Intrafamilial Variability of Ocular Manifestations of von Hippel-Lindau Disease

Abstract

In this retrospective cohort study, we describe intrafamilial phenotypic variability of retinal hemangioblastoma (RH) in families with von Hippel-Lindau (VHL) disease. Patients with molecularly confirmed VHL evaluated at our institution were identified, and records were reviewed. For individuals with sufficient follow-up and imaging (n=27), the number and location of RHs at the initial and most recent follow-up visits were recorded along with treatment method and systemic manifestations. A strategy for zonal classification of RH location was used. Intrafamilial phenotypic variation was identified in 3 families. Intrafamilial phenotypic variability of RH exists between family members with VHL with the same genetic mutation.

Von Hippel-Lindau (VHL) disease is an autosomal dominant condition caused by mutations in the VHL gene, which acts as a tumor suppressor.^1,2 The VHL gene product is expressed in multiple cell types and functions to target hypoxia-inducible factors for degradation.³ Loss-of-function mutations in VHL result in excessive production of proangiogenic factors and tumor formation.³

Among the many challenges in managing patients with VHL is the wide spectrum of clinical manifestations due to phenotypic variability. Several specific genotype-phenotype correlations have been reported, with patients having greater predisposition to specific tumors depending on the type and location of their VHL mutation.^3,4 Like many dominant disorders, there is often marked phenotypic variability between individuals within the same family who share the same disease-causing genotype. This type of intrafamilial variability suggests that genetic modifiers may influence disease severity.⁴ In this study, we describe the clinical examination and histories of several VHL families. We describe the range of phenotypic variability in individuals who share the same genetic mutation and characterize the frequency of retinal hemangioblastomas (RHs) in different regions of the retina using a zonal classification system.

This study was approved by the Institutional Review Board of the University of Iowa and adhered to the ethical principles of the Declaration of Helsinki. All participants provided informed consent. A retrospective review of patients with molecularly confirmed VHL who presented from January 1975 to October 2020 was performed. Patients with at least 1 follow-up visit and with imaging or records that documented the number of RHs were included. The following were recorded at both the initial and final follow up visits: best-corrected visual acuity, total number of RHs, location of RHs, and treatment method. Presence of systemic disease was also recorded.

The locations of RHs were categorized into 3 zones based on criteria modified from monitoring cytomegalovirus retinitis, which accounts for the proximity of posterior retinal lesions to the optic nerve and fovea and peripheral lesions to the equator and ora serrata (Fig S1a, available at www.ophthalmologyretina.org).⁵ Data were collected from color fundus photographs, fluorescein angiography, retinal drawings, and descriptive ocular examination annotation. Individuals were categorized as having mild or severe ocular disease at last follow-up (Table S1). Severe disease was defined as individuals with 10 or more RHs or a history of enucleation due to RH complications.

We identified 36 patients with VHL who were seen during the study period. Twenty-seven of these patients belonged to 5 families and met the described inclusion criteria. Patients were followed for a median of 133 months. Overall distribution of RHs included 19 (13%) in zone 1, of which 10 (7%) were in zone 1a and 9 (6%) were in zone 1b. There were 82 (58%) RHs in zone 2 and 41 (29%) RHs in zone 3 (Fig S1a, Table S2a, available at www.ophthalmologyretina.org).

Of the 5 families, members of families 1 (Tyr98His), 2 (Asn78Ser), and 3 (Arg167Gln) had marked variation in ocular phenotype (Fig S1b, available at www.ophthalmologyretina.org). Specifically, some family members had severe ocular disease, while others of similar or greater age had mild ocular disease (Fig 1; Fig S1b, Table S1, available at www.ophthalmologyretina.org). Members of families 4 (Leu63 codon deletion 4ctGCGC) and 5 (Val130Leu) all had mild disease.

Figure 1.

Family 1 with mild phenotype in the right eye of a 61-year-old (A) compared with severe phenotype in the right eye of a 60-year-old (B). The patient in B had coexisting retinopathy of prematurity resulting in temporally dragged fovea; however, the chorioretinal scars in the figure are due to treated retinal hemangioblastomas (RHs). Family 2 with mild phenotype in the left eye of a 53-year-old (C) compared with severe phenotype in the left eye of a 33-year-old (D). Family 3 with mild phenotype in the right eye of a 52-year-old (E) compared with severe phenotype in the right eye of a 48-year-old who underwent pars plana vitrectomy with placement of silicone oil due to RH-induced tractional retinal detachment (F).

Family 1 had 13 individuals. Average age at presentation was 44 years (range, 6–71) and at last follow-up was 57 years (range, 24–80). Family 2 had 7 individuals. Average age at presentation was 28 years (range, 13–55) and at last follow-up was 41 years (range, 33–55). Family 3 had 2 individuals. Average age at presentation was 35 years and at last follow-up was 50 years (range, 48–52). Eyes with no angiomas were observed. For eyes requiring treatment, laser photoco-agulation, cryotherapy, photodynamic therapy, transpupillary thermotherapy, scleral buckle, pars plana vitrectomy, and enucleation were performed depending on the clinical scenario (Table S1, available at www.ophthalmologyretina.org). Within the 5 families, there were different patterns of systemic disease by genotype (Table S2b, available at www.ophthalmologyretina.org).

We show that for 3 families carrying 3 distinct missense mutations in the VHL gene, there is marked variability in ocular disease severity among family members. Similar phenotypic variability has been described in patients with germline mutations in the retinoblastoma gene, indicating that factors other than the disease-causing mutations are important in individuals with hereditary tumor predisposition syndromes.⁶ This phenotypic variability despite shared genotype supports work by Webster et al⁴ suggesting that variables such as modifier genes or environmental factors contribute to disease severity. Webster et al⁴ found that within families with VHL, there was a correlation between the number of RHs in pairs of relatives who were closely related (sharing half of their genome) but not in more distantly related family members. Our study did not have a sample size of sibling/parent-child pairs to perform similar statistical analysis, but this is a phenomenon that we hope to investigate further using blood samples from our cohort in the future.

A strength of this study is the long-term follow-up for many patients and accounting for patient age. It has previously been shown that presentation with RHs at a younger age is associated with worse ocular prognosis.⁷ This work, along with our own, highlights the need for close follow-up of young individuals presenting with severe disease. Given the retrospective study, there is bias toward those individuals presenting for follow-up likely having more severe disease, which could be addressed with future prospective studies.

The selection of a 10 angioma cutoff/enucleation for mild versus severe disease has limitations. However, these criteria correlated well with visual acuity outcomes and gave a practical measure of disease severity. We also applied a novel strategy describing the location of RH using a modified cytomegalovirus retinitis staging system. This provides a useful way to compare severity given previous work showing that eyes with juxtapapillary RHs or RHs in more than 1 quadrant have a greater risk for vision loss.⁷ Many of our patients were examined before the development of widefield angiography, and its use may help to detect more zone 3 lesions in the future.

Although the primary focus of our analysis was ocular disease phenotype, we confirm previously documented trends in the presence of systemic disease by genetic mutation. It is important to consider the role for potential modifier genes in other VHL-associated tumors because the underlying molecular mechanisms may overlap and provide insight into treatment.

Intrafamilial variability in severity of ocular disease exists within family members who share the same VHL mutation. This suggests a role for modifier genes in modulating ocular disease severity in these patients. Zonal classification can be used to describe the location of RH.