Discovery of the Hedgehog Pathway Inhibitor Pipinib that Targets PI4KIIIß

Abstract The Hedgehog (Hh) signaling pathway is crucial for vertebrate embryonic development, tissue homeostasis and regeneration. Hh signaling is upregulated in basal cell carcinoma and medulloblastoma and Hh pathway inhibitors targeting the Smoothened (SMO) protein are in clinical use. However, the signaling cascade is incompletely understood and novel druggable proteins in the pathway are in high demand. We describe the discovery of the Hh‐pathway modulator Pipinib by means of cell‐based screening. Target identification and validation revealed that Pipinib selectively inhibits phosphatidylinositol 4‐kinase IIIβ (PI4KB) and suppresses GLI‐mediated transcription and Hh target gene expression by impairing SMO translocation to the cilium. Therefore, inhibition of PI4KB and, consequently, reduction in phosphatidyl‐4‐phosphate levels may be considered an alternative approach to inhibit SMO function and thus, Hedgehog signaling.

a C3H10T1/2 cells were incubated with 1.5 µM Purmorphamine and the compounds or DMSO as a control for 96 h. The activity of alkaline phosphatase, an enzyme that is expressed upon induction of osteogenesis, was assessed by means of a luminescence readout. Data were measured in triplicates. Viability was measured in the same setting using a CellTiter Glo Kit (Promega). IC50 values > 10 µM were obtained when the inhibition achieved up to a concentration of 10 µM was not sufficient to calculate an IC50 value.
Supporting Table S2 (related to Figure 3) Results of the kinase panel with 10 µM Pipinib. Data can be found in the included Excel file.f Supporting Table S3. Kinase panel hits. Kinases for which inhibition or displacement by 10 µM Pipinib was clearly above 50 % were considered hits. Assays were performed at SelectScreen TM , Thermo Fisher Scientific.

S17
Supporting Figure S7 (related to Figure 5). SMO trafficking. NIH/3T3 cells were serum starved for 24 h to induce ciliation prior to incubation with 10 µM Pipinib or DMSO as a control for 24 h, followed by co-incubation with 2 µg/ml ShhN for another 24 h. Cilia were then stained with an antibody against acetylated tubulin (ac-tubulin, green), SMO was stained with an anti-SMO antibody (red) and the nucleus was marked with DAPI (blue). Full-size images belonging to Figure 5a, representative of three independent experiments (n=3, scale bar: 50 µm).
Supporting Figure S8 (related to Figure 6). Dose dependency of Pipinib-mediated inhibition of target gene expression after siRNA knockdown of PI4KB. NIH/3T3 cells were incubated with 30 nM PI4KB siRNA or non-targeting siRNA as a control for 48 h prior to treatment with Pipinib for another 48 h. Cells were subjected to RNA isolation, cDNA preparation and quantitative PCR using primers specific for Ptch1 and Gli1 and Gapdh as a reference gene. Values for Purmorphamine-DMSO-treated cells were set to 1. Data are mean values ± SD of three independent experiments (n=3). a. Expression of Ptch1. b. Expression of Gli1. IC50 values for suppression of Ptch1 (c) or Gli1 (d) expression by Pipinib were compared for cells with or without PI4KB depletion (from a and b). Statistical analysis was performed using two-tailed t-test. *: p<0.05; ns: not significant.
Supporting Figure S9 (related to Figure 5 serum-reduced medium and were again incubated for 24 h. Afterwards, the ShhN-conditioned medium was harvested for a second time, as described above. For assays, both fractions were combined and diluted with fresh serum-reduced medium in a ratio of 3:1.

Osteogenesis Assay
The osteogenesis assay was performed as published before. [

Reporter gene assay
The reporter gene assay was performed as published before [10]

Reverse Transcription Quantitative PCR (RT-qPCR)
The RT-qPCR was performed as published before 3

Smoothened trafficking assay (Ciliary localization of SMO)
The in PBS, once in PBS and mounted on glass slides. Imaging was performed with a Leica SP8 or SP5 confocal microscope, equipped with 63x 1.4 NA objective (Leica Microsystems CMS GmbH, Mannheim, Germany). Images were acquired as Z-sections and converted to maximal intensity projections using the software ImageJ 7 for illustrative purposes. The staining protocol was adopted (with modifications) from Rohatgi et al. 8 . Quantification was carried out employing ImageJ 7 . For this, acetylated tubulin-positive cilia were marked and the intensity of the anti- Smoothened antibody was measured within this area. Statistical significance was assessed using an unpaired t-test (GraphPad Prism 6, GraphPad Software, USA) with a confidence level of 95% (*=p<0.05, **=p<0.005, ***=p<0.0005, ****=p<0.0001). Data of three independent experiments was used to generate the data presented.

Smoothened binding assay
Microscopy analysis of BODIPY-Cyclopamine labeling of cells was performed as described by Sinha et al. 9 . Briefly, HEK293T cells were seeded on cover slips (Thermo Scientific, 12 mm, #10062491) at 1.

End-point evaluation of PI4P levels
For detection via immunofluorescence staining, NIH/3T3 cells were seeded into 24-well plates

Live cell evaluation of PI4P levels
For transient expression, Lyn-CFP plasmid was designed as previously described 11

CRISPR/Cas9 mediated PI4KB knockdown
The gateway cloning protocol of the Zhang lab was followed 13   to only obtain successfully transfected cells. The top 10% of GFP-positive cells were re-plated into a 10 cm dish at a very high dilution to allow growth of single clones. Cells at a higher density were seeded into a 6-well plate to use these bulk cells for optimization of screening primers.
Selected colonies were trypsinized, expanded and analysed for successful PI4KB knockout using an anti-PI4KB antibody after cell lysis and immunoblotting. Clones that were validated on protein level were expanded, cryo-conserved and used for further experiments.

