Patients/study approval

All subjects, and/or their legal guardians, gave written informed consent to genetic investigations, which were carried out with approval by the institutional review boards of the University Hospital Regensburg, Germany (17-854-101), University Hospital of Bern, Switzerland (2017-01960), and Boston Children’s Hospital, Boston, MA, USA (IRB-P00025772).

Genetic testing

RIT1 was tested in a total of 691 samples by the partners’ laboratories (235 in Magdeburg, 114 in Bern, 342 in Boston), including all types of vascular anomalies. Out of these samples, 118 were submitted for sequencing with the diagnosis of “AVM” (58 to Magdeburg, including by non-specialized centers, 35 to Bern, 25 to Boston). Brain AVMs were not included in the submitted samples. From these samples, one sample at each center harbored a RIT1 mutation. Since sequencing results at Magdeburg also revealed mutations (e.g. in TEK or GNAQ) that according to current knowledge do not occur in AVMs, we conclude that phenotyping by non-specialized centers was in part incorrect and estimate that roughly 30–40 true AVM samples were sequenced in Marburg. This would be in line with a recently published cohort from Germany that included 29 patients with mutations in the RAS pathway [33]. We thus estimate the prevalence of RIT1 mutations in AVMs at roughly 1 in 30 patients.

A tissue biopsy of the AVM of P1 was submitted to the Institute of Human Genetics, University Hospital Magdeburg, and the genomic DNA was extracted. Assuming a mosaic mutation as the cause of the disorder, ultradeep sequencing and enrichment using an Agilent SureSelect XT HS2 Custom Enrichment Panel with molecular barcoding (UMIs, 3 bp duplex) (Agilent Technologies) were performed. The library was sequenced on a NextSeq550 instrument (Illumina), 2 × 150 bp paired-end reads. The target regions had a mean coverage of > 3000× after demultiplexing. The varvis 1.20.0 analysis software (Limbus Medical Technologies GmbH) was used for analysis.

P2 has been included in the Bernese Congenital Vascular Malformation Registry, a prospective cohort of congenital extracranial/extraspinal vascular malformations that have been enrolling consecutive patients since 2008 [34]. As of October 2020, genetic testing is performed on tissue available from diagnostic biopsies of vascular malformations, using the TruSight Oncology 500 (TSO500; Illumina) Next Generation Sequencing (NGS) gene panel.

Resected tissue from P3 underwent targeted DNA NGS testing via the OncoPanel assay at the Center for Advanced Molecular Diagnostics (CAMD) at Brigham and Women’s Hospital [35]. DNA was isolated using standard extraction methods (QIAGEN) and quantified with PicoGreen-based double-stranded DNA detection (Thermo Fisher Scientific). Indexed sequencing libraries were prepared from 50-ng sonically sheared DNA samples using Illumina TruSeq LT reagents (Illumina). Extracted DNA underwent targeted NGS using the KAPA HTP Library Preparation Kit (Roche), a custom RNA bait set (Agilent SureSelect) and sequenced with the Illumina HiSeq 2500 system.

Cell culture and western blot

Three million HEK293T cells were seeded in 10 cm cell culture plates supplemented with DMEM containing 10% fetal bovine serum (FBS) 12 h prior to transfection. At around 70% confluency levels cells were transfected using TurboFect transfection reagent (Thermo Fisher #R0532), with Flag-tagged RIT1 variants in pCDNA constructs or empty vector (EV) as the negative control. The medium was refreshed at the 24 h’ time point with MEK inhibitor (PD0325901, Selleckchem # S1036) and SHP2 inhibitor (SHP099, Selleckchem # S6388) being added to the transfected cells at 1 μM and 5 μM concentrations, respectively. At 48 h post-transfection, cells were washed in ice-cold phosphate-buffered saline (PBS) and lysed in ice-cold lysis buffer, containing 50 mM Tris/HCl pH 7.5, 5 mM MgCl2, 100 mM NaCl, 1% Igepal CA-630, 10% glycerol, 20 mM ß-glycerolphosphate, 1 mM Na-orthovanadate, EDTA-free inhibitor cocktail 1 tablet/50 ml. After the addition of Laemmli sample buffer, the samples were subjected to SDS-PAGE (12.5% polyacrylamide). Blots were detected by immunoblotting using a mouse anti-γ-Tubulin antibody (Sigma #T5326), a mouse anti-FLAG antibody (Sigma #F3165), a rabbit anti-ERK antibody (Cell signaling technology #9102), and a rabbit anti-p-ERK antibody (Cell signaling technology #4370), a rabbit anti-AKT (Cell signaling technology #9272), a rabbit anti-p-AKT-Thr308 (Cell signaling technol-ogy #2965), a rabbit anti-p-AKT-Ser473 (Cell signaling technology #4060). The immunoblots were detected using an Odyssey Fc Imaging System (LI-CORE Biosciences) and analyzed by Image Studio Lite Ver 5.2.

Zebrafish husbandry

Maintenance and breeding of zebrafish (Danio rerio) were performed in the fish facility of the Developmental Biology, Institute for Biology I, University of Freiburg under standard conditions. Only embryos up to 5 days post-fertilization were used. All experiments were carried out in accordance with German laws for animal care and the Regierungspräsidium Freiburg.

Plasmid preparation

Plasmids were designed using ApE—A plasmid editor version 3.0.8. Homo sapiens RIT1 sequence was obtained from the online database Ensembl (Transcript ID: ENST00000368323.8), minimally codon optimized for Danio rerio and ordered as a plasmid including Tol2 sites, a UAS promoter, RIT1P2, and P2A-GFP from Twist Bioscience (South San Francisco, CA, USA). Plasmids were purified using Wizard Plus SC Minipreps DNA Purification Systems (Promega, Walldorf, Germany, A1330) according to the manufacturer’s instructions.

