Selective Aurora A-TPX2 Interaction Inhibitors Have In Vivo Efficacy as Targeted Antimitotic Agents

Aurora A kinase, a cell division regulator, is frequently overexpressed in various cancers, provoking genome instability and resistance to antimitotic chemotherapy. Localization and enzymatic activity of Aurora A are regulated by its interaction with the spindle assembly factor TPX2. We have used fragment-based, structure-guided lead discovery to develop small molecule inhibitors of the Aurora A-TPX2 protein–protein interaction (PPI). Our lead compound, CAM2602, inhibits Aurora A:TPX2 interaction, binding Aurora A with 19 nM affinity. CAM2602 exhibits oral bioavailability, causes pharmacodynamic biomarker modulation, and arrests the growth of tumor xenografts. CAM2602 acts by a novel mechanism compared to ATP-competitive inhibitors and is highly specific to Aurora A over Aurora B. Consistent with our finding that Aurora A overexpression drives taxane resistance, these inhibitors synergize with paclitaxel to suppress the outgrowth of pancreatic cancer cells. Our results provide a blueprint for targeting the Aurora A-TPX2 PPI for cancer therapy and suggest a promising clinical utility for this mode of action.


Table S1. High content toxicology analysis.
Summary of high content toxicology from 72h dose response experiment up to 100 µM of 7 on HepG3 cells.MEC: Minimum effective concentration that significantly crosses vehicle control threshold; AC50: The concentration at which 50% maximum effect is observed for each cell health parameter.First Signal: the cell health feature which responds at the lowest observed dose (marked by •).NR: No response observed.

Figure S12. Electron densities of ligands
Electron densities of ligands 2-10 and CAM2602 in complex with Aurora A after final refinement, all contoured at 1σ level.

Figure S1 .
Figure S1.LO-NMR screening and Fluorescence polarisation anisotropy assay.Top panel: CPMG ligand-observed NMR experiment with analogue of compound 2. Top: 1D spectrum of the compound.Middle: CPMG experiment in the presence of Aurora A and ATP-site binding compound JNJ-7706621.Bottom: As in the middle but with competition with Tyr-pocket binding TPX2 peptide comprising of residues 7-22.Bottom panels: Competitive FP assay with selected compounds.AlexaFluor 488-labelled TPX2 peptide was displaced from Aurora A by increasing concentration of inhibitors.pKD values in each graph represent values from individual experiments in display, while KD values in Figure 2 are averages from 2-3 replicates.

Figure S2 .
Figure S2.ITC analysis of selectivity for Aurora A vs. Aurora B for compounds 7,8, and 9.

Figure S3 .
Figure S3.DiscoverX KINOMEscan specificity screen of compound 9. Compound 9 was applied at 10 µM to proprietary cell-free assays, which estimate kinase target inhibition as a product of blocking association with immobilised ATP.(Left) TREEspot TM plots of the 97 kinases in the screen organised by kinase family where circles indicate target dendrogram position and size/colour indicate degree of inhibition.(Right) List of the 97 kinases screened alongside percent-inhibition detected.

Figure S4 .
Figure S4.Mitotic spindle abnormalities in cells treated with 6. (A) HeLa cells were treated with 50 µM (1x GI50) 6 or DMSO for 6 hours prior to being fixed, fluorescently stained for the indicated proteins or DNA and imaged using confocal microscopy.Two representative fields containing mitotic cells are shown for both treatment conditions.Mitotic cells were enumerated to exhibit spindle abnormalities if they demonstrated unaligned chromosomes and/or non-bipolarity, examples of which are indicated by the solid and outline arrowhead, respectively.(B) Relative proportions of normal and abnormal spindle classes across all imaged mitotic cells for both DMSO and 5 treated cells (>100 mitotic cell observations).Error bars show standard deviations from the mean (n=3 image sets per condition).

Figure S5 .
Figure S5.Toxicity in non-cycling cells.Viable, non-cycling peripheral blood mononuclear cells were grown for 72 hours in the presence of increasing concentrations of the indicated compounds, Alisertib or staurosporine as a positive control.After 72 hours the cell media was supplemented with CellTiter Blue dye to fluorescently quantify the viable cells under each treatment condition.Viability values were normalised against concurrently performed vehicle controls (DMSO for CAM and Alisertib, EtOAc for staurosporine).Concentrations of compounds 5-8; 0.8-400 uM; Alisertib 0.2-100 nM; staurosporine 8 nM-4 uM.

