Insights into Epigenetic Remodeling in VHL-Deficient Clear Cell Renal Cell Carcinoma

Renal cell carcinoma (RCC) affects nearly 300,000 people worldwide annually and is responsible for nearly 100,000 deaths each year. The most common type of RCC, clear cell renal cell carcinoma (ccRCC), is found to have loss of the von Hippel–Lindau tumor suppressor gene (VHL) function by mutation, deletion, or methylation in up to 90% of tumors (1, 2). VHL is encoded at 3p25.3, a region that is telomeric of a fragile site on chromosome 3p, and loss of chromosome 3p is observed in >90% of all ccRCC (1, 2). In combination within chromosomal loss of one VHL allele, the remaining allele is frequently mutated or methylated, resulting in complete loss of activity for the VHL protein. The critical importance of VHL loss to ccRCC is reinforced by the fact that germline mutation of the VHL gene results in von Hippel–Lindau syndrome associated with angiomatosis, hemangioblastomas, pheochromocytoma, and bilateral, multifocal ccRCC (3).

Increased understanding of the function of the VHL pathway provides the foundation for the development of effective therapies for ccRCC. VHL has been shown to be an essential component of an E3 ubiquitin ligase complex, in combination with elongin C, elongin B, CUL2, and other proteins, that targets the hypoxia-inducible factors HIF1α and HIF2α for ubiquitin-mediated degradation in an oxygen-dependent manner (4, 5). In normoxia, prolyl hydroxylases transfer hydroxyl groups to the oxygen-dependent domain of HIF1α and HIF2α, enabling targeted binding of the VHL complex and subsequent degradation. When cells are hypoxic, the prolyl hydroxylases are inhibited, resulting in HIF1α and HIF2α stabilization allowing them to heterodimerize with their shared binding partner, ARNT (HIF1β), and activate a hypoxia response. Inactivation of VHL in ccRCC results in a pseudohypoxic response, as HIF1α and HIF2α are stabilized while the cells are still in normoxia. The pseudohypoxic response results in altered transcriptional regulation of numerous HIF target genes, such as increased expression of the genes encoding VEGF, platelet-derived growth factor (PDGF), the glucose transporter GLUT1, and transforming growth factor α (TGFα). This HIF-dependent transcriptional rewiring promotes angiogenesis, increases glucose uptake, alters metabolic pathways, increases cell proliferation and survival, alters cell-cycle progression, dysregulates lipid metabolism, and promotes cellular migration (4, 5).

Although activation of the HIF pathway is considered a critical tumorigenic event resulting from VHL loss, there are several HIF-independent consequences of VHL loss that dysregulate assembly and regulation of extracellular matrix, microtubule stabilization and maintenance of primary cilium, regulation of apoptosis, control of cell senescence, and HIF-independent transcriptional regulation. Increasing our understanding of how the stabilization and activation of HIF1α and HIF2α alters the transcriptional landscape within ccRCC cells and the extent of the HIF-independent effects of VHL loss on ccRCC mRNA expression profiles could prove essential in further elucidating the mechanisms of ccRCC tumorigenesis and the development of new therapeutic options for patients with ccRCC.

In addition to the effect of VHL loss on HIF-dependent transcription in ccRCC, recent studies of ccRCC have identified frequent mutation of chromosome 3p chromatin-remodeling genes, such as PBRM1, SETD2, and BAP1, that are mutated in approximately 40%, 15%, and 10% of ccRCC, respectively (2). PBRM1 encodes a component of the SWI/SNF-B (PBAF) chromatin-remodeling complex, whereas SETD2 encodes a histone 3 lysine-36 methyltransferase, and BAP1 encodes the catalytic subunit of the Polycomb repressive deubiquitinase (PR-DUB) complex that regulates histone 2A ubiquitination. This suggests that chromatin-remodeling gene alteration is another potentially important driver in the tumorigenesis and evolution of ccRCC tumors.

The study by Yao and colleagues elucidates the substantial chromatin level changes that occur within ccRCC tumors and how loss of VHL contributes to chromatin alteration in both an HIF-dependent as well as independent manner (6). Yao and colleagues analyze the differences in chromatin profiles between 10 primary tumor/normal pairs (9 harboring VHL loss) and 7 ccRCC-derived cell lines (6 harboring VHL loss), in comparison with 2 normal kidney lines. The authors focused on three specific histone modifications: H3K4me3 associated with promoters; H3K4me1 associated with enhancers; and H3K27ac associated with active elements and distinguished between active promoters, marked by H3K4me3 and H3K27ac, and active enhancers, marked by H3K4me1 and H3K27ac (Fig. 1). The comparison of the primary tumor/normal pairs identified numerous common tumor-specific chromatin alterations, including 4,719 gained promoters, 592 lost promoters, 4,906 gained enhancers, 5,654 lost enhancers, and 1,157 gained superenhancers. Evaluation of the genes associated with gained promoters demonstrated enrichment for general cancer-related processes, such as the cell cycle, transcription, and RNA metabolism, whereas the gained enhancers were associated with genes enriched in ccRCC disease-specific processes, such as the HIF and proangiogenic pathways, and metabolism, including glycolysis, glutamine intake, and lipid storage. Investigation of the genes associated the somatically altered superenhancers (otherwise known as stretch enhancers) located near master regulator genes identified gained superenhancers for well-known oncogenes, such as VEGFA and EPAS1 (HIF2A), as well as less known genes, such as ZNF395 and SMPDL3A (Fig. 1). Increased expression of ZNF395 had been previously identified as a potential ccRCC biomarker, and this study demonstrated that knockdown of ZNF395 significantly decreased cell proliferation and viability both in vitro and in vivo in two cell line models of ccRCC, A-498 and 786-O, which highlights the importance of this master regulator and its potential as a therapeutic target (6).

