Additional methods to Duso, Messuti et al “NF1 modulates microtubule repair and sensitivity to antibody-drug conjugates” Nature Cancer 2026
Genetic engineering
CRISPR/Cas9.
Guides are cloned into pSpCas9(BB) 2A-GFP (PX458) plasmid421, a gift from Feng Zhang (Addgene #48138). A DNA fragment containing 248nt of the NF1 coding region of interest is PCR-amplified from single-strand cDNA obtained from parental and clonal cells using QIAquick Gel Extraction Kit (QIAGEN).
Cell lines are transfected with the plasmid containing the NF1 gRNA in parallel with empty vector controls using Lipofectamine 2000 according to manufacturer’s instructions.
Seventy-two hours later, GFP-positive cells are single-cell sorted in 96-well plates.
For ribonucleoprotein-based engineering, the Gene Knockout Kit v2 (Synthego) was used. The Cas9 protein wa purified at IEO, gift of L Rizzuti (Department of Experimental Oncology, European Institute of Oncology, Milan).
Incorporation was confirmed by Sanger sequencing.
NF1 mammalian expression.
The pLVX-puro vector was kindly provided by A Polazzi (Department of Experimental Oncology, European Institute of Oncology, Milan) and used for cloning of wild type and mutant NF1 (L134R, L1196F and S2137, K1444E). The pLVX-puro-NF1 lentiviral vector was transfected in HEK293T cells by Calcium Chloride. Viral supernatants were collected 2 days post-transfection and filtered through 0.45 mm filters. Cells were infected in the presence of 8 μg/mL Polybrene (Sigma-Aldrich), selected by puromycin and single-cell sorted in 96-well plates.
RasG12V overexpression.
KRASG12V from pDONR223_KRAS_p.G12V (Addgene #81665) was cloned on the lentiviral backbone pLenti CMV hygro dest (Addgene #17454). HEK293T cell transfection and infection of target cells was performed by Calcium Chloride followed by hygromycin B (InvivoGen) 0.5 mg/mL selection for 8 days. RAS activation was assessed by Active Ras Detection Kit (#8821, Cell Signalling Technology, following manufacturer’s protocol.
Luciferase and H2B-GFP/mCherry expressing cells.
Tet-Off-H2B-GFP lentiviral vector422 and the H2B-mCherry retroviral vector423 were kindly provided by N Roda (Department of Experimental Oncology, European Institute of Oncology, Milan). The pLenti CMV Puro LUC (w168-1) was purchased from Addgene (#17477) and used in all in vivo experiments. Cells were transduced with the pLenti CMV Puro LUC reporter vector at MOI = 5 in the presence of 8 μg/mL of Polybrene (Sigma-Aldrich). Infected cells were selected by puromycin (2.5 μg/mL for 72h; Vinci-Biochem) and allowed to expand for additional 8 days post selection.
Drug screen
Cells are seeded (1 x 104) in 96-well, white-bottom plates (Corning), in triplicates, and let grown overnight. Then, specified compounds are added in 9 dilutions plus their respective vehicle.
After 72h, BrdU reagent is incorporated for additional 24h followed by fixation and detection according to manufacturer’s instructions (Cell Signalling Technology, #5492).
Luminescence is read using a PHERAstar FSX Microplate Reader under a 425nm wavelength.
The relative response is corrected compound-wise to the average vehicle response for each replicate.
Data are analysed in GraphPad Prism 9, and IC50 values were generated from best-fit curves. Data represent the mean ± SD.
Growth curves
Cells were plated in triplicate at a density of 2×105 cells/well for BT-474 and 1 × 105 cells/well for HCC1954 in 12-well plates. After 24h, cells were treated with control (rituximab 10 μg/mL or DMSO), trastuzumab (100 μg/mL for BT-474 and 10 μg/mL for HCC1954), T-DM1 (1 μg/mL for BT-474 and 0.1 μg/mL for HCC1954), DM1 (25nM for BT-474 and 5 nM for HCC1954) or T-DXd (1 μg/mL for BT-474 and 0.1 μg/mL for HCC1954). For time point 0 (24h after seeding), and subsequently for all other time points (day 2, day 4, day 6, and occasionally day 8), cells were washed twice with PBS, trypsinised, counted using a hemocytometer and then treated again with fresh media containing the respective drugs.
