Discovering cell-active BCL6 inhibitors: effectively combining biochemical HTS with multiple biophysical techniques, X-ray crystallography and cell-based assays

Abstract

By suppressing gene transcription through the recruitment of corepressor proteins, B-cell lymphoma 6 (BCL6) protein controls a transcriptional network required for the formation and maintenance of B-cell germinal centres. As BCL6 deregulation is implicated in the development of Diffuse Large B-Cell Lymphoma, we sought to discover novel small molecule inhibitors that disrupt the BCL6-corepressor protein-protein interaction (PPI). Here we report our hit finding and compound optimisation strategies, which provide insight into the multi-faceted orthogonal approaches that are needed to tackle this challenging PPI with small molecule inhibitors. Using a 1536-well plate fluorescence polarisation high throughput screen we identified multiple hit series, which were followed up by hit confirmation using a thermal shift assay, surface plasmon resonance and ligand-observed NMR. We determined X-ray structures of BCL6 bound to compounds from nine different series, enabling a structure-based drug design approach to improve their weak biochemical potency. We developed a time-resolved fluorescence energy transfer biochemical assay and a nano bioluminescence resonance energy transfer cellular assay to monitor cellular activity during compound optimisation. This workflow led to the discovery of novel inhibitors with respective biochemical and cellular potencies (IC_50s) in the sub-micromolar and low micromolar range.

Figure 1

Summary of HTS campaign and hit validation by orthogonal biophysical techniques. An HTS was carried out using an FP assay with an FP-probe based on a BCOR corepressor peptide. The primary HTS hits were subjected to a first confirmation step at three concentrations in the FP assay. Hits confirmed in this step were progressed to IC₅₀ determination in the FP assay using repurchased material. Active hits were characterised in orthogonal biophysical techniques including TSA, SPR and LO-NMR using a ¹⁹F-labelled WVIP reporter peptide. Compound selection for the respective biophysical assays was heavily based on experimental aqueous solubility experiments on the confirmed hits using NMR and HPLC. Hits confirmed to bind the BCL6-BTB domain in one or more biophysical methods were selected for binding mode elucidation by X-ray crystallography which yielded 12 protein–ligand structures representing 9 different chemical series.

Figure 2

SPR results for the two hits showing the highest affinity for the BCL6 BTB dimer. (A) Sensorgrams of compound 17 at different concentrations. (B) Sensorgrams of compound 21 at different concentrations. (C) Corresponding binding curves and calculated K_Ds for compounds 17 and 21.

Figure 3

Hit validation in ¹⁹F Ligand-Observed NMR. The TFA-WVIP-NH₂ tetrapeptide was used as a reporter in a ¹⁹F-LO-NMR peptide displacement assay. (A) The SMRT peptide and compound 21 (bright green and purple traces), but not the negative control 4-chlorobenzoic acid (ClBA, blue trace), were able to displace the BCL6-bound TFA-WVIP-NH₂ peptide, as indicated by the recovery of the peptide fluorine signal. (B) The LO-NMR data correlated well with biochemical FP potency, with the 11 strongest FP hits yielding a peptide displacement greater than 50% in LO-NMR.

Figure 4

Natural corepressors, WVIP peptide, and HTS hits bind to the same area on the surface of BCL6 BTB domain dimer. (A) Overlay of the BCL6 BTB domain dimer bound to the SMRT corepressor peptide (dark green, PDB ID 1R2B), the BCOR corepressor peptide (light green, PDB ID 3BIM) and the WVIP peptide bound at site 1 (magenta) and site 2 (blue). The BCL6 dimer is shown in ribbon representation with the two individual monomers coloured in grey and cyan with a grey semi-transparent surface superimposed. (B) Sequence alignment reflecting the structural overlay of the corepressors and the two molecules of the WVIP peptide bound at site 1 and site 2, respectively. (C) Overlay of BCL6 BTB domain dimer bound to the 12 structurally confirmed HTS hits (magenta). The colour scheme and surface representation of the BCL6 dimer are the same as in panel (A). The HTS hits are shown as magenta ball and sticks. The N-terminal uncleaved TEV sequence present in the Flag-TEV-BCL6 construct, used for some of the BCL6-inhibitor structures and binding at the second WVIP binding site (site 2), is shown as a yellow ribbon.

Figure 5

TR-FRET assay development and correlation with the FP assay. (A) Schematic representation of the BCL6 BTB – BCOR peptide interaction as measured by TR-FRET assay format. (B) Unlabelled BCOR peptide competed with the Alexa-633 conjugated BCOR peptide in the TR-FRET reaction with an IC₅₀ of 8 ± 0.8 µM, n = 20. (C) The two most potent HTS hits compound 17 and 21 showed respective IC₅₀ values of 54 and 70 µM in the TR-FRET assay. (D) Correlation between the TR-FRET (x-axis) and the FP (y-axis) assays for the benzimidazolone series of BCL6 inhibitors. The BCL6 BTB concentration was 3 µM in the FP and 10 nM in the TR-FRET assay. The two biochemical assays correlated with a R² of 0.82, but most compounds displayed higher potencies in the TR-FRET assay compared to the FP assay (red dotted trendline above the black dotted unity line). As compound potencies approached the tight binding limit of the FP assay (IC₅₀ < 5 µM), the correlation between the two assays decreased.

Figure 6

InCELL Hunter™ cellular target engagement assay. (A) HEK293T cells, transfected with DNA plasmids expressing complete C-terminally ePL-tagged BCL6 BTB (amino-acids 1–135) or truncated BCL6 BTB (amino-acids 14–135), were incubated with the oxindole compound 25 (30 µM) for 6 h before luminescence was detected on a plate reader (error bars represents standard deviation from n = 3); (B) HEK293T cells, transfected with C-terminally ePL-tagged truncated BCL6 BTB 14–135, were incubated with increasing concentrations of the benzimidazolone compound 27 for 6 h before luminescence detection (EC₅₀ 24 ± 9 µM, error bars represents standard deviation from n = 3).

Figure 7

NanoBRET assay sensitivity and activity of benzimidazolone compounds. (A) Effect of the NanoBRET configuration on the assay sensitivity to compound inhibition in HEK293T cells. The blue histograms show the NanoBRET ratio for the C-terminally NanoLuc tagged BCL6 and HaloTag tagged SMRT with and without the BCL6 inhibitor 27. The red histograms show the ratio for the N-terminally HaloTag tagged BCL6 and NanoLuc tagged SMRT with and without 27. We selected the former because of its higher sensitivity to compound inhibition after 6 h treatment with 25 µM 27. (B) Concentration responses for 27 in both NanoBRET configurations. An IC₅₀ value could be determined with C-terminally NanoLuc tagged BCL6 (blue triangles), but not when BCL6 was tagged N-terminally with the Halo-Tag (red diamonds). (C) Schematic representation of the optimal NanoBRET configuration for maximal sensitivity to compound inhibition. (D) Plot showing the correlation (R² = 0.62, n = 175) between biochemical (TR-FRET IC₅₀ on x-axis) and cellular (NanoBRET IC₅₀ on y-axis) activities of inhibitors from the benzimidazolone series. Data points are above the unity line (black), indicating cellular activities were on average an order of magnitude lower than biochemical potencies.