molecular genetic analysis of von hippel–lindau disease

Journal of Internal Medicine 1998; 243: 527–533

© 1998 Blackwell Science Ltd 527

Molecular genetic analysis of von Hippel–Lindau disease

F. M. RICHARDS a , A. R. WEBSTER b , R. MCMAHON c , E . R. WOODWARD d , S. ROSE e & E. R. MAHER a

From the aDivision of Medical and Molecular Genetics, University of Birmingham Department of Paediatrics and Child Health, The Medical School,Birmingham; the bDepartment of Ophthalmology and cMolecular Genetics Laboratory, Addenbrooke’s Hospital, Cambridge; the dCambridge UniversityDepartment of Pathology, Cambridge; and the eRegional Molecular Genetics Laboratory, Birmingham Women’s Hospital, Birmingham, UK

MINISYMPOSIUM MEN & VHL

Abstract. Richards FM, Webster AR, McMahon R,Woodward ER, Rose S, Maher ER (BirminghamUniversity; Birmingham Women’s Hospital,Birmingham; Addenbrooke’s Hospital, Cambridge;Cambridge University, Cambridge, UK). Moleculargenetic analysis of von Hippel–Lindau disease(Minisymposium: MEN & VHL). J Intern Med 1998;243: 527–33.

Von Hippel–Lindau (VHL) disease is a dominantlyinherited multisystem family cancer syndromepredisposing to retinal and central nervous systemhaemangioblastomas, renal carcinoma, phaeochro-mocytoma, pancreatic islet cell tumours and endolym-phatic sac tumours. In addition, renal, pancreatic andepididymal cysts occur. Morbidity and mortality fromVHL disease can be reduced by the identification andsurveillance of affected individuals and at-risk rela-tives so that compli-cations are diagnosed at an earlypresymptomatic stage. The detailed mapping and sub-sequent isolation of the VHL tumour suppressor genehas enabled molecular genetic analysis in families and patients with definite or possible VHL disease.Initially, linked DNA markers were used in informativefamilies to modify individual risks and then to makeappropriate alterations in surveillance programs.However, currently most DNA analysis involves thecharacterisation of germline mutations. World-wide,mutations have been identified in almost 500 families(including 132 in our laboratory). These studies haverevealed considerable heterogeneity both in the type

and in the location of mutations within the VHL gene.In our experience, most recurrent mutations resultfrom de novo mutations at hypermutable sequences,although a founder effect for the Tyr98His (‘BlackForest’) mutation has been reported in German andAmerican families. Although many mutations arepredicted to impair the ability of pVHL to combinewith the elongin regulatory subunits, analysis ofgenotype–phenotype relationships suggests that theVHL protein has multiple and tissue specific functions.Calculation of tumour risks for different classes ofVHL mutations has provided important prognosticinformation especially with respect to the likelihood ofphaeochromocytoma. However, there is evidence thatretinal involvement does not correlate with allelic het-erogeneity, but that the variability in retinalangiomatosis is influenced by modifier gene effects.VHL gene mutation analysis also provides a basis for investigating the genetic basis of familialphaeochromocytoma and renal cell carcinoma, andapparently isolated retinal angiomas. Results to date suggest that a substantial proportion ofpatients with familial pheochromocytoma have VHL gene mutations but in contrast, most familialclusters of clear cell renal cell carcinoma (RCC)without evidence of VHL do not have germline VHL mutations.

Keywords: haemangioblastoma, phaeochromo-cytoma, renal cell carcinoma, VHL tumour suppres-sor gene, von Hippel–Lindau disease.

