Laboratory of Molecular Otology, Epstein Laboratories, San Francisco, California
Division of Otology, Neurotology, Skull Base Surgery, San Francisco, California
Anil K. Lalwani, MD, Division of Otology, Neurotology, and Skull Base Surgery, Department of Otolaryngology—Head and Neck Surgery, University of California San Francisco, 400 Parnassus Avenue, Room A-730, San Francisco, CA 94143-0342, U.S.A.
Hypothesis Similar to familial tumors, sporadic head and neck paragangliomas are associated with chromosomal deletions at either 11q13 or 11q22-23.
Background Familial paragangliomas are inherited in an autosomal dominant pattern with genomic imprinting of the maternal allele. Genetic studies of familial paragangliomas have localized the causative genetic defect to two separate loci: 11q13.1 and 11q22-23. The molecular pathogenesis of sporadic head and neck paragangliomas has not been studied.
Methods Blood and tumor samples from patients with sporadic head and neck paragangliomas were screened for deletions on chromosome 11 using DNA microsatellite markers and polymerase chain reaction. Polymerase chain reaction–amplified alleles from tumor specimens were compared with those from the blood of eight patients. A greater than 50% reduction in band intensity (as determined by densitometric analysis) between blood and tumor sample was indicative of a chromosomal deletion.
Results Three of the eight patients were found to have deletions at chromosome 11q: two at chromosome 11q22-23 and one at 11q13.
Conclusions Sporadic head and neck paragangliomas are associated with deletions at chromosome 11q13 and 11q22-23. It is thus likely that sporadic and familial paragangliomas share a similar molecular pathogenesis.
Nonchromaffin paragangliomas are slow-growing, usually benign tumors arising from neural crest cells. The incidence of paragangliomas in whites is approximately 1:30,000. In the head and neck, these most commonly arise from the carotid body (carotid body tumors), jugulotympanic region (glomus tympanicum), jugular bulb (glomus jugulare), and vagal nerve (glomus vagale).
Although head and neck paragangliomas may most commonly present in sporadic form, the molecular genetic basis for tumorigenesis has been only studied in familial paragangliomas. Transmission of familial tumors occurs through autosomal dominant inheritance with genomic imprinting of the maternal allele. To develop tumors, the individual must inherit the mutated gene from the father. Independent analysis has localized the gene to chromosome 11q. 1–3 The two loci responsible for paraganglioma formation, designated PGL1 and PGL2, have been mapped to chromosome 11q22-q23 and 11q13.1, respectively. 1–3 Most studies have implicated 11q22-q23 as the predominant locus for familiar paraganglioma, but in one family the genetic defect has been linked to 11q13. 3
At present no studies have reported on the molecular basis of sporadic paraganglioma formation. As has been the case for other cranial base tumors such as neurofibromatosis type II, the genetic loci implicated in familial tumors may also be involved in the formation of sporadic paragangliomas. Therefore, we screened eight patients with sporadic head and neck paragangliomas for deletions at chromosome 11q13 and 11q22-23. Our understanding of the molecular pathogenesis in sporadic head and neck paraganglioma formation is discussed.
MATERIALS AND METHODS
Eight individuals with sporadic head neck paragangliomas were enrolled in our study at the University of California at San Francisco. The human study protocol was approved by the University of California at San Francisco Institutional Review Board. Clinical records were obtained from all individuals. Blood (20–40 mL) and tumor samples were obtained from all individuals. Genomic DNA was isolated from blood and tumor samples. A sample was also obtained from a normal volunteer for use as a positive control.
Polymerase Chain Reaction
DNA polymorphic markers, previously associated with deletions in familial glomus tumors, were selected from chromosome 11q13 and 11q22-23. 4 Polymerase chain reaction (PCR) amplification of these markers was carried out on the eight participating individuals. PCR primers were end-labeled with [α-32P] dATP. PCR reactions were performed in a final volume of 10 μL containing 0.2 μg of genomic DNA, 200 μmol/L 4 dNTP mix, 1X PCR buffer (Perkin Elmer Cetus, Norwalk, CT), 0.5 U of AmpliTaq DNA polymerase (Perkin Elmer Cetus, Norwalk, CT), and 100 nM of each primer. The PCR reaction parameters were 5 minutes at 95°C, followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 55°C for 1 minute, and extension at 72°C for 7 minutes, for all primers. Amplified fragments were run on standard 6% denaturing polyacrylamide gels. Gels were dried for 1 hour.
Assessment of Allelic Imbalance
Gels were exposed to a Storm 860 PhosphorImager (Molecular Dynamics, Sunnyvale, CA) to assess for allelic imbalance. The intensity of 32P emission for each allelic band was quantified by the Storm 860 PhosphorImager. A numerical comparison of 32P emission from each amplified allelic band was obtained for blood versus tumor sample (Fig. 1). If there existed a greater than 50% reduction in intensity of tumor versus blood signal, a significant allelic imbalance was considered to be present. This indicated that a deletion at that DNA marker site was present, as described in previous studies. 2 The PhosphorImager readings of 32P emission were confirmed for linearity. This was accomplished by serial dilutions of a concentrated 32P control. The PhosphorImager readings of the 32P serial dilutions spanned the full range of signal intensity read from the allelic bands of all eight patients.
