Familial Paragangliomas: Linkage to Chromosome11q23 and Clinical Implications

Jeff Milunsky,1 Anita L. DeStefano,2 Xin-Li Huang,1 Clinton T. Baldwin,1 Virginia V. Michels,3Geza Jako,1 and Aubrey Milunsky1*1Center for Human Genetics and Department of Pediatrics, Boston University School of Medicine,Boston, Massachusetts

2Departments of Neurology and Epidemiology and Biostatistics, Boston University School of Medicine,Boston, Massachusetts

3Department of Medical Genetics, Mayo Clinic, Rochester, Minnesota

Familial paragangliomas (PGL), or glomustumors, are slow-growing, highly vascular,generally benign neoplasms usually of thehead and neck that arise from neural crestcells. This rare autosomal-dominant disor-der is highly penetrant and influenced bygenomic imprinting through paternal trans-mission. Timely detection of these tumorsaffords the affected individual the opportu-nity to avoid the potential morbidity associ-ated with surgical removal, and mortalitythat may accompany local and distant me-tastases. Linkage to two distinct chromo-somal loci, 11q13.1 and 11q22.3–q23, hasbeen reported, suggesting heterogeneity.We evaluated three multigenerational fami-lies with hereditary PGL, including 19 af-fected, and 59 unaffected and potentially at-risk individuals. Numerous microsatellitemarkers corresponding to each candidateregion were tested in all members of thethree families. Confirmation of linkage to11q23 was established in all three families.The inheritance pattern was consistentwith genetic imprinting. Using these data,we were able to provide presymptomatic di-agnosis with subsequent removal of tumorfrom one individual, and to start severalothers on an MRI surveillance protocol. Am.J. Med. Genet. 72:66–70, 1997.© 1997 Wiley-Liss, Inc.

KEY WORDS: paraganglioma; chromosome11q23; genetic imprinting

INTRODUCTIONFamilial paragangliomas (PGL) (glomus tumors or

chemodectomas) are slow-growing, highly vascular,generally benign neoplasms usually of the head andneck that arise from neural crest cells. Inherited as anautosomal-dominant, highly penetrant disorder, theoccurrence of these tumors is influenced by genomicimprinting through paternal transmission. The inci-dence of PGL ranges from 1:30,000–1:100,000 [Mari-man et al., 1993; Lack et al., 1979]. The average age atdiagnosis has been 34.5 years. The tumors are classi-fied by their location, most of which involve the carotidbodies. Typically, the tumor presents as an asymptom-atic neck mass. There is increased morbidity with in-creased tumor size. Timely detection of these tumorsmay afford the affected individual the opportunity toavoid the potential morbidity associated with surgicalremoval and the mortality that may accompany localand distant metastases. Several studies have indicatedthat head and neck magnetic resonance imaging (MRI)is the most sensitive screening tool in families withfamilial PGL [van Giles et al., 1992, 1994; McCaffrey etal., 1994]. It has been recommended that individuals atrisk for PGL be screened after age 16 with head andneck MRI. The optimal interval between MRI screen-ings remains unclear. Identification of the PGL genewould enable those relatives at risk to be tested andscreened with MRI if positive, while those who testnegative would not need to be screened and would bereleased from chronic anxiety.

Evidence supporting genetic linkage to two distinctloci, 11q13.1 and 11q22.3–q23, suggests heterogeneity[Heutink et al., 1992, 1994; Mariman et al., 1993]. Weused numerous microsatellite markers, correspondingto each candidate region, to evaluate three multigen-erational families with hereditary PGL.

In this paper we describe our study, which alsoaimed at determining the location, and eventually thestructure, of the PGL gene.

MATERIALS AND METHODSThis evaluation of three multigenerational families

with hereditary PGL included 19 affected, and 59 un-

*Correspondence to: Aubrey Milunsky, M.D., D.Sc., Center forHuman Genetics, Boston University School of Medicine, 80 E.Concord St., W-409B, Boston, MA 02118.

