Therefore, the predicted disruption of the PCI-32765 Ubc13-Uev2 heterodimer should be associated with a compromise in tolerance to DNA damage by radiation or radiomimetic drugs in mammalian cells. Additional mechanisms, not explored here but possibly also involved in the chemosensitization caused by compound Ia, could be related to the regulation by Ubc13 of double-strand DNA damage recognition and repair through its interaction with the ubiquitin ligase RNF8. The fact that we have observed inhibition by compound Ia of K63 polyubiquitylation of PCNA only at high concentrations of the compound may suggest either that the compound, although it enters the cells, does not reach the nucleus efficiently, or that K63 polyubiquitylation of PCNA can be catalyzed in mammalian cells by other ubiquitin conjugating enzymes in addition to Ubc13. This may also be the case for K63 polyubiquitylation associated with damage foci in response to DNA double-strand breaks. Indeed, in immunofluorescent cH2AX focus assays, the same batches of compound Ia that inhibited NF-kB activation at low micromolar concentrations only modestly inhibited the maintenance of c-H2AX in ionizing radiation-induced foci. Given the limited effects of compound Ia on both PCNA K63-linked polyubiquitylation and on DNA damage focus formation and resolution, it is possible that the chemosensitization to doxorubicin and etoposide observed in PC-3 and HeLa cells may be better explained by its inhibitory effects on NF-kB signaling. We have observed that compound Ia exerts a direct antitumoral activity in a PC-3 mouse xenograft tumor model. This compound was not directly antiproliferative in vitro for a variety of cell lines tested, but it inhibited the invasiveness of PC-3 cells through extracellular matrix in Boyden chamber experiments, and also inhibited the formation of colonies in 3-dimensional soft-agar cultures. The NF-kB pathway is known to play a prominent role in promoting invasiveness, being constitutively active in PC-3 cells, and thus the observed inhibition of in vitro invasiveness by compound Ia could be one of the consequences of the inhibition of NF-kB activation by this compound. Clonogenicity in soft agar is associated with the capacity of cells for self-renewal, and tends to correlate well with tumorigenicity in vivo. This property, exhibited by distinct cellular subpopulations in some tumors, is not necessarily positively correlated with NF-kB activity, and thus the inhibition by compound Ia of the clonogenicity of PC-3 cells could reflect a requirement 
for Ubc13 activity in other pathways regulating the self-renewal capacity of these cells. In any case, the sum of both activities of compound Ia could GDC-0449 explain at least part of the observed direct antitumoral effect. In summary, we have developed specific and potent small molecule antagonists of the Ubc13-Uev1 interaction that inhibit the enzymatic activity of this heterodimer, K63 polyubiquitylation, and we have shown that one of these molecules produces significant effects in the activation of NF-kB by TNF-a, and in invasiveness and clonogenicity in vitro and tumorigenicity of cancer cells in vivo. Based on these activities, we anticipate that tese compounds should be useful to probe other biochemical pathways and cellular processes regulated by K63 polyubiquitylation and to test their effects in relevant models of human pathologies in which these processes are dysregulated. Many anticancer drugs used in the clinic inhibit cell division as tumors are characterized by uncontrolled proliferation.