Inactivating germline von Hippel–Lindau (VHL) mutations

| The von Hippel–Lindau disease is caused by inactivating germline mutations of the VHL tumour suppressor gene and is associated with an increased risk of a variety of tumours in an allele-specific manner. The role of the heterodimeric transcription factor hypoxia-inducible factor (HIF) in the pathogenesis of VHL-defective tumours has been more firmly established during the past 5 years. In addition, there is now a greater appreciation of HIF-independent VHL functions that are relevant to tumour development, including maintenance of the primary cilium, regulation of extracellular matrix formation and turnover, and modulation of cell death in certain cell types following growth factor withdrawal or in response to other forms of stress. H y p o x I a a n d m e ta b o l I s m R E V I E W S NATuRE REVIEwS | cancer VOLuME 8 | NOVEMBER 2008 | 865 perturbations in O2 availability over a physiologically relevant range. In addition, these enzymes are inhibited by reactive oxygen species and certain metabolites that compete with 2­oxoglutarate, including succinate — one of the by­products of the hydroxylation reaction19. All three Phd family members, as well as P4H­TM, a related enzyme found in the endoplasmic reticulum (ER), can hydroxylate proline on HIFα in biochemical assays conducted in vitro, although PHD2 (EGLN1) appears to be the primary HIF prolyl hydroxylase in intact cells under normal conditions20. Consistent with this, mice lacking PHD1 (EGLN2) or PHD3 (EGLN3) are viable and grossly normal21. By contrast, mice lacking PHD2, like mice lacking other components of the VHL–HIF pathway, are not viable21. Conditional inactivation of PHD2 in mice leads to polycythaemia owing to HIF sta­ bilization and increased transcription of HIF­responsive erythropoietic genes such as erythropoietin22,23. Likewise, polycythaemia has been reported in people with hypo­ morphic PHD2 mutations24. Therefore PHD2 activity couples HIF stability to O2 availability in vivo. Both PHD2 and PHD3 are induced by hypoxia and appear to dampen the HIF response to chronic hypoxia25–29. Interestingly, some cases of familial polycythaemia have also been linked to homozygous, or compound heterozygous, hypomorphic VHL mutations30–32. These families do not develop the usual features of VHL dis­ ease. This probably reflects the degree to which spe­ cific VHL functions are affected by the alleles linked to familial polycythaemia compared with those linked to VHL disease. Moreover, familial polycythaemia is due to a field defect, wherein every cell capable of produc­ ing erythropoietin shares the same genetic abnormality, whereas the manifestations of VHL are linked to rare cells that have lost the remaining wild­type VHL allele (FIG. 3). The fact that familial polycythaemia patients with PHD2 or VHL mutations do not share the tumour risk of VHL disease suggests that the degree of HIF activation required to promote tumorigenesis exceeds that required to stimulate red blood cell production. Alternatively, VHL alleles linked to polycythaemia might retain crucial, HIF­independent tumour suppressor functions. Hydroxylation of a conserved asparaginyl residue within the CTAD by factor inhibiting HIF (FIH1, also known as HIF1AN), which is another 2­oxoglutarate­ and iron­dependent dioxygenase, prevents the recruit­ ment of the co­activators p300 and CREB­binding protein (CBP)33,34 (FIG. 2). FIH1 has a lower O2 Km than the Phd family members35,36. As a result, FIH1 should remain active in cells at intermediate levels of hypoxia that are sufficient to promote HIFα stabilization, and should thus serve to modulate the HIF transcriptional response. The degree to which a given HIF target gene is activated by hypoxia will therefore be influenced by a number of factors, including its requirement for CTAD function (as opposed to NTAD function) and the effect of neighbouring, cis­acting transcription factors on the ability of HIF to activate transcription. Role of HIF in VHl-defective tumorigenesis There are currently no human VHL–/– haemangioblas­ toma or pheochromocytoma cell lines with which to interrogate the functions of VHL and, in particular, the contribution of HIF. By contrast, a number of VHL–/– renal carcinoma cell lines have been established from spo­ radic tumours. Restoration of VHL function in such cells, when tested, suppresses tumorigenesis in nude mice and promotes cell­cycle exit and differentiation under spe­ cific cell culture conditions37–41. Several lines of evidence suggest that deregulation of HIFα, and especially HIF2α, contributes to VHL­defective renal carcinogenesis. First, VHL–/– renal carcinomas produce either both HIF1α and HIF2α or HIF2α alone. Second, over­ production of HIF2α (for example, using a stabilized version of HIF2α lacking its prolyl hydroxylation sites), but not HIF1α, is sufficient to override the tumour suppressor function of VHL in xenograft studies42–44. Moreover, short hairpin RNA­mediated downregulation of HIF2α is sufficient to promote cell­cycle exit of VHL–/– renal carcinoma cells under low serum conditions and suppresses tumour growth in vivo45–47. Third, HIF2α lev­ els are highest in cells engineered to produce type 1 and type 2B VHL mutants (which are associated with a high risk of renal carcinoma), are substantially lower in cells engineered to produce type 2A VHL mutants (which are associated with a low risk of renal carcinoma), and are essentially normal in cells producing type 2C VHL mutants (which are not associated with renal carci­ noma)48–50 (FIG. 1). These results suggest that deregu­ lation of HIF2α has a causal role in VHL–/– renal carcinogenesis. They further suggest that the HIF2α threshold for the development of renal carcinoma is higher than that for the development of haemangio­ blastoma (FIG. 1). Some cell types appear to produce a titratable repressor of HIF2α transactivation function51. Conceivably, this putative repressor molecule is lacking in the normal precursor cells that give rise to clear cell renal carcinomas. at a glance • Germline mutations that inactivate the VHL tumour suppressor gene cause a variety of tumours including clear cell renal carcinomas, haemangioblastomas and pheochromocytomas. VHL mutations are also common in sporadic clear cell renal carcinomas and haemangioblastomas. • The product of VHL has multiple functions, including directing the polyubiquitylation of hypoxia-inducible factor-α (HIFα). Recognition by VHL requires that the HIFα subunit be modified by O 2 -dependent prolyl hydroxylase (Phd) family members. • Inappropriate accumulation of HIFα, and especially the HIF2α subunit, has a causal role in VHL renal carcinomas and its involvement is suspected in VHL haemangioblastomas. Genotype–phenotype correlations suggest that these two tumours differ with respect to the level of HIFα activation required for tumorigenesis. • Hypomorphic VHL, hypomorphic PHD2 and hypermorphic HIF2α mutations have been linked to familial polycythaemia. • VHL binds to microtubules and is required for maintenance of a specialized structure called the primary cilium. Loss of this activity probably contributes to the development of visceral cysts in VHL disease. • Pheochromocytomas are intra-adrenal paragangliomas (sympathetic nervous system tumours). The genes linked to familial paraganglioma, including VHL, NF1, RET and succinate dehydrogenase subunit genes, encode proteins that regulate neuronal apoptosis in response to loss of growth factors such as nerve growth factor. R E V I E W S 866 | NOVEMBER 2008 | VOLuME 8 www.nature.com/reviews/cancer Nature Reviews | Cancer HIF activity