The clinical characteristics of the eight patients is summarized in Table I. The average age at presentation was 58.8 years old. Seven patients were female and one was male. The distribution of tumor type included four glomus jugulare tumors, three carotid body tumors, and one glomus tympanicum tumor. No patient presented with multiple paragangliomas and all tumors were benign by histopathological analysis.
Table Table 1.. Clinical Characteristics of Patients With Sporadic Paragangliomas.
The degree of allelic imbalance in the eight screened tumor samples is shown in Figure 2. Two patients were found to have significant allelic imbalance for chromosome 11q22-23. P1 (carotid body tumor) was found to have a 50% to 70% reduction in intensity spanning at least a 0.8 MB at chromosome 11q23. Reduction in intensity of the allele was at 50% to 70% for three consecutive DNA markers. Although P1 had a minor allelic imbalance identified at D11S2000, this did not reach significant levels. P2 (glomus jugulare) (Fig. 1) was found to have a 70% to 100% allelic imbalance at DNA marker D11S560. No other patients were observed to have allelic imbalance at chromosome 11q22-23.
One patient (P3, glomus jugulare) was observed to have a 50% to 70% allelic imbalance at DNA marker D11S480 at chromosome 11q13. While no other patient was observed to have significant allelic imbalance at chromosome 11q13, P5 (glomus jugulare) was found to have a 30% to 50% reduction of allelic intensity for DNA marker D11S480.
Recent genetic studies of familial head and neck paragangliomas have significantly increased our understanding of the molecular mechanisms responsible for tumorigenesis. Linkage studies of families with familial paragangliomas have identified two separate locations or loci (PGL1/PGL2) on chromosome 11 that harbor the gene responsible for paraganglioma formation. 1–3 Further studies have shown that tumor formation is associated with loss of DNA at these loci. 2,5 This phenomena is termed allelic imbalance. Two studies have reported on deletional analysis of familial paragangliomas, but no study has reported on deletions in sporadic head and neck paragangliomas.
In the present study we postulated that sporadic tumors have deletions at the same or closely related loci as their familial counterparts. In screening for chromosome 11q13 and 11q22-23 deletions, we were able to establish allelic imbalance in three of eight tumor samples. In all three cases, the degree of allelic imbalance was quite severe (>50%), indicating that a deletion had taken place in these tumor specimens. This conservative threshold of allelic imbalance as an indicator for chromosomal deletions is well described in other studies. 2 Two samples had deletions at chromosome 11q22-23, with the largest span being at least 0.8 MB in distance. A third tumor was found to harbor a deletion at chromosome 11q13. These data strongly suggest that deletions at PGL1 and PGL2 contribute to the formation of these tumors, being closely related to the loci previously identified in familial paragangliomas.
Previous studies have reported that the incidence of deletions in familial tumors ranges from 26% to 82%. 2,5 The rate of detection in our study is within the reported range from familial data. This wide range of reported deletions may be explained by variability in the number of DNA markers used for tumor screening: as more DNA markers are used in the screening of the tumor samples, the likelihood of detection of deletions increases. In our study we employed DNA microsatellite markers encompassing the full region of chromosome 11q22-23, with less intensive screening of chromosome 11q13. The remaining tumors without identified deletions may have other mutations at PGL1/PGL2 not screened for in our study, such as point mutations or mutations in unscreened areas such as intronic segments.
A variety of reports have recently been published on the genetic defects involved in other neuroendocrine tumors. 6,7 The gene responsible for the multiple endocrine neoplasia syndrome 1 (MEN1) has been recently cloned and found to function as a tumor suppressor gene at chromosome 11q13. As gastric neuroendocrine tumors are associated with MEN1, D'Adda et al. 7 recently screened for deletions at chromosome 11q13-14 for gastric tumors not associated with MEN1. Deletions were found for certain subtypes of gastric neuroendocrine tumors but not others. Genetic defects in pheochromocytomas have also been recently studied. Vargas et al. 6 found deletions at chromosome 1p and 3p in pheochromocytomas. A recent study of a malignant retroperitoneal paragangliomas by Blasius et al. 8 found chromosome 1p deletions in this tumor. Therefore, a number of possible loci may be involved in the molecular pathogenesis of neuroendocrine tumors. The mutations in some tumors localize to chromosome 11q, but others are located on entirely different chromosomes. In understanding genetic basis of the formation of neuroendocrine tumors, it is clear that a complex set of mutations and their interactions is needed to explain the very diverse types and behavior of this class of tumors.
The understanding of the molecular mechanisms involved in paragangliomas formation will have a wide variety of clinical implications. We have recently reported on the use of genetic screening by linkage analysis to diagnose familial paragangliomas in individuals before symptoms occur. 9 Presymptomatic diagnosis may lead to earlier detection of these tumors and thus reduce the morbidity and mortality associated with head and neck paragangliomas. Once the PGL1 and PGL2 genes are cloned, genotype-phenotype correlation studies will shed light on which types of mutations predispose to more aggressive or malignant paragangliomas and thus provide more accurate prognostic information for individuals with familial or sporadic paragangliomas. Molecular regulation of tumor behavior, however is likely to be very complex, involving multiple genes and making simple correlations less likely.