Gregor Hutter, M.D. Ph.D1,2, #, Martin Sailer, M.D. Ph.D. 1#, Tej Deepak Azad, BA2#, André O. von Bueren, M.D., Ph.D.3,4,5, Peter Nollau, M.D.6, Stephan Frank, M.D.7, Cristobal Tostado, M.Sc1, DurgaSarvepalli, Ph.D.8, Arkasubhra Ghosh, Ph.D.8, Marie-Françoise Ritz, Ph.D.1, Jean-Louis Boulay, Ph.D.1, Luigi Mariani, M.D. 1
Reverse Phase Protein Arrays Enable Glioblastoma Molecular Subtyping
Correspondence
Gregor Hutter, M. D., Ph. D.
Department of Neurosurgery
University Hospital Basel
Spitalstrasse 21
CH-4031 Basel
Phone + 41 61 265 25 25
Switzerland
Email:
Supplementary Experimental Procedures
Quantitative Real time PCR analysis
RNA extraction and cDNA synthesis: The preserved tissue was lysed and RNA extracted using the LEV Total RNA Purification Kit (Cat.No. AS1220, Maxwell). RNA quality was assessed using an Agilent bioanalyzer (Agilent Technologies, USA). RIN values ranged from 7.2 to 9.7. cDNA synthesis was performed using the iScript cDNA Synthesis Kit (Cat.No. 170-8891, BIO-RAD) according to the manufacturer’s instructions. Quantitative real-time PCR (qPCR): PCR primers were synthesized at Microsynth (Switzerland) with the sequences in 5’-3’ direction listed in supplementary table 2.
QPCR reactions were run on a CFX96 real time cycling system (Biorad, USA) in optical 96-well plates using 10ng cDNA per reaction and SsoFastEvaGreenSupermix (Cat.No. 172-5201, Biorad, USA). Gene expression was calculated according to the ddCT method, normalized to GAPDH and reference human brain RNA (LifeTechnologies, USA).
Cell culture and reagents
Glioma cell lines U373 and U251 were purchased at ATCC (Manassas, VA, USA) cultured in Dulbecco’s modified Eagle medium (DMEM) (Invitrogen, Carlsbad, CA, USA), supplemented with antibiotic and antimycotic solution (Penicillin, 100 units/ml; Streptomycin, 100 μg/mL and Amphotericin B, 25 ng/ml; Himedia Labs, Mumbai, India) and 10% (v/v) heat-inactivated fetal bovine serum (FBS) (Invitrogen). Cells were incubated at 37 °C in a humidified atmosphere containing 5% CO2.
Factors. The following substances/peptides were used: XAV939 (Wnt inhibitor, HY-15147, Medchem), 10 mM; AZD2858 (Wnt agonist/GSK 3b inhibitor, HY-15761, Medchem), 1 mM; Doxorubicin hydrochloride (Sigma) with concentrations up to 8 mM; 5 - Fluorouracil (Biochem, India) with concentrations up to 2 mM.
Cell death assay
100,000 cells were seeded into 12 well plates in duplicates. At 60% confluency, cells were treated with agonists and inhibitors in the presence and/or absence of Doxorubicin or 5 - FU and the drugs were replenished every 24 hours. Drug effectswere studied for 48 hours. After 48 hours, one set of cells were trypsinized (0.25% Trypsin, Gibco) and % cell death for every treatment was measured by trypan blue exclusion assay, where 10 ml of re-suspended cells were mixed 1:1 with 10 mL trypan blue (0.4% trypan blue, Sigma)(1:1). Stained dead cells and bright live cells were counted using hemocytometer.
Migration assay
2D migration assay: 150,000 cells were seeded in a 12-well plate and at 85% confluency, a scratch assay was performed using a 200 μl pipette tip. Cells were treated with agonists and antagonists alone or in combination and scratchthe closure was measured using ImageJ® software.
Supplementary Figure Legend
Supplementary Figure 1
Quantitative real-time PCR validation of proteomic analysis.
(A)Relative mRNA expression analysis of CREB, SMAD2 and 4EBP1 in clusters 1-3 compared to control samples (normalized to GAPDH and a normal brain RNA as reference).
