The expression of PHD2/3, the main regulators of HIF-α has not been investigated in primary human ccRCC using double immunohistochemical staining to detect these proteins simultaneously in consecutive sections of the same tumors. In this study, we have demonstrated low incidence, distribution and staining intensity of PHD2, deficient PHD3 protein, and high HIF-α incidence, distribution and intensity in 88 primary ccRCC cancers compared to head & neck and colorectal cancers (Figure A, B and C). Furthermore, like clinical samples, the two ccRCC cell lines (RC2 and 786–0) used for mechanistic studies were deficient in PHD3 protein (Figure D) but not mRNA (Figure B). The high incidence of HIF-α in ccRCC has been partially linked to the mutation of VHL gene. The VHL gene mutation incidence varies from 19.6 to 89.4% in ccRCC [32
] and the majority of reports show 30-60% mutation incidence [34
]. Furthermore, the up-regulation of both HIF-1α (88.2%) and HIF-2α (100%) with only 39.1% VHL mutations was found in ccRCC showing the VHL independent up-regulation of HIF-α in many cases [34
]. Our results suggest a role for PHD2/3 in addition to the well documented VHL mutations in the constitutive expression of HIF-α in ccRCC. A recent report showed the silencing of PHD3 expression by CpG methylation in the promoter region of human cancer cell lines including renal cancer, prostate, breast and melanoma [12
], and in plasma cells and B-cell lymphoma, suggesting PHD3 as a potential biomarker [35
]. In addition, Astuli et al., [36
] found the absence of pathogenic mutations in PHD1, 2 and 3 that could cause renal cell carcinoma. Our western blot analysis showed very weak expression of PHD3 protein compared to PHD2 (Figure C) in two representative primary tumor cases. This weak signal may be derived from the normal stromal cells expressing PHD3 [9
]. These results suggest that there may be some epigenetic regulation of PHD3 expression in ccRCC that might lead to the degradation or inhibition of PHD3 protein. A recent clinical study showed a positive correlation between decreased PHD3 expression and aggressive type of breast tumors [37
]. Similarly, the lack of expression or low incidence/intensity of PHD3 may contribute to the aggressiveness of ccRCC tumors. Thus, the agents that enhance HIF-α degradation by PHD2, independent of PHD3 expression may offer treatment modality that could affect resistance and clinical outcome.
This laboratory is the first to show that therapeutic dose of selenium as highly effective inhibitor of both constitutively expressed HIF-1α, HIF-2α in ccRCC (Figure A and B) and hypoxia induced HIF-1α in head & neck cancer [22
]. Consistent with our data, published results show the degradation of constitutively expressed HIF-1α in prostate cancer [38
] and hypoxia induced HIF-1α in B-cell lymphoma [39
] by selenium. These findings show that both hypoxia induced and constitutively expressed HIF-α are inhibited by selenium suggesting that selenium could inhibit growth of tumors expressing HIF-1α, HIF-2α or both. HIF-α transcriptionally regulated gene, VEGF, is regulated by MSA in renal cancer cells (Figure B). MSA treatment leads to the down-regulation of secreted VEGF in HIF-1α expressing RC2. The lack of MSA effects on secreted VEGF in 786–0 cells could be due to low levels of secreted VEGF in these cells. To our surprise we did not see difference in cytotoxic effects of MSA in RC2 and RC2VHL cells even though there is a marked difference in HIF-1α levels in these cells under normoxic culture conditions. This may be due to the other effects of MSA in these particular cells with VHL transfection. VHL being a multifunctional adaptor molecule involved in the inhibition of HIF-α independent and dependent cellular processes [40
]. The cytotoxic effects of MSA in RC2VHL cells may be through VHL interacting proteins. Our data demonstrate that selenium main target HIF-α is degraded by PHD dependent and VHL independent, but some of our unexpected findings with VHL transfected RC2 cells indicate that VHL transfection may influence the cytotoxic effects of MSA independent of HIF-1α by currently unclear molecular mechanism.
We have demonstrated HIF-α inhibition by selenium as a post-translational degradation mechanism. As shown in the Figure A and B, MSA did not affect HIF-α protein synthesis. In a separate experiment, we have demonstrated that the overall protein synthesis was not altered by MSA using the 35
S-Methionine incorporation studies (Figure C and D). The proteasome inhibitor MG132 reversed the degradation of HIF-α by MSA in FaDu cells (Figure E) demonstrating the proteasome dependent degradation. In contrast, in RC2 cells proteasome inhibition did not reverse the degradation of HIF-1α by MSA suggest that in VHL mutant cells MSA may be degrading HIF-1α through proteasome independent pathway. Further detailed mechanistic studies need to be performed to investigate how MSA is degrading HIF-α in the absence of VHL in ccRCC. Our results also show that MSA is unable to degrade HIF-1α stabilized by DMOG, an inhibitor of PHDs activity (Figure A). DMOG inhibits PHD activity by competing with 2-oxoglutarate, a cofactor for PHDs activity. In addition, gene specific inhibition of PHD2 also prevented the degradation of HIF-1α by MSA (Figure B). Furthermore, we have confirmed VHL independent degradation of HIF-1α by silencing of VHL with siRNA in VHL positive FaDu cells (Figure C). As reported in the literature, VHL knockdown did not lead an increase of HIF-1α in FaDu cells under hypoxic conditions [41
]. These results indicate that selenium utilizes a unique pathway for HIF-1α degradation through PHD2 dependent and VHL independent degradation mechanism. Future studies are warranted to investigate specific function of PHD2 that might be altered by selenium leading to the degradation of HIF-α through another ligase independent of VHL.
Our recent report [22
] and study by Sinha et al., [38
] demonstrated stabilization of PHDs by MSA leads to the degradation of HIF-1α. HIF-1α degradation through VHL dependent and independent pathways is known. Under aerobic conditions, HIF-1α is hydroxylated at 402 and 564 proline molecules by PHDs and recognized by VHL and further degraded by proteasome [42
]. HIF-1α is also degraded without PHD through a small ubiquitin-like modifier (SUMO)ylation that allows the binding of VHL to further degrade HIF-1α by proteasome [44
]. There has been growing evidence for VHL independent degradation of HIF-1α through histone deacetylases (HDACs) inhibition [28
], heat shock protein 90 (HSP90) [45
], the hypoxia associated factor (HAF) [47
] and an undescribed cullin-independent proteasome degradation pathway [29
Based on the demonstrated low incidence of PHD2, lack of PHD3 protein and high incidence of HIF-α in ccRCC, we expect that HIF-α mediated drug resistance is particularly important in this type of cancer. Therefore, decreasing HIF-α expression in ccRCC cells seems to be an important new strategy in order to sensitize tumor cells to the currently used standard therapy. We found MSA treatment lead to 786–0 tumor growth inhibition which correlated with reduced HIF-2α protein levels (Figure ). It is important to indicate that although HIF-1α role in drug resistance has been widely evaluated [49
], to date, efforts have been focused on the development of agents that would effectively inhibit HIF-1α synthesis [50
]. MSC represents a new type of HIF-α inhibitor by enhancing the degradation, but not affecting the synthesis of HIF-α. Currently, it is difficult to predict what approach of HIF-α inhibition combined with chemotherapy will improve the cancer therapy. Furthermore, utilization of clinically more relevant orthotopic imageable mouse models [53
] would be more appropriate for further development of MSC as HIF-α inhibitor in ccRCC.