Fig 2: ADAM8 interacts with the actin-based motor protein Myo1f via SH3 domains and modulates cell motility.(A) Domain structure of ADAM8 and Myo1f as interaction partner identified by a yeast-2-hybrid screen using a leukocyte library. ADAM8 CD (aa 678–824) was used as prey and the reporter gene indicated interaction (“+”) via the SH3 domain of Myo1f (clone BP19, aa 1032–1098). (B) Co-IP of activated human PMN (hPMN) and (C) HL-60 cells using polyclonal ADAM8 (A8) antibody coupled to magnetic beads. Negative controls, beads bound to control IgG (IgG); unconjugated beads (beads). (D) Immunofluorescence staining against Myo1f (green) and ADAM8 (red) of hPMN and HL-60 cells. Overlay (yellow) and colocalized pixel map (Coloc Map) indicate colocalization in cell protrusions. Scale bar, 10 µm. (E) Co-IP of hPMNs + BK-1361 or CP during activation. (F) Relative pixel intensity of Myo1f coprecipitated with ADAM8 normalized to input using ImageJ software (NIH). *, P < 0.05, Student’s t test. (G) Colocalization map of hPMNs + BK-1361 or CP as described in D. Scale bar, 10 µm. (H) Effect of BK-1361 or SH3 domain blockade (SH3-Inh) on fMLP-induced transendothelial migration of hPMNs (n = 4 donors). (I) Migration across a 3D gel matrix, respectively (n = 3 donors). (H and I) Data are mean ± SD, **, P < 0.01; 2-way ANOVA followed by Holm-Šídák multiple-comparison test. (J) 3D chemotactic migration of hPMNs preincubated with BK-1361 or CP toward an IL-8 (1 µg/mL) gradient using chemotaxis µ-Slides. Representative trajectory plots. Triangles indicate orientation of the gradient. Violin plots of (K) migration velocity and (L) mean Euclidian and (M) accumulated distance; BK-1361 (white), SH3 inhibitor (gray), control (black). (K–M) Three independent experiments, 30 cells per experiment were analyzed. *, P < 0.05; **, P < 0.01; ***, P < 0.001; 2-way ANOVA followed by Holm-Šídák multiple-comparison test. fMLP, N-formyl-methionyl-leucyl-phenylalanine.

Fig 3: Genetic deletion and pharmacological inhibition of Adam8 protect mice against severe P. aeruginosa infection.(A) Adam8–/– (white bars) or littermate controls (Adam8+/+, black bars) were intranasally instilled with P. aeruginosa (strain PA103) or vehicle control (PBS), then sacrificed 12 hours after infection. (B) Mouse Clinical Score for Sepsis (M-CASS) and (C) temperature scoring were determined in Adam8–/– (white) and Adam8+/+ (black) mice for P. aeruginosa–infected and control animals. (D) BAL neutrophils and (E) MPO levels. (F) WBCs and neutrophils in the peripheral blood of Adam8+/+ and Adam8–/– mice infected with P. aeruginosa (blue) or instilled with PBS (gray). (G) Representative histologic tissue sections of Adam8+/+ and Adam8–/– mice after 12 hours’ P. aeruginosa challenge. Arrowheads, neutrophils in the airspaces; scale bar, 50 µm. (See enlarged sections in Supplemental Figure 12.) (H) BAL TNF-a and (I) total protein. CFU in (J) BAL, (K) blood, and (L) spleen, respectively. (A and B) Box-and-whisker plots; (C–L) mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001, 2-way ANOVA followed by Holm-Šídák multiple-comparison test. (M) C57BL/6J mice were inoculated with P. aeruginosa and treated with either cyclic ADAM8 inhibitor peptide (BK-1361, dotted white bars) or control peptide (CP, dotted black bars) at 2 and 12 hours after infection, and sacrificed at 24 hours. (N) M-CASS and (O) temperature scoring. (P) BAL neutrophils and (Q) total protein. (R) Representative histologic tissue sections of mice treated with CP or BK-1361 during P. aeruginosa challenge. Arrowheads, neutrophils in the airspaces; scale bar, 50 µm. (See enlarged sections in Supplemental Figure 12.) CFU in (S) BAL, (T) blood, and (U) spleen. (N and O) Box-and-whisker plots; (Q–U) mean ± SD; individual data points for each animal are plotted as gray dots. * P < 0.05; ** P < 0.01; Student’s t test.

