Aims—Phaeochromocytomas and paragangliomas are uncommon. The aims of this study were to analyse the characteristics and the possible roles of p53, Rb, and mdm2 alterations in these tumours.
Methods—The clinicopathological features of 65 patients (31 men, 34 women) with phaeochromocytomas or paragangliomas were analysed. The tumours were studied for the expression of p53, Rb, and mdm2 by immunohistochemical methods.
Results—Thirty nine of the patients had phaeochromocytomas and 26 had paragangliomas. Bilateral tumours were noted in eight of the patients and malignant tumours were seen in 13. Paragangliomas were often small, non-functional, and presented incidentally, whereas phaeochromocytomas were usually large, functional, and symptomatic. p53 overexpression, loss of Rb expression, and mdm2 overexpression were seen in four, 43, and 37 of the patients, respectively. Three of the four patients with p53 overexpression had bilateral tumours. Loss of Rb expression was often found in phaeochromocytomas, whereas mdm2 overexpression was more frequently seen in paragangliomas. The 10 year survival rate of patients with malignant tumours was 45%. Two patients died of tumour metastases more than 10 years after resection of the primary tumours.
Conclusions—Phaeochromocytomas and paragangliomas had distinctive clinical features and genetic alterations. The prognosis of patients with these tumours was related to the malignant potential. p53 overexpression, more common in bilateral phaeochromocytomas and paragangliomas, could be a marker for this tumour subgroup.
Statistics from Altmetric.com
The products of the p53, retinoblastoma (Rb), and mouse double minute 2 (mdm2) genes are important modulators of the cell cycle.1–3 The p53 and Rb genes are tumour suppressor genes, whereas the mdm2 gene is a proto-oncogene, which produces a protein that inhibits the function of the p53 and Rb proteins.1–3 Wild-type p53 protein is labile, whereas mutant forms have longer half lives and can be detected by immunohistochemistry. In contrast, the Rb protein can be detected in non-neoplastic tissue, and mutations of the Rb gene could result in the loss of Rb staining. Lastly, amplification of the mdm2 gene may lead to mdm2 protein overexpression.
The role of p53 alterations in phaeochromocytomas/paragangliomas has been investigated in several studies.4–14 However, to the best of our knowledge, the importance of Rb and mdm2 aberrations in human phaeochromocytomas/paragangliomas has never been studied. In our study, we analysed the clinicopathological features of patients with phaeochromocytomas and paragangliomas. The possible roles of alterations in the p53, Rb, and mdm2 genes and their interaction in these tumours were also studied by means of immunohistochemistry. Although the immunohistochemical analysis of these cell cycle proteins does not necessarily reflect genetic changes, it allows genetic alterations to be localised and correlated with other clinicopathological parameters.
Materials and methods
The patients who were chosen for our study had phaeochromocytomas or paragangliomas confirmed by histological examination during a 25 year study period (1973–97) at Queen Mary Hospital. The age, sex, clinical presentation, urinary catecholamine values, and survival data of these patients were taken from the clinical records. Macroscopically, the location and size (maximum length) of the tumours were recorded. The tumours were categorised as either benign or malignant. Malignant phaeochromocytomas/paragangliomas were defined as tumours with unequivocal evidence of metastatic disease(s).
PATHOLOGICAL REVIEW AND SELECTION OF TISSUES
The histological sections of the tumours were reviewed. The diagnoses were confirmed by positivity to the neuroendocrine markers (neurone specific enolase, chromogrannin, and synaptophysin). Electron microscopic examination was performed to detect the neurosecretory granules if the diagnosis was uncertain. Representative paraffin wax blocks from the tumours were chosen for immunohistochemical studies.
