Compounds

CLR457 and SHP099 were obtained from Novartis (Basel, Switzerland and Cambridge, USA), and Sunitinib from Pfizer Inc (Sutent). Compounds were prepared as 10 mM stock solutions in DMSO and stored protected from light at –20 °C. CLR457 (20 mg/kg), SHP099 (100 mg/kg) and Sunitinib (60 mg/kg) were freshly formulated in methylcellulose / Tween-80 (0.5% / 0.5%) and administered to mice by oral gavage at 5 ml/kg.

Animal experiments

All in vivo experiments were performed in accordance with the Swiss animal welfare ordinance and approved by the cantonal veterinary office Basel Stadt. Female severe combined NOD-scid IL2rγnull (NSG) and Balb/c animals were maintained in the Friedrich Miescher Institute for Biomedical Research and the University Department of Biomedicine animal facilities in accordance with Swiss guidelines on animal experimentation. For orthotopic engraftment of cell lines, 0.3 × 106 4T1, 2 × 106 MDA-MB-436, and 2 × 106 MDA-MB-468 cells were suspended in 50 µL PBS and injected into mammary fat pad number 4 of 8-week-old mice. Tumor-bearing mice were randomized based on tumor volume prior to the initiation of treatment, which started when average tumor volume was at least 80 mm3. CLR457 was administered twice a day and SHP099 once daily. Tumors were measured every 3 – 4 days and tumor volumes calculated by the formula 0.5 x (larger diameter) x (smaller diameter)2. End-point tumor sizes were analyzed for synergism using the formula AB/C < A/C x B/C, where C is tumor volume Vehicle, A is tumor volume compound 1, B is tumor volume compound 2, and AB is tumor volume combination [46]. For survival studies, day 0 corresponds to tumor removal and animals were sacrificed as soon as they showed any signs of distress (e.g., breathing disorders, weight loss, or immobility).

Cells, cell culture and reagents

SUM159 were propagated in Nutrient Mixture F-12 supplemented with 5% fetal calf serum, 0.5 μg/ml hydrocortisone, and 10 μg/ml insulin (all from Sigma), 100 IU/ml penicillin, 100 μg/ml streptomycin and 100 μg/ml Normocin (InvivoGen). Balb/c tumor-derived mammary cancer lines 4T1 were propagated in DMEM, with 10% fetal calf serum (all from Sigma), 100 IU/ml penicillin, 100 μg/ml streptomycin and 100 μg/ml Normocin (InvivoGen). MCF10A-HER2/HER3 [4] were propagated in DMEM/F12 medium (Invitrogen) supplemented with 5% horse serum (Hyclone), 20 ng/ml EGF (Peprotech), 0.5 μg/ml hydrocortisone, 100 ng/ml cholera toxin, and 10 μg/ml insulin (all from Sigma), 100 IU/ml penicillin, 100 μg/ml streptomycin and 100 μg/ml Normocin (InvivoGen). All other cell lines were obtained from and were cultured according to the protocols of the American Type Culture Collection. Profiling of human cell lines used highly-polymorphic short tandem repeat loci sequencing(STRs) (Microsynth). For treatment with inhibitor(s), cells were synchronized with 0.5% serum for 6 h to avoid masking effects of growth factors present under full-serum conditions.

Cell number assay

Different cell lines were cultured overnight in 96-well plates at 1,000 to 5,000 cells/well before culture medium containing 0.5% FCS (or HS) and the inhibitor(s) described above were added. The culture medium with inhibitor(s) was renewed 48 h after initial treatment and cells were fixed 24 h later. Cell fixation, staining and quantification were performed using the Sulforhodamine B colorimetric assay. Values measured using the Sulforhodamine B colorimetric assay were analyzed for synergism using the formula AB/C < A/C x B/C, where C is value in DMSO condition, A is the value of inhibitor 1, B is the value of inhibitor 2, and AB is the inhibitors combination value.

