Version July 2025
INTRODUCTION —
Von Hippel-Lindau (VHL) disease is an inherited, autosomal dominant disease manifested by a variety of benign and malignant tumors. VHL is diagnosed in approximately 1 in 36,000 people [1-3].
The initial manifestations of disease can occur in childhood, adolescence, or later (mean age approximately 26 years). The spectrum of VHL disease-associated tumors includes:
●Hemangioblastoma of the central nervous system
●Retinal hemangioblastoma
●Clear cell renal cell carcinoma
●Pheochromocytoma
●Endolymphatic sac tumors of the middle ear
●Serous cystadenoma and neuroendocrine tumors of the pancreas
●Papillary cystadenoma of the epididymis and broad ligament
The molecular biology and pathogenesis of VHL disease, and the relationship of specific gene pathogenic variants to the clinical manifestations of the disease are reviewed here.
The clinical presentation and diagnosis as well as the surveillance and management of VHL disease are discussed separately. (See "Clinical presentation and diagnosis of von Hippel-Lindau disease" and "Surveillance and management of von Hippel-Lindau disease".)
An approach to discussing VHL genetic test results is also presented separately. (See "Gene test interpretation: VHL (von Hippel-Lindau) gene".)
MOLECULAR BIOLOGY AND PATHOGENESIS —
The VHL gene was mapped to chromosome 3p25 and cloned in the early 1990s [4], with further research into the understanding of VHL gene function extending over the next 20 years [5]. Its gene product, VHL, functions as a tumor suppressor protein [6].
As with pathogenic variants in certain other tumor suppressor genes (eg, the retinoblastoma 1 [RB1] gene), a "two-hit" model has been validated for VHL disease in which a germline loss of function variant inactivates one copy of the VHL gene in all cells. For VHL disease-associated tumors to develop, there must be loss of expression of the second normal allele through either somatic changes or deletion of the second allele, or through hypermethylation of its promoter. In sporadic kidney cancers, inactivation of VHL through somatic changes of both alleles is very common.
Major advances have been made over the last two decades in understanding the biology that underlies the formation of VHL-associated tumors [6-10]. The VHL protein forms a stable complex with several other proteins, including elongin B, elongin C, and cullin 2. This complex targets several proteins for proteasomal degradation, thereby regulating their levels within the cell [8-10]. The VHL component of this complex functions as an E3 ubiquitin ligase for the target molecules. Once bound to the VHL complex, the target molecules are covalently bound to ubiquitin, facilitating degradation by the proteasome.
In addition to its function as an E3 ubiquitin ligase, VHL performs several other important cellular functions, including maintenance of the primary cilium, regulation of cytokinesis, control of microtubule function, extracellular matrix integrity, and regulation of the cell cycle.
A distinct set of pathogenic variants in the VHL gene have also been associated with an autosomal recessive disorder that results in polycythemia without VHL disease [11]. (See "Molecular pathogenesis of congenital erythrocytoses and polycythemia vera", section on 'Chuvash erythrocytosis'.)
Hypoxia-inducible factor 1 and 2 — Hypoxia-inducible factor 1 alpha (HIF1A) and 2 alpha (HIF2A) are two of the major proteins regulated by VHL (figure 1) [12]. The 2019 Nobel Prize in Physiology or Medicine was awarded jointly to Dr. William G Kaelin, Dr. Peter J Ratcliffe, and Dr. Gregg L Semenza for their discovery of this molecular pathway [13].
HIF1A and HIF2A are transcription factors that regulate a number of key cellular processes.
●Both HIF1A and HIF2A regulate glucose transport, lipid metabolism, pH homeostasis, and angiogenesis [12].
●HIF1A and HIF2A regulate erythropoietin, vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF) beta, and transforming growth factor (TGF) alpha [6,9,10].
●Skewing of the ratio of HIF1A and HIF2A towards HIF2A may alter cell signaling and upregulate Myc activity in cells [14]. The ratio switching is likely due to several factors.
•Hypoxia-associated factor (HAF) was shown to increase HIF2A transactivation [15] and HIF1A instability [16], resulting in a more aggressive cellular phenotype.
•In renal cell carcinoma tissue, preferential loss of chromosome 14q, the locus for the HIF1A gene, results in decreased levels of HIF1A [17].
HIF1A and HIF2A also have unique targets. For example, HIF1A is a key transcription factor for glycolysis [18], whereas HIF2A is involved in erythropoiesis through its ability to induce transcription of messenger RNA coding for erythropoietin [19,20].
The role of HIF2 in erythropoiesis is discussed in detail separately, but it will be briefly reviewed here to permit understanding of the events that occur in VHL disease. (See "Regulation of erythropoiesis", section on 'Hypoxia-inducible factor and the response to hypoxia'.)