The immunohistochemical study was conducted on 5 μm thick paraffin wax sections using the avidin–biotin–peroxidase complex method. The antibodies against p53 (microwave pretreated; NCL-p53-DO7 at a dilution of 1/50), Rb (microwave pretreated; NCL-RB1 at a dilution of 1/50), and mdm2 (trypsin pretreated; NCL-MDM2 at a dilution of 1/50) were from Novocastra Laboratories Ltd (Newcastle upon Tyne, UK). Paraffin wax blocks of oesophageal squamous cell carcinoma (known to be strongly positive for the p53 and Rb proteins) and breast carcinoma (known to be strongly positive for mdm2) were used as positive controls.4–7 Brown nuclear staining was regarded as a positive signal, whereas cytoplasmic staining (non-specific background staining) was considered to be negative. In addition, tumours with less than 5% of tumour cells with a positive signal were defined as negative.
Statistical analysis was performed using the Student's t test (continuous variables), with Yates correction and χ2/Fisher exact test (categorical variables). The actuarial survival rate of the patients was measured from the date of resection of these tumours to the date of death or of the last follow up. Survival was calculated using the Kaplan–Meier method, whereas the impact of survival by various factors was examined by means of the log rank test.
Sixty five patients (31 men, 34 women) with phaeochromocytomas/paragangliomas were found (62 Chinese, two white, and one Indian). The clinicopathological features of patients are summarised in table 1 and some of the cases had been described previously.15–17
The mean age of the patients was 43 (range, 2–85). Thirty nine had phaeochromocytomas and 26 had paragangliomas. Approximately half (33) of the patients had functional tumours (with raised urinary catecholamine values) and 46 had been diagnosed correctly before pathological examination. Thirty one of the patients with phaeochromocytomas and two of the patients with paragangliomas had functional tumours. Paragangliomas studied comprised: carotid body paragangliomas (12), paragangliomas from the sympathetic trunk (four in the retroperitoneum and two in the bladder), middle ear (five), jaw (two), and base of the skull (one).
Thirty six of the patients had lesions on the right side of body, 21 had left sided tumours, and eight had bilateral tumours. The last group comprised four patients suffering from bilateral phaeochromocytomas and four with bilateral carotid body tumours. Thus, we analysed a total of 73 phaeochromocytomas and paragangliomas. On gross examination, the tumours usually appeared light brown in colour. The mean diameter of these tumours was 5 cm (range, 0.2–17). Malignancy was noted in 13 of the patients.
Phaeochromocytomas were often functional, usually leading to a correct diagnosis before histological examinations were conducted, whereas paragangliomas were normally non-functional tumours and detected incidentally (p = 0.0001 and p = 0.025, respectively). The mean diameter of the phaeochromocytomas was significantly larger than paragangliomas (6 v 3.5 cm; p = 0.002). Phaeochromocytomas were detected more often on the right side (right to left ratio, 1.9 : 1), whereas paragangliomas did not seem to have a predilection for either side (right to left ratio, 1.1 : 1). Nevertheless, the difference between phaeochromocytomas and paragangliomas with respect to side predilection was not significant (p = 0.32). Bilateral paragangliomas were limited to the location at carotid bodies. One third (four of 12) of the patients with carotid body tumours suffered from bilateral growth whereas no patients with other paragangliomas had bilateral lesions (p = 0.033). There was no significant difference between the phaeochromocytomas and paragangliomas with respect to the age (p = 0.45; Student's t test) or sex (p = 1) of the patients or malignant potential of the tumours (p = 1).
Four patients had p53 overexpression: three with phaeochromocytomas and one with carotid body tumours. Bilateral phaeochromocytomas were found in two of the three patients with p53 overexpression. Among them, one had multiple endocrine neoplasia type 2 (MEN2) and another died of bilateral phaeochromocytomas with metastases (no necropsy or genetic screening was performed to verify the presence of MEN2). The third patient with a p53 positive phaeochromocytoma was a 2 year old boy (the youngest patient in our series). He was lost to follow up. The patient with p53 positive bilateral carotid body tumours also had familial tumours and her brother also suffered from bilateral carotid body paragangliomas. The difference in p53 overexpression in the patients with and without bilateral tumours was significant (p = 0.003). The patients with p53 positive tumours did not differ significantly from those with p53 negative ones with respect to age at presentation (p = 0.16), sex (p = 1), clinical presentation (p = 1), and functional status (p = 0.61). Furthermore, there was no difference between p53 positive and p53 negative tumours in relation to their size (p = 0.36), location (adrenal medulla v extra-adrenal; p = 0.64), and malignant potential (p = 1).
