 10) had a median age of 20. There were no patients aged 46 years. Thirty-eight percent had a positive family history of PPGL. Five percent had multicentric tumours and 57% had sporadic presentation. Nineteen per cent had extra-adrenal tumours, and 32% of individuals had bilateral disease. Ten percent had a malignant tumour. Clinical characteristics and SDHB, SDHD or VHL mutation prediction Our findings that the likelihood of detecting a SDHB, SDHD or VHL mutation was increased by the presence of a positive family history of PPGL or HNPGL, multiple tumours, extra-adrenal location and younger age at onset were consistent with previous

5 902 M. Jafri et al. reports (see Jafri and Maher 3 and references within). However, whilst there is a consensus about the risk factors for the presence of a mutation, there is little consensus about the age up to which patients with sporadic phaeochromocytoma should be offered genetic testing. We therefore applied different selection models to our data set to investigate the sensitivity and efficiency of different age cut-offs. Thus in Model A, all patients with a positive family history, multiple tumours, extra-adrenal location and/or malignant PPGL plus those with sporadic adrenal phaeochromocytomas aged <45 years would be tested. Applying these criteria to our data and undertaking logistic regression analysis revealed that application of Model A would have detected 69% of all mutationpositive cases (97% of SDHD mutations, 95% of VHL mutations and 76% of SDHB mutations). We then investigated the effects of increasing the age cut-off for sporadic phaeochromocytoma cases to 50 years (Model B) and 60 years (Model C); proportions of mutation-positive cases detected would increase to 691% (Model B) and 764% (Model C), respectively. The ROCs for each model are shown in (Fig. 3). Although increasing the age cut-off improves the mutation detection rate, there is an increase in the number of tests undertaken. Thus, assuming an indicative rate of 1500 ($2300) for mutation analysis of SDHB, SDHD and VHL, the cost per mutation detected for analysing all PPGL cases is 4802 (approximately $7360). Using Model A, the cost is 4170 (approximately $6380), 4245 for Model B (approximately $6490) and 4500 for Model C (approximately $6885). Clinical and mutational data for individuals undergoing susceptibility testing with HNPGL Eighty-eight probands presented with HNPGL and their median age at diagnosis was 40 years (range 8 75 years). Forty-one percent of the cohort had a positive family history of either HNPGL or PPGL. Eight of 52 patients without a family history presented with multiple tumours. Median age at presentation was 38 years in those with a positive family history (n = 36), Area under the curve a Test result variable(s) Area Std. error Asymptotic Sig.b Asymptotic 95% confidence interval Lower bound Less_ Less_ Less_ Upper bound The test result variable(s): less_45, less_50, Less_60 has at least one tie between the positive actual state group and the negative actual state group. Statistics may be biased. a. Under the nonparametric assumption b. Null hypothesis: true area = 0 5 Fig. 3 Receiver operating characteristic curves if testing occurs in all individuals with family history, extra-adrenal, multiple and malignant phaeochromocytoma and aged 45, 50, 60 years old (Models A, B and C, labelled less_45, less_50 and Less_60, respectively).

