Ethical approval
The collection of PBMCs from healthy donors from the Blood Bank was approved by the Institutional Review Board (Comitato Etico Aziendale) of the AOU Città della Salute e della Scienza, Torino, Italy (approval number: #993/2023; date: 09/27/2023). The collection and analysis of patient-derived NSCLC cells and PBMCs was approved by the Institutional Review Board (Comitato Etico Aziendale) of San Luigi Gonzaga Hospital, Orbassano, Italy (protocol: #16685; approval number: #73/2018; date: 11/15/2018; protocol renewal: #2866; approval number: #73/2018; date: 02/22/2023). The study was conducted in accordance with the Declaration of Helsinki. The experimental protocol on mice was approved by the Ethical Committee of the University of Torino (OPBA-Organismo per il Benessere Animale) and by the Italian Ministry of Health (approval number: #627/2018-PR; date: 08/10/2018; protocol renewal:798/2023-PR; date: 09/19/2023). The experimental procedures were performed in accordance with the ARRIVE guidelines.
Chemicals and materials
Plasticware for cell culture was obtained from Falcon (Becton Dickinson, Franklin Lakes, NJ), and the culture medium was from Invitrogen (Milan, Italy). Fetal bovine serum (FBS), cis-diammineplatinum (II) dichloride and all the reagents, unless otherwise specified, were purchased from Sigma-Aldrich-Merck (St. Louis, MO).
Cells
Human NSCLC cells (NCI-H1650, NCI-H1385, NCI-H460, NCI-H522, NCI-H661, NCI-H2126, NCI-H23, NCI-H1703, NCI-H1435, NCI-H596, NCI-H2286, NCI-H1437, NCI-H1651, NCI-H2085, NCI-H2342, NCI-H2073, NCI-H1793, NCI-H2170, NCI-H1299, NCI-H2066, NCI-H2347, NCI-H1734, NCI-H1563, NCI-H441, NCI-H1975, A549, Calu-3, and NCI-H2228) were purchased from ATCC (Manassas, VA). Mycoplasma spp. contamination was checked every 4 weeks using a MycoStrip® 100 detection kit (InvivoGen, San Diego, CA). The contaminated cells were discharged. Primary human NSCLC cells (#24, #636, #778, #670, #206, and #607) were obtained from diagnostic thoracoscopies/pleuroscopies or surgical materials from San Luigi Gonzaga Hospital, Orbassano, Department of Oncology, University of Torino, Italy. Each sample was anonymized and indicated as an identification number. Tumor cells were obtained from enzymatically digested tissues or pleural effusion, grown in Ham’s F10 medium supplemented with 10% v/v FBS and 1% v/v penicillin‒streptomycin for 48 h,53 and vitally frozen in FBS supplemented with 10% DMSO. For the ex vivo coculture systems, 1 × 105 viable cells were used 24 h after being thawed. Patient characteristics are reported in Supplementary Table 2.
IRAK1 silencing and overexpression
A total of 1 × 106 NCI-H2228 cells were transfected with 1 μg of human green fluorescence protein (GFP)-IRAK1 shRNA plasmid (#abx952448; Abbexa Ltd., Cambridge, UK) with three different sequences targeting IRAK1 or scrambled shRNA (#abx991273; Abbexa). Stably silenced clones were generated by selecting cells with 1 μg/ml puromycin (Sigma-Aldrich) for 4 weeks. For IRAK1 overexpression, 1 × 106 NCI-H1650 cells were transfected with an IRAK1 expression vector (#sc-400264-ACT, Santa Cruz Biotechnology, Santa Cruz, CA) or the control CRISPR activation plasmid (#sc-437275, Santa Cruz Biotechnology). Stably overexpressing clones were generated by selecting cells with 0.1 μg/mL puromycin for 4 weeks. The efficacy of silencing and overexpression was verified by RT‒PCR and immunoblotting.