Transcriptional activation by HIF requires the heterodimerization and nuclear translocation of alpha and beta subunits [21]. The beta subunit is not influenced by oxygen tension and is not bound by the VHL protein complex. By contrast, the alpha subunits are sensitive to oxygen levels and are a substrate for the VHL protein complex.
●With normal oxygen tension, HIF1A and HIF2A are enzymatically hydroxylated [21]. The hydroxylated HIF subunits are bound by the VHL protein complex and covalently linked to ubiquitin. Once this occurs, HIFA subunits are rapidly degraded by proteasomes (figure 2).
●Under conditions of hypoxia, hydroxylation does not occur, and HIF1A and HIF2A are not bound to the VHL protein complex and cannot undergo ubiquitination. The levels of HIF1A and HIF2A rise, resulting in increased messenger RNA transcription of a variety of proteins, thus inducing a physiologic angiogenic response.
In patients with VHL disease, loss of the sole functioning VHL allele in somatic tissues causes a situation analogous to hypoxia, despite the presence of normal oxygen tension [6,7,10,22]. Pathogenic variants in the VHL gene can result in VHL failing to form the necessary protein complex to bind HIF1A and HIF2A, as many of the variants in type 1 VHL disease result in complete loss of VHL expression through nonsense-mediated decay. The types of VHL disease are discussed separately. (See "Clinical presentation and diagnosis of von Hippel-Lindau disease", section on 'Types of VHL disease'.)
Alternatively, dysfunctional VHL protein may result in failure of the binding site on the complex to recognize HIFA proteins [9]. In either case, HIF1A and HIF2A are not linked to ubiquitin and are not degraded in proteasomes. Elevated levels of HIF1A and HIF2A can then induce abnormal production of the same factors that would be produced in conditions of physiologic hypoxia.
In addition, VHL affects several other factors potentially involved in tumorigenesis that are not regulated through the HIF1A system. These targets include matrix metalloproteinases (MMPs) such as MMP1, MMP inhibitors, and atypical protein kinase C [6,9].
Although the mechanism of tumorigenesis remains unproven, the combined effect of various angiogenic factors and other growth factors may create an autocrine loop that provides an uncontrolled growth stimulus consistent with the highly vascular central nervous system tumors found in VHL patients [23]. Additionally, VHL regulates several key cellular processes of which disruption may result in a malignant phenotype. These processes include extracellular matrix control, microtubule regulation, cilia centrosome cycle control, and cell cycle control.
Extracellular matrix control — Presence of functional VHL is required to maintain proper assembly of an extracellular fibronectin matrix. VHL binds to and regulates fibronectin in a phosphorylation-dependent fashion [24]. Pathogenic variants that preserve HIF regulation but lack collagen IV binding affect tumor angiogenesis in some experimental models [25].
Cilia centrosome regulation and microtubule control — The primary cilium is a nonmotile organelle involved in mechanosensing, cell signaling, and regulation of cellular entry into mitosis [26]. Loss of ciliary proteins or of signaling via canonical or noncanonical Wnt signaling pathways disrupts regulation of planar cell polarity and results in cyst formation [27].
Coordinate inactivation of VHL and glycogen synthase kinase 3 beta (GSK3B) is sufficient to induce loss of primary cilium [28] and, in VHL and PTEN knockout animal models, increases cyst formation [29]. (See 'Animal models of VHL disease' below.)
VHL was found to bind to and stabilize microtubules [30]. The binding of VHL to microtubules is regulated by glycogen synthase 3, which phosphorylates VHL at serine 68 and requires a priming phosphorylation event on serine 72 by casein kinase 1 [31]. Loss of VHL or expression of abnormal VHL in cells results in unstable astral microtubules, dysregulation of the spindle assembly checkpoint, and an increase in aneuploidy [32].
Cell cycle control — Adding back VHL to the VHL deficient 786-0 cell line results in reacquisition of cell cycle arrest upon serum withdrawal, with concomitant upregulation of p27 cyclin-dependent kinase inhibitor 1B (also referred to as p27) [33]. Nuclear localization and intensity of p27 is inversely associated with tumor grade [34]. VHL is responsible for S-phase kinase-associated protein 2 (SKP2) destabilization and concomitant upregulation of p27 after deoxyribonucleic acid (DNA) damage [35].
Emerging data suggest that VHL may also regulate p53. It has been shown that p53 is an important regulator of mitotic checkpoints, and loss of p53 permits aneuploid cells to survive [36]. VHL has been shown to bind to, stabilize, and transactivate p53 [37], and this binding may be regulated by phosphorylation [35,38]. Further work is required to dissect out the significance of these findings and their role in driving tumorigenesis.
Animal models of VHL disease — Animal model studies are used to study the necessary molecular steps required to replicate the human von Hippel-Lindau (VHL) disease phenotype.