Received 26 November 1996; Accepted 19 February 1997

American Journal of Medical Genetics 72:66–70 (1997)

© 1997 Wiley-Liss, Inc.

affected and potentially at-risk individuals (Fig. 1). Theaffected individuals were diagnosed with PGL due toeither an asymptomatic neck mass or symptoms local-izing or radiating to the ear. In light of their familyhistory, the presence of tumor was confirmed by a neckultrasound study or MRI.

Blood samples were processed at the Center for Hu-man Genetics at Boston University, and DNA was pre-pared using Puregene reagents (Gentra Systems, Inc.,Research Triangle Park, NC). High-resolution chromo-somes using G-banding from the proposita in family 1(III-22) were normal. Numerous highly polymorphicmicrosatellite markers [Dib et al., 1996], correspondingto 11q13.1 and 11q22.3–q23, were tested in all avail-able relatives of the three families.

Two-point linkage analysis was performed using theMLINK program of the FASTLINK (version 3) [Cot-tingham et al., 1993] package. An age-dependent pen-etrance function was assumed (Table I). In order tomodel genetic imprinting, offspring of affected or at-risk females were assigned to a liability class with 0penetrance. Alleles within each marker were assumedto have equal frequency.

Additional linkage analyses were performed to ex-amine some of the assumptions made. For markersshowing significant evidence of linkage, allele frequen-cies were varied to determine the robustness of theLod-score results. A genetic model which did not incor-porate imprinting was also used. For this model, thoseindividuals assigned to liability class 6 (Table I) werereassigned to liability classes 1–5 based on age at lastexamination, regardless of the sex of the gene-carrierparent.

RESULTS

In these three families (Fig. 1), the range of the ageof tumor diagnoses was between 16–49 years (Table II).In family 2, an earlier age at diagnosis was observed inthe younger generation. Linkage analysis yielded acombined maximum lod score of 5.5 at u 4 0 withmarker D11S1340 (Table III). Affected individuals (IV-6, IV-8) in one branch of family 3 do not share anyalleles in common with other affected relatives atD11S1356. We infer that their father (III-4), who doesnot express the disease due to imprinting, is a recom-binant individual. An affected individual recombinantat D11S295 was also observed in family 2. Given thesedata, we conclude that the PGL gene is proximal tomarker D11S1356. Linkage to the chromosome11q13.1 region was excluded in these three families.Varying allele frequencies did not have a large impacton the lod scores. At marker D11S1340, for example,varying the frequency of the apparently linked allelefrom .1–.5 yielded a combined LOD score of 5.7–5.0.Significant evidence of linkage to the 11q23 region wasalso observed when a nonimprinting model was as-sumed (Zmax 4 5.34 at u 4 0 at D11S1340). Lod scoreswere generally slightly lower than those obtained whenimprinting was incorporated into the genetic model (re-sults not shown).

As a result of this study, we identified several at-riskindividuals based on haplotype analysis. Genetic coun-

seling was provided and head and neck MRI was usedas a screening tool for these individuals. In family 1, wewere able to provide presymptomatic diagnosis withsubsequent removal of tumor from one individual (III-17).

DISCUSSION

Most PGL cases are sporadic, with an estimated 10%being familial [Sobol and Dailey, 1990]. However, thetrue proportion of inherited cases may be underesti-mated, since the gene may be passed through severalgenerations of females before tumors become manifest,thus obscuring the family history from many genera-tions before. PGL are typically classified by their loca-tion: carotid body, jugular (arising from the glomusjugulare), vagal body, orbital, and laryngeal. The mostcommon of the paragangliomas, the carotid body tu-mor, usually presents as an asymptomatic neck mass[McCaffrey et al., 1994]. Enlarging tumors often pro-duce compressive symptoms resulting in hearing loss,tinnitus, or facial nerve paralysis. Most of our patientshad/have carotid body tumors, with several having glo-mus jugulare tumors either alone or together with acarotid body tumor. Most presented with an asymp-tomatic neck mass. Several had hearing loss and cra-nial nerve palsies, and a few have had recurrences ofthe carotid body tumors, some 10–15 years after initialsurgical removal. One of the individuals in family 1(III-23) had central sleep apnea documented by a sleepstudy before removal of a recurrent carotid body tumor.Obstructive sleep apnea due to a carotid body paragan-glioma has been reported [Metersky et al., 1995]. Bi-lateral carotid body paraganglioma and central alveo-lar hypoventilation have also been reported [Roncoroniet al., 1993].

