Fig 2: Afatinib reduces VEGF-A secretion, angiogenesis, and lesion size.(A) Afatinib decreased EGFR/ErbB1-phosphorylation in rhTGFA-stimulated HPF-c cells. Total epidermal growth factor receptor (EGFR) was used to control equal loading of the samples. (B) VEGF secretion measured by ELISA from rhTGFA-stimulated control fibroblasts (HPF-c) with or without afatinib treatment. Two independent experiments done in triplicates are presented as mean and SEM. Brown-Forsythe and Welch ANOVA with Dunnet T3 post-hoc test. **p<0.01. (C, D) In the fibrin bead assay with HUVECs and HPF-c, afatinib inhibited endothelial cell (EC) sprouting induced by co-stimulation with rhVEGF-A and rhTGFA. Representative images of each group are presented at d7 (C) ECs are labeled with phalloidin (red) and nuclei with DAPI (blue). Scale bars, 100 µm. Quantitative analysis for the number of sprouts per bead (D) was performed with ImageJ software (30 beads/group). Afatinib treatment was started at d3 after HUVECs had already formed angiogenic sprouts. The data from two independent experiments done in triplicates are presented as mean and SEM. One-way ANOVA with Bonferroni post-hoc test. *p<0.05; ***p<0.001. Figure 6—source data 1.Raw data for western blot images.

Fig 3: Genes involved in ErbB signaling pathway are upregulated in patient-derived endothelial cells (ECs) in vascular lesions with the venous components.(A) Magnetic resonance images of an angiomatosis of soft tissue (AST) lesion (arrows) in the soleus muscle show the replacement of the normal muscle by dilated venous channels, diffusely enhancing small vessels and adipose tissue. (i) Axial T1-weighted fat-saturated contrast-enhanced image. (ii) Axial T2-weighted image. (iii) Sagittal T2-weighted fat-saturated image. (B) 1mm-thick whole immunomounts were prepared from patient lesions, immunolabeled, and imaged by laser scanning confocal microscopy. Images of the vasculature in the AST lesion located in the shin of a 16-year-old female are shown. Endothelial cells are immunolabeled with CD31 antibody (red). Vasculature of the same lesion in the intramuscular area (i), with abundant fat (ii) and next to resection margin (iii) are presented. Longitudinal vessels are seen. Scale bars, 100 µm. (C) Heatmap of normalized RPKM values (-1–1) of the differentially regulated genes in patient-derived ECs compared to HUVEC and human saphenous vein endothelial cells (HsaVEC) control cells detected by bulk RNA-seq. Clustering was performed using Spearman’s rank correlation. Biological processes and cell signaling pathways detected by gene ontology analysis in patient-derived ECs. (D) Scatter plot of the fold changes in gene expression comparing patient-derived and control ECs. Selected genes involved in PIK3CA, VEGFR2, and ErbB1-4 signaling are highlighted in red (upregulated) and blue (downregulated). Pearson correlation value (r2) is shown. (E) Changes in mRNA expression levels of ErbB ligands (transforming growth factor A (TGFA), with two different assays; amphiregulin, AREG; epigen, EPGN1; neuregulin 1, NRG1) were validated with RT-qPCR from patient-derived and control ECs. Mean and SEM is presented (HsaVEC and HUVEC, n=3; patient ECs, n=5). Kruskal-Wallis test with two-stage step-up method of Benjamini, Krieger, and Yekutieli to control FDR [TGFA (1), TGFA (2), AREG; data not normally distributed] or Brown-Forsythe and Welch ANOVA with Dunnet T3 post hoc test (NRG1, EPGN; data normally distributed but unequal variances). *p<0.05. (F) Sequence motifs associated with differentially regulated genes in patient-derived ECs.

Fig 4: Epidermal growth factor receptor (EGFR)/ErbB1 ligand transforming growth factor A (TGFA) is upregulated in venous malformation (VM) and angiomatosis of soft tissue (AST) patient tissue samples.(A) RT-qPCR analysis showed significantly higher expression of TGFA mRNA (with two different assays) in VM and AST tissue than in the control group. Mean and SEM are presented (lesions n=11; control group, n=6), analyzed using two-tailed Mann-Whitney U test. *, p<0.05. (B) Representative images of TGFA expression in AST and VM patient samples. See Figure 2—figure supplement 2 for normal skeletal muscle control. TGFA signal was detected in lesions in the endothelium, pericytes, and intrastromal cells by immunohistochemistry. Asterisks point out the largest vascular lumens which, especially in AST, are commonly tightly packed with erythrocytes. Scale bars, 200 µm. (C) Representative whole immunomount images of CD31-labeled endothelium (green) and TGFA expression (red) in a patient diagnosed with intramuscular AST. Nuclei are stained with DAPI (blue). Longitudinal vessels are seen. Scale bars, 100 µm. (D) Representative confocal images of phosphorylated epidermal growth factor receptor (EGFR) (red) expression in AST lesion. Endothelium is labeled with CD31 antibody (green), and nuclei with DAPI. Vascular lumens are indicated with white asterisks in the cross-sections. Scale bars, 50 µm. (E) Heatmap of TGFA and pEGFR protein expression and presence of oncogenic PIK3CA variants in AST patients. Level of protein expression was scored (0–3) based on the detected signal in immunocytochemistry (0, none; 1, low; 2, medium; and 3, high). (F) RT-qPCR analyses of TGFA expression in patient-derived endothelial cells (ECs) and intervascular stromal cells (SCs). Selection of ECs was performed by CD31 MicroBead Kit. Stromal cells were characterized by western blot and showed to be negative for EC marker CD31, and positive for fibroblast and smooth muscle cell marker vimentin (see Figure 2—figure supplement 2). The data is presented as relative mean fold change to HsaVEC control group and SEM (Human saphenous vein endothelial cells:HsaVEC, HUVEC, and HPF-c, n=3; patient ECs, n=5; patient SCs, n=6). Statistical significance was analyzed using Kruskal-Wallis test with two-stage step-up method of Benjamini, Krieger and Yekutieli to control FDR (ECs) or two-tailed Mann-Whitney U test (SCs; in both analyses, data not normally distributed). *p<0.05. (G) TGFA was shown to be secreted from patient-derived intervascular stromal cells (SCs) by ELISA (patient SC n=5; HPF-c n=3). Two-tailed unpaired t-test with Welch’s correction (data normally distributed but unequal variances). **p<0.005.