6 SDHB, SDHD and VHL mutations in phaeochromocytoma and HNPGL years in those with sporadic multiple HNPGL (n = 43) and 36 years in those with a single sporadic HNPGL (n = 40). SDHB and SDHD mutation data in HNPGL Overall, 63% (n = 55) of probands with HNPGL had a detectable mutation in SDHB (27%, n = 24), or SDHD (35%, n = 31). Pathogenic mutations in SDHB consisted of deletions of one or more exons (n = 2), nonsense mutations (n = 3), frameshift mutations (n = 1), splice site mutations (n = 3) and missense mutations (n = 15). Pathogenic mutations in SDHD consisted of deletions of one or more exons (n = 4), nonsense mutations (n = 2), frameshift mutations (n = 7), splice site mutations (n = 0) and missense mutations (n = 18). Penetrance of SDHB mutations by method of ascertainment In addition to the 121 SDHB mutation-positive probands with PPGL or HNPGL, a further 187 of their relatives underwent genetic testing and were subsequently demonstrated to have inherited a SDHB mutation. Only 35 of the relatives were clinically affected at the time of genetic testing and 152 were asymptomatic. Comparison of the penetrance of PPGL/HNPGL and PPGL only in probands and non-proband mutation carriers revealed significantly lower penetrances in the latter groups (v 2 = 1165 P < and v 2 = 915 P < 00001, respectively) (see Fig. 4). Characteristics of mutation-positive and mutationnegative HNPGL probands Family history. Thirty-three of 36 (92%) probands with a positive family history of PPGL or HNPGL had a detectable mutation in SDHB (n = 15), SDHD (n = 18). Two patients had multicentric tumours and two had malignant tumours. Multiple tumours. Eight of 9 (89%) probands with multiple tumours and no family history of PPGL/HNPGL had a detectable mutation in SDHB (n = 2) and SDHD (n = 6). The median age at diagnosis for those with multiple tumours and a mutation was 385 years compared with 62 years for those without a mutation. Malignancy. Only three patients (two of whom had a positive family history) were diagnosed with a malignant HNPGL. Two had truncating mutations (one SDHB nonsense mutation (c268c>t, p.arg 90X) and one a SDHD frameshift mutation (c. 94_95del TC, p Ala33 Ile fs*35). No mutation in SDHD, SDHB or VHL was detected in the third case. Single sporadic HNPGL. Sixteen of 44 probands of individuals with a sporadic presentation had a detectable mutation in SDHB (n = 7) and SDHD (n = 9). Characteristics for mutation-positive HNPGL cases for specific genes SDHB mutation carriers. The median age of SDHB mutation carriers was 46 years. Sixty-three percent (15/24) of SDHB mutation carriers had a family history and 8% (2/24) had multicentric tumours at presentation. Twenty-nine percent (7/24) of individuals with an SDHB mutation had a sporadic presentation. SDHD mutation carriers. The median age of individuals with SDHD mutations was 385 years. Fifty-eight percent (18/31) had a positive family history, 19% (6/31) had multicentric disease and 29% (9/31) had a sporadic presentation. Two patients had both a positive family history and multiple tumours. Fig. 4 (a) Penetrance of SDHB mutations in paraganglioma/phaeochromocytoma and HNPGL in probands with (dashed line) and their mutation-positive relatives (solid line). (b) Penetrance of SDHB mutations inparaganglioma/phaeochromocytoma in probands with SDHB mutations (dashed line) and their mutation-positive relatives (solid line). HNPGL, head and neck paraganglioma.

7 904 M. Jafri et al. Discussion Currently, mutations in 10 genes (SDHB, SDHC, SDHD, VHL, NF-1, RET, TMEM127, SDHAF, SDHA and MAX) are accepted to cause inherited PPGL and/or HNPGL (see Jafri and Maher 3 and references within). A further two possible inherited PPGL genes have been reported in single reports (PHD2/EGLN1 and KIF1b. 3,14,15 ) To date, the contribution to inherited PPGL/ HNPGL from mutations in MAX, TMEM127, SDHAF, SDHA and SDHC appear to be relatively small and/or not well defined Among non-syndromic PPGL cases, SDHB, SDHD and VHL account for most cases and; although VHL mutations may occasionally be associated with HNPGL, SDHD and SDHB are the major HNPGL genes. 