Von Hippel–Lindau disease: insights into oxygen sensing, protein degradation, and cancer

My mentor, the late David M. Livingston (139), taught me and his other mentees that every good experiment starts with the question it is aimed at addressing. Studying the von Hippel–Lindau tumor suppressor gene shed light on a number of questions: What are the earliest steps in the development of ccRCC, and how should we treat these tumors based on that knowledge? How do tumors regulate angiogenesis, which is a conspicuous feature of VHL-associated tumors, and, more broadly, how do cells and tissues sense oxygen and couple that information to changes in gene expression? The VHL gene product, pVHL, proved to be a key node in the oxygen sensing mechanism, targeting the HIF transcription factor for destruction when oxygen is plentiful. Drugs that inhibit HIF2, and HIF-responsive gene products such as VEGF, are now cornerstones of kidney cancer therapy, while HIF agonists, which block oxygen-dependent prolyl hydroxylation of the HIFα subunits, appear promising for the treatment of anemia and ischemia.

A number of questions and mysteries remain. For example, it remains unclear why VHL loss is intimately linked to ccRCC, but not other common epithelial cancers. It is potentially relevant in this regard that mammalian kidneys are hypoxic at rest, which might lead to epigenetic changes that allow certain renal cells to proliferate (for example, in response to injury) in a hypoxic environment. Consistent with this idea, HIF lowers cyclin D1 levels and proliferation in many cell types, but increases cyclin D1 and proliferation in the cells capable of giving rise to ccRCC (140). Nor do we completely understand the genotype-phenotype correlations in VHL disease, although the degree of HIF dysregulation almost certainly plays a role here.

It is also unclear how HIF1 and HIF2, which appear to oppose one another with respect to ccRCC proliferation, achieve their paralog-specific effects. Prior reports suggest that this is achieved, at least in part, through paralog-specific binding to specific HIF response elements (141143). These initial reports, however, might have been confounded by technical factors such as the use of different antibodies or reliance on overexpression systems.

We do not fully understand how, mechanistically, stereotypical non-allelic mutations, such as of BAP1 and SETD2, cooperate with VHL loss to cause ccRCC. Nor do we know whether these mutations play roles in tumor maintenance, as opposed to tumor initiation and progression, and whether they engender any specific therapeutic vulnerabilities.

It is not known why the response to VEGF inhibitors and, based on early data, HIF2 inhibitors is variable in ccRCC. With respect to the latter, the percentage of ccRCCs exhibiting measurable tumor shrinkage is much higher in the setting of VHL disease patients whose ccRCCs were previously untreated than in the setting of metastatic disease patients who had been heavily pretreated. This suggests that all ccRCCs are initially HIF2 dependent, but can evolve toward HIF2 independence over time under the selection pressure created by standard-of-care agents such as VEGF inhibitors and immune checkpoint inhibitors.

Our knowledge of the effects of HIF on the immune system is incomplete. This knowledge might influence the outcome of combining HIF2 inhibitors with other anticancer drugs and could prove valuable in exploring the therapeutic utility of HIF agonists. In this regard, we are just beginning to understand the benefits and risks of acutely or chronically inactivating HIF for the treatment of anemia and other disorders such as ischemic diseases. It is perhaps noteworthy that the most advanced HIF agonists inhibit the EglN prolyl hydroxylases, but not the FIH1 asparaginyl hydroxylase (100). Accordingly, these drugs can induce EPO without inducing VEGF, which relies on the FIH1-responsive HIF1α C-terminal transactivation domain in most tissues (144). This was initially viewed as fortuitous, insofar as it was feared VEGF would induce angiogenesis and possibly stimulate latent tumors. On the other hand, there is no evidence that systemic (rather than focal) VEGF induces angiogenesis, presumably because an effective gradient to stimulate endothelial cells is not established, and a recent study showed that increasing systemic VEGF levels in adult mice caused tissue rejuvenation and increased lifespan (145). It is also theoretically possible that a slight increase in VEGF would improve endothelial cell health and reduce, for example, the risk of thrombosis, especially as decreasing VEGF clearly has the opposite effect (146). Combined inhibition of EglN and FIH1 should more faithfully mimic the effects of true hypoxia, such as life at very high altitude. It would therefore be interesting to test combined EglN1 and FIH1 inhibition preclinically and clinically.

留言 (0)

沒有登入
gif