Design and Development of Small Molecules for Specific Targeted Therapy of Diffuse Large B-Cell Lymphoma.

BCL6 is the most commonly involved oncogene in diffuse large B-cell lymphoma (DLBCL). BCL6 mediates lymphomagenesis through direct transcriptional repression of target genes including p53 and ATR. BCL6 represses these genes by recruiting three corepressors: SMRT, N-CoR and BCoR. This interaction is mediated by an 18 amino acid corepressor sequence with a unique “lateral groove” motif located in the BCL6 BTB domain. We previously showed that a peptide mimic of the corepressor BCL6-binding sequence could inhibit BCL6 and kill DLBCL cells in vitro and in vivo, without toxicity to normal tissues. Although, transcription factors such as BCL6 that function through protein-protein interactions were traditionally regarded as “undruggable”, we hypothesized that a detailed structural and biochemical analysis might facilitate the design of small molecule inhibitors. Based on our crystal structure of the BCL6-corepressor complex, we used a computational strategy to screen one million small molecules for their ability to potentially dock to a specific lateral groove region. The chemistry of protein contacts in this region led us to predict that small molecules that could dock to this site would have the best chance of destabilizing the BCL6-corepressor complex. Among the top-scoring 100 molecules from this screen, we identified 10 compounds that could specifically inhibit the repressor activity of the BCL6 BTB domain in reporter assays and that displayed direct binding to purified BCL6 BTB domains. Using these leads as molecular scaffolds we generated small libraries of molecules derived from each parental compound. The most active of these families was called “the 57 series”. Series 57 compounds could all specifically block BCL6 repression in reporter assays, and disrupt corepressor/BCL6 complexes at low micromolar concentrations as shown in fluorescence polarization assays. X-ray crystallography of the most active member of the 57 family (called “57-6”) showed that the small molecule docked as predicted in the critical region of the lateral groove. Moreover, 57-6 induced an allosteric conformational change in the entire lateral groove that explains how these small molecules so effectively disrupt the BCL6/corepressor complex. 57-6 was also biologically active, since it could induce expression of BCL6 target genes including p53 and ATR in BCL6-positive DLBCL cells as shown by QPCR. 57-6 had no effect on negative control genes nor in BCL6-negative DLBCL cells. The mechanism of action was confirmed in ChIP assays showed that 57-6 abrogated BCL6 mediated corepressor recruitment to BCL6 target genes but had no effect on negative control genes. Most importantly, 57-6 specifically killed BCL6-positive DLBCL cells but had no effect on BCL6-negative DLBCL cells. A dose escalation experiment in mice revealed no toxic effects. In xenotransplantation experiments, 57-6 potently inhibited the growth of already established human DLBCL tumors in mice, again without toxicity to other organs. In summary, we used a rational approach to design specific and potent small molecule inhibitors of BCL6, which could serve as targeted agents for DLBCL in clinical trials. Our data show that transcription factors are druggable targets that can be harnessed to potentially improve cancer therapy.