Immunoblotting and Phospho-RTK arrays

Cells were lysed with RIPA buffer (50 mM Tris–HCl pH 8, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with 1 × protease inhibitor cocktail (Complete Mini, Roche), 0.2 mM sodium orthovanadate, 20 mM sodium fluoride and 1 mM phenylmethylsulfonyl fluoride. Lysates from xenografts were prepared by lysing kryo-homogenized tumor powder in RIPA buffer. Whole cell lysates (30–80 μg) were subjected to SDS-PAGE, transferred to PVDF membranes (Immobilon-P, Millipore) and blocked for 1 h at room temperature with 5% milk in PBS-0.1% Tween-20. Membranes were then incubated overnight with primary antibodies as indicated and exposed to secondary HRP-coupled anti-mouse or anti-rabbit antibodies at 1:7,500 for 2 h at room temperature. The following antibodies were used: anti-pAKT (Ser473, Cell Signaling, #4060), anti-pERK1/2 (Thr202/Tyr204, Cell Signaling, #4377), anti-pPDGFRβ (Tyr751, Thermo Scientific, MA-14823), anti-ERK2 (Santa Cruz, sc-1647), anti-pS6 (Ser235/236, Cell Signaling, #2211), anti-PDGFRβ (Cell Signaling, #3169), anti-AKT (Santa Cruz, sc-5298) and anti-Vinculin (Thermo Scientific, #14–9777-80). Phospho-RTK arrays on tumor lysates were performed using the Proteome Profiler Mouse Phospho-RTK Array Kit (R&D systems) according to the manufacturer’s protocol.

Co-immunoprecipitation

4T1 cells were plated in 10 cm dishes, starved for 6 h in 1% Serum medium and treated for 24 h with DMSO or SHP099 at 5 μM. Cells were washed in ice-cold PBS and lysed in RIPA Buffer (50 mM Tris–HCl pH 8, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with 1 × protease inhibitor cocktail (Complete Mini, Roche), 0.2 mM sodium orthovanadate, 20 mM sodium fluoride and 1 mM phenylmethylsulfonyl fluoride, for 30 min at 4 °C. Cell lysates were sonicated and protein abundance was quantified. Input fraction was recovered and 500 μg of proteins were subjected to immuno-precipitation with 2 μg of anti-p85 antibodies (Cell Signaling, #4257) or 2 μg of control IgG (Cell Signaling, #3900), and 100 μl of binding buffer (Thermo Scientific, #10007D). After overnight incubation at 4 °C on a turning wheel, 50 ul of protein A magnetic beads (Dynabeads, Thermo Scientific, #10002D) was added to the mixture for 4 h more. Immune-complexes were washed 3 times using washing buffer (Thermo Scientific, #10007D) on a magnetic rack, and extracted using 20 μl of RIPA buffer supplemented with 2 × Laemmli buffer. Input and immune-complexes fractions were subjected to proteins analysis by western-blot using anti-p85 and anti-PDGFRβ antibodies.

Proximity ligation assay

The DUOLink in situ Detection Reagent Red kit (Sigma – DUO92008) was used to detect and quantify PDGFRβ and p85 interactions. DMSO- or SHP099-treated 4T1 and MDA-MB-468 cells seeded on glass coverslips (24 well-plate) were fixed in 4% PFA for 15 min at RT, permeabilized with PBS 0.1% Triton for 30 min at RT and blocked with Duolink blocking buffer for 1 h at 37 °C. Cells were then incubated overnight at 4 °C with primary antibodies directly coupled to the DUOLink DNA probes using the in situ Probemaker PLUS kit (Sigma – DUO92009) diluted in DUOLink Probemaker PLA probe diluent at 1/100. The PLA reaction was then performed according to the manufacturer’s instructions. Briefly, cells were washed in buffer A (provided) and the two steps of ligation of the DNA probe and rolling-circle amplification (RCA) were carried out for 30 min and 100 min respectively in a humid dark chamber at 37 °C. Cells were then washed in buffer B (provided) and the coverslips mounted on slides using the Duolink Mounting Medium with DAPI. Z-stacked images were acquired with a Nikon Ti2 microscope at 40 × magnification using Dapi and 555 channels. Images were analyzed using the FIJI software and represented as a number of dots (interactions) per cell, with hundreds of cells analyzed per experiment.

