We established that DIM inhibited HIF-1α accumulation and HIF-1 activity in hypoxic cells and that the inhibitors of prolylhydroxylase, CPX, DPL and DMOG, and of the proteasome, MG132, could reverse this inhibition. DIM inhibited the expression of major HIF-1-regulated endogenous genes, most notably the pro-angiogenic factors, furin and VEGF, in a concentration-dependent manner in hypoxic tumor cells. We confirmed that DIM is a strong inhibitor of mitochondrial F1F0-ATPase and determined that DIM binds to the oligomycin-binding site in the transmembrane F0 domain. This inhibitory effect on respiration was accompanied by an increase in intracellular levels of O2 and very strong increases in ROS production with exposure to the higher concentrations of DIM. Although the ROS were efficiently quenched by co-treatment with antioxidants, including DTT, as well as NAC, ascorbate, GSH, vitamin E and vitamin C (data not shown), the antioxidant treatments did not reverse the effects of DIM on HIF-1α accumulation or HIF-1 activities. In addition, inhibitors of the stress kinases, PI3K and p38, as well as catalase over-expression (data not shown), did not counter the effects of DIM. These results suggest that the effects of DIM, especially on HIF-1α degradation, are not likely to result from ROS release and related oxidative stress induced signaling.
The more likely mechanism for the lower concentrations of DIM is the reactivation of the prolyl hydroxylation degradation pathway as a result of an increase in levels of cellular O2
. Indeed, the inhibitory effects of DIM were accompanied by an approximately 2-fold increase in O2
levels in hypoxic cells. Furthermore, previous studies have shown that several established inhibitors of mitochondrial electron transport, including myxothiazol, sodium azide, antimycin A, rotenone, nitric oxide and oligomycin, all can inhibit HIF-1α accumulation in hypoxic cells and that this effect is not countered by co-treatments with any of several antioxidants [30
]. These published studies further showed that cotreatments with prolyl hydroxylase inhibitors (DMO and DFO) or the proteasome inhibitor (MG-132) largely reversed the stabilizing effects of the respiratory inhibitors on HIF-1α levels. Thus, the results we present for DIM are consistent with these previous reports on the effects of other respiration inhibitors and provide strong support for the hypothesis that the inhibitory effects of the lower concentrations of DIM on F1F0-ATPase with resultant accumulation of cellular O2
are responsible for the reactivation of the prolyl hydroxylase-dependent degradation pathway of HIF-1α.
Our results show further that exposure of cells to the higher concentrations of DIM (>25 μM) results in a near total oblation of HIF-1α expression. In this concentration range, the rate of HIF-1α synthesis continues to decline, while the rate of HIF-1α degradation shows no further increase, (c.f. .c.) Although the mechanism of this aspect of DIM activity is yet to be fully understood, this effect is consistent with reports of activities of other inhibitors of mitochondrial respiration. Indeed, studies with mitochondrial thioredoxin (Trx2) have shown that over expression of this protein can inhibit translation of HIF-1α by a mechanism that involves the attenuation of activities of Akt, p70S6K and eIF-4E, which together mediate HIF-1α translation [34
] It is suggested that the observed down-regulation of Akt signaling is a result of ROS production arising from a disruption of mitochondrial membrane potential by over-expressed Trx2. We have shown that DIM can produce an increase in mitochondrial membrane protential as reported for Trx2 [27
]. In addition, we and others have observed that DIM can inhibit Akt signaling as does Trx2 over expression, which is consistent with similar modes of action of high concentrations of DIM and Trx2 on HIF-1α translation [35
]. Although our results indicated a strong increase in ROS production along with a strong decrease in HIF-1α synthesis with the higher concentrations of DIM, the specific role of ROS in this inhibition is yet to be determined. Although we observed no effect of soluble antioxidants on HIF-1α oblation by high concentrations of DIM, the possibility remains that the effects might arise from a compartmentalized production of ROS in mitochondria that might not be quenched by the soluble antioxidants, as was suggested for the Trx2 studies [34
The oxygenation effect of DIM may have consequences for tumor therapy beyond those resulting from a decrease in HIF-1α accumulation in hypoxic tumor tissue. Oxygenation of tumors ranges from 30% to less than 10% the oxygenation level of normal tissues [36
]. A major result of these low oxygen levels is a resistance of tumor tissue to radiation therapy [37
]. Although there is considerable interest in developing means of increasing tumor sensitivity to radiation, only a few substances are reported to be successful in this regard. Recent studies have shown that one such substance, nelfinavir, exhibits antiangiogenic activity by a mechanism that involves inhibition Akt signaling and VEGF expression, and down-regulation of HIF-1α expression in hypoxic tumor cells [36
]. In addition, this drug was shown to increase oxygenation of xenograft tumor tissue and increase the sensitivity of the tumors to radiation. In light of the similarities in the effects of DIM and nelfinavir on Akt and HIF-1 signaling and angiogenesis, it is reasonable to suggest that DIM could increase tumor oxygenation and sensitivity to radiation, as well.
HIF-1 has become an attractive target for the development of anti-cancer drugs [38
]. Studies in many laboratories have shown that potential therapeutic agents can disrupt the HIF-1 signaling pathway through a variety of mechanisms, including the inhibition of HIF-1α protein synthesis, nuclear translocation, HIF-1 transactivation of target genes, and stabilization [39
]. The inhibitory effects of DIM on mitochondrial respiration through inhibition of F1F0-ATPase appear to be primarily responsible for the increase in HIF-1α degradation and decrease in HIF-1α transcription that we observed. It is of interest to note that a group of cancer protective phytochemicals, including resveratrol, have been shown to both down-regulate HIF-1α in hypoxic cells and to inhibit mitochondrial F1F0-ATPase [40
]. Thus, F1F0-ATPase and other components of the mitochondrial respiratory chain, appear to be important molecular targets for a relatively large group of potentially useful antiangiogenic agents from plants.