Colony staining
Plates are washed twice with ice-cold PBS and fixed with ice-cold methanol for 15 min.
0.5% crystal violet solution (Sigma V5265) is added to the plates and incubated at room temperature for 15 min. Plates are then washed with ddH2O, until the unbound crystal violet is removed, and dried at room temperature. Images are processed using the ColonyArea ImageJ plugin250.
Beta-galactosidase assay
Cells were seeded at a density of 5×105 cells/well into 6-well plates and treated for 5 days with either vehicle (DMSO), T-DM1 (0.2 μg/mL) or doxorubicin (200nM). The Cell Signaling Technology β-Galactosidase staining kit (#9860) was employed for the detection of senescence, in accordance with the producer's guidelines. Images were captured via an Invitrogen EVOS XL Core 36microscope with a 10X 0.4 NA air objective and analyzed though ImageJ. The Colour Deconvolution2 plugin was utilized to obtain the stained cells signal and the Integrated density was assessed for each Field of View.
Mass photometry
NF1 molecular mass was measured at a final concentration of 10nM in 25mM Hepes pH7,5, 300mM NaCl, 5% Glycerol, 1mM DTT with the TwoMP instrument (Refeyn). The calibration curve was generated using the two peaks corresponding to 66 kDa and 132 kDa resulting from a BSA measurement.
Microtubule co-sedimentation
20µM Tubulin (Cytoskeleton Inc) was polymerized into stable MTs in General Tubulin Buffer (80 mM PIPES pH 6.8, 1 mM MgCl2, 1 mM EGTA), supplemented with 1 mM GTP and 50 µM Paclitaxel at 37°C for 20 minutes. For MT-binding reactions, MTs were diluted to a final concentration of 9µM in General Tubulin Buffer supplemented with 1 mM GTP, 50µM Paclitaxel and 60mM NaCl. 0,6 µM NF1 and 5 µM NuMA (a gift from Marina Mapelli, Department of Experimental Oncology, European Institute of Oncology, Milan) proteins were added to a final volume of 50µl. Reactions were incubated at room temperature for 15 minutes, transferred onto 100 µl of cushion buffer (80 mM PIPES pH 6.8, 1 mM MgCl2, 1 mM EGTA, 50 µM Paclitaxel, 50 % glycerol) and ultracentrifuged for 15 minutes at 400,000 g at 25 °C in a Beckman TLA100 rotor. The supernatant (50µl) was transferred from each centrifuge tube in SDS loading buffer. The cushion buffer was discarded very gently. The pellet was resuspended in SDS loading buffer 1x. Pellets and supernatants were analysed by SDS-PAGE.
GTPase activity
RAS-GAP activity was assessed through the GTPase-Glo™ assay (Promega), following manufacturer’s instructions, and detection through GloMax®-Multi+ Detection System.
Turbidity assay
3 mg/mL Tubulin was mixed on ice in a 96 well plate with 0,6 µM NF1 and/or indicated drugs in a final volume of 100μL in Tubulin Polymerization Buffer (80 mM PIPES pH 6.9, 2mM MgCl2, 0.5mM EGTA, 1mM GTP, 10% glycerol). The polymerization reaction was started by the increase in temperature from 4°C to 37°C upon transfer of the reaction to pre-warmed microtiter spectrophotometer. Tubulin assembly was monitored for one hour at 340 nm in kinetic mode of 61 cycles for each sample and the readings were recorded at an interval of 1 min (Infinite 200 PRO-Tecan plate reader).
