Comment on the role of FIH in the inhibitory effect of bortezomib on hypoxia-inducible factor-1

To the editor:

Recently, Shin et al reported that the proteasomal inhibitor bortezomib inhibits the hypoxic response by stimulating the factor-inhibiting HIF-1 (FIH)–mediated repression of hypoxia-inducible factor-1 (HIF-1).¹  The essence of this study is that bortezomib enhances the FIH function and the repression is dependent on asparagine 803 (N803). Both conclusions are in contradiction to previously published data,^2,3  Shin et al suggested that this contradiction could be the consequence of a cell-type specific effect and/or the higher concentrations of bortezomib used in these previous studies.

We tested the effect of various concentrations of bortezomib on the activity of the wild type and N803A mutant Gal4⁻ HIF-1α C-terminal activation domain (CAD) in various cell lines. No inhibitory effect was observed at subnanomolar concentrations in any of 4 cell lines tested, including the HEK 293 cell line used in the Shin et al studies (data not shown). More importantly, we found that in all 4 cell lines activities of the wild type and the N803A mutant HIF-1α CAD constructs were strictly coregulated by bortezomib (Figure 1). Both constructs were invariably inhibited by higher concentrations (≥ 10 nM), whereas 1 nM had a moderate cell type specific effect: inhibition (HEK 293 and M006) and activation (JEG and MCF-7; Figure 1).

Mechanistically, Shin et al concluded that the inhibitory effect of bortezomib was due to stimulation of the physical interaction between FIH and HIF-1α CAD. However, bortezomib treatment had no effect on cellular localization of HIF-1α (mainly nuclear) and FIH (cytoplasmic).¹  In our opinion, transiently enhanced interaction between 2 proteins that eventually end up in different cellular compartments cannot explain the potent inhibitory effect. Should FIH be implicated, before being released, HIF-1α would have to be inactivated (presumably permanently) by hydroxylation of N803. It is not clear, however, how this hydroxylation could efficiently proceed under conditions of hypoxia, the major physiologic inhibitor of FIH, without a significant up-regulation of FIH. Shin et al failed to discuss the conclusion of Birle and Hedley that FIH does not play a role in the inhibitory effect of bortezomib.³  In addition, their statement that “demonstration that CAD activity was regulated by FIH expression or knock-down even at an O₂ tension of 1%, indicating that FIH regulates CAD even in hypoxia” was misassigned to our studies.² 

Their statement that “bortezomib inhibited HIF-1α more so than proteasome” cannot be justified exclusively by the absence of stabilized HIF-1α or CITED2 (cAMP-responsive element–binding protein [CBP]/p300-interacting transactivators with glutamic acid [E] and aspartic acid [D]–rich tail 2) in bortezomib-treated samples (bortezomib inhibits chymotrypsin-like activity of the proteasome with K_i = 0.6 nM).⁴  Nevertheless, Shin et al speculated that the HIF-inhibitory effect of subnanomolar concentrations of bortezomib “may not be attributable to proteasome inhibition.” Bortezomib is considered a highly specific inhibitor of proteasome and before being approved for clinical use, extensive screening found no other intracellular targets.⁴ 

Although we are not clear about the underlying cause(s), the discrepancies brought up in this Letter argue against the universal appeal of the mechanism outlined by Shin et al. More work is required before the mechanism of inhibition of HIF function by bortezomib is fully understood.