Figure
Figure S6.PH3 levels evaluation with compound 7. (A) Western blot analysis of PH3 levels in Jurkat cells treated with the indicated fold-GI50 equivalents of alisertib (7 or 35 nM) and 7 (20 or 100 µM).(B) Flow cytometric analysis of Jurkat cells treated with a range of fold-GI50 concentrations of 7 (1x GI50 = 20 µM) for 8 h.The cells were stained for DNA, PH3 and P-Thr288 Aurora A and were analysed to determine the proportion of mitotic cells (having both 4n DNA and PH3 positivity); additionally, the proportion of cells positive for P-Thr288 within the mitotic population was also measured per treatment condition.Data is plotted as normalised values relative to the untreated control.

Figure S7 .
Figure S7.Flow cytometry gating strategy to detect Aurora A inhibition biomarkers.(A) Flow cytometric data as summarised in main Fig. 5. (B) Flow cytometric analysis of Jurkat cells treated for 8 hours with either DMSO or 2x GI50 concentrations of 7. From left to right, the four scatter plots show the sequential analysis used to detect the Aurora A biomarker changes resulting from the compound treatment applied using flow cytometry.Each scatter plot represents only those cells within the gated population (box) of the previous plot to the left.The leftmost two panels

Figure S8 .
Figure S8.Pharmacokinetics of CAM2602.CAM2602 was administered at three separate doses in female CD-1 mice and measuring the total concentration of compound in plasma over time.

Figure S10 .
Figure S10.Analysis of blood samples upon conclusion of efficacy study.On the final day of the study, blood samples from each mouse were assessed to determine relative population sizes of each blood cell type per sample.Units are shown adjacent to categories: absolute counts (millions or thousands per microliter, M/μl, K/μl) or weight per volume (g/dL) or percentage parent population, where indicated.Categories: red blood cells (RBC); haematocrit (HCT); haemoglobin (HGB); reticulocytes (RETIC); white blood cells (WBC); neutrophils (NEU); lymphocytes (LYM).Values are means and standard deviations taken from 5 biological replicates.

Figure S11 .
Figure S11.Aurora A:TPX2 PPI inhibitors synergise with paclitaxel in PANC-1 cells.(A) and (B): PANC-1 cells were dosed with a matrix of concentrations of Paclitaxel and 6, including single agent and vehicle controls for all concentrations tested.72 hours following treatment, the cells were assayed for remaining viability relative to vehicle controls.(C) Table showing effective decrease in 6 GI50 in PANC-1 cells when combined with increasing concentrations of paclitaxel.Also shown are the corresponding viability changes effected by paclitaxel if applied as a single agent.(D)The vehicle-normalised viability assay data were processed using SynergyFinder webserver (https://synergyfinder.org/) 57 , producing a heatmap indicating the presence of synergy (red) or antagonism (green) between the two drugging agents when compared to modelled predictions of additivity (E) Chart comparing vehicle-normalised 72-hour viability assay values between single agent and combined treatments of the concentrations of paclitaxel and 6 yielding the greatest synergic effect.The single-agent inhibition values for paclitaxel alone or 6 alone were used to calculate a drug combination surface under the assumption of an additive effect using SynergyFinder, which is shown as the 'predicted' value.Bars show standard deviations from the mean (n=4).

Figure S13 .S25FigureS26FigureFigure S25 .
Figure S13.High-content assay to assess cellular Aurora A engagement by PPI inhibitors.(A) Experimental structure for the HCS assay -timelines proceed from seeding cells to 96-well plates through to the final 2 hour treatment of the cells with compound titrations in the presence of Velcade.The control wells of each plate received either DMSO or doxycycline (for mCherry-TPX2-1-43 peptide induction) immediately following cell seeding, 24 hours prior to the drugging of the compound wells.All wells receive Velcade in the last 2 hours of the assay.(B) Example Aurora A mislocalisation assay images of vehicle control and compound treated HeLa cells after 2 hours of treatment.Left panels show merged-channel images of the fixed cells fluorescently stained for DNA (blue), TPX2 (red) and Aurora A (green).The middle panels show the Aurora A and TPX2 fluorescent stains superimposed onto the software-generated mask of nuclei positions generated from the DNA fluorescence data.Only the TPX2-positive, mitotic nuclei are retained for phenotypic assessment.The right panels show histograms of the two example cell populations where the x-axis shows the observed fluorescence intensity values for Aurora A immunostaining within the limits of the TPX2immunostained mitotic spindle for every mitotic cell.The red line represents the assay threshold, calculated per plate, to the left of which cells exhibit loss of Aurora A from the spindle.