(B)Relative mRNA expression levels of NOTCH1 and target genes HES1 and HEY1
(C)Relative mRNA expression levels of NFKBIA and WNT-target genes AXIN2 and SNAI1. Values are plotted as mean +/- SD over all the biological replicates within the respective groups.
Supplementary Figure 2
(A)Array layout of 64 samples within one of the six arrays of a chip. In total, 384 sampels in 4 dilution spots were accommodated on one chip. Red dots indicate constant fluorescence reference spots. Control spots on each array are spotting buffer, rabbit IgG, and a mixture of mouse IgG1, IgG2a and IgG2b.
(B)A typical reverse assay image was taken the fluorescent reader. This array was incubated with anti-GSK3 antibody. The image shows homogenous and circular spot morphology for all samples. The signal integration time was 10 s.
(C)Automatic grid alignment and spot finding: the center of each evaluated area is positioned on the center of gravity of the spot.
(D)Background determination: for each sample spot, 4 additional areas were defined between the spots and placed at equal distances from the spot. These areas were used to determine the mean local background fluorescence signal. Their combined areas equal the size of the area of interest used for signal quantification of the spots. Net fluorescent intensity of each spot was calculated by subtraction of the mean local background fluorescence signal from the mean spot signal. Net fluorescence intensity was used for all further calculations. Referenced fluorescence intensities were determined at the center concentration (0.125 mg/mL) of a weighted linear fit curve (4 data points per sample).
Supplementary Tables
Supplementary Table 1. Description of relevant pathways investigated in this study.
4EBP1 / p4EBP1, the direct downstream effector of mTORC1, dissociates from EIF4E and allowscap-dependent protein synthesis. In active PI3K-AKT-mTOR signaling, 4EBPI is hyperphosphorylatedleading to increased protein synthesis and cell proliferation.SMAD2 / TGF and MAPK phosphorylate SMAD2, which in turn builds a complex with SMAD4 and exerts antimitogenic signals in the nucleus 16. Hypophosphorylated SMAD2 therefore results in activation of cell-cycle promoting proteins such as Cyclin D, E and cyclin-dependent kinases.
CREB1 / The transcription factor CREB1 has only recently been implicated in gliomagenesis15 and integrates signals from both the MAPK and PI3K pathways to modulate expression of three key cell cycle factors, Cyclin B, D andproliferating cell nuclear antigen (PCNA). CREB is active in neuronal and especially epileptogenic brain regions 18 which may explain the higher expression in the examined neural cortices.
NFKBIA / Phosphorylation of NFKBIA leads to its ubiquitination, which releases active NFB complex to the nucleus. It has been reported that primarily mesenchymal and proneural subtypes of GBM have higher NFB activity. NFB interacts with CBP/p300 at the same region as phosphorylated CREB, and it has been proposed that NFB activity is inhibited by activated CREB through competition for limiting amounts of CBP/p300 19
Supplementary Table 2. Primers used for quantitative real-time PCR.
Gene ID / Forward primer (5’-3’) / Reverse primer(5’-3’)AXIN2 / AGCCAAAGCGATCTACAAAAGG / GGTAGGCATTTTCCTCCATCAC
CREB1 / CCAGCAGAGTGGAGATGCAG / GTTACGGTGGGAGCAGATGAT
EIF4BP1 / GTTCTCACGGCAGGAACCGAAG / GATGCGCCTTATTGCAGTCAC
HES1 / ATGGAGAAAAATTCCTCGTCCC / TTCAGAGCATCCAAAATCAGTGT
Hey1 / GAAACTTGAGTTCGGCTCTAGG / GCTTAGCAGATCCTTGCTCCAT
NFKBIA / CTCCGAGACTTTCGAGGAAATAC / GCCATTGTAGTTGGTAGCCTTCA
NOTCH1 / CACACACGGAGGCATCCTAC / GCTGGAGCATCTTCTTCGGAA
SMAD2 / ATGCGTCCTTATGGGGTTAAAC / AGGCCATGTATTTCCTCACAAG
SNAI1 / AATCGGAAGCCTAACTACAGCG / GTCCCAGATGAGCATTGGCA