F. M. RICHARDS et al.: MOLECULAR GENETIC ANALYSIS OF VHL DISEASE528

© 1998 Blackwell Science Ltd Journal of Internal Medicine 243: 527–533

Introduction

Von Hippel–Lindau (VHL) disease is an autosomaldominant inherited familial cancer syndrome predis-posing to retinal and central nervous system hae-mangioblastomas, renal cell carcinoma (RCC) andrenal cysts, phaeochromocytoma and pancreaticcysts and tumours, epididymal cysts and endolym-phatic sac tumours [1, 2]. The birth incidence ofVHL disease is thought to be one in 36 000 per year[3] with an estimated de novo mutation rate of 4.4 31026 gametes per generation [3, 4]. VHL diseaseshows variable expression and interfamilial differ-ences in phenotype, most notably for the frequency ofphaeochromocytoma [1, 5–7]. The relative frequen-cies and mean ages at clinical diagnosis of the majorcomplications of VHL disease are listed in Table 1.Early detection and treatment of VHL complications,particularly retinal haemangioblastomas and RCC,can reduce morbidity and mortality from VHL dis-ease. Hence all VHL patients and at-risk relativesrequire careful follow-up and regular clinical andradiological surveillance to detect complicationsearly. We and others have described appropriate sur-veillance programs [1, 8]. As a result of increasedsurveillance there is a trend towards earlier diagnosisand the mean age at diagnosis of subclinical lesionsis earlier than those described in Table 1. The leadingcause of death in VHL disease is RCC. The lifetimerisk of RCC in VHL disease is .70% and bilateralrenal tumours are frequent. However, the prognosisof patients with presymptomatically diagnosed renaltumours is significantly better than that of patientswith symptomatic lesions (see Fig. 1).

Surveillance of individuals at risk of VHL diseasecommences in childhood and is continued annuallyuntil the 7th decade for relatives without evidence of VHL disease. The detailed genetic mapping andsubsequent identification of the VHL tumoursuppressor gene at chromosome 3p25–p26 in 1993

[9] has enabled direct genetic testing in VHL families.Initially such testing was performed using linkedDNA markers but more recently direct mutationanalysis has become the standard approach. For at-risk relatives who are shown not to be genecarriers, surveillance can be discontinued. Hencemolecular genetic diagnosis enables screeninginvestigations to be targeted to those individuals mostat risk and prevents many years of screening inrelatives who are shown not to be gene carriers. Theuptake of predictive DNA testing in VHL families ishigh, in contrast to that of other familial cancersyndromes (e.g. Li–Fraumeni syndrome) for whichthe benefits of screening are perceived to be lessestablished [10].

Linkage analysis

The VHL gene was mapped to the short arm ofchromosome 3 in 1988 [11]. Subsequently, we and others localized the gene to 3p25–p26 [12–15] and identified flanking markers D3S1038and D3S18 [16]. The VHL gene was then mappedfurther by identification of VHL patients withgermline deletions detected by Pulsed-Field GelElectrophoresis [17, 18], and the VHL tumoursuppressor gene was cloned from the region of dele-tion in 1993 [9].

Polymorphic markers close to the VHL gene cannow be used to identify at risk individuals in a VHL family by genetic linkage analysis (Fig. 2). The most informative microsatellite markers areD3S1317 and D3S1038, D3S587 and D3S1435 all of which are thought to be within 500 kb ofthe VHL gene [4]. There are also two intragenic VHL polymorphisms which can be assayed bypolymerase chain reaction (PCR); VHL 1149 A/G[19], VHL19 A/G [20], and a TaqI or PstI polymor-phism which can be assayed by Southern blotting[21].

Table 1 Relative frequencies and mean age at diagnosis of VHL complications. Data from a large UK series (Maher et al. [1]) and from aliterature review (Lamiell et al. [36])

UK series Literature review Mean age of onset (years)Lesion (n 5 152) (n 5 554) (Maher et al. [1])

Retinal angioma 89 (59%) 317 (57%) 25.4 6 12.7Cerebellar haemangioblastoma 89 (59%) 304 (55%) 29.0 6 10.0Spinal cord haemangioblastoma 20 (13%) 079 (14%) 33.9 6 12.6Renal cell carcinoma 43 (28%) 133 (24%) 44.0 6 10.9Phaeochromocytoma 11 (7%) 106 (19%) 20.2 6 7.6

MINISYMPOSIUM: MEN & VHL 529


Presymptomatic diagnosis using genetic linkage analysis

Genetic linkage analysis has been used widely forpresymptomatic diagnosis in VHL families. Initiallystudies were restricted to large families collected forresearch studies [22, 23]. Using the best flankingmarkers available in 1992, we were able to study thecarrier risk in a panel of 23 asymptomatic subjectsfrom VHL families who were at 50% prior risk of VHLdisease, and reduced the risk in 11 of the subjects to,2%. More recently Olschwang et al. [24] used themicrosatellite markers flanking the VHL gene toanalyse individuals from 26 families and were able toimprove the accuracy of risk assessment in 90% of99 asymptomatic individuals, with a confidence limit.0.98 in 80% of cases. These results demonstratethe effectiveness of using genetic linkage forpresymptomatic diagnosis in VHL families in whichthe germline mutation has yet to be characterized.

