MATERIALS AND METHODS

Tumoral and normal tissue

Tumoral samples from 35 patients with pheochromocytoma were collected anonymously with no information surrounding the existence of familial antecedents of the disease. This clinical information was subsequently obtained, along with normal tissue from the cases with variants. High molecular weight DNA was extracted from fresh frozen samples following standard procedures¹⁰ and also from paraffin embedded tumours using either the DNA Easy Tissue kit (Qiagen, Chatsworth, CA, USA) or a modification of a previously described protocol.¹¹ Normal tissue specimens used in the methylation study were collected from autopsy material.

Amplification and sequencing analysis

The analysis was carried out by genomic DNA amplification by PCR and direct sequencing of all eight exons of SDHB, all four exons of SDHD, all three exons of VHL, and exons 10, 11, and 16 of RET, using primers and PCR conditions as previously described.^12–15 By using DNA from 150–200 unrelated and unaffected individuals as a control population for single strand conformation polymorphism (SSCP) analysis,¹⁶ the pathogenic or polymorphic character of variants with unknown significance was defined. PCR products displaying mobility shift were subsequently sequenced using an automatic sequencer (ABI PRISM™ 3700; Perkin Elmer Applied Biosystems, Foster City, CA, USA).

Methylation specific polymerase chain reaction

The methylation status of the SDHB, SDHD, and VHL CpG islands was analysed using the methylation specific polymerase chain reaction (MSP) technique.¹⁷ This assay distinguishes between unmethylated and methylated alleles in a given gene on the basis of sequence changes induced by sodium bisulphite treatment of DNA, which converts all unmethylated, but not methylated cytosines to uracil. Subsequently, the DNA region of interest is amplified with primer pairs specific for methylated versus unmethylated DNA. Normal lymphocytes, and placental DNA that had methylated in vitro with SSSI bacterial methylase were used as negative and positive control for methylation, respectively. The MSP primers used are summarised in table 1. In selected samples, the DNA methylation status was confirmed by bisulphite genomic sequencing, as previously described,¹⁸ using the primers summarised in table 1.

RESULTS

We screened 35 pheochromocytomas for the presence of somatic mutations in VHL, RET, SDHB and SDHD, using amplification analysis followed by PCR products sequencing. All findings in primary tumours are summarised in table 2.

We identified three RET variants in three tumours: p.Cys611Phe in exon 10, p.Cys634Tyr in exon 11, and p.Met918Thr in exon 16. The mutational analysis of normal patient tissue with mutations revealed that p.Cys611Phe and p.Cys634Tyr had a germline origin, while p.Met918Thr resulted from a somatic change. By verifying the clinical data of the p.Cys611Phe and p.Cys634Tyr patients we found antecedents compatible with MEN2A. The first variation, p.Cys611Phe (table 2), was detected in a tumour sample of a female patient who had been diagnosed at 58 years of age with pheochromocytoma and C cell hyperplasia. The p.Cys634Tyr change was present in the primary tumour of a male subject diagnosed at 35 years of age with pheochromocytoma and signs of invasion. The clinical records of this patient revealed familial antecedents of MEN2A syndrome. The patient’s sister was 45 years old when she was operated for unilateral pheochromocytoma and she had developed medullary thyroid carcinoma at 44 years of age. The patient’s father had bilateral pheochromocytoma and there were three cases of medullary thyroid carcinoma among his grandchildren. Finally, the p.Met918Thr was found in a sporadic tumour of a female patient diagnosed with non-familial pheochromocytoma at 36 years of age.

The analysis of the succinate dehydrogenase genes revealed four new alterations: a missense substitution (p.Ser163Pro) and a 3 bp deletion (p.Ser195del) in exons 5 and 6 of SDHB respectively, and two missense substitutions in SDHD, p.Glu42Ala and p.Ala90Thr, both in exon 3. Neither of the patients had familial antecedents. Ages of onset, sex, and tumour characteristics are summarised in table 2. All the substitutions were found in constitutional tissue of the patients, supporting their germline origin. Given that neither variant had previously been described, we performed a study in a control population; the p.Ser163Pro variant was found in four individuals of the control population tested by SSCP, while the remaining changes were not found in controls. We also detected six polymorphic variants: g.129586_7insCTTCTT in SDHB,¹³ p.Gly148Gly, p.Ser68Ser,⁵ p.His30His, p.Gly12Ser (19, 20) and g.112759G→A in SDHD.¹³ Their respective frequencies in the control population are summarised in table 2. There was no relation between the genetic alterations found in this study and the clinical characteristics and behaviour of tumours in terms of invasion, metastatic status, or malignancy.

DNA methylation patterns analysis at the CpG islands of VHL, SDHB, and SDHD did not show any hypermethylation event occurring in the pheochromocytomas studied or in normal adrenal medulla samples (figs 1 and 2).

