Fig 2: Inhibition of VCAM-1/ICAM-1 mediated integrin signaling in vivo down-regulates IL-1α serum levels of JAK2VF/+ mice. A 10 weeks old Vav1-Cre x Jak2+/+ (grey; n = 8) and Vav1-Cre x Jak2VF/+ mice (dark blue; n = 7) were injected with a single intraperitoneal dose of PBS (200 µl) to control for injection-dependent artefacts. Results were compared with Vav1-Cre x Jak2.VF/+ mice that had received a single dose (200 µl) of anti-VLA4/β2-integrin antibody (red; 200 µg, n = 8) or the corresponding IgG control (light blue; 200 µg, n = 8) (Created in BioRender. Haage, R. (2025) https://BioRender.com/v80w629). On day 8 upon injections, serum was collected and cytokine serum analysis was performed. B IL-1α serum concentrations are shown as mean + SEM. Ordinary One-Way ANOVA P* < 0.05, P** < 0.01

Fig 3: Characterization of integrin outside-in signaling nodes involved in VCAM-1/ICAM-1-induced mRNA expression of IL-1α and IL-1β. A Cartoon depicting the experimental design (Created in BioRender. Haage, R. (2025) https://BioRender.com/w21u020). B In 32D JAK2-VF cells, Il1a and Il1b mRNA expression is induced by the VLA4-VCAM-1 interaction (n = 4) but not by the β2-integrin-ICAM-1 interaction (n = 4). Specificity was tested by pre-incubation using 5 µg/ml anti-VLA4 or anti-β2-integrin antibody (n = 4). The stimulation of VCAM-1 and ICAM-1 in combination resulted in a further increase in adhesion-induced mRNA expression (orange; n = 3). C-D 32D JAK2-WT (blue) and 32D JAK2-VF (red) cells were stimulated with VCAM-1/ICAM-1 and various inhibitors were applied. The expression levels of (C) Il1a (n = 3) and (D) Il1b mRNA (n = 3) were measured upon inhibition of signaling molecules, including SRC (5 µM PP2), FAK (25 µM Y15), SYK (30 µM Piceatannol (Pic)), PI3K (50 nM Wortmannin (Wort)) and NFκB (50 µM JSH-23). E–F Clinical therapeutic drugs such as Ruxolitinib (2 µM RUX; orange) and Hydroxyurea (0.5 mM HU; blue) were tested for their effects on adhesion-induced (E) Il1a and (F) Il1b mRNA expression. To exclude the possibility of autocrine IL-1 signaling, IL1Ra was applied (dark grey). mRNA expression was measured by qPCR upon 3 h of stimulation and the resulting data are shown as fold change (2−ΔΔCT) as mean + SEM (n = 3–4). B Mixed-effects analysis (uncorrected Fisher´s LSD), C-D Friedman test (uncorrected Dunn´s test), E–F RM one-way ANOVA (uncorrected Fisher´s LSD) with P* < 0.05, P** < 0.01

Fig 4: RNASeq in primary granulocytes shows a VCAM-1/ICAM-1 activated inflammatory cytokine signature. A Granulocytes isolated from Vav1-Cre x Jak2+/+ and Vav1-Cre x Jak2VF/+ mice were stimulated for 1 h using immobilized VCAM-1/ICAM-1 or BSA (control). RNA sequencing (n = 4) and DNA-methylation analysis (n = 3) were performed (Created in BioRender. Haage, R. (2025) https://BioRender.com/n22t547). B Gene set enrichment analysis (GSEA) profiles of cytokine activity in Jak2-WT granulocytes (left) and in Jak2-V617F mutated granulocytes (right) upon adhesion to VCAM-1/ICAM-1 vs. unstimulated control (BSA) are shown. The normalized enrichment score (NES), the p-value and the false discovery rate (FDR) are given. C Changes in mRNA expression (log2 fold change) of 18 representative inflammatory genes of the GSEA gene set are depicted and shown as heatmap normalized to unstimulated (BSA stimulated) Jak2-WT granulocytes. The scale ranges from blue (representing 1) to bright red, indicating the highest observed increase in log2 fold change. Unstimulated (BSA) and VCAM-1/ICAM-1 stimulated granulocytes of Vav1-Cre x Jak2+/+ and Vav1-Cre x Jak2VF/+ mice were compared. Differentially expressed genes (deseq2, log2 fold > 1, padj < 0.05) are indicated with an asterisk *. The unstimulated (BSA) controls have been partly depicted in a previous publication [78]. Please note that in Fig. 5 C the log2 fold change covers a range from 0 up to > 40. D A DNA-methylation array was performed upon VCAM-1/ICAM-1 stimulation (1 h) of isolated granulocytes from Vav1-Cre x Jak2+/+ (JAK2_WT; n = 3) and Vav1-Cre x Jak2.VF/+ (JAK2_VF; n = 3) mice. The CpG methylation islands of the genes shown in (C) were depicted in a heatmap (blue represents 0—not methylated, red represents 1 – fully methylated)

Fig 5: Outside-in signaling of β1/β2 integrins is activated upon adhesion to VCAM-1/ICAM-1. A-B Adhesion of 32D JAK2-VF cells on VCAM-1/ICAM-1 (15 min; blue; n = 4) induces phosphorylation of canonical integrin outside-in signaling in a VLA4/β2-integrin-specific manner. Pre-incubation with 5 µg/ml anti-VLA4/β2-integrin antibody (orange; n = 4) led to a reduction in adhesion-induced phosphorylation, while the corresponding IgG isotype control (5 µg/ml; green; n = 4) remained largely unaltered. A Representative Western blots of n = 4 replicates are shown. B Semi-quantification of phospho-signals derived from Western blots (n = 4) is depicted as phosphorylated (p) to total (t) protein signals normalized to the unstimulated control (BSA). 2-Way ANOVA P* < 0.05, P** < 0.01, P*** < 0.001 (C) Model of outside-in signaling upon adhesion to VCAM-1/ICAM-1. For characterization of outside-in signaling upon VCAM-1/ICAM-1 stimulation, a panel of representative signaling molecules (SRC, SYK, FAK, PI3K and NFκB) was investigated using the following inhibitors: PP2 (5 µM), Piceatannol (30 µM), Y15 (25 µM), Wortmannin (50 nM) and JSH-23 (50 µM), respectively. Results are given in Table 1. The proposed signaling model derived from this data is depicted as a cartoon (Created in BioRender. Haage, R. (2025) https://BioRender.com/q22h778). D-E VCAM-1/ICAM-1 stimulation (15 min) using total bone marrow cells of Vav1-Cre x Jak2+/+ and Vav1-Cre x Jak2.VF/+ mice induced the phosphorylation of STAT3 (Ser727), SYK (Tyr519/520) and JNK (Tyr183/185) in comparison to unstimulated cells (BSA). D Representative Western blots (n = 2–3) and (E) semi-quantification (n = 2–3) are shown. Data are shown as mean + SEM. 2-Way ANOVA P* < 0.05, P** < 0.01, P*** < 0.001