Fig 4: ADAM8 protease is expressed in human leukocytes and detected in lung fluids of patients with ARDS.(A) Workflow for protease profiling of PMNs at baseline and during activation with TNF-a (20 ng/mL) using 7 FRET-based peptide substrates (PEPDab 005, 008, 010, 011, 013, 014, and 052). To deduce a profile of specific MMP and ADAM activities, peptide cleavage patterns from 3 independent donors were calculated using a nonlinear kinetic model; proteolytic signatures were compared to a reference matrix of catalytic efficiencies to deconvolute protease identities. (B) Proteolytic profiling infers increased activity of known neutrophil MMPs and significant levels of active ADAM8; *, P < 0.05; ****, P < 0.0001; 2-way ANOVA followed by Holm-Šídák multiple-comparison test. (C) Representative images of ADAM8-positive leukocytes (arrowheads, brown color) in lung tissue of patients with ARDS and healthy controls. Scale bar, 50 µm. (See enlarged sections in Supplemental Figure 12.) (D) Representative time-lapse fluorimetry of healthy control BAL (gray) and BAL from patients with ARDS from pneumonia (black; ARDS survivor, red; ARDS nonsurvivor) using the most ADAM8-specific FRET reporter, PEPDab013 (mean ± SD of 3 technical replicates); right, box-and-whisker plots of inferred ADAM8 activity in BAL samples of control patients (gray, n = 4), ARDS survivors (black, n = 5), and ARDS nonsurvivors (red, n = 6). *, P < 0.05; ***, P < 0.001; 1-way ANOVA followed by Tukey’s multiple-comparison test. (E) Representative time-lapse fluorimetry of BAL of patients with mild (cyan) or severe PGD (blue) using PEPDab 013 (mean ± SD of 3 technical replicates); right, box-and-whisker plots of inferred ADAM8 activity in BAL samples of mild PGD (cyan, n = 16) and severe PGD (blue, n = 16). **, P < 0.005, Student’s t test with Welch’s correction. nd, not detectable.

Fig 5: Genetic deletion and pharmacological inhibition of Adam8 impair neutrophil motility in vivo.(A) Schema for cremaster IVM after i.sc. TNF-a. (B) Number of cells adherent to endothelium, (C) rolling velocities, and (D) quantification of transmigrated Adam8-/- and Adam8+/+ neutrophils (right). Representative fields showing leukocyte extravasation (left); arrow indicates direction of emigration. Scale bar, 25 µm. Data for B–D are shown as mean ± SD. **, P < 0.005; Student’s t test. (E) Schema for ADAM8 inhibitor experiments after i.t. LPS. (F) BAL neutrophils and cytospins (left); scale bar, 25 µm. (G) BAL MPO ELISA. (H) Quantification of lung tissue and peripheral blood neutrophils. Data for F–H are mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; 2-way ANOVA followed by Holm-Šídák multiple-comparison test. (I) Immunostaining of neutrophils (S100A8, green) in 100 µm lung sections. *, large airways. Scale bar, 2 mm. (J) Quantification and (K) representative images of intravascular (red) and total (green) neutrophils in lung sections at baseline and after i.t. LPS + BK1361 or CP. Ly6G antibody was injected i.v. 10 minutes prior to euthanasia to label intravascular neutrophils; total lung neutrophils were visualized by S100A8 staining (green) and nuclei by DAPI (blue). Neutrophils were quantified from 7 fields/animal (n = 5). Data in J are mean ± SD, 2-way ANOVA followed by Holm-Šídák multiple comparison test; LPS+CP vs. LPS+BK1361 (P = 0.0025); LPS+CP vs. PBS (P = 0.008); LPS+BK1361 vs. PBS (NS). Scale bar, 100 µm (K). (L) Two-photon lung IVM of MRP8-Cre mTmG mice + BK1361 or CP 3 hours after LPS challenge. Representative images of neutrophil influx (MRP8+, green). Arrowheads, neutrophil cluster. Scale bar, 50 µm. (M) Quantification of neutrophils circulating (track duration < 60 s, left graph) and migrating (track duration > 60 s, right graph) through the lung, n = 3. Blue, baseline; gray, LPS+BK1361; black, LPS+CP. **, P < 0.01; 2-way ANOVA. i.sc., intrascrotal injection.