1,2,4,7,21 24 The frequency of VHL, SDHB and SDHD mutations among PPGL/HNPGL is variable because of the presence of founder mutations in some populations. To our knowledge, this is the first extensive study describing mutation analysis of PPGL and HNPGL cohorts from the United Kingdom and is similar in size to recent large studies from the French COMETE Group 1 and Spain. 23 In addition, in contrast to earlier studies, 7 all patients in the current study were tested for germline exonic deletions/duplications in SDHB, SDHD and VHL. Two limitations of the current study were that RET germline mutation analysis was not performed as part of the genetic testing service and that the cohort represents a referral-based series and not a population-based series. Nevertheless, many of the findings of the current study are consistent with those of large population and referral-based studies from outside the United Kingdom in that a positive family history, multiple tumours, malignant disease, extra-adrenal location and younger age at diagnosis were all indicators of an increased mutation detection rate. Our cohort had a larger representation of malignant PPGL than the cohort described by Erlic et al., 4 (125% compared with 4%). Furthermore, the frequency of germline VHL gene mutations detected was less than that reported by Neumann et al. 7 This latter finding might reflect more intensive investigation to detect subclinical features of VHL disease before genetic testing in the United Kingdom or some referral bias. However, the age distribution of sporadic cases in our series was not markedly dissimilar to that reported in population-based series (Fig. 1). 7 Recommendations on genetic testing derived from the first international phaeochromocytoma symposium were published in and recommended testing all those aged under 50, or with familial tumours or multiple tumours. Individuals with malignant tumours were recommended to be tested for SDHB and VHL mutations and those with bilateral tumours for VHL, RET and SDHB although other international series have advocated testing all cases. 7 Currently, there are no UK national referral guidelines for genetic testing in PPGL or HNPGL cases. Our referral-based series therefore reflects the investigative pattern of a large number of different referring clinicians, and the age profile of our cohort suggests that universal testing of all PPGL is not widespread. We found that 34% of those with single HNPGL tumours had a germline mutation, and overall, the mutation detection rate in our HNPGL cohort (63%) was higher than that seen in some other studies (e.g. 55% of HNPGL cases had a mutation in SDHB, SDHC or SDHD in a large study from France. 1 Such differences may reflect local referral patterns (although 30% of our HNPGL cases tested aged over 50 years) or frequency of specific mutations in the population. Head and neck paraganglioma is mostly associated with mutations of SDHB and SDHD, 4 although SDHC and rarely TMEM127, SDHAF and VHL mutations may be detected. 16,17,26,27 A minority of patients had a malignant HNPGL (3%); this proportion was similar to that seen by Boedeker et al. 28 who identified 36% of 195 individuals tested for HNPGL had evidence of distant metastases. All malignant HNPGL in Boedeker and colleagues study had SDHB mutations. In our cohort, one patient had an SDHB mutation, one patient had an SDHD mutation and one patient had no detectable mutation in SDHB, VHL or SDHD. In addition to PPGL and HNPGL, a number of other tumours have been reported to be associated, or potentially associated, with SDHB/D mutations. The best defined is an increased risk of developing renal cancer, 29 but SDHB or SDHD mutations have been reported in patients with pituitary adenoma 30 and c-kit/pdgfra mutation-negative gastrointestinal stromal tumours. 31 In addition, SDHB and SDHD variants were reported in patients with PTEN mutation-negative Cowden-like syndrome who had an increased risk of breast cancer and papillary thyroid cancer. 