Knockout of the VHL homologous gene in mice (Vhlh) does not cause renal cell carcinoma formation or the development of hemangioblastomas [39]. Attempts to generate murine homologues of human phenotypes have resulted in a replication of Chuvash polycythemia [40], but a VHL p.Arg167Gln homologue did not generate renal cell carcinoma [41].
The combination of Vhlh and PTEN inactivation resulted in accelerated cyst formation [29]. Similarly, a murine Vhlh conditional knockout, using a homeobox-B7-driven Cre driver specific for collecting ducts and a subset of distal tubules, exhibited widespread kidney epithelial disruption, interstitial inflammation, and cystic lesions with severe fibrosis and significant hyperplasia [42].
The addition of further gene knockouts associated with human renal cell carcinoma to the Vhlh knockout may result in mouse models that more accurately replicate the human VHL disease phenotype [43-45]. Examples of such genes include Bap1 [43], a gene encoding a deubiquitinase lost in human renal cell carcinoma, and Pbrm1, a gene associated with chromatin remodeling [44]. For example, mouse models with compound losses of Vhlh with either Bap1 or Pbrm1 have demonstrated neoplastic lesions [43,44]. This study evaluated the effects of either a Bap1 or a Pbrm1 knockout in a Vhl null background [44]. Pax8-Cre;VhlF/F;Bap1F/F mice developed early neoplastic lesions but died at three months due to failure to thrive, whereas Pax8-Cre;VhlF/F;Bap1F/+ mice lived approximately 15 months but also developed neoplastic kidney lesions. Similarly, Pax8-Cre;VhlF/F;Pbrm1F/F mice had a 12-month life expectancy and developed large homogeneous tumors, whereas Pax8-Cre;VhlF/F;Pbrm1F/+ mice had a 15-month life expectancy but did not develop tumors. This study also demonstrated that Pax8-Cre;VhlF/F control mice (without knockouts in the Bap1 or Pbrm1 genes) did not develop any neoplastic lesions and had a life expectancy of 16 to 18 months.
The knockout of Vhl and Trp53 using a Ksp Cre resulted in kidney cysts and neoplasms after a fairly long latency period [46]. Knockout of Vhl, Trp53, and Rb1 using a Ksp Cre resulted in the development of more aggressive kidney tumors in a 30- to 47-week period [45]. Tumor formation in this model was primarily dependent on HIF1A and not HIF2A [47]. In the same model, HIF2A-deficient tumors were characterized by increased antigen presentation, interferon signaling, and CD8-positive T cell infiltration and activation.
While advances have been made, further work is required to develop optimal animal models of renal cell carcinoma as well as models for other manifestations of VHL disease.
PATHOGENIC VARIANTS AND CLINICAL MANIFESTATIONS OF DISEASE —
VHL variants are classified according to standard criteria as pathogenic, likely pathogenic, variant of uncertain significance (VUS), likely benign, and benign [48]. (See "Gene test interpretation: VHL (von Hippel-Lindau) gene", section on 'Classification of variants'.)
A Clinical Genome Resource (ClinGen) VHL variant curation expert panel has provided specifications for classifying germline VHL variants per their US Food and Drug Administration-recognized process [49].
Specific germline disease-associated pathogenic variants in the VHL gene can influence the clinical manifestations of VHL disease [6,50-52]. Large datasets and machine learning tools have been used to better understand genotype-phenotype correlations [52]. VHL deletions and nonsense and frameshift mutations appear to be more common in type I disease, while missense mutations may be more common in type II disease, although these differences are relative and not absolute [8]. The types of VHL disease are discussed separately. (See "Clinical presentation and diagnosis of von Hippel-Lindau disease", section on 'Types of VHL disease'.)
These relationships can be illustrated by the following observations:
●Pheochromocytomas and paragangliomas – In a study of 138 families with VHL disease, pathogenic variants in the VHL gene predicted to result in large deletions or protein truncations were associated with a much lower risk of pheochromocytoma than missense mutations, which create a single amino acid substitution (6 versus 40 and 9 versus 59 percent at 30 and 50 years, respectively) [53]. In particular, missense mutations at codon 167 were associated with a particularly high risk of pheochromocytoma (over 80 percent by age 50).
Similar findings have also been seen in other analyses that included data on VHL disease-associated pheochromocytoma and paraganglioma [52]. It is important to note that, although the risk for pheochromocytoma and paraganglioma is diminished with deletions or other loss-of-function variants, the residual risk is high enough to warrant continued screening for these tumors in patients with VHL disease.
●Renal cell carcinoma – Data are conflicting on the association between specific germline variants and the development of renal cell carcinoma. As examples, in the study above, the cumulative probability of renal cell carcinoma in 138 families with VHL disease was similar in those with large deletions and intragenic loss of function VHL mutations compared with those with missense mutations [53].