In vivo experiments
Bioluminescence imaging (BLI) acquisition and analysis. The images were acquired using PerkinElmer’s IVIS Lumina Series III instrument wavelengths (600- 800 nm). Scans were taken with an integrated CCD camera (Andor, Belfast, UK) supercooled down to −80 °C, with a 25 mm focal length lens (Navitar, Rochester, NY). The camera pointed straight down and was focused 10 mm above the imaging membrane. The F-number was kept at f/0.95 throughout the study. Prior to imaging, 200 µL of D-Luciferin (10 mg/mL, XenoLight, Perkin Elmer) were injected intraperitoneally and mice were anesthetised in induction chambers with 1-4% isoflurane. With animals in the supine position, BLI images were acquired after 10 min after luciferin injection. An exposure time of 2s and binning of 4 was used at the beginning of the study, and imaging parameters were updated as the tumors became brighter throughout the study, in order to maximize sensitivity of bioluminescence and avoid pixel saturation. Tumor burden was represented as total flux (photons/s), which is average radiance (flux per unit area and unit solid angle) integrated over the region of interest, using the Living Image v4.7.3 in vivo software package (Perkin Elmer Inc).
Cell imaging
Immunohistochemistry.
Formalin fixed and paraffine embedded (FFPE) tumour sections of 3µm thickness were cut and left for an overnight incubation at 37°C before staining. All immunohistochemistry (IHC) were performed using Bond III IHC autostainer (Leica Biosystems). For a full automated IHC staining, the antibodies rabbit monoclonal anti-Caspase 3 Activated (#9661 - Cell Signaling, final dilution of 1:250), rabbit anti-Cleaved Caspase 7 (#9491 - Cell Signaling, final dilution 1:100), rabbit monoclonal anti-cleaved Parp (#5625 - Cell Signaling, final dilution 1:50), and mouse monoclonal antibody anti-BCL-2 (#ab692 - Abcam, final diluition 1:100) were unmasked with EDTA Ph9 (Bond Epitope Retrieval Sol2 Leica AR9640). All the antibodies were diluted with Bond Primary Antibody Diluent (AR9352 - Leica Biosystems), and BOND IHC Polymer Detection Kit (DS9800 - Leica Biosystems) was used to detect antibodies with DAB chromogen and counterstain tissues with haematoxylin. H&E staining was performed with automated Leica Biosystems ST 5020 instrument, using Kit Infinity 2000 Test (3801698-Leica Biosystems) and pictures of stained sections were acquired with the Aperio ScanScope XT instrument.
Co-localization immunofluorescence of NF1 and microtubules
Cells were seeded (1 x 105 in 12-well plates) on coverslips coated with 0,5% gelatin and allowed to attach overnight. Cells were synchronised and fixed as previously described. Fixed cells were then blocked with 5% Donkey Serum in IF buffer for 1 hour and subsequently incubated with primary antibodies mouse anti-Neurofibromin (MilliporeSigma, NF1-A 376G3, #MABN2557) diluted 1:100 and rabbit anti-alpha-Tubulin (Abcam, Ab18251) diluted 1:200 in IF buffer for 5 hours at RT. Primary antibodies were detected using anti-mouse Alexa Fluor-555 and anti-rabbit Alexa Fluor-488 secondary antibodies (Life Technologies). DAPI (Sigma, D9542) was used for nucleic staining. Images were acquired with CrestOptics confocal spinning disk X-light V3 with a 100x/1.49 NA objective lens and analysed with ImageJ.
Proximity Ligation Assay
Cells were synchronised and fixed as previously described. Fixed cells were then blocked with 5% Donkey Serum in IF buffer for 1 hour and subsequently incubated with mouse anti-Neurofibromin (MilliporeSigma, NF1-A 376G3, #MABN2557, 1:100 and rabbit anti-alpha-Tubulin (Abcam, Ab18251, 1:200 for 2,5 hours at RT. PLA staining was performed with Duolink® flowPLA Detection Kit – Orange (Sigma-Aldrich, #92102).
Chromosome alignment
Cells were seeded in late G2 with RO3306 (Sigma-Aldrich) at 9 µM for 39h (BT-474) and 20h (HCC1954) at 37°C. Cells were then washed three times with warm medium to release from G2 arrest and put back at 37°C. After 1 hour, cells were treated with MG-132 (Tocris) 10 µM for 90 min at 37°C. Finally, cells were fixed with cold methanol at −20°C for 3 min followed by rehydration in PBS. Fixed cells were then blocked with 5% Donkey Serum in IF buffer (1x TBS + 0,1% Triton X-100 + 2% BSA + 0,1% Sodium Azide in PBS) for 1 hour.