However, genetic linkage-based DNA testing is notpossible for relatives of isolated cases, or in familieswhere the necessary samples are not available. In apopulation based study of VHL disease, Maddock etal. [25] reported that the family structure was suit-able for linkage-based analysis in only 14 of 23 VHLfamilies.

Germline mutations and mutationscreening procedures

The cloning of the VHL tumour suppressor gene [9]has enabled direct mutation screening to be offeredto VHL families. To date, germline VHL gene muta-tions have been characterized in .300 unrelatedkindreds [26], including 132 in our laboratory.Germline VHL gene mutations are heterogeneous inboth type and position, ranging from very large dele-tions to point mutations. Our current mutationscreening protocols have identified the germlinemutation in up to 81% of VHL families [27], but thevariation in mutation types requires the use of anumber of different protocols to achieve this figure.Routinely we screen VHL-affected individuals forgermline VHL mutations using a combination of sin-gle-stranded conformation polymorphism (SSCP)analysis, DNA sequencing, and Southern blotting. Ifno mutation is identified, at-risk individuals in a fam-ily are investigated for genetic linkage studies withflanking polymorphic markers and intragenic VHLpolymorphisms (see above).

Of the 132 VHL families with germline VHL genemutations characterized by our group, 33 families(25% of those with identified mutations) have largedeletions of at least 2 kb. We have previously shownthat these deletions encompass at least one of thethree exons of the VHL gene [28]. These deletions

Fig. 1 Comparison of survival of patients with renal cellcarcinoma (RCC) and von Hippel–Lindau (VHL) disease. j, patients in whom RCC was diagnosed presymptomatically (n 5 37); h, patients in whom RCC was symptomatic at diagnosis(n 5 15).

Fig. 2 Map of polymorphic markersin the VHL region of chromosome3p25–p26.



were detected by Southern blotting of EcoRI-digestedgenomic DNA, probed with a VHL cDNA [9] to detectaltered fragments when compared to the normal 20kb band. The VHL gene is contained within this 20kb fragment and an EcoRI Southern analysis detectsdeletions of either end or the middle of the VHL gene.EcoRI is the enzyme of choice; analysis of a panel of116 patients with BglII did not identify any furtherdeletions, and DraI, HindIII and PvuII only con-firmed the presence of a deletion in a subset of thoseidentified by EcoRI [28]. No other restriction endonu-clease has therefore identified deletions by Southernblotting in patients who appeared normal on EcoRIdigestion, with the exception of the small number ofpatients with deletions of the entire gene. Deletionsof more than 50 kb (which were detected by pulsedfield gel electrophoresis using MluI- or NruI-digestedDNA) were identified in 2 of 80 VHL patientsanalysed [17]. Such large deletions might be bestdetected by fluorescence in situ hybridization (FISH)using a genomic clone containing the VHL gene as aprobe, rather than by the time-consuming method ofpulsed-field gel electrophoresis. The presence of alarge deletion may also be revealed by failure toinherit an allele in family studies using an intragenicpolymorphism.

Many small intragenic mutations may be detectedby amplification of each of the three VHL exons fromgenomic DNA using PCR followed by SSCP analysis.Any DNA fragments in which a band shift wasobserved are then sequenced on both strands to iden-tify the mutation. This procedure detected mutations

in approximately 50% of VHL families [27, 29]. Forany individuals with no detectable mutations bySSCP analysis or Southern blotting, direct DNAsequencing was performed of all three exons, whichincreased our overall mutation detection rate from73 to 81% of all VHL families [27]. Using thesemethods we found that 20 families (15% of thosewith identified mutations) had microdeletions of 1–9bp, eight families (6%) had microinsertions of 1–4 bpand 71 families (54%) had point mutations. Thesemutations were spread throughout the 855nucleotide VHL coding region (213 codons),although none were identified before nucleotide 378(codon 55) or after nucleotide 796 (codon 195), inthose regions of sequence which are not highly con-served between human, mouse and rat. The distribu-tion and frequencies of these 99 intragenic germlineVHL mutations are shown in Fig. 3. The microdele-tions and insertions result in either in-frame dele-tions of 1–3 amino acids (seven families, 5%), orframe-shifts resulting in truncated protein (20 fami-lies, 15%); the point mutations are either nonsensemutations resulting in truncated proteins (23 fami-lies, 17%), missense mutations (46 families, 35%), orpredicted splice-junction mutations (three families,2%).