ActiveX affinity chromatography
The enrichment of ATP binding proteins was carried out using the Pierce™ Kinase Enrichment The mass accuracy was set to 20 ppm for the first and 6 ppm for the second search. The false discovery rates for peptide and protein identification were set to 0.01. Only proteins for which at least two peptides were quantified were chosen for further validation. Relative quantification of proteins was carried out using the label-free quantification algorithm implemented in MaxQuant.
All experiments were performed in technical triplicates. Statistical data analysis was performed using Perseus 17 , Label-free quantification (LFQ) intensities were logarithmized (log2) and samples resulting from affinity purification using the small molecule as a competitor were grouped together and samples resulting from affinity purification using DMSO as solvent control as well. Proteins which were not at least three times quantified in at least one of the groups were filtered off. Missing values were imputed using small normal distributed values and a t-test was performed. Proteins which were statistically significant (FDR: 0.05) outliers and enriched in the samples treated with DMSO over the ones treated with compound competitor were considered as hits.

Cellular Thermal Shift Assay (CETSA)
The compound (or DMSO as a control) was added to 900 µg NIH/3T3 lysate at a concentration of 50 µM for 10 min at room temperature. The lysate was then aliquoted in nine PCR tubes and

Biochemical kinase measurements
If not stated otherwise, all kinase measurements were performed by SelectScreen (Thermo Fisher Scientific). All lipid kinases were measured in the Adapta activity assay that monitors ADP formation via a fluorescence resonance energy transfer (FRET)-based ATP-depletion reaction. The majority of kinases measured in the panel was subjected to the Z'Lyte activity assay that measures kinase activity by monitoring the conversion of a synthetic peptide substrate in a FRET-based manner. The Lantha binding assay is a FRET based displacement assay. For MYLK, only the Lantha binding assay was available at SelectScreen. To assess compound influence on kinase activity, we employed the radiometric activity assay offered by Reaction biology. In this assay, 33 P-ATP is added to the kinase reaction and after 2 h at RT, the reaction mixture is spotted onto P81 ion exchange paper to separate kinase substrate, which has been phosphorylated with radioactive phosphate, and 33 P-ATP. The amount of radiolabeled substrate is then measured by scintillation counting. For all kinase assays, data were normalized to DMSO treated samples.

Mode of inhibition study
The luminescent ADP-Glo assay was performed at SignalChem to measure ATP consumption as per the manufacturer's instructions. Full-length recombinant human PI4KB was co-expressed at SignalChem by baculovirus in Sf9 insect cells using an N-terminal GST tag. For the assay, different concentrations of compound were titrated against different ATP concentrations.

S36
Relative light units (RLU) were converted to ATP conversion rate using standard ATP conversion curves. Reaction velocities were calculated based on the amount of ATP transferred in the reactions. For determination of the mode of inhibition, Km and Vmax were determined using the equation Y= Vmax*X/(Km+X) using GraphPad Prism. Alternatively, the inverse of the ATP concentration was plotted against the inverse velocity in a Lineweaver-Burk plot.

Experimental procedures -Chemical synthesis
The chemicals used were purchased from Alfa Aesar and Sigma Aldrich and were used as
Then, DMF was removed and the resulting oily product was precipitated by adding 5 mL water.
The solid material was washed 3 x 5 mL MeOH and dried to yield a pale yellow solid (yield = 56%) which was subsequently used in the next step without any further purification. Compound identity has been assessed using LC-MS. Entries 1-11 have been synthesized following previously published methods. [9] In our hands, changing the solvent to THF lead to superior yields and therefore followed the general procedure: to i2 (28.6 mg, 0.1 mmol, 1 eq) has been added the corresponding primary or secondary amine (1 mmol, 10 eq) in THF (0.3 mL). The mixture was let to react in a sealed tube at 94 o C for 36-48 h while monitoring the reaction progression using TLC. The reaction was stopped, concentrated in vacuo and purified using column chromatography on silica (eluent: EtOAc : PE = 1:2 to EtOAc : PE = 2:1). The same experimental procedure has been applied to entries 12-16. Note: entries 1 and 12 were synthesized in the absence of THF, using neat isopropylamine.

PI4KA inhibitor
AZ7 was obtained as a kind donation by Astra Zeneca [18] and was later on purchased from Ximbio.

N 2 -(3,5-dimethoxyphenyl)pyrazine-2,3-diamine
A suspension of 3-chloropyrazin-2-amine (500 mg, 3.86 mmol) and 3,5-dimethoxyaniline (621 mg, 4.05 mmol) in 20 mL HCl (aq., 0.1 M) was heated in the micro wave to 120°C for 2 hours. After bringing the reaction mixture to room temperature it was diluted and thereafter stirred for 5 minutes with 20 mL EtOAC before the two layers were separated. The aqueous layer was extracted three times with in each case 20 mL EtOAc and the combined organic layers were dried over anhydrous magnesium sulfate. The crude was purified by flash column chromatography on silica gel (10-60% EtOAc / Cyclohexane) to yield the product (293 mg, 1.19 mmol, 31%) as an off-white wax.