T-DM1 Internalisation
Cells were seeded on 0.5% gelatin-coated coverslips (7.5 x 105 each) and allowed to attach overnight. After a 15 min T-DM1 pulse (1.5 μg/mL), the drug was washed off and cells were either immediately fixed with PFA 4% for 15 min RT or received fresh warm media and incubated at 37°C for 24 h, with or without chloroquine (CQ, 5 μM).
Cells were then fixed permeabilised with Triton X-100 0.5% in PBS and a sheep polyclonal anti-Human IgG-Cy5 (AC112S, Sigma) was used for the detection of T-DM1; a FITC-phalloidin (P5282, Sigma) was used for actin and nuclear counterstaining was done with DAPI. Images were acquired with an SP8 confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany) and a 63x/1.4NA oil immersion objective lens. Multichannel, Z-stack images were acquired with a voxel size of 72x72x200 nm3 by 3 different PMT detectors.
HER2 expression.
Cells were blocked with 5% BSA in PBS for 30 minutes and then incubated with primary antibody anti-HER2 Alexa Fluor-488 conjugate (#FAB1129G-025, Biotechne) diluted in 1% BSA in PBS for 1 hour at 4° C in agitation. Flow cytometry was performed on BD FACS Celesta and all data were analysed using FlowJo 10.
Live imaging of chromosomal and microtubule dynamics.
BT-474WT-H2B-GFP and BT-474KO-H2B-mCherry cells were seeded (4 x 104 cells) in µ-Slide 8-well glass bottom chambers (80827, Ibidi). 24 hours later, growing cells were stained for 1.5h before acquisition with the SiR-tubulin probe (#SC002, Spirochrome; λabs 652 nm/λfl 674 nm) at 1 μM. Time-lapse microscopy was performed using an inverted microscope (Nikon Eclipse Ti) with a 20X/0.75NA objective. The microscope was equipped with an incubation chamber maintained at 37°C in an atmosphere of 5% CO2 and acquisitions were made every 4 min. A total of 28 cells were analysed for BT-474WT-H2B-GFP and 38 for BT-474KO-H2B-mCherry.
Microtubule tip tracking by EB3-GFP
1,5 x 104 HCC1954WT and HCC1954KO cells were cultured onto µ-slide 8 well (ibidi Cat. No. 80826) and infected with LentiBrite™ EB3-GFP (Sigma, Cat. No. 17-10208) according to the manufacturer’s protocol. Fresh media containing DM1 500 pM or DMSO were added 72 h after infection and 30 min before acquisition. Images were acquired on a Yokogawa spinning disk field scanning confocal system CSU-W1 (Nikon Europe B.V., Stroombaan 14, 1181 VX Amstelveen, The Netherlands), equipped with a 488 nm laser line and Photometrics Prime BSI camera. Images were acquired at a single focal plane with a time interval of 1 second between frames for a total duration of 1 minute per each field of view. A Plan Apo λ 100X/1.45NA oil immersion objective was used and a binning 2x2 was applied, to increase the signal to noise ratio and avoid phototoxicity, for a final pixel size of 0.1358x0.1358 um2. To improve image quality, each timelapse movie was denoised with the built-in function Denoise.ai of Nis-Elements AR software (v. 5.42.01). For image analysis, cells were first segmented using Cellpose80 on GFP fluorescence of the first frame of each field of view, considering that cells were not migrating during the total duration of the acquisition (1 min). Per each field of view, the label image resulting from the segmentation was merged to the stack with EB3 fluorescence, in order to establish the cellular hierarchy of EB3 movements. Following segmentation, a customized script that execute TrackMate79 software (v. 7.10.2) in Fiji81 (v. 2.9.0/1.53t) was employed to detect and track EB3 tips throughout the time series: Laplacian of Gaussian detector was used to identify the EB3 spots, Kalman tracker was used to create tracks, and filters on Track duration (> 4 seconds) were used to discard uninformative tracks. The label images from the Cellpose segmentation were integrated into the TrackMate analysis to identify the parent cell of each EB3 track, allowing us to associate specific EB3 dynamics with individual cells. EB3 dynamics was quantified by the Track Mean Speed parameter. Cell surface was calculated by Cellpose.
