Stereo fundus photograph of the right eye of a 49-year-old woman with a tyrosine-to-histidine change at codon 98 (Tyr98His) in the VHL gene. A large retinal angioma with a prominent feeding arteriole and draining venule is present in the superotemporal quadrant.
Stereo fundus photograph of the right eye of a 23-year-old woman with a tyrosine-to-histidine change at codon 98 (Tyr98His) mutation in the VHL gene. An elevated optic disc angioma involves more than three fourths of the optic disc.
Pedigree structure of the family. Affected individuals are represented by solid symbols; unaffected individuals, by open symbols. Shaded symbols represent individuals with an unknown clinical diagnosis.
Allen RC, Webster AR, Sui R, Brown J, Taylor CM, Stone EM. Molecular Characterization and Ophthalmic Investigation of a Large Family With Type 2A von Hippel–Lindau Disease. Arch Ophthalmol. 2001;119(11):1659-1665. doi:10.1001/archopht.119.11.1659
EDWIN M.STONEMD, PhD
Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001
Von Hippel–Lindau (VHL) disease is a dominantly inherited cancer syndrome. Since the identification of the VHL gene, at least 3 clinical-genetic subtypes of the disease have been recognized.
To identify the specific abnormality in the VHL gene and to correlate it with the prevalence and severity of ocular involvement in a large family with VHL disease.
A longitudinal clinical study and DNA analysis of 24 family members.
All 14 affected family members exhibited a thymine-to-cysteine change at nucleotide 505 (T505C) in exon 1 of the VHL gene, consistent with the clinical diagnosis of VHL disease subtype 2A. Two asymptomatic gene carriers were also identified. Seventy-five percent (12/16) of the gene carriers had 1 or more ocular angiomas. The mean number of ocular angiomas per gene carrier was 3.3. Six eyes had optic disc angioma. Five gene carriers(31%) had lost vision because of angiomatosis. Cerebellar hemangioblastomas were present in 4 patients (25%) and pheochromocytomas in 11 (69%). No patient was found to have a renal cell carcinoma.
The family shows a low susceptibility to renal carcinoma consistent with the clinical diagnosis of VHL disease type 2A. The prevalence and severity of ocular angiomatosis in this subtype do not significantly differ from those of the other more common subtypes of VHL. Recognition of the VHL disease 2A phenotype suggests the presence of a specific mutation (T505C) in the VHL gene. Confirmation of this genotype increases the clinician's ability to provide favorable prognostic information to affected family members.
VON HIPPEL–LINDAU (VHL) disease is a dominantly inherited cancer syndrome characterized by susceptibility to ocular angiomas, pheochromocytomas, renal cysts, renal carcinomas, hemangioblastomas of the subtentorial central nervous system, pancreatic cysts, pancreatic tumors, and cystadenomas of the epididymis. The ocular angiomas lie on the surface of the retina (Figure 1) or optic disc (Figure 2). Retinal lesions always exhibit feeder vessels and hyperfluoresce on fluorescein angiography. These frequently cause visual loss through intraretinal exudation, exudative retinal detachment, hemorrhage, and epiretinal fibrosis.1- 4 Heterozygosity for a disease-causing sequence variation in the VHL gene occurs approximately once in 36 000 live births, and the overall penetrance is 97% by age 60 years.5,6 Renal cell carcinoma is the most common cause of death, and the overall median survival is only 49 years.6 Since ocular lesions occur so frequently and so early in the course of VHL disease, ophthalmologists play a central role in the diagnosis, screening, and management of the disorder. Early (especially presymptomatic) detection of VHL lesions reduces the morbidity and mortality associated with the disease, and, for this reason, defined screening protocols for affected patients and their at-risk relatives have been developed.6- 8