32 Nevertheless, although the number of cases were small, we did not find any patients with PPGL and an additional tumour type (breast, thyroid or pituitary) who had a detectable SDHB or SDHD mutation. The high mutation detection rate in our series compared to population-based series 7 may reflect a combination of ascertainment bias and MLPA analysis to detect SDHB and SDHD (and SDHC) exonic deletions. As universal genetic testing of all patients with PPGL using conventional sequencing technology is expensive (even for just the major inherited PPGL genes; see below), a number of groups have suggested clinical stratification strategies to increase mutation detection efficiency and cut costs. Thus, Erlic et al. 4 analysed mutation detection predictors in 989 cases of non-syndromic phaeochromocytoma and suggested that germline mutation testing for RET, SDHB, SDHD and VHL could be most efficiently focused on patients with one or more of the following: family history of PPGL/HNPGL, age <45 years, multiple tumours, extra-adrenal location and previous HNPGL. Though there is a consensus that patients with a positive family history or multiple tumours should be offered genetic testing and most would suggest at least SDHB and SDHD mutation testing in patients with extra-adrenal or malignant tumours, there is less agreement regarding age cut-offs for targeting genetic testing of patients with a single benign phaeochromocytoma and no family history. Thus, in a series from Spain, only one of 65 sporadic probands with a single phaeochromocytoma tumour aged over 40 had a detectable mutation 23 but the higher mutation detection rate in our series could reflect differences in ascertainment, in the frequency of specific mutations in the relevant population or the use of MLPA to detect SDHB/D deletions in our cohort (of seven of 80 sporadic phaeochromocytoma cases aged >40 years with a

8 SDHB, SDHD and VHL mutations in phaeochromocytoma and HNPGL 905 detectable patients mutation, one had a SDHB deletion). Application of the Erlic et al. 4 testing criteria to our cohort would detect most SDHB, SDHD and VHL mutations in the cohort. However, a significant minority of cases would not be detected. Increasing the age cut-off to 60 years would have increased the proportion of mutations detected but increased genetic testing costs (e.g. for patients with sporadic benign phaeochromocytoma and age cut-offs of 40, 50 and 60 years would result in the detection of 52%, 72% and 93%, respectively, of all mutations, but would have involved testing 52%, 69% and 89%, respectively, of all sporadic benign cases). An age cut-off of 45 or 50 years might appear to offer the best balance between mutation detection and genetic testing costs and we note the majority of mutations missed by an age cut-off of 45/50 years were in the SDHB gene. Thus, there is a clear rationale for suggesting that if an age cut-off of 45/50 years is instigated, then this could be combined with SDHB immunostaining 33,34 to identify those older patients who could benefit from mutation analysis. Although clinical and immunohistochemical criteria can be used to prioritize patients for genetic testing, it seems likely that, as the number of inherited PPGL genes increase (and the proportion of cases with a genetic basis), the demand for more extensive genetic testing will increase. The cost of VHL, RET, SDHB and SDHD by conventional (Sanger) sequencing technology has been estimated to be approximately 1800 ($2700), 2 but the advent of high-throughput parallel sequencing technologies (second-generation sequencing) offers the potential for faster and less expensive testing of multiple genes simultaneously. Thus, preliminary findings have suggested that it could be possible to test 9 PPGL/HNPGL genes (SDHB, SDHC, SDHD, RET, VHL, MAX, TMEM127, SDHAF2 and SDHA) for approximately 500 (Rattenberry E., Macdonald F., Maher E.R., unpub. obs.). Hence, we anticipate that comprehensive mutation analysis of all proven PPGL and HNPGL genes (excepting NF1) in all patients will emerge as the standard testing strategy, and therefore, patients who do not satisfy current local criteria for genetic testing should have DNA stored for future analysis. A major advantage of a comprehensive highthroughput second-generation sequencing testing strategy is that all genes are tested simultaneously rather than serially and so the time to mutation detection is reduced. We would however emphasize the importance of evaluating the results of genetic testing in service-based settings. Thus, it is important that recommendations for genetic testing are kept under review and reflect the characteristics of the relevant population. For example, in our series, a significant proportion of patients did not have a detectable mutation, whereas series of HNPGL and PPGL from the Netherlands have revealed that six founder mutations accounted for 888% of all mutations. 