To calculate dynamic parameters, we selected tracks with confinement ratio>0.8 and track displacement >1 µm. Catastrophe events were calculated as events/µm2/min by dividing the number of tracks per cell by its area. Rescue events were calculated as events/µm2/min and were defined as events starting £ 1 µm from the ending coordinate of the previous event and £ 5 seconds after the previous event, divided by the cell area.
Cell cycle and ploidy.
Cells were pelleted at 1200rpm for 5 minutes and fixed with 90% cold Ethanol for 30 minutes on ice, washed once with 1% BSA in PBS and stained with Propidium Iodide (PI) 2,5 ug/ml + RNase 0,25 mg/ml overnight at 4˚C. Flow cytometry was performed on BD FACS Celesta and cell cycle gates were adjusted as needed to encompass the G0/G1, S, and G2/M populations.
In some experiments, BrdU was pulsed for 1 h on treated cells with T-DM1 or vehicle for 36 h. Cells were fixed and stained according to manufacturer’s instructions (FITC BrdU Flow Kit, BD Biosciences).
Sequencing
DNA
Genomic DNA extraction was performed using the DNeasy Blood & Tissue Kit (Qiagen), in accordance with the manufacturer’s instructions. DNA concentrations were carried out using Qubit dsDNA Broad Range quantification assay kit (Thermo Fisher Scientific).
For WES, Target regions were captured with Twist Comprehensive Exome Panel probes followed by PCR amplification and purification of the enriched library.
Quantification of enriched libraries was performed with Qubit dsDNA High Sensitivity quantification assay kit (Thermo Fisher Scientific) and library size distribution was measured with Bioanalyzer 2100 and High Sensitivity DNA Kit (Agilent Technologies). Final DNA libraries sequencing was performed in Illumina NovaSeq 6000 platform using the NovaSeq 6000 S1 Reagent Kit 300 cycles (2 x 150 paired-end reads) (Illumina). Sequencing data were mapped against hg38 genome and analysed with Illumina DRAGEN Bio-IT Platform v4.0 using proprietary pipelines for variant calling. The resulting VCF with detected variants files were annotated and classified with the GATK-Funcotator.
RNA
mRNA-seq libraries were prepared according to the TruSeq low sample protocol (Illumina, San Diego, CA, USA), starting with 1 µg of total RNA per sample. RNA-seq libraries were pair-end sequenced on an Illumina NovaSeq 6000 sequencing platform. RNA-seq data were mapped using the STAR aligner version 2.7.1a against the human genome (hg19). Read counts were calculated with htseq-counts with the standard UCSC reference GFF (https://hgdownload.soe.ucsc.edu/goldenPath/hg19/bigZips/genes/hg19.refGene.gtf.gz). Differential expression analysis was done with DESeq2 R package version 1.36.0. Genes of interest were selected using a false discovery rate (FDR) cut-off of 1x10-4
Functional enrichment analysis war performed using EnrichR R package version 3.1 on differentially expressed genes with log2foldchange < -1 in NF1 knock-out T-DM1 treated vs NF1 knock-out and log2foldchange > -1 and log2foldchange < 1 in wild-type T-DM1 treated vs wild-type. Queries were performed over four different molecular signatures: Gene Ontology Biological Process, Gene Ontology Molecular Function, Gene Ontology Cellular Component (https://www.gsea-msigdb.org/gsea/msigdb/download_file.jsp?filePath=/msigdb/release/2022.1.Hs/c5.go.v2022.1.Hs.symbols.gmt) and Hallmark (https://www.gsea-msigdb.org/gsea/msigdb/download_file.jsp?filePath=/msigdb/release/2022.1.Hs/h.all.v2022.1.Hs.symbols.gmt) (10.1073/pnas.0506580102, 10.1093/bioinformatics/btr260, 10.1016/j.cels.2015.12.004). Enriched terms were selected using a standard FDR cut-off of 1x10-2.