Patient-derived organoids (PDOs)
Organoids HCM-WCMC-0789-C34 (ATCC PDM-685™) and HCM-WCMC-0749-C34 (ATCC PDM-683™) were purchased from ATCC and were cultured in organoid media formulation #10 according to the manufacturer’s instructions. Up to five drops of Matrigel/cell suspension per well (100 μl volume) were distributed into a six-well ultralow-attachment culture plate. The drops were solidified by a 30-min incubation in a cell culture incubator at 37 °C and 5% CO2. After solid drops formed, 2 mL of organoid medium was added. Fresh culture media was added to each well every three days, and 10 μM ROCK inhibitor (Y-27632) was added for the first 3 days. Organoids ~300 to 500 μm in size were subcultured, and single cells and small cell clusters were replated as described above. Total RNA was extracted, and IRAK1 levels were measured. Organoids were fixed in 4% v/v paraformaldehyde for 1 h, permeabilized in PBS with 0.1% Triton X-100 for 4 h and stained with the following antibodies: anti-IRAK1 (Abcam; 1:100) for 18 h, followed by a secondary FITC-conjugated antibody for 18 h. The nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and analyzed by a Leica SP8 confocal microscope.
Flow cytometry
A total of 1 × 104 cells were washed in PBS (pH 7.2, 0.5% bovine serum albumin (BSA) and 2 mM EDTA), centrifuged at 300×g for 10 min, incubated for 10 min at room temperature in the dark with 250 μL Inside Fix Kit (Miltenyi Biotec, Bergisch Gladbach, Germany), centrifuged at 300×g for 5 min, washed with 1 mL Inside Perm (Miltenyi Biotec), centrifuged at 300×g for 5 min, and incubated 10 min at room temperature with the following antibodies (all from Miltenyi Biotec): anti-ABCB1 (PE-Vio® 770-conjugated); anti-ABCC1 (PE-conjugated); anti-ABCA1 (DyLight 488-conjugated); anti-TCR Vγ9 (REA470, APC-conjugated); anti-CD3 (REA613 VioBright B515-conjugated); anti-CD4 (REA623 PerCP-Vio700-conjugated); anti-CD45RA (REA1047 VioBright720-conjugated); and anti-CD62L (REA615 VioBright V600-conjugated). The cells were subsequently washed with 1 mL of Inside Perm reagent, centrifuged at 300×g for 5 min, and read using a BD FACSCelesta™ Cell Analyzer (BD Biosciences, Franklin Lakes, NJ).
Cell viability
A total of 5 × 103 cells were seeded in a 96-well plate, incubated for 72 h with increasing concentrations (from 1 × 10−9 to 1 × 10−5 M) of cisplatin and docetaxel for 72 h and stained with a WST-1 kit (Roche, Basel, Switzerland). PDOs were seeded in a 24-well plate, incubated for 48 h with 100 μM cisplatin, and then treated with a 3D CellTiterGlo kit (Promega, Madison, WI). Plates were read using a Synergy HT Multi-Detection Microplate Reader (BioTek Instruments, Winooski, VT, USA). The absorbance/chemiluminescence of the untreated cells/PDOs was considered to indicate 100% viability; the results are expressed as a percentage of viable cells versus control cells.
T lymphocyte activation and tumor killing
PBMCs were isolated by Ficoll-Hypaque density gradient centrifugation, used immediately or vitally frozen and used 24 h after thawing. For Vγ9Vδ2 T lymphocyte activation, PBMCs were stimulated for 7 days with 1 μM zoledronic acid and 10 IU/mL IL-2 to expand Vγ9Vδ2 T lymphocytes, after which Vγ9Vδ2 T cells were purified with a TCRγ/δ+T cell isolation kit (Miltenyi Biotec).54 The purity of Vγ9Vδ2 T lymphocytes was confirmed by staining 5 × 105 isolated cells with anti-TCR Vγ9 (REA470, APC-conjugated; Miltenyi Biotec) and anti-CD3 (REA613 VioBright B515-conjugated, Miltenyi Biotec) antibodies. Samples with >80% Vγ9+/CD3+ cells were cultured for 72 h with NSCLC cells at a 1:2 ratio. To analyze IFN-γ secretion and degranulation activity (CD107a level), brefeldin A (#420601; BioLegend, San Diego, CA) was added during the final 4 h of coculture. The cells were then stained with anti-Ki67 (REA183 PE-conjugated), anti-CD107a (REA792 PE-conjugated), and anti-INF-γ (REA600 APC-conjugated) antibodies (Miltenyi Biotec) and quantified with a BD FACSCelesta™ Cell Analyzer. The results are expressed as the fold change in Vγ9+Ki67+/CD107a+/IFNγ+ cells compared with Vγ9+ cells.