9 Patients included in research studies are not necessarily representative of those in the general population and the phenotype associated with a particular gene may emerge over time. For example, TMEM127 mutations were initially thought to be confined to adrenal phaeochromocytoma cases, often aged >40 years, but were subsequently reported in patients with PGL and HNPGL. 35 In addition, we and others have found a high penetrance (at least 70% at age 60 years) of SDHB and SDHD mutations in a research setting, 23,24,29,33,36 whereas comprehensive investigation of single families or multiple families (with correction for method of ascertainment) have suggested much lower penetrances. 37,38 In the current study, we clearly demonstrated that there is a substantial difference between the penetrance in probands (families are invariably ascertained through an affected individual) compared to their mutation-positive relatives (the majority of whom were asymptomatic). This information is important for determining the benefits and costs of intensive surveillance in asymptomatic mutation carriers; multicentre collaborative studies will be required to further define the gene- and mutation-specific tumour risks in these individuals. Acknowledgements MJ was supported by the Birmingham Women s Hospital Springboard Fellowship and MRC Clinical Training Fellowship. ER was supported by the NIHR Clinical Training Fellowship. DGE and FL were supported by the Manchester BRC. References 1 Burnichon, N., Rohmer, V., Amar, L. et al. (2009) The succinate dehydrogenase genetic testing in a large prospective series of patients with paragangliomas. The Journal of Clinical Endocrinology and Metabolism, 94, Neumann, H.P., Erlic, Z., Boedeker, C.C. et al. (2009) Clinical predictors for germline mutations in head & neck paraganglioma patients: cost reduction strategy in genetic diagnostic process as fall-out. Cancer Research, 69, Jafri, M. & Maher, E.R. (2012) GENETICS IN ENDOCRINOL- OGY: the genetics of phaeochromocytoma: using clinical features to guide genetic testing. European Journal of Endocrinology, 166, Erlic, Z., Rybicki, L., Peczkowska, M. et al. (2009) Clinical predictors and algorithm for the genetic diagnosis of pheochromocytoma patients. Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 15, Astuti, D., Douglas, F., Lennard, T.W. et al. (2001) Germline SDHD mutation in familial phaeochromocytoma. Lancet, 357, Astuti, D., Latif, F., Dallol, A. et al. (2001) Gene mutations in the succinate dehydrogenase subunit SDHB cause susceptibility to familial pheochromocytoma and to familial paraganglioma. The American Journal of Human Genetics, 69, Neumann, H.P.H., Bausch, B., McWhinney, S.R. et al. (2002) Germ-line mutations in nonsyndromic pheochromocytoma. New England Journal of Medicine, 346, Brauch, H., Kishida, T., Glavac, D. et al. (1995) Von Hippel- Lindau (VHL) disease with pheochromocytoma in the Black Forest region of Germany: evidence for a founder effect. Human Genetics, 95, Hensen, E.F., van Duinen, N., Jansen, J.C. et al. (2012) High prevalence of founder mutations of the succinate dehydrogenase genes in the Netherlands. Clinical Genetics, 81, Ricketts, C.J., Forman, J.R., Rattenberry, E. et al. (2010) Tumor risks and genotype-phenotype-proteotype analysis in 358 patients

9 906 M. Jafri et al. with germline mutations in SDHB and SDHD. Human Mutation, 31, Bayley, J.-P., Devilee, P. & Taschner, P. (2005) The SDH mutation database: an online resource for succinate dehydrogenase sequence variants involved in pheochromocytoma, paraganglioma and mitochondrial complex II deficiency. BMC Medical Genetics, 6, Burnichon, N., Vescovo, L., Amar, L. et al. (2011) Integrative genomic analysis reveals somatic mutations in pheochromocytoma and paraganglioma. Human Molecular Genetics, 20, Henderson, A., Douglas, F., Perros, P. & Maher, E.R. SDHB associated renal oncocytoma suggests broadening of the renal phenotype in heriditary paragangliomas Familial Cancer, 8(5), Ladroue, C., Carcenac, R., Leporrier, M. et al. (2008) PHD2 mutation and congenital erythrocytosis with paraganglioma. New England Journal of Medicine, 359, Yeh, I.T., Lenci, R.E., Qin, Y. et