Fig 2: Platelet-specific C5a receptor deletion inhibits revascularization but does not alter platelet deposition in vivo.Pf4-cre+ C5ar1fl/fl mice were generated as described in the “Methods” section. Hindlimb ischemia was induced, and Pf4-cre- C5ar1fl/fl mice served as controls. a Platelet-specific C5a receptor 1 knockout mice showed increased revascularization. Data are presented as the mean ± SEM (n = 5 animals per group) and are displayed as the percentage of the perfusion in the contralateral control limb. *p < 0.05. b Representative LDI images of mouse hindlimbs after femoral artery ligation illustrate increased revascularization in Pf4-cre+ C5ar1fl/fl mice compared with littermate control animals. c Vessel density in the representative gastrocnemius muscle sections (vessel marker IB4 in green, nuclei in blue) of the ischemic limbs of Pf4-cre+ C5ar1fl/fl showed significantly higher vessel density than Pf4-cre- C5ar1fl/fl controls. Scale bars represent 200 µm. d Quantification of the IB4-positive area fraction in the muscle sections reveals higher vessel abundance in ischemic Pf4-cre+ C5ar1fl/fl hindlimbs at ×200 magnification. Data are shown as the mean ± SEM (n = 10 whole-muscle sections per group) and are displayed as the percentage of control. The IB4-positive area fraction in Pf4-cre- C5ar1fl/fl hindlimbs represents 100%. *p < 0.05. e By staining for CD42b (red), platelet deposition was quantified in whole-muscle sections of Pf4-cre+ C5ar1fl/fl or of Pf4-cre- C5ar1fl/fl mice, as described in the “Methods” section. No difference in the number of platelets per area or microthrombi size was detected. Displayed are representative images for both genotypes, IB4 staining (green) depicts vascular structures; DAPI (blue) depicts nuclei. Scale bars represent 100 µm. f CD42b-positive single platelets were quantified by size using automated digital image analysis as described in the “Methods” part in whole-muscle sections. There was no significant difference between Pf4-cre+ C5ar1fl/fl and Pf4-cre- C5ar1fl/fl mice both with respect of the number of platelets per muscle area as well as the size of the platelets/microthrombi. Data are shown as the mean ± SEM (n = 6 whole-muscle sections per group) and are displayed as the percentage of control. The readings in Pf4-cre- C5ar1fl/fl hindlimbs represent 100% in both graphs. *p < 0.05. Two-way ANOVA with Bonferroni’s post hoc test in a. Two-sided Student’s t test in d, f.

Fig 3: Platelet C5aR1 mediates release of the paracrine effector CXCL4 in an Akt- and PKC-dependent fashion.a Washed murine WT platelets carefully isolated with inhibitors of platelet activation were stimulated with C5a. The supernatant was analyzed by a membrane-based antibody array. Specific mediators, such as CXCL4 (red circles), were upregulated after stimulation with C5a (bottom) compared with the vehicle control (top). b Results of four array repeats were quantified. For most factors, C5a stimulation (gray) induced a slight upregulation compared to vehicle control (black). The strongest increase was observed for CXCL4. Data are shown as the mean ± SEM (n = 4 independent experiments) and are displayed as intensity values. *p < 0.05. Bonferroni’s post hoc analysis was performed; results are displayed in Supplementary Fig. 9. c Single-cell staining revealed that murine platelets are activated upon C5a stimulation and release CXCL4. Upon stimulation with ADP and C5a, a ring-like staining pattern was observed, indicating platelet activation (arrows). However, C5a-stimulated platelets exhibit reduced CXCL4 content after stimulation with C5a, which cannot be observed to the same extent in the ADP-stimulated group. ×630 magnification, scale bars represent 5 µm. Images are representative of four independent experiments. d Quantification of P-selectin intensity in each platelet reveals upregulation, indicating platelet activation upon stimulation with ADP and C5a. However, no significant difference was detected between ADP and C5a. Data are shown as the mean ± SEM (n = 5 images analyzed) and are displayed as the percentage of control. The P-selectin intensity in vehicle control-stimulated WT platelets represents 100%. *p < 0.05 e Quantification of CXCL4 intensity in relation to P-selectin intensity in each platelet