These chromatin profile alterations distinguished tumor tissue from surrounding normal tissues, but could result from numerous alterations associated with tumorigenesis, including the loss of the above-mentioned chromatin-remodeling genes due to mutation and chromosome 3p loss. The influence of VHL loss on these chromatin alterations was evaluated by comparing four tumor-derived cell lines with and without reexpression of wild-type VHL. VHL restoration resulted in the depletion of chromatin signals from 12% of gained enhancers and 6.5% of gained promoters across the cell lines, and nearly a third (32%) of all the gained enhancers were depleted within at least one VHL-restored cell line. The genes associated with these enhancers were enriched for the HIF pathway, and the enhancers demonstrated preferential binding of HIF2α, whereas HIF1α showed a preferential occupancy at promoter regions, with some enhancers exclusively binding HIF2α. This demonstrated a significant effect of VHL loss on chromatin alteration, particularly at enhancer regions that was driven in part by the stabilization of HIF1α and HIF2α. A significant degree of the chromatin profile changes remained independent of VHL loss, and the study identified mutations within several other chromatin-remodeling genes, such as PBRM1, SETD2, KDM5C, ARID1A, and KMT2C (MLL3), that could be responsible for some of the alterations. The study by Yao and colleagues proposes this is consistent with the concept that VHL loss creates an altered chromatin profile that can aid tumorigenesis and that further mutation of additional chromatin-remodeling genes could enhance these epigenetic changes.

The functional consequences of VHL loss have provided the foundation for the development of several targeted therapies, and currently nine agents targeting the HIF pathway are FDA approved for the treatment of patients with advanced kidney cancer. Six tyrosine kinase inhibitors, sorafenib, sunitinib, pazopanib, axitinib, lenvatinib, and cabozantinib, and one recombinant humanized mAb, bevacizumab, are used to target the downstream effects of HIF pathway activation by inhibiting the VEGF and PDGF receptors or directly inhibiting VEGF (Fig. 1; ref. 7). Two further agents, temsirolimus and everolimus, inhibit protein translation of de novo HIF subunits by downregulating the activity of the mTOR pathway. Although the use of these agents has produced significant results in patients with advanced ccRCC, including increased disease-free progression, increased survival, and prolonged disease stability, and the response rates can be as high as 45%, few patients are cured and most will eventually progress and die of this disease. Recently, a new therapeutic approach was developed involving agents such as PT2385/PT2399, which are direct and specific HIF2α antagonists that inhibit the interaction between HIF2α and its essential binding partner, ARNT (HIF1β), and leave the stabilized HIF2α inactive (Fig. 1; refs. 8, 9). This agent has been shown to have a significant effect on proliferation and survival of ccRCC cell lines and ccRCC patient-derived xenografts in preclinical studies (8, 9). The study by Yao and colleagues suggests that this direct inhibition of HIF2α could result in significant downregulation of activated enhancer regions within ccRCC tumors that are largely HIF2α dependent, which the study demonstrates are more specific drivers of ccRCC tumorigenesis.

Although this study provides further and growing evidence for the importance of direct inhibition of HIF2α, it also demonstrates that some chromatin profile alterations are HIF1α driven and HIF independent. The findings in this study suggest that, although HIF2α inhibition may prove a critical component in the development of an effective form of therapy for this disease, additional tumorigenic alterations remain that would not be reversed by these therapies, even in conjunction with existing therapies. The evidence for a high degree of tumor heterogeneity within advanced ccRCC, including inconstant mutation within chromatin-remodeling genes, suggests that a great degree of variation may exist within the epigenetic alterations present within a single tumor (10). The findings in this study suggest the possibility that therapeutic approaches that target epigenetic alterations in a more global manner or that target master regulators, such as ZNF395, could be advantageous in the treatment of ccRCC.

This study provides a new direction for investigation of the epigenetic dysregulations essential to ccRCC and provides further mechanistic insight into the function of VHL loss, which will hopefully aid the development of therapeutic approaches for the treatment of this disease.

No potential conflicts of interest were disclosed.

This research was supported by the Intramural Research Program of the NIH, NCI, Center for Cancer Research.