For T cell subset analysis and activation, T lymphocytes were isolated from PBMCs with a Human Pan T-Cell Isolation Kit (#130-096-535; Miltenyi Biotec). The purity of the T cells was confirmed by staining 5 × 105 isolated cells with an anti-CD3 (REA613 VioBright B515-conjugated, Miltenyi Biotec) antibody. T cell subset analysis was performed using anti-CD45RA (REA1047 VioBright R720-conjugated), anti-CD62L (REA615 VioBright R600-conjugated), anti-CD8 (REA734 VioBlue-conjugated), anti-CD4 (REA623 PerCP-Vio700-conjugated) and anti-CD3 (REA613 VioBright 515-conjugated) antibodies (Miltenyi Biotec). T cells were cultured for 72 h with NSCLC cells at a 1:2 ratio. To analyze proliferation and activation, the cells were stained with anti-Ki67 (REA183 PE-conjugated), anti-CD107a (REA792 PE-conjugated), and anti-INF-γ (REA600 APC-conjugated) antibodies (Miltenyi Biotec) and quantified with a BD FACSCelesta™ Cell Analyzer. The results are expressed as the fold change in the number of Ki67+/CD107a+/INFγ+ cells compared with that of CD3+ cells.
Tumor killing was evaluated by staining adherent NSCLC cells using a (1) MEBCYTO® Apoptosis Kit (MBL JSR Life Science Company, Sunnyvale, CA), and the percentage of Annexin V/PI+ cells was calculated by flow cytometry; (2) 0.1% v/v crystal violet dye was used to quantify the absorbance with a Synergy HT Multi-Detection Microplate Reader. The absorbance of the untreated cells was considered 100%; the results are expressed as a percentage of viable cells versus control cells.
CRISPR-KO kinome library
A total of 13 × 106 NCI-H2228 SpCas9 (BNT)_A6 clonal cells were expanded to obtain ≥30 × 106 cells for two transduction replicates. A total of 15 × 106 NCI-H2228 cells were transduced per replicate with 500 µL of the Kinome sgRNA Lentivirus Library (Oxford Genetics, Oxford, UK) using the spin-infection method. After 24 h, the cells were counted using a Vi-Cell XR cell counter and selected with 1 µg/mL puromycin for 6 days. Next, 5 × 106 cells per replicate were stored for next-generation sequencing (NGS), and 5 × 106 cells were expanded for 11 days for cell sorting. Cells were stained with PE-conjugated ABCC1 and DyLight 488-conjugated ABCA1 antibodies and sorted using a SONY Cell Sorter SH800S (Minato, Japan). PCR products were measured using a Nanodrop and Illumina NEBnext Library Quantification Kit for Illumina (New England Biolabs, Ipswich, MA). PCR samples were sequenced by GENEWIZ (Azenta Life Sciences, Leipzig, Germany) using an Illumina HiSeq4000. Data were analyzed by MAGeck (Model-based Analysis of Genome-wide CRISPR. Cas9 knockout, version 0.5.9.2) at Oxford Genetics, Oxford, UK.
Immunoblotting
Cells were lysed in RIPA buffer supplemented with protease inhibitor cocktail set III (Sigma-Aldrich), 2 mM phenylmethylsulfonyl fluoride (PMSF) and 1 mM Na3VO4, sonicated (10 bursts of 10 s, 4 °C, 100 W, Labsonic sonicator, Hielscher, Teltow) and centrifuged at 13000×g for 10 min at 4 °C. Twenty micrograms of protein was subjected to immunoblotting (4–20% gradient SDS‒PAGE) and probed with the following antibodies: ABCB1 (sc-13131, Santa Cruz), ABCC1 (sc-18835 Santa Cruz), ABCA1 (ab307534, Abcam, Cambridge), IL1-R (ab154524 Abcam), MyD88 (ab33995 Abcam), IRAK1 (ab180747, Abcam), IRAK2 (sc-515885, Santa Cruz), IRAK3 (sc-100389, Santa Cruz), IKKγ (#2695, Cell Signaling Technology, Danvers, MA) phospho(Ser176/180)IKKβ/α (#2697, Cell Signaling Technology), phospho(Ser32)IkBα (#2859, Cell Signaling Technology), IkBα (sc-371, Santa Cruz), NFkBp50 (sc-8414, Santa Cruz), NFkBp65 (sc-8008, Santa Cruz), ERK1/2 (ABS44, Merck Rahway, New Jersey), phospho(Thr202/Tyr204, Thr185/Tyr187)ERK1/2. The proteins were detected by enhanced chemiluminescence (ChemiDocTM Touch Imaging System device; Bio-Rad Laboratories, Hercules, CA).