Twenty two of the patients showed Rb staining. Loss of Rb expression was noted in 35 of the 39 patients with phaeochromocytomas and eight of the 18 patients with paragangliomas. Therefore, loss of Rb was more frequently detected in phaeochromocytomas than in paragangliomas (p = 0.0001). In addition, Rb positive tumours were often non-functional (p = 0.0001) and of smaller size (p = 0.001). There was no significant difference between patients with Rb positive and Rb negative tumours with respect to age at presentation (p = 0.58), sex (p = 1), clinical presentation (p = 0.78), bilaterality (p = 1), or malignant potential (p = 0.5).
Thirty seven of the patients had mdm2 nuclear staining. Mdm2 staining was found in 20 of the 26 patients with paragangliomas and 17 of the 39 patients with phaeochromocytomas. Thus, the incidence of mdm2 staining was significantly higher in paragangliomas than in phaeochromocytomas (p = 0.011). The staining was also detected frequently in non-functional tumours (p = 0.024) and those of a smaller size (p = 0.038). However, the patients with mdm2 positive tumours did not differ from those with mdm2 negative tumours with respect to age at presentation (p = 0.85), sex (p = 1), or clinical presentation (p = 0.28). There was no significant difference between the two groups of tumours in terms of the malignant potential (p = 0.058) and the occurrence of bilateral tumours (p = 1). Overall, there was no significant correlation between p53, Rb, and mdm2 staining in the tumours studied.
Survival data were available for 50 of the 65 patients. Five patients had tumours detected at necropsy and two patients died shortly after surgery. The median follow up of the remaining 43 patients was 5.2 years. The survival rate was not related to the patients' sex (p = 0.12), clinical presentation (p = 0.86), functional status (p = 0.28), or tumour location (p = 0.75). However, survival was related to the malignant potential of the tumour (p = 0.01) (fig 1A). Five patients with benign tumours were detected at necropsy, whereas other patients with benign tumours survived and had no recurrence of disease during the follow up period (median follow up, 4.6 years; range, 5 months to over 16 years). In comparison, only 45% of patients with malignant tumours survived for 10 years. Although death in patients with malignant tumours occurred mainly within 10 years after the resection of the primary tumour, two patients lived for more than 10 years before dying of tumour related diseases (one with malignant phaeochromocytoma and the other with malignant carotid body tumour died of distant metastases 13 and 16 years, respectively, after surgery). The survival rate of the patients with phaeochromocytomas/paragangliomas appeared to be higher in the patients with p53 negative tumours than those with p53 overexpression (p = 0.04), even though the number of patients involved was quite small. However, survival was not related to either Rb expression (p = 0.8) or mdm2 expression (p = 0.26) (fig 1B–D).
The reported incidence of malignancy in phaeochromocytomas/paragangliomas varies among different series.18 The variations may be related to differences in the length of the follow up period and the proportions of adrenal and extra-adrenal tumours in the series. The traditional figure of 10% is close to the average of most studies.15 Paragangliomas have a slightly higher rate of malignancy, from 14% to 50%.16 In our series, which contained a large number of patients (predominately Chinese), and with longterm follow up in many cases, the rate of malignancy was found to be 20%. This rate was similar both for phaeochromocytomas and paragangliomas. The only criterion that determined the prognosis of the patients with phaeochromocytomas/paragangliomas was the biological aggressiveness of the disease. Patients with benign tumours had longer term survival, whereas the 10 year survival rate of patients with malignant tumours was only 45%. Some patients with malignant tumours died of metastatic tumours more than 10 years after the resection of their primary tumours.