By contrast, in another report of 274 individuals in 126 unrelated families, variants resulting in a truncated or absent protein, or large rearrangements led to an increased incidence of renal cell carcinoma compared with those that lead to single amino acid changes (81 versus 63 percent) [54]. However, missense mutations within two narrow cluster regions of the VHL gene were associated with higher incidence of renal cell carcinoma than missense mutations elsewhere in the VHL gene.
●Pancreatic neuroendocrine tumors – Several studies show a relationship between pancreatic neuroendocrine tumors and intragenic variants compared to large deletions, specifically with VHL exon 3 variants demonstrating hotspots at 161 and 167 [55,56]. These findings were supported by a larger dataset, demonstrating co-occurrence between pancreatic neuroendocrine tumors and paraganglioma, both with nontruncating variants [52].
●Retinal capillary hemangioblastomas – Correlations between VHL gene pathogenic variants and the frequency of retinal capillary hemangioblastomas have also been observed [57]. As an example, in a series of 196 patients, retinal capillary hemangioblastomas were twice as common among patients who had a variant that created a single amino acid change than among those who had a variant resulting in a truncated protein. Another observational cohort study of 77 patients from Korea with VHL disease and retinal capillary hemangioblastomas [58] divided the VHL pathogenic variants into HIF1A binding site missense (HM), non-HIF1A binding site missense (nHM), and truncating (TR) mutations. Compared with patients having nHM mutations (15 patients), patients with HM mutations (33 patients) or TR mutations (26 patients) presented with a greater number of eyes affected, a greater number of retinal capillary hemangioblastomas, and higher incidence of larger hemangioblastoma ≥2 disc diameters.
In patients with VHL disease-associated tumor presentations, the most common identification of pathogenic variants in the VHL gene is the result of directed genetic testing or hereditary cancer gene panels. The American College of Medical Genetics and Genomics (ACMG) publishes recommendations for the reporting of secondary findings (previously referred to as incidental findings) when performing clinical whole exome or genome sequencing [59,60]. (See "Secondary findings from genetic testing", section on 'Definitions and classification of variants'.)
The VHL gene is included on the version 3.1 (81 genes) Secondary Findings list genes evaluated for medically actionability and recommended for reporting of pathogenic or likely pathogenic variants. Thus, with the expansion of genomic sequencing, particularly in the diagnosis of young children with neurodevelopmental disorders, there has been an increase in the number of patients with VHL disease identified as a secondary finding. However, the ACMG secondary finding guidelines for reporting findings from exome or genome sequencing should not be interpreted to imply that the ACMG is suggesting that the genes listed, including VHL, should be used for general population screening without appropriate trials. One general issue with the reporting of genetic variants as a secondary finding, including VHL variants, is our understanding of the penetrance of genetic disorders in unselected populations [61,62]. The penetrance of disease genes is typically determined by studying highly affected families, but these penetrance estimates may overestimate the penetrance in unselected individuals identified through population screening. (See "Clinical presentation and diagnosis of von Hippel-Lindau disease", section on 'When to suspect the diagnosis' and "Secondary findings from genetic testing", section on 'ACMG list'.)
SUMMARY
●Clinical features of von Hippel-Lindau disease – Von Hippel-Lindau (VHL) disease is a heritable, autosomal dominant syndrome manifested by a variety of benign and malignant neoplasms, including clear cell renal cell carcinoma, hemangioblastomas, pheochromocytomas, and other rare tumors. (See "Clinical presentation and diagnosis of von Hippel-Lindau disease".)
●Molecular biology and pathogenesis – The pathogenesis of VHL disease has been linked to germline (constitutional) variants in the VHL gene. VHL, the product of the VHL gene, is a tumor suppressor protein that performs several important cellular functions, including oxygen sensing, targeting proteins for proteasomal degradation, maintaining an intact primary cilium, and regulating the extracellular matrix. (See 'Molecular biology and pathogenesis' above.)
●Pathogenic variants – The type of pathogenic variant in the VHL gene impacts the disease phenotype. (See 'Pathogenic variants and clinical manifestations of disease' above.)
●Hypoxia-inducible factor 1 and 2 alpha – The VHL protein complex binds the transcription factors hypoxia-inducible factor 1 alpha (HIF1A) and 2 alpha (HIF2A) and targets them for proteasomal degradation. In the absence of VHL-induced degradation, HIF1A and HIF2A may contribute to increased levels of erythropoietin, vascular endothelial growth factor (VEGF), and other growth factors, providing a stimulus for tumor growth and/or angiogenesis. (See 'Hypoxia-inducible factor 1 and 2' above.)