al. (2008) A germline mutation of the KIF1B beta gene on 1p36 in a family with neural and nonneural tumors. Human Genetics, 124, Abermil, N., Guillaud-Bataille, M., Burnichon, N. et al. (2012) TMEM127 screening in a large cohort of patients with pheochromocytoma and/or paraganglioma. Journal of Clinical Endocrinology & Metabolism, 97, E805 E Bayley, J.P., Kunst, H.P., Cascon, A. et al. (2010) SDHAF2 mutations in familial and sporadic paraganglioma and phaeochromocytoma. The Lancet Oncology, 11, Burnichon, N., Briere, J.J., Libe, R. et al. (2010) SDHA is a tumor suppressor gene causing paraganglioma. Human Molecular Genetics, 19, Burnichon, N., Cascón, A., Schiavi, F. et al. (2012) MAX mutations cause hereditary and sporadic pheochromocytoma and paraganglioma. Clinical Cancer Research, 18, Kunst, H.P., Rutten, M.H., de Monnink, J.P. et al. (2011) SDHAF2 (PGL2-SDH5) and hereditary head and neck paraganglioma. Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 17, Amar, L., Bertherat, J., Baudin, E. et al. (2005) Genetic testing in pheochromocytoma or functional paraganglioma. Journal of Clinical Oncology, 23, Boedeker, C.C., Erlic, Z., Richard, S. et al. (2009) Head and neck paragangliomas in von Hippel-Lindau disease and multiple endocrine neoplasia type 2. The Journal of Clinical Endocrinology and Metabolism, 94, Cascón, A., Pita, G., Burnichon, N. et al. (2009) Genetics of pheochromocytoma and paraganglioma in spanish patients. Journal of Clinical Endocrinology & Metabolism, 94, Neumann, H.P., Pawlu, C., Peczkowska, M. et al. (2004) Distinct clinical features of paraganglioma syndromes associated with SDHB & SDHD gene mutations. JAMA, 292, Pacak, K., Eisenhofer, G., Ahlman, H. et al. (2007) Pheochromocytoma: recommendations for clinical practice from the Symposium, First International. Nature Clinical Practice. Endocrinology and Metabolism, 3, Qin, Y., Yao, L., King, E.E. et al. (2010) Germline mutations in TMEM127 confer susceptibility to pheochromocytoma. Nature Genetics, 42, Niemann MU, S. (2000) Mutations in SDHC cause autosomal dominant paraganglioma, type 3. Nature Genetics, [Support, Research & Non-Gov t, U.S.] 26, Boedeker, C.C., Neumann, H.P.H., Maier, W. et al. (2007) Malignant head and neck paragangliomas in SDHB mutation carriers. Otolaryngology and Head and Neck Surgery, 137, Ricketts, C., Woodward, E.R., Killick, P. et al. (2008) Germline, SDHB mutations & familial renal cell carcinoma. Journal of the National Cancer Institute, 100, Xekouki, P. & Stratakis C.A. (2012) Succinate dehydrogenase (SDHx) mutations in pituitary tumors: could this be a new role for mitochondrial complex II and/or Krebs cycle defects? Endocrine-Related Cancer, 19, C McWhinney, S.R., Pasini, B. & Stratakis, C.A. (2007) Familial gastrointestinal stromal tumors and germ-line mutations. New England Journal of Medicine, 357, Ni, Y., He, X., Chen, J. et al. (2012) Germline SDHx variants modify breast and thyroid cancer risks in Cowden and Cowdenlike syndrome via FAD/NAD-dependant destabilization of p53. Human Molecular Genetics, 21, Korpershoek, E., Favier, J., Gaal, J. et al. (2011) SDHA immunohistochemistry detects germline SDHA gene mutations in apparently sporadic paragangliomas and pheochromocytomas. The Journal of Clinical Endocrinology and Metabolism, 96, E1472 E van Nederveen, F.H., Gaal, J., Favier, J. et al. (2009) An immunohistochemical procedure to detect patients with paraganglioma and phaeochromocytoma with germline SDHB, SDHC, or SDHD gene mutations: a retrospective and prospective analysis. The Lancet Oncology, 10, Neumann, H.P., Sullivan, M., Winter, A. et al. (2011) Germline mutations of the TMEM127 gene in patients with paraganglioma of head and neck and extraadrenal abdominal sites. The Journal of Clinical Endocrinology and Metabolism, 96, E1279 E Benn, D.E., Gimenez-Roqueplo, A.P., Reilly, J.R. et al. (2006) Clinical presentation and penetrance of pheochromocytoma/ paraganglioma syndromes. The Journal of Clinical Endocrinology and Metabolism, 91, Hes, F.J., Weiss, M.M., Woortman, S.A. et al. (2010) Low penetrance of a SDHB mutation in a large Dutch paraganglioma family. BMC Medical Genetics, 11, Schiavi, F., Milne, R.L., Anda, E. et al. (2010) Are we overestimating the penetrance of mutations in SDHB? Human Mutation, 31,