Immunohistochemistry

Tumors were fixed in 10% neutral buffered formalin (NBF) for 24 h at 4 °C, washed with 70% EtOH, and embedded in paraffin. Sections of 2.5 µm were cut and processed for hematoxylin and eosin (H&E) staining and immunohistochemistry. Prior to fixation, dissected lungs were inflated by injecting 5 mL of PBS through the trachea, then inflated with 5 mL of 10% NBF and gently released into a tube filled with 10% NBF. Immunohistochemical staining was performed on formalin-fixed, paraffin-embedded tissue sections using a Discovery XT (Ventana) fully automated system for anti-pAKT (Ser473), anti-pERK1/2 (Thr202/Tyr204) and anti-cleaved Caspase-3 (Asp175). Algorithms for quantitative analysis of immunostained-positive areas and the areas of lung metastases were designed in Halo software that allowed assessment of the relative fractions of positive areas.

Apoptosis assay

Cells were synchronized with DMEM 0.5% serum overnight and then supplemented with medium containing inhibitor(s). Fresh inhibitors were added after 48 h and cells (floating and adherent) were collected 24 h later using trypsin–EDTA, resuspended in growth medium and counted. For Annexin V/propidium iodide staining, cells were washed twice with cold Cell Staining Buffer (BioLegend, #420,201) and resuspended in Annexin V Binding Buffer (BioLegend, #422,201) at a concentration of 1 × 106 cells/mL. Aliquots of Alexa Fluor 647 Annexin V (5 µL) (BioLegend, #640,911) and of propidium iodide (10 µL) (BioLegend, #421,301) were added to 100 µL of this suspension, which was then incubated for 15 min at room temperature in the dark. After addition of 400 µl of Annexin V Binding Buffer to each tube, samples were analyzed by flow cytometry.

Transcriptomic analysis

Total RNA was extracted from frozen tumors using the RNeasy Plus Mini Kit (Qiagen, #74,136) and sample quality was controlled on an Agilent 2100 Bioanalyzer system with the RNA6000 Nano kit (Agilent, #5067–1511). mRNA isolation was performed with the NEBNext Poly(A) mRNA magnetic isolation module (NEB, #E7490) and libraries were prepared with the NEBNext Ultra II Directional RNA Library Prep kit (NEB, #E7765) according to the manufacturer’s recommendations. Samples were individually barcoded during library preparation using NEBNext Multiplex Oligos for Illumina Index Primers Sets 1 and 2 (NEB, #E7335 and #E7500). Library quality control was performed with the DNA1000 kit (Agilent, #5067–1504) on the Agilent 2100 Bioanalyzer system. Finally, libraries were sequenced on an Illumina NextSeq 500 that generated paired-end 75-bp reads. Adaptor trimming was performed using cutadapt. Trimmed reads were aligned to the GRCm38 genome using the two-pass approach of STAR. A median of 53 million reads (range 44–60) were aligned per sample. qCount from QuasR was then used to obtain counts at the gene level. Differential gene expression analysis was performed using edgeR [47]. A cutoff of log2 fold change > 1 and adjusted to P < 0.05 (corrected by the Benjamini–Hochberg algorithm method) was applied to selected genes. Network analysis was performed on the cBioportal website using the breast cancer dataset METABRIC [48].

3D 4T1 cell culture

For in vitro drug treatments, 4T1 cells were seeded in DMEM containing 10% FCS and 30% Matrigel Growth Factor Reduced (Corning, 356,231) at 300 cells per well in 384-well plates in quadruplicates. After three days, 3D colonies were treated with DMEM containing 1% FCS together with CLR457, SHP009, Sunitinib or combinations thereof at the indicated concentrations. Two days later, 50% of the culture medium was exchanged with medium containing drugs at 200% higher concentrations. Cells were kept under treatment for a further two days. At treatment day 5, the viability of cells was assessed by the CellTiter-Glo 3D Cell Viability Assay (Promega, G9618) according to the manufacturer’s instructions. In brief, after removing the culture medium, cells were lysed in 25 µl CellTiter-Glo 3D Reagent. After a 30-min incubation at room temperature on a horizontal shaker, luminescence was recorded for 0.5 s with an ELISA-reader.

Statistical analysis

Unless stated differently in the figure legends, all results shown represent at least three independent experiments and are reported as means ± STDEV. Data were tested for normal distribution and ANOVAs tests were applied. GraphPad Prism 7.04 was used for Kaplan–Meier survival analysis and log rank Mantel-Cox tests were applied to test statistical significance (SAS), as well as for all other statistical tests (SAS). The P values < 0.05 were considered statistically significant.

Data deposition

Transcriptomic data are available on the GEO database, reference GSE128051 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE128051).