The disease was linked to chromosome 3p25-26 in 1988,9 and the VHL gene was identified in 1993.10 This 3-exon gene encodes a novel protein of 213 amino acids and is ubiquitously expressed in fetal and adult human tissues. Germline mutations can be identified in the majority of VHL kindreds and range from deletions of all or part of the VHL gene to small intragenic insertions and deletions and even single base substitutions.10- 14 In keeping with its role as a tumor suppressor gene, somatic inactivation of the gene has been found in tumors from patients with VHL disease,15 sporadic (ie, non-VHL) clear cell renal carcinomas,16- 18 sporadic cerebellar hemangioblastomas,19 and sporadic cystadenomas of the epididymis.20 Since the gene's discovery, it has been determined that there is a reciprocal relationship between the presence of VHL protein and vascular endothelial growth factor in cultured cells21 and that VHL protein interacts with proteins that affect transcription22,23 and the cell cycle.24,25
The germline mutations that occur in families with VHL disease are extremely varied, and many specific mutations have been found in only 1 or 2 families.26 As a result, in many cases the entire VHL gene must be carefully screened (with multiple screening modalities), which is a relatively laborious and, hence, expensive task. However, a few clear genotype-phenotype correlations have been identified that can suggest specific mutations to look for in specific families and shed light on some of the important functional regions of the gene.11,12 Families with VHL disease without pheochromocytomas (classified as type 1) typically have germline VHL deletions or protein-truncating mutations, while families with VHL disease in which 1 or more family members exhibit pheochromocytomas (classified as type 2) typically have germline missense mutations. Families with VHL disease type 2 can be further classified into those in which some members are affected by renal carcinoma (in addition to their ocular, cerebellar, and renal lesions; type 2B) and those in which renal carcinoma is very rare (type 2A). Types 1 and 2B are the most common subtypes and appear to show a similar susceptibility to ocular angiomatosis.4,11,12 Type 2A has been described in a number of families from Germany and the United States and is associated with a thymine-to-cysteine change at nucleotide 505 (T505C), which results in a tyrosine-to-histidine change at codon 98 (Tyr98His). There is evidence that these families share a common ancestor.27 However, the ocular phenotype in this subtype of VHL disease has not been studied in detail, and it is not known whether the susceptibility to ocular lesions differs from that seen in the other types of VHL disease.
In this study, we investigated a large family with VHL disease from Iowa at both the clinical and molecular levels. The clinical diagnosis of VHL disease subtype 2A led to a rapid determination of the family's VHL germline mutation, and this, in turn, allowed us to perform a clinical study of the ocular phenotype of this least common subtype of the disease.
A computer search of the records of the Molecular Ophthalmology Laboratory at The University of Iowa Hospitals and Clinics, Iowa City, disclosed a total of 74 individuals who were suspected to have or at risk for having VHL disease. Of these, 34 individuals were confirmed to have VHL disease, and 14 were found to belong to a single large family that is the basis of this investigation. Patients were diagnosed as having VHL disease if they had 2 or more hemangioblastomas(including ocular lesions), 1 major VHL disease complication, and a family history of hemangioblastoma or 2 major VHL disease complications as previously described.6,28 All individuals gave informed consent for participation in this study according to the principles outlined in the Universal Declaration of Helsinki.
Case notes were available for all family members who were found to be gene carriers, and these documented a mean period of follow-up of 21 years(range, 2-42 years). All patients had been regularly screened for ocular and systemic VHL disease lesions, including annual clinical examination and blood pressure measurement, annual abdominal ultrasonography, twice-yearly abdominal magnetic resonance imaging (from age 15 years), baseline brain magnetic resonance imaging with further neuroradiologic investigation of neurologic symptoms or headache, annual urinary vanillylmandelic acid analysis, and annual ocular examination. In some cases, fluorescein angiography was used to confirm the diagnosis of ocular angiomatous lesions. All retinal angiomas had been treated with argon laser or cryotherapy as needed to inactivate the lesion. All patients had been examined in The University of Iowa Hospitals and Clinics ophthalmology clinic within 1 year of this study.
