Clinical diagnostic criteria introduced in 1964 [12] enabled the diagnosis of VHL disease in sporadic patients who had two manifestations (such as two HGBs or a HGB and a visceral tumor), and in patients who had only a single simple manifestation (a CNS HGB or a visceral lesion) but with family history of VHL disease. Molecular genetic testing for early identification of the patients improves diagnostic certainty and erases the psychological burden of at-risk family members who have not inherited the pathogenic variant. In the present study, using Sanger sequencing, we successfully identified a novel nonsense variant, c.351G>A (p.Trp117Ter), in the second exon of VHL, which was heterozygous in 6 VHL-diagnosed members (III-1, III-3, III-8, Ⅵ-2, Ⅵ-3 and Ⅵ-4) and 1 currently phenotype-normal mutation carrier (Ⅵ-1) in this pedigree.
From the bioinformatics analysis, we found that the c.351G>A variant is absent from public databases, and predicted to be deleterious by bioinformatics tools. The residue p.Trp117 in pVHL which is located within the β-domain (Figure 3C) and maps to hydrophobic core residue important for the structural integrity of the β sandwich [13], is evolutionarily conserved, suggesting that this amino acid is important for maintaining the protein’s structure and function.
pVHL contains two tightly coupled functional domains, the α-domain and the β-domain, held together by two short polypeptide linkers (residues 154 to 156 and 189 to 194) and a polar interface that is stabilized by hydrogen-bond networks from the H1 helix, the β sandwich, and Elongin C [13]. The α-domain is responsible for directly binding to Elongin C, which consists of three α-helices (H1, H2 and H3) located at amino acid residues 155-192, whereas the β-domain is the substrate recognition region of pVHL, which contains seven-stranded β sandwiches (residues 63 to 154) and an α-helix (H4; residues 193 to 204) that packs against one of the β-sheets through hydrophobic interactions [13]. The pVHL–Elongin C complex nucleates a complex containing Elongin B, CUL2 and RBX1, forming the VCB–CR complex (Figure 4), which is thus resistant to proteasomal degradation through their interactions with each other [8]. The α-domain has an important role in the maintenance of the spatial conformation stability of pVHL [14]; the β-domain binds directly to HIF-α (HIF-1α or HIF-2α) and participates in the degradation of the HIF subunit under aerobic conditions. Previous data shows that the HIFα peptide binds exclusively to the β-domain of pVHL [15]. A six-residue NH2-terminal segment (residues 561 to 566) that is centered on Hyp564 (Figure 4 in blue), a three-letter code for hydroxyproline, is central to the binding of HIF-1α to pVHL β-domain [9]. The pVHL residues that interact with Hyp564, including W117 which is spatially close to Hyp564 (3.5A , Figure 4), are highly conserved. And W117R missense mutation of pVHL has been shown to abolish HIF-1α binding [16].
Considering that the mutation c.351G>A introduced a premature stop codon which results in the replacement of tryptophane (TGG) with a stop codon (TGA) at codon 117 (p.Trp117Ter), either it can lead to the production of a truncated protein missing 45% of its residues including the predicted downstream α-domain (residues 155–192) and the α-helix (H4; residues 193 to 204) part of β-domain, failling to bind to Elongin C, Elongin B, CUL2 and RBX1 to form the VCB–CR complex, or the protein may be entirely absent due to the Nonsense-Mediated mRNA Decay (NMD), a process that typically degrades transcripts containing premature termination codons (PTCs) in order to prevent translation of unnecessary or aberrant transcripts. According to the ClinGen Haploinsufficiency Score and the prediction of the aforementioned bioinformatics tools, it is likely that aberrant VHL transcripts with the nonsense mutation p.Trp117Ter undergo NMD, thus no protein will be synthesized from the mutant allele. The haploinsufficiency of VHL expression will lead to the loss of function (LoF) of the pVHL, then the accumulation of HIFα and subsequent overexpression of HIF target genes, including VEGF, PDGF β, TGF α, CyclinD1 and EPO, which play a key role in the process of tumorigenesis, and consequently, results in VHL-associated tumors [17].
Two different nonsense mutations of residue 117 have previously been reported and enrolled in Human Gene Mutation Database (HGMD) (Table 3): a somatic c.350G>A (p.Trp117Ter) mutation was detected in a 64 years female sporadic RCC patient [18], while a somatic c.351G>A (p.Trp117Ter) mutation was found in cell lines UOK163 derived from tumor tissue from patients with renal cell carcinomas [19], and a germline c.351G>A (p.Trp117Ter) mutation was discovered in a kindred with VHL disease without phaeochromocytoma phenotype [20].
Taken together, this variant is classified as pathogenic and the supporting evidence for the pathogenicity of VHL c.351G>A according to the American College of Medical Genetics and Genomics (ACMG)(2015) was as follows: (1) PVS1 (very strong pathogenicity 1): variation may lead to loss of gene function; (2) PS1 (strong pathogenicity 1): known disease mutation at this position enrolled in HGMD [18-20]; (3) PM2 (moderate pathogenicity 2): absent from public databases; (4) PP1 (supporting pathogenicity 1): cosegregation with disease in all affected family members in the VHL gene that was definitely known to cause the VHL disease; (5) PP3 (supporting pathogenicity 3): PROVEAN, Gerp++, PhyloP, PhastCons, Polyphen-2, MutationTaster, CADD and ClinGen Haploinsufficiency Score predicted that the variant affects the gene products and is harmful in the conservativeness and structure of the protein. (6) PP4 (supporting pathogenicity 4): patient’s phenotype and the family history is highly specific for the VHL disease with a single genetic etiology.
All of the diagnosed patients examined in this study were classified as type 1 VHL, in accordance with the fact that missense mutations are associated with the development of type 2 VHL disease, whereas deletions or mutations that lead to truncation of the VHL protein (pVHL) are primarily associated with the development of type 1 VHL disease [21]. In addition, previous studies have indicated that VHL deletions and protein truncating mutations appear to confer a higher risk of CNS HGBs than missense mutations [20, 22, 23]. In line with this observation, individuals in the family examined in our study all presented with CNS HGBs, but no manifestation of phaeochromocytoma. At least 4 out of 8 (50%) of our patients developed retinal angiomas (RA) diagnosed at an early age (Ⅲ-3:9, Ⅵ-4:14, Ⅲ-8:16, Ⅵ-4:23), which is younger than the mean age (25 years old) of RA diagnosis compared to VHL patients in general [21]. This frequency is much higher than those of retinal lesions in VHL patients with intragenic mutations and partial deletions, suggesting that nonsense c.351G>A mutation may confer to a high risk of early onset of RA, which is in contrast to Maher’s observation that the risk of RA is slightly higher (45% vs 37%) in the missense mutation group than in the deletion/protein truncation group [20]. With regard to the other manifestations, a single pancreatic cyst was detected in the proband, while RC or RCC was diagnosed in the proband’s father, aunt and grandfather, suggesting a relatively lower incidence of visceral organs lesion.
This is the first elaborately studied VHL family caused by p.Trp117Ter mutation. Further functional evidence research remains to be conducted to reveal the pathogenesis of p.Trp117Ter.