reveals relatively stronger secretion of CXCL4 upon C5a stimulation compared to ADP. Data are shown as the mean ± SEM (n = 5 images analyzed) and are displayed as the percentage of control. The CXCL4 intensity in relation to P-selectin intensity in vehicle control-stimulated WT platelets represents 100%. *p < 0.05. f Single platelets from mice were stimulated with C5a or vehicle control. Granules were quantified by calculating the area of predominantly green staining within the platelets (P-selectin-predominant), predominantly red areas (CXCL4-predominant) as well as overlayed yellow areas (P-selectin-CXCL4 double positive). ×630 magnification, scale bars represent 1 µm. Images are representative of >100 analyzed single platelets. g As described in Supplementary Fig. 10, granule area fractions were quantified. We observed a distinct distribution pattern of granules, with a similar amount of CXCL4-predominant as well as P-selectin-CXCL4 double-positive granules and a significantly lower number of P-selectin-predominant granules, which contain P-selectin over CXCL4. Data are shown as the mean ± SEM (n = 112 single platelets analyzed) and are displayed as the area fraction. p (CXCL4 versus double positive) < 0.001, p (CD62P versus double positive) = 0.0022. h Upon stimulation with C5a, analysis revealed a reduction in CXCL4-predominant granules, while the area fraction of double-positive granules remains unchanged, while the relative area fraction of P-selectin predominant granules increases, thus demonstrating C5a-induced secretion of a subset of a-granules, which contain CXCL4 over P-selectin. Data are shown as the mean ± SEM (n = 166 single platelets analyzed) and are displayed as the area fraction. p (CD62P versus CXCL4) < 0.001. i Conventional ELISA confirmed a significant dose-dependent increase in CXCL4 secretion from human platelets after C5a stimulation. Maximum CXCL4 release is reached at a C5a concentration of 2000 ng/ml. Data are shown as the mean ± SEM (n = 8 independent experiments containing separate donors) and are displayed as the percentage of control. The CXCL4 protein level of vehicle-stimulated platelet supernatant represents 100%. *p < 0.05. j Murine megakaryocytes from WT versus C5ar1-/- mice were assessed for distribution of CXCL4-predominant granules (red). Displayed are representative images, nuclei are shown in green. ×630 magnification, scale bars represent 10 µm. k As described in Supplementary Fig. 10, granule area fractions were quantified in megakaryocytes from WT versus C5ar1-/- mice. No significant difference could be detected for CXCL4-predominant granules. Data are shown as the mean ± SEM (n = 10–12 single megakaryocytes analyzed) and are displayed as the area fraction. l RT-PCR analysis detected similar mRNA levels of CXCL4 in megakaryocytes from WT versus C5ar1-/- mice. Data are presented as the mean ± SEM (n = 5 independent experiments) and are shown as relative expression in relation to GAPDH. m Citrated whole blood from WT and C5ar1-/- mice was stimulated using different concentrations of C5a (20, 200, 2000 ng/ml) and vehicle control and assessed for platelet activation markers by flow cytometry. For the gating strategy, please refer to the “Methods” section. Activated platelets are detected as CD62PJONA++, whereas JONA detects activated GPIIbIIIa. C5a induced platelet activation in WT platelets but not in C5ar1-/- platelets. Data are shown as the mean ± SEM (n = 4 independent experiments) and are displayed as the percentage of control. The percentage gated CD62PJONA++ platelets in the vehicle-stimulated group represents 100%. *p < 0.05. n In order to uncover the signaling downstream of C5aR1 leading to CXCL4 secretion, lysates of WT platelets were generated after vehicle control or C5a stimulation and samples were probed at equal protein concentrations for phospho-proteins as well as non-phosphorylated controls. Platelet C5aR1 ligation induced reproducible PKC substrate phosphorylation (PKC phosphor Ser). As a control, we used the PKC activator PMA. PKC activation was quantified by dividing the phosphorylation signal by the total protein signal as well as unphosphorylated PKC a (Supplementary Fig. 23). Displayed are representative images of at least four independent experiments. o We also