Blood specimens were obtained from all affected members of the family as well as asymptomatic relatives at 50% risk. Genomic DNA was extracted by means of standard procedures. Because of the clinical findings consistent with VHL disease type 2A, the polymerase chain reaction (PCR) was used to amplify nucleotides 397 to 553 of exon 1, which surrounds the T505C (Tyr98His) mutation. The following primers were used: forward 5′-GTGCTGCGCTCGGTGAACTC and reverse 5′-ACCCTGGATGTGTCCTGCCTCAA. Fragments were amplified in a 48-well thermal cycler (PerkinElmer Cetus, Wellesley, Mass) in 30-µL reaction volumes consisting of 1mM magnesium chloride, 0.2mM deoxyribonucleoside 5′-triphosphates (dATP, dGTP, dTTP, and dCTP), 0.4 U of Taq polymerase, 0.8mM concentrations of each primer, and 100 ng of genomic DNA as a template. Thirty-five cycles of amplification (1 minute at 94°C, 1 minute at 62°C, and 1 minute at 72°C) were performed with a terminal extension of 5 minutes. The PCR product was purified by means of purification columns (Qiagen, Bothell, Wash) as recommended by the manufacturer. Ten microliters of nonpurified PCR product was denatured at 94°C for 3 minutes before loading on a 6% polyacrylamide, 5% glycerol gel for single-strand conformation analysis as previously described.29 Single-strand conformational polymorphism (SSCP) gels were stained with silver nitrate and developed by means of sodium carbonate–formamide. In addition, 2 µL of the purified amplification product was used as template with each of the above primers for sequencing on an automated sequencer(ABI 373 Fluroescent DNA Sequencer; Applied Biosystems, San Jose, Calif). Normal control DNA samples were run in parallel during SSCP analysis and also were sequenced and compared with study samples by means of Sequencher software(Gene Codes Corporation, Ann Arbor, Mich).
The SSCP analysis disclosed 16 individuals from the pedigree (Figure 3) to harbor an identical band shift when compared with control samples. Direct sequencing of the PCR product showed a heterozygous T to C change at nucleotide 505. This change is identical to the mutation reported previously for type 2A VHL disease in families who trace their lineage to the Black Forest region of Germany.27
Family history investigation of several affected individuals allowed several small VHL kindreds from Iowa to be assembled into a single large pedigree with an extensive history of VHL (Figure 3). More than 250 individuals were ultimately identified who shared a common ancestor who had immigrated to Iowa from Baden-Baden, Germany, in 1832.
Twelve (75%) of 16 gene carriers were found to have ocular hemangioblastomas(Figure 1 and Figure 2), and 7 of these had bilateral disease (Table 1). The mean number of angiomas was 3.3. However, 1 individual had 22 angiomas, most of which were in an area of previous exudative retinal detachment (see the "Comment" section) (Table 2). Exclusion of this individual resulted in a mean number of angiomas of 2.1. Six eyes had optic disc angiomas (Figure 2). Six of the 19 eyes with ocular hemangioblastomas had visual acuity less than 20/40. Two patients had no light perception in one or both eyes. The average age at diagnosis of ocular angiomatosis was 37 years.
Eleven (69%) of 16 gene carriers were diagnosed as having adrenal pheochromocytomas, and 4 of these patients had bilateral disease (Table 1). The average age at diagnosis of pheochromocytoma was 31 years. Four (25%) of 16 gene carriers were found to have central nervous system hemangioblastomas, with an average age at diagnosis of 40 years. None of the gene carriers was diagnosed as having renal cell carcinoma.