Although a wide variety of germline VHL genemutations are described, a number of recurrentmutations have been identified (Fig. 3), e.g. a non-sense mutation C694T (Arg161Stop) was detected inseven of the 132 UK VHL families with identifiedmutations (5%) [27, 29] and the missense mutations

Fig. 3 Distribution of intragenicVHL germline mutations in 99families. MS/SD (h), missensemutation or small in-framedeletion; FS/NS (j), frameshift ornonsense mutation.



C712T and G713 A (Arg167Trp/Gln), identified in12 of 132 UK families (9%), in 12 of 85 (14%) of USand two of 27 (7%) of Japanese VHL families [20, 29,30]. We have shown that this is not due to a foundereffect as families which share these mutations do notusually share a conserved haplotype around the VHLgene [4]. Each of these mutations occurs at a hyper-mutable CpG dinucleotide. However, there is one spe-cific VHL gene mutation which has been shown byothers to be due to a founder mutation: the ‘BlackForest’ mutation T505C (Tyr98His), which causes aless severe form of VHL disease lacking RCC [31].This mutation has been observed in many familiesfrom Germany and in some from Pennsylvania, USA,who are thought to originate from the Black Forestregion of Germany.

Genotype–phenotype correlations

In addition to allowing accurate predictive testing,the characterisation of germline VHL mutations mayalso provide some information on the likely pheno-type. We and others have identified significant geno-type–phenotype correlations in VHL disease andconsequently the identification of the mutation in aparticular family may have implications for subse-quent screening and prognosis [20, 27, 29].

Large deletions and mutations predicted to cause atruncated protein are associated with a lower risk ofphaeochromocytoma (6 and 9% at 30 and 50 years,respectively) than missense mutations (40 and 59%,respectively). Missense mutations at codon 167 inparticular are associated with a high risk ofphaeochromocytoma (53 and 82% at 30 and 50years, respectively) [27, 29]. Mutations at this codoncan easily be identified by the loss of an MspI restric-tion enzyme site in a PCR amplified fragment of exon3. As codon 167 mutations occur relatively frequent-ly, testing directly for this mutation using this simpleassay is helpful in VHL families with phaeochromo-cytoma. In our experience the risk of RCC with mis-sense mutations characterized in our laboratory issimilar to that associated with large deletions or pro-tein truncating mutations. However, two missensemutations (Tyr98His and Tyr111His) have beenreported for which there is a high risk of phaeochro-mocytoma and a low risk of RCC [31, 32].

Although the characterisation of a VHL mutationmay provide some indication as to the risk ofphaeochromocytoma, this tumour may occur in fami-

lies with apparently low risk mutations (deletions,nonsense and frameshift mutations) and all patientsrequire surveillance for phaeochromocytoma, partic-ularly prior to surgery. The risks of retinal and cerebel-lar haemangioblastomas in VHL disease have not beencorrelated with allelic heterogeneity. However, thevariable severity of retinal and central nervous systeminvolvement appears to be influenced by environmen-tal or genetic modifiers [A.R. Webster et al., submit-ted]. Hence although the identification of a germlineVHL mutation enables an accurate assessment of car-rier status for at-risk relatives, it is not possible to reli-ably predict disease severity in gene carriers.

Cryptic VHL disease

Molecular genetic studies have revealed that approxi-mately 50% of patients with apparently isolatedfamilial phaeochromocytoma or bilateral phaeochro-mocytoma may have germline VHL gene mutations[33, 34]. In some cases the VHL mutation identified ispredicted to predispose to other manifestations of VHLdisease also (e.g. codon 167 mutations). However, insome cases there is the suggestion that specific mis-sense mutations (e.g. Leu188Val, Val84Leu andSer80Gly) may predispose to phaeochromocytomawithout any of the other manifestations of VHL dis-ease [33–35]. These findings suggest that VHL protein(pVHL) has tissue-specific functions which may be dif-ferentially affected by specific missense mutations.Individuals with bilateral or familial phaeochromocy-toma should therefore be investigated for mutations inthe VHL gene as well as for evidence of multipleendocrine neoplasia type 2 and neurofibromatosistype 1. However, unless there is an early age at onset,germline VHL gene mutations appear to be rare in iso-lated cases of phaeochromocytoma.

Molecular genetic testing for VHL disease is alsoindicated in patients with early onset retinal or cere-bellar haemangioblastoma (,30 years) or familialclear cell RCC. However, we and others find that mostfamilial clear cell RCC is not allelic with VHL disease(E.R.W. et al., in preparation).

Acknowledgements

We thank Action Research, the Cancer ResearchCampaign, Guide Dogs for the Blind and the NationalKidney Research Fund for supporting our researchon VHL disease.