DISCUSSION

The presence of positive familial antecedents of pheochromocytoma and the presence of bilateral or multiple tumours are the main parameters associated with RET, VHL, SDHD, or SDHB germline mutations. Recent studies, however, report that about 20% of sporadic cases without such clinical features are associated with these genes, so the screening of mutations should be considered essential in all patients independent of their family history or the presence of tumour multiplicity and bilaterality. On the other hand, little is known about which genes are involved in the development of sporadic tumours, therefore in this study we firstly intended to characterise the mutations of the genes commonly involved in the sporadic form of the disease.

We searched for mutations in VHL, RET, SDHB, and SDHD in 35 pheochromocytomas (table 2) and we identified six tumours (17%) with pathogenic changes: three mutations of RET, one of SDHB, and two of SDHD. The three mutations found in RET (~8.4%) had been described previously.^21–23 The p.Cys611Phe and p.Cys634Tyr changes were found in normal tissue, confirming their germline origin. They affect two conserved amino acids often mutated in MEN2A. The p.Met918Thr variant was somatic and results in the replacement of methionine with threonine within the catalytic core region of the tyrosine kinase domain.²¹ This mutation, affecting substrate interactions, has been strongly linked (95%) to MEN2B, and has also been reported in some sporadic pheochromocytomas.^24,25

One tumour (~2.8%) was found to have an SDHB mutation, and was also germline. The change, a 3 bp deletion, eliminates a serine residue of the conserved and functional 4Fe-4S ferredoxin iron–sulphur binding domain of SDHB.⁶ The p.Glu42Ala and p.Ala90Thr SDHD substitutions (~5.7%) were germline (not previously described) and they did not affect conserved amino acids of the protein. We had no available material to perform further functional experiments, but as they were not detected in controls we considered these variants to be pathogenic changes.

Aside from other substitutions (g.129586_7insCTTCTT, p.Gly12Ser, p.Ser68Ser, g.112759G→A) previously described as polymorphisms,^12,13,20 we found a new variant, p.Ser163Pro. This substitution was found in 2.3% of the control population, so it could be a rare polymorphism.

We found 17% of tumours to have mutations in some of the genes. All the identified changes were also found in the corresponding normal tissue with the exception of one somatic RET mutation. We have therefore found 14% of tumours to have germline mutations in some of the genes studied. This result differs from a previous study⁸ where 24% of apparently sporadic pheochromocytomas with germline mutations were found in one of the genes related to the development of these neoplasias (VHL, RET, SDHB, SDHD), although our study was of both a genetic and epigenetic nature. In addition, no more than one somatic RET change in tumours was found, nor were there somatic mutations in either the complex II genes or the VHL gene. It has been previously reported that somatic RET or VHL mutations are rare in sporadic pheochromocytomas,²⁵ and SDHD mutations have been described as rare in neuroendocrine tumours.²⁶ Our results support these findings, and also confirm the absence of somatic SDHB mutations in sporadic pheochromocytomas.²⁷

As one of the classical inactivation mechanisms of tumour suppressor genes is the hypermethylation of promoter CpG islands,²⁸ we attempted to discover whether this mechanism was involved in gene silencing. To date, the methylation status of SDHB remains unknown, and this study suggests that hypermethylation of the upstream CpG island of this gene is not a frequent event in sporadic pheochromocytomas. On the other hand, it has been reported that VHL promoter methylation is not a common event in pheochromocytomas,²⁹ and in this report we confirm that epigenetic silencing of the VHL gene in sporadic pheochromocytomas is not a prominent mechanism, in spite of its relevance in renal cell carcinoma.^30,31

SDHD, which is believed to be maternally imprinted by means of its inheritance pattern in PGL families, was not methylated at its canonical promoter CpG island in our pheochromocytomas. We also confirmed this finding in three different types of normal adrenal tissue. It has been suggested⁵ that it could be an epigenetic modification specifically affecting the carotid body in PGL families, although this event has not yet been molecularly demonstrated. Our results suggest that if methylation related imprinting exists for this gene, it can occur in another region. Moreover, the absence of methylation of this gene in both pheochromocytomas and normal adrenal tissue could indicate that tissue specific methylation is not the underlying mechanism of this pattern of inheritance, and therefore a different molecular mechanism must account for this model.

As no relation was found between the genetic alterations and the clinical characteristics of our series of pheochromocytomas, and given that methylation was not involved in the development of these tumours, other genes may be involved in the pathogenesis of these types of tumours. Further analysis is required in order to identify new candidate genes responsible for tumour pathogenesis of sporadic pheochromocytomas and to estimate the hereditary susceptibility to developing this neoplasia.