Previously, we have shown that Siomycin A and thiostrepton inhibit FoxM1 transcriptional activity
[13],
[17],
[18]. We also detected decrease in FoxM1 protein levels after thiopeptide treatment
[13],
[17],
[18] as a result of FoxM1 positive autoregulatory loop
[19]. However, when we evaluated the expression of other cellular proteins following exposure to thiazole antibiotics we found that Siomycin A treatment led to an opposite effect, predominantly to the stabilization of a variety of proteins, such as p21, Mcl-1, p53 and hdm2 (). It has been shown earlier that these proteins are usually upregulated by proteasome inhibitors
[2],
[20] and we found that Siomycin A and proteasome inhibitor MG132 stabilize the expression of these proteins in a similar manner (). This was the first evidence suggesting that thiazole antibiotics may also inhibit proteasome activity.
Since proteasome inhibitors usually hinder the activity of NF-κB via the stabilization of its negative regulator IkB-α
[1],
[2] we examined whether Siomycin A and thiostrepton inhibit NF-kB activity. 293T cells stably expressing an NF-kB-Luc reporter construct were treated with TNF-α and the next day with Siomycin A or thiostrepton, and the luciferase activity was measured. We found that both thiostrepton and Siomycin A suppress NF-kB transcriptional activity () suggesting that thiazole antibiotics inhibit NF-kB activity similarly to proteasome inhibitors
[21].
To directly test whether the thiazole antibiotics inhibit proteasome activity in vitro we compared them with well-known proteasome inhibitors MG132 and lactacystin against the 20S proteasome using the Proteasome Activity Assay Kit (Millipore/Chemicon) (see
materials and methods). We found that thiazole antibiotics are indeed proteasome inhibitors, although not as potent as bona-fide proteasome inhibitors, MG132 and lactacystin (). Since we demonstrated before that the thiopeptides inhibit the transcriptional activity and the expression of FoxM1
[13],
[18] and here we showed that they also act as proteasome inhibitors (), we decided to test the notion that other known proteasome inhibitors may inhibit FoxM1 as well. To evaluate how proteasome inhibitors affect FoxM1 transcriptional activity we used a derivative of U2OS osteosarcoma cell line, C3-Luc cells with a doxycycline-inducible FoxM1-GFP fusion protein and a firefly luciferase under the control of multiple FoxM1 response elements
[13]. Cells were treated with a combination of doxycycline and proteasome inhibitors, and 24 hours later the luciferase activity was measured. We confirmed that all tested proteasome inhibitors strongly inhibited FoxM1 transcriptional activity () similarly to the thiazole antibiotics
[13],
[17],
[18].
In addition, we investigated if proteasome inhibitors suppress FoxM1 protein expression as well. We treated U266 and RPMI8226 multiple myeloma cells, HL-60 leukemia cells and human U2OS-C3 osteosarcoma cells with proteasome inhibitors MG115, MG132 and bortezomib () and analyzed the cell lysates for the levels of FoxM1 by immunoblotting. We found that proteasome inhibitors repress FoxM1 protein expression, suggesting that they antagonize not only the transactivation ability of FoxM1, but they also inhibit its expression () very similarly to Siomycin A and thiostrepton
[13],
[17],
[18]. Moreover, we found by immunoblotting for cleaved caspase-3 and also by annexin V staining that proteasome inhibitors induce apoptosis in a variety of human cancer cell lines (), which correlated with the suppression of FoxM1 ().
To further examine the role of FoxM1 in apoptosis induced by proteasome inhibitors we utilized a U2OS-C3-osteosarcoma cell line with a doxycycline-inducible FoxM1-GFP fusion protein
[13] and tested how proteasome inhibitors induce cell death in the presence and the absence of exogenous FoxM1. FoxM1 expression was induced with the addition of doxycycline and the following day the cells were treated with different concentrations of bortezomib for 24 hours (). We observed that overexpression of FoxM1 protected cells against cell death induced by bortezomib as detected by immunoblotting for cleaved caspase-3 (). In contrast, we found that FoxM1 does not protect against doxorubicin-induced apoptosis (), suggesting that FoxM1 specifically protects cells against proteasome inhibitor-induced apoptosis. Since proteasome inhibitors down-regulate FoxM1 and FoxM1 overexpression protects against cell death mediated by proteasome inhibitors, suppression of FoxM1 may be required for the anticancer activity of these drugs. Further experiments are needed to clarify the mechanisms of FoxM1 down-regulation by proteasome inhibitors and the importance of FoxM1 suppression in the activity of proteasome inhibitors as anticancer drugs.