Generally, familial paragangliomas are benign neo-plasms with between 4–10% manifesting local and dis-tant metastases. There is increased morbidity with in-creased tumor size. With larger tumors, injury to thecranial nerves and baroreceptor failure secondary tobilateral loss of carotid sinus function have been re-ported after resection of these head and neck paragan-gliomas [Netterville et al., 1995]. Most affected indi-viduals in our kindreds have been left with variouscranial nerve injuries, hoarseness, deafness, and baro-receptor failure. It is clear that timely detection ofthese tumors before they either enlarge or metastasizeis of paramount importance to reduce associated mor-bidity and mortality.

Previous analyses [Heutink et al., 1992, 1994; Meri-man et al., 1993, 1995; Devilee et al., 1994; Baysal etal., 1997] established genetic linkage to two distinctloci, 11q13.1 and 11q22.3–q23, suggesting heterogene-ity. The results from these three families confirm link-age of PGL to the chromosome 11q23 region and ex-clude linkage to the 11q13.1 region. Evidence for pa-ternal genetic imprinting has been presented [Van derMey et al., 1989] and was inferred by Herrmann [1977]in families with PGL. Our results also indicate that theinheritance of paraganglioma in these families is con-sistent with genetic imprinting. Offspring who inher-ited the disease haplotype from their mother do not

Familial Paragangliomas 67

Fig

.1.

Ped

igre

esof

the

PG

Lfa

mil

ies.

Top

:Fam

ily

1;M

idd

le:f

amil

y2;

Bot

tom

:fam

ily

3.A

geat

diag

nos

isis

show

nbe

low

the

sym

bols

for

affe

cted

indi

vidu

als,

ifkn

own

.Pro

posi

tus

isin

dica

ted

byan

arro

w.

manifest the disease. However, male offspring mayhave affected children. This is clearly demonstrated infamily 2, where no offspring of II-2 manifest the dis-ease, yet 3 of these individuals have affected offspring.

Anticipation, or earlier onset of disease in subse-quent generations, has been suggested in familial para-ganglioma [Baysal et al., 1997]. This trend was ob-served in at least one of our families. Whether this isdue to true genetic anticipation or represents ascer-tainment bias cannot be determined from these data.

Screening MRI has been recommended for patientswith a family history of paraganglioma [McCaffrey etal., 1994; van Giles et al., 1994]. In these families, hap-lotype analysis has proven critical for identifying indi-viduals who carry the disease gene and those who donot among those who are at-risk. Those with the dis-ease haplotype can benefit from an MRI surveillanceprotocol. As a result of this study, we identified severalat-risk individuals based on haplotype analysis. Aftergenetic counseling, MRI was used as a screening toolfor these individuals. In family 1, we were able to pro-vide presymptomatic diagnosis with subsequent re-moval of tumor from one individual. First experienceswith genetic counseling based on predictive DNA diag-nosis for paragangliomas have been reported [Ooster-wijk et al., 1996]. Until the PGL gene is cloned, use ofDNA haplotype analysis, with careful consideration offamily history and genetic imprinting, will guide the

clinical management of these patients. An MRI surveil-lance protocol will benefit those not subject to imprint-ing with the disease haplotype. Those individuals atrisk without the disease haplotype may be monitoredwith careful physical examinations while being sparedthe expense of MRI screening and the continued anxi-ety associated with this diagnosis.

These confirmatory linkage studies, taken togetherwith other published studies, represent an importantstep in identifying the gene causing paraganglioma inthese families. This will permit the precise diagnosis ofthe disease, facilitate presymptomatic diagnosis, andallow for the determination of genotype/phenotype cor-relations. Successful cloning of the gene will enable thedesign of an animal model in which therapies can betested.

ACKNOWLEDGMENTS

We thank Dr. James Guerrini and Kiran Milunskyfor their help in the collection of blood samples fromfamily 1.

