Vascular surveillance by haptotactic blood platelets in inflammation and infection

Abstract

Breakdown of vascular barriers is a major complication of inflammatory diseases. Anucleate platelets form blood-clots during thrombosis, but also play a crucial role in inflammation. While spatio-temporal dynamics of clot formation are well characterized, the cell-biological mechanisms of platelet recruitment to inflammatory micro-environments remain incompletely understood. Here we identify Arp2/3-dependent lamellipodia formation as a prominent morphological feature of immune-responsive platelets. Platelets use lamellipodia to scan for fibrin(ogen) deposited on the inflamed vasculature and to directionally spread, to polarize and to govern haptotactic migration along gradients of the adhesive ligand. Platelet-specific abrogation of Arp2/3 interferes with haptotactic repositioning of platelets to microlesions, thus impairing vascular sealing and provoking inflammatory microbleeding. During infection, haptotaxis promotes capture of bacteria and prevents hematogenic dissemination, rendering platelets gate-keepers of the inflamed microvasculature. Consequently, these findings identify haptotaxis as a key effector function of immune-responsive platelets.

Fig. 1. Platelets recruited to inflamed blood vessels are migratory and form lamellipodia.

a, b Platelet recruitment and migration in inflamed microvessels. a Representative micrograph and b platelet migratory behavior, exemplary observed patterns, tracks (white) and cell velocities (n = 39 cells in five mice). White arrows: platelets. c, f Platelet phenotype and microenvironment in LPS-induced inflammation and ferric chloride-induced thrombosis in the cremaster microvasculature. c Fibrin(ogen) deposition and d associated platelet phenotype in inflammatory settings. Asterisks: lamellipodia. e Platelet phenotype and f fibrin(ogen) deposition in thrombosis. White arrows: filopodia. c, f Right lower insets: rendered 3D reconstructions. g Directionality of in vivo migrating (n = 39, see b compared to retracting cells (mesenteric thrombus formation, n = 59 cells in three mice), Mann–Whitney. h Analysis of in vivo platelet shape in the cremaster microvasculature, n = 26 (thrombosis), n = 33 (inflammation) in three mice, Mann–Whitney. Scale bars =5 µm. All statistical tests are two-sided. Source data are provided as a Source Data file.

Fig. 2. Mechanical properties of fibrin(ogen) are sufficient to instruct platelet shape and motility.

a–d In vitro reconstitution of in vivo microenvironments. a Substrate clearance and morphology of platelets on monomeric and cross-linked substrate. b Representative micrograph of migrating and retracting cell on fibrinogen and cross-linked fibrin substrate and c mean shape outlines and comparison of the aspect ratio of cells on fibrinogen leading to migration (n = 43 cells) and cells on cross-linked fibrin leading to retraction (n = 47 cells, Mann–Whitney). d p-MLC and F-actin distribution in migrating and retracting human platelets. (Left) Exemplary cells and (right) FMI profiles along short axes in n = 50 (migrating) and n = 47 (retracting) cells from three experiments. e, f Plasmin treatment of cross-linked substrate after human platelet addition. e (Left) Experimental setup and cleavage of fibers after plasmin treatment indicated by loss of fluorescence. (Center) Representative micrographs. (Right) Color-coded (time) representative cell outlines. f Migrating cell percentage after plasmin/control treatment (n = 5 experiments, Mann–Whitney), and aspect ratio prior and after treatment (n = 46 cells from four experiments, Wilcoxon’s signed-rank test). g Integrin ligands of tunable mechanical stability. Left: stable: RGD (Arg-Gly-Asp)-peptides are covalently bound (stable) to a PLL (poly-l-lysine)-PEG (polyethylene glycol) backbone immobilized on a glass coverslip, fragile: RGD-biotin is bound to a PLL-PEG-biotin backbone via a neutravidin-FITC (NA) bridge, middle: exemplary micrograph showing that platelet pulling forces rupture NA–biotin bonds as indicated by the loss of fluorescence (FITC) and right: percentage of spreading and migrating platelets, n = 3 independent experiments, t test. Scale bars = 5 µm. Error bars = s.e.m. All statistical tests are two-sided. Source data are provided as a Source Data file.

