Patients

Samples of patients who had surgery on vascular malformations between 2018 and 2022 were analyzed. The study was conducted at the Department of Otorhinolaryngology, University Medical Center Regensburg, Germany, according to the principles of Helsinki and approved by the Local Ethics Committee (No. 17-854-101). All subjects, and/or their legal guardians, gave written informed consent.

RNA isolation and PCR array

RNA was extracted from formalin-fixed and paraffin-embedded (FFPE) samples using the RNeasy FFPE Kit (Qiagen, Hilden, Germany), as described by the manufacturer. A pool of ten fast-flow malformation patients (arteriovenous malformations) was compared with a pool of ten slow-flow malformation patients (venous (n = 6) and lymphatic malformations (n = 4)). Expression of genes relevant for angiogenesis was determined with a RT2 Profiler PCR Array (PAHS-024Z, Qiagen) using RT2 SYBR Green qPCR Mastermix (Qiagen) and the LightCycler 480 Real-Time PCR System (Roche Diagnostics, Mannheim, Germany).

Immunohistochemistry

FFPE samples (5 µm) were deparaffinized and immersed in an antigen retrieval solution (Citrate-EDTA buffer: 10 mM Citric Acid, 2 mM EDTA, 0.05% Tween-20, pH 6.2) for 20 min at 95 °C–99 °C. Sections were subsequently blocked for 30 min in TNB Blocking buffer (PerkinElmer, Boston, MA) followed by incubation with human-specific TGF-β polyclonal antibody (1:400, rabbit anti-human, Abcam, Cambridge, UK, ab92486). Immunohistochemical staining for VEGFR-2 was provided by the Department of Pathology, University Medical Center Regensburg. (Benchmark ultra, Roche, Mannheim, Germany; 1:400, rabbit anti-human, medac, Wedel, Germany, E3710). Each slide was manually scanned with a microscope at 10X, 25X or 40X magnification. We captured five areas per slide (five high-power fields, HPFs) with each showing a characteristic staining of the whole slide. Positive stained cells were counted via ImageJ (National Institutes of Health, USA) cell counter function by two independent examiners. The H-score is a method of assessing the extent of immunoreactivity, applicable to VEGF receptors. The score is obtained by the formula: 3 × percentage of strongly staining + 2 × percentage of moderately staining + percentage of weakly staining, giving a range of 0–300 [39].

Cell isolation and culture

Single-cell suspensions were prepared from three different tissue samples of AVM patients with disease stage Schobinger III. The AVM diagnosis was confirmed by the Institute of Pathology at the University Hospital Regensburg. In all three AVM patients KRAS mosaic mutations were detected. Endothelial cells (EC) were selected using anti-CD31-coated magnetic beads (Miltenyi Biotec, Auburn, CA) respectively and expanded. Testing for mycoplasma contamination by qPCR was performed when cells were thawed and every 4–6 weeks thereafter. Cells were cultured on fibronectin-coated (0.1 µg/cm2; EMD Millipore, Billerica, MA) plates with Endothelial Cell Growth Medium-2 (EGM-2; Lonza, Allendale, NJ), which consists of Endothelial Cell Growth Basal Medium-2 (EBM-2; Lonza), SingleQuot supplements (all except hydrocortisone; Lonza), 10% heat-inactivated fetal bovine serum (FBS; Hyclone, South Logan, UT) and 1X GPS (292 mg/mL Glutamine, 100 U/milliliter (mL) penicillin, 100 mg/mL streptomycin; Mediatech Inc, Manassas, VA). Cells were cultured at 37 °C in a humidified incubator with 5% CO2.

Cyclic mechanical stretching

To mimic the mechanical stress on high-flow malformations a mechanical stretcher was designed by the department for Science Laboratory Technology of the University Regensburg. The mechanical stretcher performs uniaxial cyclic mechanical stretch on cells in tissue culture (Supporting Fig. 1B). AVM endothelial cells (2 × 104 cells/cm2 in 2000μL EGM-2 culture medium per well) were seeded in duplicates into 6-well plates with flexible floor membranes (BioFlex Culture Plate Untreated 6-Well, Flexcell International Corporation, Burlington, USA) for 24 h. Then, the medium was removed and either fresh EGM-2 culture medium or respective treatment (bevacizumab 1000 µg/ml or thalidomide 20 µM in EGM-2 medium) was added. The flexible membranes were stretched by a generated negative pressure of − 0,3 bar every minute. Once the threshold pressure was reached every minute pressure increased to 0 bar. The cells attached to the membrane were exposed to cyclic mechanical stretching every minute for 24 or 48 h. During the experiment the mechanical stretcher was placed at 37 °C in a humidified incubator with 5% CO2. A 6-well plate with flexible floor membranes placed in the incubator without stretching served as control.