Quantitative real-time PCR (qRT‒PCR)
Total RNA was extracted and reverse-transcribed using the iScript™ cDNA Synthesis Kit (Bio-Rad Laboratories). qRT‒PCR was performed with the IQ SYBR Green Supermix (Bio-Rad Laboratories). The primer sequences (Supplementary Table 4) of the ABCB1, ABCC1, ABCA1, IRAK1 and S14 (housekeeping genes) genes were designed using the qPrimerDepot database (http://primerdepot.nci.nih.gov). Their quantification was performed using the Bio-Rad Software Gene Expression Quantification.
Chromatin immunoprecipitation (ChIP)
Chromatin immunoprecipitation assays were performed using 5 × 10⁶ cells as indicated.27 Promoter regions of the ABCB1, ABCC1, and ABCA1 genes were identified using the Eukaryotic Promoter Database (EPD; https://epd.expasy.org/epd/). Putative binding sites for AP-1, NFκB, and LXRα were predicted using the JASPAR database (https://jaspar.elixir.no/). ChIP-enriched DNA was analyzed by qRT‒PCR using primer pairs designed by PrimerBlast (https://www.ncbi.nlm.nih.gov/tools/primer-blast) (Supplementary Table 4).
NFκB and AP-1 transcriptional activity
NFκB and AP-1 transcriptional activity was quantified using TransAM® NFκB and TransAM® AP-1 Assay Kits (Active Motif) according to the manufacturer’s instructions. Nuclear extracts were prepared using the Nuclear Extract Kit (Active Motif). For each assay, 10 µg of nuclear protein was incubated in 96-well plates precoated with consensus oligonucleotide sequences containing NFκB or AP-1 binding sites. The absorbance was measured at 450 nm with a reference wavelength of 655 nm using a BioTek Synergy HT plate reader. The data were normalized to the background signal and are expressed as optical density (OD450) values.
TCGA analysis and patient retrospective analyses
The transcriptome profile of the TCGA-LUAD cohort was imported to the R working environment (version 4.2.2) and analyzed as detailed previously.27 The distributions of IRAK1, ABCA1 and ABCC1 in tumor tissue were evaluated to estimate “high” and “low” gene expression. Patient OS was extracted and correlated with the coexpression of IRAK1, ABCA1 and ABCC1 as previously reported.27
The transcriptome profiles of 247 treatment-naïve, I-IIIA-stage NSCLC patients26 were analyzed for IRAK1 expression. According to the median value of IRAK1 expression, the patients were categorized into “high-” and “low-expressing” groups and analyzed for OS. Two retrospective analyses were conducted on patients enrolled at San Luigi Gonzaga Hospital, Department of Oncology, University of Torino, Italy. Each patient was anonymized and indicated by an identification number. First, IRAK1 expression was compared between 81 NSCLC tissues and 48 adjacent normal lung tissues from the same cohort. Patients’ clinical and molecular features are described previously25 and have been deposited in the European Genome-phenome Archive (EGA) under accession number EGAS00001007219. The datasets used for all analysis steps are available under accession number EGAD00001010838. Second, the levels of IRAK1, ABCB1, ABCC1, and ABCA1 were measured by qRT‒PCR and normalized to those of the housekeeping gene β-2-microglobulin (B2M) in formalin-fixed, paraffin-embedded (FFPE) samples from 77 treatment-naïve patients at stage IIIB-IV,27 36 patients received cisplatin as first-line therapy (chemotherapy group), and 41 patients received pembrolizumab as first-line therapy (immunotherapy group). IRAK1, ABCB1, ABCC1, and ABCA1 levels were reassessed in tumor cells obtained from the pleural effusion53 of patients who relapsed after treatment (8 patients in the chemotherapy group and 13 patients in the immunotherapy group). Patients were categorized into “high-” and “low-expressing” groups, and TTP and OS were analyzed. The pathological features, smoking habits, TTP and OS are reported in Supplementary Table 2. Patients’ FFPE samples at diagnosis were also stained with an antibody recognizing IRAK1 (Abcam), followed by a peroxidase-conjugated horseradish antibody (Dako, Glostrup, Denmark). Nuclei were counterstained with hematoxylin (Sigma-Aldrich).