p53 alterations in phaeochromocytomas and paragangliomas have been studied in a few series (table 2).4–14 The number of patients investigated in these series was usually small (range, one to 36). In many of these studies, no p53 aberration was found. On the contrary, Lin et al in Taiwan and Yoshimoto et al in Japan and China detected p53 mutations in phaeochromocytomas.7, 14 However, the mutation patterns identified in these two studies were different. Both geographical and ethnic factors may have affected the mutational spectrum of p53 in different populations. Our investigation, which studied a larger number of patients than the other studies, showed the presence of p53 overexpression in four of the 65 phaeochromocytomas and paragangliomas (including three patients with phaeochromocytomas and one with paraganglioma). Apart from our case, the other p53 positive paragangliomas reviewed in the English literature include a sporadic paraganglioma reported by Dahia et al and a lung paraganglioma, which occurred after radiation exposure, described by Hagemeyer et al.8, 9
Hereditary factors have been reported in phaeochromocytomas and paragangliomas. Phaeochromocytoma can occur in the setting of MEN2, an autosomal dominant disease of disordered development and tumour formation that principally affects thyroid C cells, the adrenal medulla, and the parathyroids. Mutations in the RET proto-oncogene, located on chromosome 10, have been related to the MEN2 syndrome. The gene encodes a receptor tyrosine kinase expressed in tissues derived from the neural crest. Familial paragangliomas have been reported and were also noted in our study. Recently, germ line mutations in SDHD, a mitochondrial complex II gene, were found to be responsible for this syndrome.19 Other genetic alterations may be important in the pathogenesis of phaeochromocytomas and paragangliomas.
Bilateral phaeochromocytomas were often noted in the patients with MEN2. Lin et al found that p53 positive patients had either multiple or malignant phaeochromocytomas.7 In agreement with this finding, we noted that the p53 positive tumours were found in the patients with bilateral tumours. The p53 positive tumours occurred not only in a patient with bilateral phaeochromocytomas (in MEN2) but also in a patient with bilateral carotid body tumours. Thus, p53 overexpression might play a role in the pathogenesis of multiple phaeochromocytomas and paragangliomas. It could act as a marker for this subtype of tumour and might be associated with poor prognosis.
Although paragangliomas and phaeochromocytomas had identical histological features, we found that they were different in terms of their clinical and genetic attributes. Clinically, paragangliomas were often small, non-functional, presented as incidental findings, and did not show the side predilection seen in phaeochromocytomas. In the patients with paragangliomas, the carotid body was the only location with bilateral tumours.
Our study represents the first examination of phaeochromocytomas/paragangliomas for alterations in the Rb and mdm2 genes. The high overall rate of loss of Rb staining (43 of 65) and mdm2 expression (37 of 65) emphasised their importance in the development of these tumours. Loss of Rb staining was seen more often in phaeochromocytomas (35 of the 39 patients), whereas the presence of mdm2 expression was seen more frequently in paragangliomas (noted in 20 of the 26 patients). This implies that the tumour suppressor Rb and the oncoprotein mdm2 play different roles in the pathogenesis of phaeochromocytomas and paragangliomas. Nevertheless, the alterations in expression of these proteins did not seem to affect the biological aggressiveness and survival of the patients with these tumours.
Mdm2 has recently been shown to interact with tumour suppressor gene products (p53 and Rb) and inhibit their function.3 In our study, no significant relation was found between the alterations of mdm2, p53, and Rb. Recently, it has been noted that mdm2 might function differently in p53 dependent and p53 independent pathways to regulate cellular proliferation.4 In the former pathway, loss of p53 leads to a lack of mdm2 and thus to p53 protein accumulation, and vice versa. In our patients, mdm2 staining occurred in tumours with and without p53 overexpression. Many patients with paragangliomas had mdm2 expression but no p53 protein overexpression. In addition, the propensity of Rb and p53 alterations to occur together in samples from the same patient was noted previously in some tumours.20 In our study, Rb and p53 alterations in phaeochromocytomas and paragangliomas usually occurred in the absence of each other.
Phaeochromocytomas and paragangliomas have distinctive clinical features and genetic alterations. They should not be treated as a single entity although they have identical histological features. The expression profile of Rb and mdm2 might help distinguish the two lesions in some cases. In addition, the expression of p53 could be used as a marker for familial phaeochromocytomas and paragangliomas, and might be useful in assessing the prognosis of these patients.