RNA isolation and quantitative reverse transcriptase PCR (qPCR)

Total cellular RNA was extracted from cultured cells with a RNeasy Micro extraction kit (Qiagen, Valencia, CA). Reverse transcriptase reactions were performed using an iScript™cDNA synthesis kit (BioRad, Hercules, CA). QPCR was performed using KAPA SYBR® FAST ABI Prism 2 × qPCR Master Mix (KAPA BioSystems, Wilmington, MA). Amplification was carried out in a StepOne™ Real-Time PCR System (Applied Biosystems, Foster City, CA). A relative standard curve for each gene amplification was generated to determine the amplification efficiency, with greater 90% considered acceptable. Fold increases in gene expression were calculated according to two delta CT method, with each amplification reaction performed in duplicates or triplicates. GAPDH was used as housekeeping gene expression reference. Primer sequences are shown in Table 1.

Enzyme-linked immunosorbent assay (ELISA)

Soluble cytokine production in the supernatants of cultured AVM endothelial cells was tested by ELISA (DuoSet ELISA Development Systems; R&D Systems). Prior to use, the cell culture supernatants were centrifuged at 1200 rpm for 5 min. The DuoSet kit human VEGF (DY293B) was used according to the manufacturer’s instructions. Measurements were obtained in triplicates.

Proliferation assay

CD31+ endothelial AVM cells (8 × 104 cells in 3000 μL of the respective culture medium per well) were seeded in triplicates into 6-well plates. Four doses of bevacizumab and onartuzumab (250 µg/ml, 500 µg/ml, 750 µg/ml und 1000 µg/ml) and three doses of thalidomide (10 µM, 20 µM and 40 µM) in EGM-2-medium were analyzed. Concentrations were selected in accordance with current literature [12,13,14,15]. After 4 h and 24 h cells were trypsinized and counted. Proliferation was assessed by the number of cells detected after 24 h of growth.

Angiogenesis assay

25 ng/ml VEGF-A was added to the control group and the two treatment groups (1000 µg/ml bevacizumab and 20 µM thalidomide) one hour prior to the angiogenesis assay. Wells were precoated with Matrigel (Sigma-Aldrich, St. Louis, USA) and incubated for 30 min at 37 °C. AVM endothelial cells and HDMEC, as positive control, were seeded at a density of 4 × 104 cells/cm2 in 500 μL of EBM-2/0.1%FBS. 25 ng/ml VEGF-A and respective treatment (1000 µg/ml bevacizumab and 20 µM thalidomide) were added to the VEGF-A control and treatment conditions. After 6 h, pictures were taken with an inverted microscope (Echo Rebel Inverted Brightfield Microscope, Echo, San Diego, CA). The number of circular networks was counted per nine squares per high power field. The area of the circular networks was measured in pixels. Fiji ImageJ software (NIH) was used for analysis.

Bevacizumab treatment of AVM patients

After interdisciplinary and multicentric discussion three patients were treated with bevacizumab. Other therapeutic options, like embolization and surgery, did not result in symptom control in all three patients. The patients were informed about the off-label use of bevacizumab in AVM treatment and possible side-effects, especially gastrointestinal, renal, and cardiovascular risks (for a detailed list of possible side-effects see AVASTIN® (bevacizumab) full prescribing information provided online by the manufacturer; https://www.gene.com/medical-professionals/medicines/avastin). The patients gave informed consent to the individual patient treatment. Before the first administration of bevacizumab blood pressure was measured, a blood test with complete blood count, a comprehensive metabolic panel, C-reactive protein test, thyroid function test, coagulation tests, B-type natriuretic peptide test, a urinalysis with urine protein test and an echocardiography was performed. Immediately before systemic bevacizumab administration, 8 mg dexamethasone were given intravenously. The day after the treatment blood and urine tests were repeated and the patient was discharged from the hospital.

A 25-year-old female patient with an AVM of the right ear and face was treated with 5 mg bevacizumab per kg/bodyweight systemically every 14 days (395 mg bevacizumab dissolved in 250 mL of 0.9% sodium chloride intravenously every 14 days) over a period of eight months. The dosage was chosen as previously described for the systemic treatment of HHT [16, 40]. A 66-year-old female patient with residual disease of an AVM of the left thumb and a 35-year-old male patient with an AVM of the upper lip received treatment with 45 mL bevacizumab (concentration 1,5 mg/mL, in total 67,5 mg bevacizumab, dissolved in 0.9% sodium chloride) and 3 mL bevacizumab (concentration 3,75 mg/mL, in total 11,25 mg bevacizumab, dissolved in 0.9% sodium chloride) by intralesional injection. The volume of bevacizumab was determined by the volume of contrast agent used to visualize the lesion during digital subtraction angiography. As the second patient showed a small lesion with low volume, a higher concentration of bevacizumab was used. The concentration was chosen as previously described for the intralesional treatment of arteriovenous malformations of the mucosa of the nose in HHT [41]. A review of the literature on intralesional use of bevacizumab in HHT showed a significant improvement of the bleeding burden (42).

Statistics

Data was analyzed and plotted by using GraphPad Prism 9.5 (GraphPad Software). For experiments in which cells were exposed to mechanical stress or treated with drugs, the differences were assessed by one-way analysis of variance (ANOVA) followed by the post-hoc Šidák or Bonferroni test for multiple comparisons of different treatment modalities. For comparisons between fast-flow malformations and slow-flow malformations Mann Whitney test was applied. p-values ≤ 0 0.05 are considered significant. Means and standard deviations are shown in all graphs.