We compared the degree and severity of ocular angiomatosis in this cohort of patients with type 2A VHL disease with cohorts of patients with types 1 and 2B VHL disease, described by Webster et al.4 In the latter study, 183 VHL gene carriers were systematically examined with methods that were the same as those of the present study. None of the 183 patients carried the Tyr98His mutation, and none of the families was suggestive of the type 2A phenotype. The 2 studies had a similar median age at examination (38.5 years in this study, 34 years in the study by Webster et al). In addition, the prevalence of angiomatosis in the 2 studies was very similar (75% this study, 68% in the study by Webster et al). Finally, the median number of angiomas was not significantly different with or without inclusion of the patient in the present study with 22 angiomas (Wilcoxon rank-sum test: P = .57 including the outlying patient, P = .87 excluding him). However, because of the small sample of patients in this study, we cannot exclude subtle differences in the susceptibility to angiomatosis in these 2 categories of VHL disease (type 2A vs types 1 and 2B).
The characterization of the VHL gene was an important contribution to the understanding of the molecular pathologic features of inherited and sporadic tumors. However, because of the wide spectrum of different mutations that can give rise to the VHL phenotype, identification of the underlying germline mutation in most affected families is not straightforward. Previous work, using a battery of molecular genetic techniques including pulse-field gel electrophoresis, Southern blot analysis, SSCP, and direct sequencing, achieved a sensitivity of 80% in the detection of mutations in known VHL disease pedigrees.30 More recently, the inclusion of a quantitative Southern analysis technique has increased this sensitivity further.31 In some cases, recognized correlations between certain genotypes and phenotypes allow us to direct the molecular genetic testing of a family on the basis of the clinical phenotype that they exhibit. In the case of the family in this study, the absence of renal disease among 16 clinically affected individuals strongly suggested the presence of the Tyr98His mutation. Similarly, the most appropriate initial investigation for a family with a low rate of pheochromocytoma (type 1 VHL disease) would be Southern analysis, as this single test has a high sensitivity in families with this phenotype.11,12,30 Finally, a family with a high rate of both renal and adrenal disease (type 2B VHL disease) would be most appropriately studied by a PCR-based restriction enzyme assay (or other allele-specific assay) for a mutation at codon 167, the specific mutation most commonly associated with this phenotype.11,12,30 Further subtle phenotype-genotype relationships may yet remain to be discovered in VHL disease. Such discoveries may allow clinicians to increase the efficiency of the molecular screening even further.
The phenotype exhibited by the family in this study is similar to that previously reported for families with the same underlying mutation in which the most striking feature is a very low risk of renal carcinoma.27 There are 2 potential explanations for the apparent organ-specific effect of this specific mutation. First, the mutation might cause a general reduction in the penetrance of disease, in which all susceptible organs are less prone to tumorigenesis. This effect might be reasonably expected to manifest itself clinically as a relative absence of the tumor that occurs latest in the course of disease—renal carcinoma. Such low-penetrance mutations have been described previously in another ophthalmic inherited cancer syndrome, retinoblastoma,32 and the low level of function of the protein product of these specific low-penetrance retinoblastoma alleles has been confirmed by in vitro studies.33 However, our study provides some evidence against this hypothesis. When studied in detail, ocular angiomas in this subtype of VHL disease are no less prevalent or severe than that seen in other families with VHL disease, suggesting that this mutation does not have a globally lower penetrance than other disease-causing mutations. The second and more likely explanation is that the VHL gene has organ-specific functions, and that these can be selectively abrogated when a specific domain of the VHL gene product is altered in a specific fashion. Such tissue- or organ-specific functions are not explained by selective expression of the gene, as it has been found to be expressed in all adult and fetal human tissues that have been tested.34,35 Since the families with the Tyr98His mutation appear to share a common ancestor, and, by genealogy, the family in this report seems likely to share this same ancestor, it is possible that another DNA change, for example in a control region of the gene, exists in phase with the Tyr98His substitution and is responsible for these organ-specific effects.