IRAK1 immunoprecipitation and enzyme activity
NCI-H2228 cells (2 × 106) were incubated with the IRAK1 inhibitor JW30 at concentrations of 10 and 100 nM for up to 72 h. One milligram of protein from whole-cell extracts was immunoprecipitated with an anti-IRAK1 antibody (ab180747; Abcam) using PureProteome protein A and protein G magnetic beads (Millipore, Billerica, MA). The eluted protein was subjected to immunoblotting analysis to verify the specificity for IRAK1. IRAK1 kinase activity was quantified using the ADP-Glo™ Kinase Assay (Promega, Madison, WI) following the manufacturer’s guidelines. Recombinant IRAK1 protein (50 ng; SignalChem Lifescience, Richmond, Canada) was incubated with the JW30 inhibitor, and IRAK1 immunoprecipitated from cell lysates was incubated with the specific substrate (modified AKT, 0.2 µg/µL) and ATP (25 µM) in kinase reaction buffer (40 mM Tris-HCl (pH 7.5), 20 mM MgCl₂, 0.1 mg/mL BSA, 40 µM DTT). Reactions were performed in 384-well plates with a final volume of 20 µL and incubated at room temperature for 60 min. The kinase reaction was stopped by the addition of ADP-Glo reagent, which converted the generated ADP into a luminescent signal proportional to kinase activity. Luminescence was measured after a 30-min incubation with Kinase Detection Reagent at room temperature using a BioTek Cytation 3 microplate reader (BioTek Instruments). The results are expressed as a percentage of enzyme activity versus activity in treated cells, which was considered 100%.
Immunoxenografts
A total of 2 × 106 wild-type and shIRAK1 NCI-H2228 cells mixed with 100 μL of Matrigel were injected subcutaneously into female NOD SCID-γ (NSG) mice engrafted with human hematopoietic CD34+ cells (Hu-CD34+; The Jackson Laboratories, Bar Harbor, MA). The mice were housed (5 per cage) under a 12 h light/dark cycle, with food and water provided ad libitum. Tumor growth was measured daily by calipers according to the equation (LxW2)/2, where L = tumor length and W = tumor width. When the tumors reached a volume of 50 mm3, the animals (5/group) were randomized. In the first experiment, mice bearing wild-type and shIRAK1 NCI-H2228 tumors were treated for 3 weeks as follows: the vehicle group was treated with 0.1 mL of saline solution intraperitoneally (i.p.) once a week, and the cisplatin group was treated with 2 mg/kg cisplatin i.p. once a week. In a second experiment, mice bearing wild-type NCI-H2228 tumors were treated as follows: vehicle group, treated with 0.1 mL saline solution i.p., twice a week; cisplatin group, treated with 2 mg/kg cisplatin i.p., once a week; JW30 group, treated with 10 mg/kg IRAK1 inhibitor JW30 i.p., twice a week; and JW30+cisplatin group. Tumor volumes were monitored by calipers, and the animals were euthanized on day 28 after randomization with zolazepam (0.2 mL/kg) and xylazine (16 mg/kg). The tumors were excised, weighed, and photographed. Tumor tissues were fixed in 4% v/v paraformaldehyde and stained with hematoxylin/eosin and an anti-Ki67 antibody (AB9260; Merck), followed by a peroxidase-conjugated secondary antibody (Dako). The nuclei were counterstained with hematoxylin. The tumor sections were also stained with an in situ cell death detection kit (TUNEL Assay; Roche, Basel, Switzerland) or an anti-CD8 antibody (MA5-44283; Invitrogen) followed by an Alexa Fluor488-conjugated secondary antibody (A32731TR; Invitrogen) and DAPI. Sections were examined with a Cytation 5 Cell Imaging Multimode Reader (BioTek Instruments). Immunostaining quantification was performed using ImageJ software, and the results are expressed as a percentage of Ki67/CD8- or TUNEL-positive nuclei/nuclei counted in each field. In each sample, ten fields were analyzed. To evaluate the intratumor infiltration, tumors were digested with 1 mg/mL collagenase and 0.2 mg/mL hyaluronidase for 1 h at 37 °C and filtered through a 70-μm cell strainer to obtain a single-cell suspension. Infiltrating immune cells were collected, isolated and immunostained with anti-TCR Vγ9 (REA470 APC-conjugated, Miltenyi Biotec), washed with 1 mL of Inside Perm reagent, and centrifuged at 300×g for 5 min, after which the cells were read using a BD FACSCelesta™ Cell Analyzer.