REFERENCES

Baysal BE, Farr JE, Rubinstein W, Galus R, Johnson K, Aston C, MyersEN, Johnson JT, Carrau R, Myssiorek D, Kirkpatrick S, Singh D, SahaS, Gollin SM, Evans GA, James MR, Richard CW III (1997): Fine map-

TABLE II. Age of Tumor Diagnosis

Pedigree Individual Age of tumor diagnosis

1 II-2 33II-7 48II-9 42III-15 28III-17 48III-22 16III-23 31IV-4 18

2 II-2 48III-1 42III-3 49III-5 42IV-7 24IV-9 28

3 III-12 20IV-1 32IV-6 28IV-8 25IV-11 19IV-18 29

TABLE I. Liability Classes Defined to Account forAge-Dependent Penetrance

Liability class Criteria Penetrance

1 <20 years of age 0.12 20–30 years of age 0.33 30–40 years of age 0.54 40–50 years of age 0.65 >50 years of age 0.96a Offspring of affected or

at-risk females0.0

aThis liability class was used only for the genetic imprinting model.

TABLE III. Two-Point lod Scores Obtained Assuming a GeneticModel With Age-Dependent Penetrance and Imprinting

Marker Family

Recombination fraction (u)

0.0 0.01 0.05 0.10 0.2 0.3

956 1 −5.18 −2.56 −1.20 −0.65 −0.22 −0.102 −1.80 −0.59 0.02 0.23 0.34 0.303 −1.12 −0.29 0.25 0.40 0.38 0.22

Total −8.10 −3.44 −0.93 −0.01 0.50 0.42527 1 −2.88 −1.44 −0.21 0.20 0.38 0.29

2 −1.69 −0.64 −0.04 0.18 0.30 0.273 −1.81 −1.25 −0.68 −0.41 −0.18 −0.07

Total −6.38 −3.33 −0.92 −0.04 0.50 0.491,327 1 0.54 0.53 0.48 0.42 0.29 0.15

2 −0.08 −0.08 −0.08 −0.07 −0.05 −0.033 0.69 0.67 0.59 0.49 0.30 0.15

Total 1.15 1.12 1.00 0.84 0.54 0.27938 1 0.52 0.51 0.44 0.37 0.21 0.08

2 0.30 0.29 0.26 0.21 0.13 0.063 0.20 0.19 0.16 0.12 0.07 0.03

Total 1.02 0.99 0.86 0.71 0.41 0.171,885 1 −0.11 −0.11 −0.10 −0.09 −0.06 −0.03

2 1.08 1.06 0.97 0.87 0.65 0.433 2.17 2.12 1.94 1.70 1.20 0.69

Total 3.13 3.06 2.81 2.48 1.80 1.10939 1 1.99 1.95 1.80 1.61 1.17 0.69

2 1.08 1.06 0.97 0.87 0.65 0.433 1.10 1.08 1.00 0.90 0.68 0.43

Total 4.17 4.09 3.78 3.37 2.50 1.551,340 1 2.58 2.53 2.35 2.11 1.58 0.97

2 0.69 0.68 0.64 0.59 0.47 0.343 2.27 2.22 2.06 1.84 1.36 0.85

Total 5.54 5.44 5.05 4.53 3.41 2.161,356 1 2.60 2.55 2.37 2.12 1.59 0.98

2 1.03 1.01 0.93 0.83 0.62 0.413 −1.06 −0.62 −0.14 0.06 0.16 0.14

Total 2.57 2.94 3.17 3.01 2.37 1.53

Familial Paragangliomas 69

ping of an imprinted gene for familial non-chromaffin paragangliomas,on chromosome 11q23. Am J Hum Genet 60:121–132.

Cottingham RW Jr, Idury RM, Schaffer AA (1993): Faster sequential link-age computations. Am J Hum Genet 53:252–263.

Devilee P, Evert M, van Schothorst EM, Bardoel AFJ, Bonsing B, Kuipers-Dijkshoorn N, James MR, Fleuren G, van der Mey AGL, Cornelisse CJ(1994): Allelotype of head and neck paragangliomas: Allelic imbalanceis confined to the long arm of chromosome 11, the site of the predis-posing locus PGL. Genes Chromosomes Cancer 11:71–78.