The accessibility of the ocular tissues to noninvasive examination facilitated the detailed assessment of this specific VHL manifestation in this family. We found the mean number of angiomas per individual and the prevalence of ocular angiomatosis to be similar to those previously reported for other clinical subtypes of VHL disease.4,28,30 As noted previously,4 the optic disc appears to be highly susceptible and the macula relatively resistant to tumorigenesis, when the relative areas of these structures are taken into account. It is also of interest that the most severely affected patients in this family presented at a relatively early age (younger than 20 years) compared with the less severely affected members, supporting the trend seen in a previous study.4 This underscores the need for ophthalmoscopic screening at a young age so that angiomas can be detected when they are small and therefore more effectively treatable. This study suggests that the ocular phenotype is not helpful in predicting a specific class of mutation in the VHL gene. However, there is evidence to suggest that the severity of the ocular disease is positively correlated with the severity of central nervous system and renal involvement.36 Our family was not large enough to confirm this correlation (Table 3).
One individual was observed to have 22 angiomas in one eye. The concentration of the angiomas was in an area of previous exudative retinal detachment. Webster et al4 previously observed that areas of previous retinal injury (ie, photocoagulation, cryotherapy, retinal detachment, etc) in a patient with VHL disease are susceptible to an increased number of retinal angiomas. The mechanism by which this has been postulated to occur is that areas of injury induce retinal endothelial and/or other vascular cell mitosis.4 This increases the chance of somatic mutations in the VHL gene, thereby increasing the chance of retinal angioma formation. This explanation is related to the theory that the number of angiomas a given patient with VHL disease will develop is determined at an early age. When retinal development is complete, mitoses are infrequent. This is supported by the observation of one individual in our pedigree who was 93 years old and free of retinal angiomatosis.
The observation that renal carcinoma is less prevalent in this subtype of VHL disease might suggest a modification of the standard screening protocols for families with the Tyr98His mutation. However, the chief complication of this subtype of VHL disease is adrenal pheochromocytoma, and these patients therefore require regular annual abdominal screening. Thus, simultaneous imaging of the kidneys is of little further inconvenience to the patients and is probably warranted given the deadly consequences of missing an early renal cell carcinoma.
This is the third family in the United States in which the Tyr98His mutation has been reported; the other 2 were ascertained in Pennsylvania.27 This raises the possibility that this mutation might be fairly prevalent in the United States. Of 34 individuals seen in our clinic with confirmed VHL, 14 were found to be members of the family described in this article. It is likely that patients with VHL disease with the Tyr98His mutation have a greater reproductive fitness than other patients with VHL disease, given that renal cell carcinoma is the prime cause of mortality in VHL disease overall. The increased fitness of this VHL subtype would explain why a founder effect is so prominent for the Tyr98His mutation, while most other VHL disease pedigrees are rather small and not related to one another.37
The correlation between the VHL disease type 2A phenotype and the Tyr98His genotype is useful in both directions. That is, the recognition of the 2A phenotype helped streamline the determination of the specific mutation in this family with VHL disease. However, the confirmation of the Tyr98His genotype allowed us also to give known gene carriers useful and reassuring prognostic information. Only by future careful and parallel genetic and clinical studies will further genotype-phenotype relationships in VHL disease and other genetic diseases become apparent. The discovery and characterization of this specific subtype of VHL disease also suggest directions for further research. Patients with type 2A VHL disease are clearly better off than those with other subtypes. By further study of the specific VHL disease type 2A allele, we might be able to determine why renal carcinoma is rare, which may in turn suggest treatment strategies to convert general VHL disease into a clinical course that resembles that of subtype 2A.
Accepted for publication May 3, 2001.
This study was supported in part by the Foundation Fighting Blindness(Owings Mills, Md), the Carver Endowment for Molecular Ophthalmology (Iowa City, Iowa), the Helen Keller Foundation for Research and Education (Birmingham, Ala), and an unrestricted grant from Research to Prevent Blindness Inc (New York, NY).
We thank Heidi Haines, MS, and Paula Moore for their excellent technical assistance.
Corresponding author and reprints: Edwin M. Stone, MD, PhD, Department of Ophthalmology and Visual Sciences, The University of Iowa Hospitals and Clinics, 200 Hawkins Dr, Iowa City, IA 52242 (e-mail: email@example.com).