Fig. 3. Arp2/3-dependent lamellipodia formation regulates platelet haptotaxis.

a–c Early shape changes of human platelets on fibrinogen surfaces. a Micrograph of spreading cell. Arrows: filopodia; asterisk: lamellipodium. b Evolution of lamellipodia/filopodia over time, and c direction of filopodia and lamellipodia in relation to the subsequent direction of migration. b, c n = 32 cells from three experiments. Lines: mean; error bands: s.d. d, e Platelet decision-making upon encountering substrate gradients. d (left) Exemplary line profile of fibrinogen gradients encountered by cells and (right) representative micrographs of the substrate (green: fibrinogen-AF488) and platelet tracks (white), 30 min after cell seeding. Scale bar = 5 µm. e Tracks of cells (n = 256 cells) in contact with interface, and (right) platelet positioning before (0 min) and after migration of cells getting in contact with the interface (45 min), n = 4 experiments, paired t test. Scale bar = 10 µm. f Arp2/3 and F-actin distribution in migrating and retracting cells, and FMI profiles along short axes in n = 50 (migrating) and n = 53 (retracting) cells pooled from three experiments. Scale bar = 5 µm. g Effect of Arp2/3 inhibition with CK666 compared to control CK689 on the number of filopodia per cell (Mann–Whitney), solidity (t test), lamellipodia formation, and F-actin distribution (compare Fig. 1d). n = 41 cells from three experiments. Scale bar = 5 µm. h, i Decision-making upon encountering substrate gradients, control (CK689) and Arp2/3 inhibited (CK666) platelets (h) representative micrograph. Scale bar = 5 µm and i direction of lamellipodia and filopodia in relation to the fibrinogen gradient, n = 168 filopodia and n = 51 lamellipodia, χ² test. Error bars = s.e.m. Scale bars = 5 µm. All statistical tests are two-sided. Source data are provided as a Source Data file.

Fig. 4. Platelet spreading and haptotactic migration preserve vascular integrity in inflammation.

a, b Analysis of Arpc2+/+ and Arpc2−/− platelet morphology in the inflamed microvasculature. a Cremaster whole mount of adherent platelets. Asterisk: lamellipodium; arrows: filopodia. b Filopodia per cell and solidity, Arpc2+/+: n = 28 cells, Arpc2−/−: n = 30 cells in three mice per group, t tests. Scale bar = 5 µm. c Motility patterns of transfused Arpc2+/+ and Arpc2−/− platelets. Quantification of patterns (n = 21 vessels from three mice), Mann–Whitney. d Distribution and MFI profile of vessel-deposited fibrinogen in relation to VE-Cadherin+ endothelial junctions. Scale bar = 5 µm. e Localization of Arpc2+/+ and Arpc2−/− platelets in relation to VE-Cadherin+ endothelial junctions. (Left) Exemplary whole mounts, arrows indicate platelets adherent to junctions, asterisks highlight platelets distant from junctions and (right) platelet distance from junctions, violin plot, Arpc2+/+: n = 135, Arpc2−/−: n = 144 cells from three mice per group, t test. Scale bar = 10 µm. f Platelet recruitment to α-SMA low regions in Arpc2+/+ and Arpc2−/− mice. (Left) Whole mount and (right) percentage of α-SMA low areas positive for platelets, Arpc2+/+: n = 16, Arpc2−/−: n = 16 vessels from three mice per group, t test. Scale bar = 5 µm. White arrows: platelets. g Bleeding assessment in the inflamed cremaster microvasculature of Arpc2+/+ and Arpc2−/− mice, representative whole mounts stained for erythrocytes (TER119) and vasculature (CD31) and quantification of extravascular TER119 signal, Arpc2+/+: n = 8, Arpc2−/−: n = 5 mice per group, t test. Scale bar = 50 µm. Error bars = s.e.m. All statistical tests are two-sided. Source data are provided as a Source Data file.

Fig. 5. Vascular surveillance by platelets prevents pulmonary bleeding.

a, b LPS-induced (sub-)acute lung injury model. a Representative micrograph of fibrinogen deposition (asterisks) and platelet morphology in the inflamed (LPS) and control lungs (NaCl) of WT animals. Asterisk: lamellipodium. Scale bar = 5 µm. b Exemplary micrograph and quantification of fibrin(ogen) deposition at sites of vascular damage (arrows) compared to control mice indicated by single extravasated erythrocytes and intact vasculature, n = 3 WT mice, paired t test. Scale bar = 5 µm. c, f LPS-induced lung injury in busulfan-treated Arpc2+/+ and Arpc2−/− mice. c Appearance, d hemoglobin content in bronchoalveolar lavage (BAL), and e lung histology 8 h after LPS challenge. Scale bar = 10 mm. f Lung HE stain. Scale bar = 30 µm. d Mann–Whitney. g, h Hyperacute lung injury model. g SpO₂ 90 min after high-dose tracheal LPS application. Center: mean; whiskers: min–max; box: 25–75th percentiles. h Kaplan–Meier survival curve, n = 8 mice per group. g t Test and h log rank. Error bars = s.e.m. All statistical tests are two-sided. Source data are provided as a Source data file.