Dib C, Faure S, Fizames C, Sampson D, Drouot N, Vignal A, Millasseau P,Marc S, Hazan J, Sebaoun E, Lathrop M, Gyapay G, Morrissette J,Weissenbach J (1996): A comprehensive genetic map of the humangenome based on 5,264 microsatellites. Nature 380:152–154.

Herrmann J (1977): Delayed mutation model: Carotid body tumors andretinoblastoma. In Mulvihill JJ, Miller RW, Fraumeni JF Jr (eds): ‘‘Ge-netics of Human Cancer.’’ New York: Raven Press, pp 417–438.

Heutink P, van der Mey AGL, Sandkuijl LA, van Giles APG, Bardoel A,Breedveld GJ, van Vliet M, van Ommen GJB, Cornelisse CJ, OostraBA, Weber JL, Devilee P (1992): A gene subject to genomic imprintingand responsible for hereditary paragangliomas maps to chromosome11q23–qter. Hum Mol Genet 1:7–10.

Heutink P, van Schothorst EM, van der Mey AGL, Bardoel A, Breedveld G,Pertis J, Sandkuijl LA, van Ommen GJB, Cornelisse CJ, Oostra BA,Devilee P (1994): Further localization of the gene for hereditary para-gangliomas and evidence for linkage in unrelated families. Eur J HumGenet 2:148–158.

Lack EE, Cubilla AL, Woodruff JM (1979): Paragangliomas of the head andneck region. A pathologic study of tumors from 71 patients. HumPathol 10:191–218.

Mariman ECM, van Beersum SEC, Cremers CWRJ, van Baars FM, RopersHH (1993): Analysis of a second family with hereditary non-chromaffinparagangliomas locates the underlying gene at the proximal region ofchromosome 11q. Hum Genet 91:357–361.

Mariman ECM, van Beerseum SEC, Cremers CWRJ, Stuycken PM, RopersHH (1995): Fine mapping of a putatively imprinted gene for familialnon-chromaffin paragangliomas to chromosome 11q13.1: Evidence forgenetic heterogeneity. Hum Genet 95:56–62.

McCaffrey TV, Meyer FB, Michels VV, Piepgras DG, Marion MS (1994):Familial paragangliomas of the head and neck. Arch Otolaryngol HeadNeck Surg 120:1211–1216.

Metersky ML, Castriotta RJ, Elnaggar A (1995): Obstructive sleep apneadue to a carotid body paraganglioma. Sleep 18:53–54.

Netterville JL, Reilly KM, Robertson D, Reiber ME, Armstrong WB, ChildsP (1995): Carotid body tumors: A review of 30 patients with 46 tumors.Laryngoscope 105:115–126.

Oosterwijk JC, Jansen JC, van Schothorst EM, Oosterhof AW, Devilee P,Bakker E, Zoeteweij MW, van der Mey AGL (1996): First experienceswith genetic counselling based on predictive DNA diagnosis in heredi-tary glomus tumors (paragangliomas). J Med Genet 33:379–383.

Roncoroni AM, Montiel GC, Semeniuk GB (1993): Bilateral carotid bodyparaganglioma and central alveolar hypoventilation. Respiration 60:243–246.

Sobol SM, Dailey JC (1990): Familial multiple cervical paragangliomas:Report of a kindred and review of the literature. Otolaryngol HeadNeck Surg 102:382–390.

Van der Mey AG, Maaswinkel-Mooy PD, Cornelisse CJ, Schmidt PH, vander Kamp JJ (1989): Genomic imprinting in hereditary glomus tu-mours; evidence for new genetic theory. Lancet ii:1291–1293.

Van Giles AP, van der Mey AG, Hoogma RP, Sandkuijl LA, Maaswinkel-Mooy PD, Falke TH, Pauwels EK (1992): MRI screening of kindred atrisk of developing paragangliomas: Support for genomic imprinting inhereditary glomus tumors. Br J Cancer 65:903–907.

Van Giles AP, van den Berg R, Falke TH, Bloem JL, Prins HJ, Dillon EH,van der Mey AG, Pauwels EK (1994): MRI diagnosis of paragangliomaof the head and neck: Value of contrast enhancement. AJR 162:147–153.

70 Milunsky et al.