We have investigated seven VHL patients from five unrelated Hungarian families. After detailed physical, ophthalmological, radiological and laboratory investigations all patients were found to have type 1 VHL disease. Molecular genetic investigations detected three novel and one previously described point mutations in the VHL gene. According to the present and a previous report [
24], there are 12 VHL families in Hungary identified so far. Nonsense, missense, frame shift mutations and exon deletions were detected in four, four, two and two families, respectively. These proportions show no major deviation from those of other populations and are comparable to proportions reported by the comprehensive analysis of Nordstrom-O’Brien et al [
7]. The novel c.232A > T (p.Asn78Tyr) missense mutation associated to type I VHL phenotype. The site is evolutionary conserved suggesting it is an important amino acid position to maintain protein structure and function. Other mutations at the same codon, namely p.Asn78His, p.Asn78Ser, p.Asn78Thr and p.Asn78Ile associated with type I disease phenotype as well [
20,
22,
23,
25‐
29]. This mutation site is buried within the protein core where mutations were found to associate with lower risk of phaeochromocytoma compared to protein surface missense mutations [
23]. N78H, N78I and N78S missense mutations were all reported to coincide with RCC, however the N78T mutation did not show this association [
20,
22,
23,
25‐
29]. According to a later publication RCC associates with high risk to missense mutations located between codon 74 and 90 [
30]. The effect of the novel p.Asn78Tyr mutation on VHL protein structure was assessed using molecular modeling. The Asn78 side chain participates in the stabilization of the turn where it can be found. The breaking of the loop-stabilizing H-bond interaction between Asn78 and Arg82 as well as the replacement of this residue with a larger one obviously destabilizes this turn (Figure
3). It can cause partial or global misfolding. The most remarkable effect of this structural reorganization on protein interactions is the high probability of dissociation of the VHL and HIF-1alpha proteins. Despite the turn deformation, disruption of the VHL-Elongin C interaction could not be observed during the molecular dynamics simulation (Figure
3). In vitro analysis of pVHL mutants showed reduced ability to bind and ubiquitinate HIF-1alpha in case of type I, IIa and IIb mutations, on contrary, mutations associated to IIc phenotype showed normal binding and ubiquitination of HIF-1alpha protein. Completely absent HIF-1alpha binding was only observed in consequence of mutations associating to RCC in this study [
31]. However, a recent computational study could not confirm the dose-dependent effect of VHL-HIF-1alpha dissociation and suggests that the canonical configuration of the wild-type beta domain is vital for the efficient functioning of the complex and that mutation of any of the residues implicated in the H-bond network in the binding site disrupts HIF binding [
32]. Taken together, mutations in the beta domain disrupting the VHL-HIF-1alpha interaction lead to type I disease phenotype with an increased risk of RCC. Based on these data and our molecular modeling results we can conclude that the novel p.Asn78Tyr mutation associates to type I disease and has a high chance to associate to RCC as well, while only minor risk to associate to phaeochromocytoma. In an attempt to assess the effect of all previously reported missense mutations in familial VHL, we applied SIFT analysis and found that in type II disease the rate of tolerable mutations is significantly higher than in type I mutations. This finding further supports that deleterious mutations disrupting protein integrity are more prone to cause type I phenotype. Besides, it suggests that SIFT analysis can be a useful tool when quick prediction of a novel mutation is necessary.
As expected, nonsense mutations associated with type I VHL phenotype. Two of the three nonsense mutations associated to RCC as well. Previous publications show that large germline deletions, nonsense and frameshift mutations associate with type I VHL phenotype [
6,
20‐
22,
25,
33]. Consistent with previous publications indicating that RCC coincides more frequently with nonsense mutations than missense mutations, RCC only associated to MLTP in our patients [
23,
34]. Nonsense mutations independent of their exonic location associate with earlier age at onset and higher age-related risk of RA and RCC compared to missense mutations and large deletions [
23]. Supporting this observation, the patient with the p.55GluX mutation was 15-years-old when bilateral retinal angioma and bilateral RCC was diagnosed, which represents the earliest occurrence of RCC in VHL disease reported so far [
1,
35]. VHL screening guidelines [
2,
8‐
10] recommend a screening schedule for VHL patients regardless of their genetic status in order to recognize VHL complications in a curable stage. The earliest occurrence of RCC was reported in a sixteen-year-old boy. Accordingly, screening guidelines recommend yearly or biannual abdominal MRI from sixteen years of age and yearly abdominal ultrasound starting at 8 years of age. The occurrence of RCC in a fifteen-years old boy with a truncating germline VHL mutation highlights that renal cell carcinoma can be detected in patients with nonsense VHL mutations younger than sixteen years of age. Taken into account that abdominal MRI scans with a higher resolution can detect RCC earlier than ultrasound, patients with nonsense mutations known to associate to RCC and early onset of the VHL disease [
23] could be considered for earlier MRI scans. In general, genetic status might be considered to help tailoring individual screening schedules in the future.