References1 Maher ER, Yates JRW, Harries R, Benjamin C, Harris R, Moore

AT, Ferguson-Smith MA. Clinical features and natural historyof von Hippel–Lindau disease. Quart J Med 1990; 77:1151–63.

2 Decker HJH, Weidt EJ, Brieger J. The von Hippel–Lindautumour suppressor gene. Cancer Genet Cytogenet 1997; 93:74–83.

3 Maher ER, Iselius L, Yates JRW, Littler M, Benjamin C, HarrisR. et al. Von Hippel–Lindau disease: a genetic study. J MedGenet 1991; 28: 443–7.

4 Richards FM, Payne SJ, Zbar B, Affara NA, Ferguson-SmithMA, Maher ER. Molecular analysis of de novo germline muta-tions in the von Hippel–Lindau disease gene. Hum Mol Genet1995; 4: 2139–43.

5 Glenn GM, Daniel LN, Choyke P, Linehan WM, Oldfield E,Gorin MB et al. Von Hippel–Lindau disease: distinct pheno-types suggest more than one mutant allele at the VHL locus.Hum Genet 1991; 87: 207–10.

6 Neumann HPH, Weistler OD. Clustering of features of vonHippel–Lindau syndrome: evidence for a complex geneticlocus. Lancet 1991; 337: 1052–4.

7 Green JS, Bowmer MI, Johnson GJ. Von Hippel–Lindau diseasein a Newfoundland kindred. Can Med Assoc J 1986; 134:133–8.

8 Maher ER, Webster AR, Moore AT. Clinical features and mole-cular genetics of von Hippel–Lindau disease. Ophthalmic Genet1995; 16: 79–84.

9 Latif F, Tory K, Gnarra J, Yao M, Duh F-M, Orcott ML et al.Identification of the von Hippel–Lindau disease tumour sup-pressor gene. Science 1993; 260: 1317–20.

10 Evans DGR, Maher ER, Macleod R, Davies DR, Craufurd D.Uptake of genetic testing for cancer predisposition. J Med Genet1997; 34: 746–8.

11 Seizinger BR, Rouleau GA, Ozelius LJ, Lane AH, Farmer GE,Lamiell JM et al. Von Hippel–Lindau disease maps to the regionof chromosome 3 associated with renal cell carcinoma. Nature1988; 332: 268–9.

12 Hosoe S, Brauch H, Latif F, Glenn G, Daniel L, Bale S et al.Localization of the von Hippel–Lindau disease gene to a smallregion of chromosome 3. Genomics 1990; 8: 634–40.

13 Maher ER, Bentley E, Yates JRW, Latif F, Lerman M, Zbar B etal. Mapping of the von Hippel–Lindau disease locus to a smallregion of chromosome 3p by genetic linkage analysis.Genomics 1991; 10: 957–60.

14 Seizinger BR, Smith DI, Filling-Katz MR, Neumann H, GreenJS, Choyke PL et al. Genetic flanking markers refine diagnosticcriteria and provide insights into the genetics of vonHippel–Lindau disease. Proc Natl Acad Sci USA 1991; 88:2864–8.

15 Richards FM, Maher ER, Latif F, Phipps ME, Tory K, Lush M etal. Detailed genetic mapping of the von Hippel–Lindau diseasetumour suppressor gene. J Med Genet 1993; 30: 104–7.

16 Crossey PA, Maher ER, Jones MH, Richards FM, Latif F, PhippsME et al. Genetic linkage between von Hippel–Lindau diseaseand three microsatellite polymorphisms refines the localisa-tion of the VHL locus. Hum Mol Genet 1993; 2: 279–82.

17 Richards FM, Phipps ME, Latif F, Yao M, Crossey PA, Foster K etal. Mapping the von Hippel–Lindau disease tumour suppressorgene: identification of germline deletions by pulsed field gelelectrophoresis. Hum Mol Genet 1993; 2: 879–82.

18 Yao M, Latif F, Orcutt ML, Kuzmin I, Stackhouse T, Zhou FW etal. Von Hippel–Lindau disease: identification of deletion muta-tions by pulsed field gel electrophoresis. Hum Genet 1993; 92:605–14.

19 Payne SJ, Richards FM, Maher ER. A PCR-generated AccIRFLP in the 39 untranslated region of the von Hippel–Lindaudisease (VHL) tumour suppressor gene. Hu Mol Genet 1994; 3:390.