Fig. 6. WASp and WAVE play a redundant role in platelet haptotaxis.

a Morphology and migration of WASp−/− platelets compared to controls, n = 3 WASp+/+ and n = 4 WASp−/− mice, t test. Scale bar = 5 µm. b Number of filopodia per cell (Mann–Whitney), solidity (t test), and representative F-actin-stained platelet of Cyfip1−/− (n = 32 cells) and Cyfip1+/+ (n = 30 cells) mice, three mice per group. c, d Representative cell body/filopodia outline and c total cell area of Arpc2−/− (n = 31 cells) and Cyfip1−/− mice (n = 31 cells), and d size of cell body, three mice per group, t test. e p-MLC and F-actin distribution of Arpc2−/− (n = 39 cells) and Cyfip1−/− mice (n = 39 cells), three mice per group. Insets: F-actin distribution between filopodia. Scale bar = 5 µm. e–g Migration of WT, Cyfip1−/−, and Arpc2−/− platelets, e representative micrographs, f percentage of cells migrating and g area cleared per cell, ANOVA and Tukey’s post hoc test, n = 3 mice per group. Scale bar = 5 µm. h Effect of additional Arp2/3 inhibition on Cyfip1−/− platelets. Representative micrographs, and analysis of percentage of cells migrating and area cleared per cell of CK689- and CK666-treated Cyfip1−/− platelets, n = 3 mice per group, paired t test. i Decision-making upon encountering substrate gradients, (left) examples of Cyfip1−/− platelets in contact with interface, and (right) platelet positioning before (0 min) and after migration (45 min), n = 18 cells from three mice. j LPS-induced lung injury in Cyfip1+/+ and Cyfip1−/− mice, quantification of hemoglobin content, erythrocytes, and leukocytes in bronchoalveolar lavage (BAL), n = 3 mice, t test. k LPS-induced lung injury in busulfan-treated, partially platelet-depleted Cyfip1+/+ and Cyfip1−/− mice. Quantification of hemoglobin content, erythrocytes, and leukocytes in bronchoalveolar lavage (BAL), n = 3 mice. Scale bars = 5 µm, t test. Error bars = s.e.m. All statistical tests are two-sided. Source data are provided as a Source data file.

Fig. 7. Migrating platelets protect from bacterial dissemination in MRSA pneumonia.

a Platelet recruitment and clearance in sterile and MRSA biofilms in vitro, n = 5 experiments, Mann–Whitney. Scale bar = 10 µm. b In vitro effect of Arp2/3 inhibition with CK666 compared to control CK689 in platelet-mediated MRSA biofilm clearance. (Left) Percentage of MRSA associated with platelets, n = 7 experiments, t test, and (right) 3D reconstruction of platelets interacting with MRSA. Scale bar = 5 µm. c Scanning electron microscopy of platelet surface-bound MRSA. Inset: bundled bacteria. Scale bar = 1 µm. d–i MRSA lung infection model of Arpc2+/+ and −/− mice, (d) (left) mice were intranasally inoculated, and (right) exemplary micrograph of platelet–MRSA interaction in the lung of Arpc2+/+ mice. Asterisks: lamellipodium; arrows: bundled bacteria. e Micrographs of platelet–MRSA localization and f percentage of MRSA associated with platelets in non-abscessing infected lung tissue (Arpc2+/+: n = 7, Arpc2−/−: n = 9 mice per group). g Percentage of mice with bacteremia was assessed 24 h (n = 16) and 48 h p.i. (Arpc2+/+: n = 11, Arpc2−/−: n = 12 mice per group). h MRSA CFUs recovered from blood at 48 h p.i. (i), percentage of mice with MRSA in the kidney and spleen (Arpc2+/+: n = 15, Arpc2−/−: n = 16 mice). d, h Mann–Whitney and g, i χ². Scale bars = 5 µm. Error bars = s.e.m. All statistical tests are two-sided. Source data are provided as a Source data file.