Immediately after euthanasia, 200 µL of blood was collected to measure the following hematochemical parameters: red blood cells (RBCs), white blood cells (WBCs), hemoglobin (Hb), and platelets (PLTs), which are indicators of bone marrow function; lactate dehydrogenase (LDH), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (AP), which are indicators of liver function; creatinine, which is an index of kidney function; creatine phosphokinase (CPK), which is an index of muscle/heart damage using commercially available kits from Beckman Coulter, Inc. Heart, lungs, liver, kidneys and spleen were collected and fixed in 4% v/v paraformaldehyde. The sections were stained with hematoxylin‒eosin and examined with a Leica DC100 microscope.
Single-cell RNA sequencing (scRNA-Seq)
A total of eight freshly isolated samples, consisting of 2 replicates for each of the four experimental conditions (shIRAK1 untreated, shIRAK1 treated with Pt, wild-type untreated, and wild-type treated with Pt), were analyzed using microfluidic-based scRNA-Seq across eight batches. Excised tumor tissues were processed as described previously.27 The library pool was sequenced using an Illumina NovaSeq 6000 (Illumina, San Diego, CA) flow cell. Raw 10x reads were demultiplexed into FASTQ files for individual libraries using Cellranger (version 7.0.0). The same tool was used to demultiplex samples based on cell multiplexing oligos (CMOs), align unique reads to the reference genome (10x Genomics prebuilt combined human GRCh38 and mouse GRCm39 genomes), assign them to individual cells via 10x Genomics barcodes, remove duplicates, and filter them for valid cells. The sparse matrices were then converted into complete matrices using the h5tocsv function from the rCASC package.55 Using the human/mouse cell annotation provided by Cellranger, only human cells were retained. Quality control was subsequently performed by assessing the number of genes per cell and the percentage of reads mapped to mitochondrial and ribosomal genes, utilizing the mitoRiboUmi function from rCASC. Finally, only cells expressing at least 100 genes with more than 3 UMIs per cell were selected using the scannobyGtf function from rCASC. All samples were merged into matrices corresponding to the four experimental conditions: shIRAK1 untreated (shc1, shc2), shIRAK1 with Pt (shp1, shp2), wild-type untreated (wtc1, wtc2), and wild-type with Pt (wtp1, wtp2), while preserving, as part of the cell barcode, information on cell origin from each replicate and sequencing batch. Each matrix was analyzed separately using Louvain modularity clustering (seuratBootstrap function from the rCASC package) with a resolution of 0.3. The resulting clusters were then annotated using scATOMIC software56 integrated into the rCASC package. All analysis step data are available on Zenodo (https://doi.org/10.5281/zenodo.15205625), and the raw data are in the GEO dataset (https://www.ncbi.nlm.nih.gov/gds; GSE294588).
Design, stability, and biodistribution of IRAK1 inhibitors
IRAK1 inhibitors were designed by virtual screening and pharmacophore screening methods. JW30 and SG5180 compounds were chemically synthesized (Supplementary Fig. 18) at the Department of Drug Science and Technology, University of Torino, Italy. Purity (Supplementary Fig. 17), plasma stability and biodistribution of the lead compound JW30 were detected by reversed-phase-performance liquid chromatography (RP-HPLC) and ultrahigh-performance liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS), respectively (see Supplemental material and methods).
Statistical analysis
All the data were presented as the means ± SDs. The results were analyzed by one-way analysis of variance (ANOVA) using GraphPad PRISM (v.10). p < 0.05 was considered statistically significant. The Kaplan‒Meier method was used to calculate the TTP (time from the beginning of treatment to the first signs of disease progression) and OS (survival from the beginning of treatment until the patients’ death). The patient and animal sample sizes were calculated using G*Power (www.gpower.hhu.de), with α ≤ 0.05 and 1-β = 0.80.