20 Chen F, Kishida T, Yao M, Hustad T, Glavac D, Dean M et al.Germline mutations in the von Hippel–Lindau disease tumorsuppressor gene: correlations with phenotype. Hum Mutation1995; 5: 66–75.

21 Richards FM, Latif F, Lerman MI, Zbar B, Maher ER. Taq I andPst I RFLPs in the von Hippel–Lindau disease gene (VHL). HuMol Genet 1993; 2: 1750.

22 Maher ER, Bentley E, Payne SJ, Latif F, Richards FM, Chiano Met al. Presymptomatic diagnosis of von Hippel–Lindau diseasewith flanking DNA markers. J Med Genet 1992; 29: 902–5.

23 Glenn GM, Linehan M, Hosoe S et al. Screening for vonHippel–Lindau disease by DNA-polymorphism analysis. JAMA1992; 267: 1226–31.

24 Olschwang S, Boisson C, Richard S, Resche F, Thomas G. DNA-based presymptomatic diagnosis for the von Hippel–Lindaudisease by linkage analysis. Eur J Hum Genet 1995; 3: 108–15.

25 Maddock IR, Moran A, Maher ER, Teare MD, Norman A,Payne SJ et al. A genetic register for von Hippel–Lindau dis-ease. J Med Genet 1996; 33: 120–27.

26 Zbar B, Kishida T, Chen F, Schmidt L, Maher ER, Richards FMet al. Germline mutations in the von Hippel–Lindau (VHL)gene in families from North America, Europe, and Japan.Human Mutation 1996; 8: 348–57.

27 Maher ER, Webster AR, Richards FM, Green JS, Crossey PA,Payne SJ, Moore AT. Phenotypic expression in vonHippel–Lindau disease: correlations with germline VHL genemutations. J Med Genet 1996; 33: 328–32.

28 Richards FM, Crossey PA, Phipps ME, Foster K, Latif F, Evans Get al. Detailed mapping of germline deletions of the vonHippel–Lindau disease tumour suppressor gene. Hu Mol Genet1994; 3: 595–8.

29 Crossey PA, Richards FM, Foster K, Green JS, Prowse A, Latif Fet al. Identification of intragenic mutations in the vonHippel–Lindau disease tumour suppressor gene and correla-tion with disease phenotype. Hu Mol Genet 1994; 3: 1303–8.

30 Clinical Research Group for VHL in Japan. Germline mutationsin the von Hippel–Lindau disease (VHL) gene in Japanese VHL.Hu Mol Genet 1995; 4: 2233–7.

31 Brauch H, Kishida T, Glavac D, Chen F, Pausch F, Hofler H et al.Von Hippel–Lindau disease with pheochromocytoma in theBlack Forest region of Germany: evidence for a founder effect.Hum Genet 1995; 95: 551–6.

32 Chen F, Slife L, Kishida T, Mulvihill J, Tisherman SE, Zbar B.Genotype-phenotype correlation in von Hippel–Lindau dis-ease: identification of a mutation associated with VHL Type2A. J Med Genet 1996; 33: 716–7.

33 Crossey PA, Eng C, Ginalska-Malinowska M, Lennard TWJ,Sampson JR, Ponder BAJ, Maher ER. Molecular genetic diag-nosis of von Hippel–Lindau disease in familial phaeochromo-cytoma. J Med Genet 1995; 32: 885–6.

34 Woodward ER, Eng C, McMahon R, Affara NA, Ponder BAJ,Maher ER. Genetic predisposition to pheochromocytoma:analysis of candidate genes GDNF, RET and VHL. Hu MolGenet 1997; 6: 1051–6.



35 Neumann HPH, Eng C, Mulligan L, Glavac D, Ponder BAJ,Maher ER et al. Consequences of direct genetic testing forgermline mutations in the clinical management of familieswith multiple endocrine neoplasia type 2. JAMA 1995; 274:1149–51.

36 Lamiell JM, Salazar FG, Hsia YE. Von Hippel–Lindau diseaseaffecting 43 members of a single kindred. Med 1989; 68: 1–29.

Received 25 November 1997; accepted 22 January 1998.

Correspondence: Professor Eamonn R. Maher, Division ofMedical Genetics, Department of Paediatrics and Child Health,University of Birmingham, The Medical School, Edgbaston,Birmingham B15 2TT, UK (fax: 144 121627 2618; e-mail: [email protected]).

molecular genetic analysis of von hippel–lindau disease

Documents