Mutagenesis

RIT1wt, empty vector (EV), as well as all other RIT1 mutations analyzed in this study were derived from the UAS:RIT1P2-P2A-GFP construct using Q5 Site-Directed Mutagenesis (New England Biolabs, E0554S). Corresponding mutagenesis primers were designed using NEBaseChanger version 1.3.3. All plasmids were sequenced by Eurofins genomics to confirm the expected sequence.

Tol2 transposase mRNA synthesis

8 µg of the plasmid that contains the transposase gene under control of the SP6 promoter were linearized using 4 µl of NotI-HF enzyme (New England Biolabs) for 1 h at 37 °C. The digested sample was purified using the QIAquick PCR Purification Kit according to the manufacturer’s protocol. Capped Tol2 transposase mRNA was synthesized from purified DNA using the mMESSAGE mMACHINE™ SP6 Transcription Kit (ThermoFisher Scientific) according to the manufacturer’s protocol. The resulting mRNA was separated into 5 µl aliquots and stored at − 20 °C to prevent freeze–thaw cycles.

Plasmid injection

The construct was then injected into Tg(fli1a:Gal4FFubs3; UAS:RFP) embryos at the one-cell stage together with Tol2 transposase mRNA [36], both at a concentration of 30 ng/µl. For better readability, Tg(fli1a:Gal4FFubs3; UAS:RFP) embryos injected with a gene of interest (e.g. UAS:RIT1P1-P2A-GFP) are abbreviated as fli1a:RIT1P1G0mosaic instead of Tg(fli1a:Gal4FFubs3; UAS:RFP) and Tg(UAS:RIT1P1-P2A-GFP)G0mosaic.

Microscopy

Imaging plates for confocal microscopy were prepared in 35 mm glass bottom dish with 1.5% agarose in egg water, 1-phenyl 2-thiourea (PTU) and tricaine mix using previously designed and 3D printed molds (Online resource 9). Embryos anesthetized with 0.168 mg/ml tricaine in egg water at room temperature and gently positioned laterally inside the trenches. Confocal microscopy images were acquired as z-stacks with ZEISS Celldiscoverer 7 with LSM 900. Images were obtained with the 488 nm and 561 nm lasers, with a slice interval of 2–4 μm with a 20X (NA 0.7) objective or as brightfield images with 5x (NA 0.35) objective unless otherwise specified. For brightfield timelapse experiments, the interval was set to 1 s.

Fluorescence microscopy images for pre and post late treatment was acquired with ZEISS Axio Examiner D.1 fixed stage fluorescence microscope. During acquisition embryos were placed in 3 ml of E3 medium with tricaine (0.168 mg/ml) at room temperature and imaged with 10X (NA 0.15) objective. During imaging both, RFP and GFP channels acquired and only RFP channel is exported for representative images.

Lightsheet microscopy was performed with ZEISS Lightsheet 7 using water immersion W Plan-Apochromat 10x (NA 0.5) M27 75 mm objective. Images were obtained with the 488 nm and 561 nm lasers using single illumination, pivot scan on, with a slice interval of 2 μm. Timelapse interval is set to 1 s.

The color cyan was assigned for GFP channel. Colored versions of the images are included in the supplementary information. Images and videos were exported as.tif and.AVI (uncompressed) files, respectively using Fiji software.

Pharmacological treatments

For pharmacological treatments, injected zebrafish embryos were randomized into control and treatment groups. From the 10-somite stage or from 48 hpf on, embryos of the treatment group were transferred in E3 Medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM Mg2SO4) containing 0.2 mM 1-phenyl 2-thiourea (PTU; Sigma, Taufkirchen, Germany, P7629) and MEK1/2 inhibitor Trametinib (MedChemExpress, GSK1120212; 10 mg) using 100× stock solutions dissolved in dimethyl sulfoxide (DMSO; Sigma, D2650). The treatment dose of trametinib was chosen at 100 nM, according to our previous publication [18], and by repeating of the toxicity assay in zebrafish embryos. Embryos of the control group were raised in E3 medium with 0.2 mM PTU and DMSO (equal amount to the treatment group). The response of the AVM-like lesion size to trametinib was calculated as follows: each embryo with an AVM-like lesion was imaged at the Axio Examiner, and the area of the lesion was divided by the area of the entire embryo to give a relative area of the malformation at 2 dpf. This was done to control for different embryo sizes and different embryo growth rates. This measurement was then repeated at 4 dpf (after treatment) and the relative area of the AVM-like lesion at 4 dpf was divided by the relative area 2 dpf, followed by multiplication with 100 to give a result in percent. A result greater than 100% showed an AVM-like lesion growing in size in relation to the embryo, a result less than 100% showed a regressing lesion.

Statistical analysis

The statistical analysis was performed using two-tailed Fisher’s exact test of significance for malformation rate analysis and early pharmacological treatments and two-tailed unpaired t-test for late pharmacological treatment experiments and one-way ANOVA for in vitro experiments in GraphPad Prism version 10.2.2. The legends of the figures include information on sample sizes and significance. P value of < 0.05 was regarded as significant. Analyzed data for zebrafish malformation rates and early pharmacological treatments was obtained from at least three independent experiments for each variant. The number of examined embryos is indicated in the figure legends for each variant. Data are presented as mean ± SD for all experiments. P value *< 0.05, **< 0.01; ***< 0.001; **** < 0.0001. The structure of the RIT1 protein is predicted using AlphaFold database [37, 38].