checked whether PKC activation also entails CXCL4 release by stimulating washed platelets with C5a and PMA. Both C5a and PMA induced significant CXCL4 release. Data are shown as the mean ± SEM (n = 5–16 independent experiments) and are displayed as the percentage of control of CXCL4 concentration in platelet supernatant measured by ELISA. The mean fluorescence intensity (MFI) of platelets in the vehicle-stimulated group represents 100%. *p < 0.01. p Furthermore, C5a induced an upregulation of phosphorylated Akt (Ser473), PI3K (Ser 47), GSK-3ß (Ser 9), p44/42 MAPK (Thr202/Tyr204), PLCß3 (Ser537). Displayed are representative images of at least four independent experiments. For quantification of phosphorylation, please refer to Supplementary Fig. 23. q Interestingly, C5a induced no regulation of phosphorylated PLC?2 (Tyr1217) and PKA (a/ß/? catalytic subunit phospho T197). Activated Rap1 (Rap1-GTP) was analyzed using a Rap1 pulldown assay. No significantly altered amounts of Rap1-GTP could be detected. Displayed are representative images of at least four independent experiments. One-way ANOVA with Bonferroni’s post hoc test in b, d, e, g–i, m. Two-sided Student’s t test in k, l, o.

Fig 4: Platelets express C5aR1 and platelet C5aR1 inhibits various endothelial functions.a Isolated washed murine platelets express C5aR1 as assessed by immune fluorescence microscopy. There was only partial colocalization of C5aR1 (red) with the a-granule marker P-selectin (green). ×630 magnification, scale bars represent 5 µm. Images are representative of four independent experiments. b Histogram showing C5aR1 expression on platelets (gray curve); the black curves show the histogram obtained with an IgG isotype. The histogram is representative of the analysis of four independent platelet samples. c Furthermore, flow cytometry revealed that C5aR1 expression on platelets is dynamic. Upon stimulation with CRP, platelet C5aR1 expression increased in WT platelets. In c–e, data are displayed as the mean ± SEM (n = 4 independent experiments) and are shown as the percentage of control. The expression in the unstimulated WT group represents 100%. *p < 0.05. d Upon stimulation with ADP, WT platelets displayed C5aR1 upregulation at higher concentrations. e Upon treatment with C5a at the indicated concentrations, no significant changes in C5aR1 expression on platelets could be detected. f Primary lung endothelial cells (MLECs) were grown to a confluent monolayer, which was wounded by scratching with a plastic pipette and coincubated with C5ar1-/- or WT platelets. The absence of C5aR1 resulted in increased endothelial migration after 17 h. Data are displayed as the mean ± SEM (n = 6 independent experiments). The area not populated with cells at the start of the experiment minus the area not populated with cells after 17 h of coincubation with WT platelets represents 100%. *p < 0.05. g Endothelial tube formation by MHEC-5T cells in vitro was significantly increased after incubation with platelets isolated from C5ar1-/- mice compared platelets isolated from WT mice. Data are shown as the mean ± SEM (n = 4 independent experiments) and displayed as the total tube length after 6.5 h. The cumulative length of endothelial tubes after incubation with WT control platelets represents 100%. *p < 0.05. h Similarly, endothelial tube formation by MLECs was increased after exposure to C5ar1-/- platelets compared with WT platelets. Data are shown as the mean ± SEM (n = 4 independent experiments). Tube formation after coincubation with WT control platelets represents 100%. *p < 0.05. i Representative images of endothelial tube formation of MHEC-5T after coincubation with WT or C5ar1-/- platelets. j The supernatant of murine WT platelets significantly inhibited tube formation from endothelial spheroids compared with the supernatant of C5ar1-/- platelets. Data are presented as the mean ± SEM (n = 3 independent experiments) and are shown as the percentage of the total tube length after 24 h in controls. Tube formation after incubation with WT control platelet supernatant represents 100%. *p < 0.05. k Representative images of sprout formation from endothelial spheroids in both groups. Scale bar represents 100 µm. l Human platelets were preincubated with a C5aR1 antagonist (PMX53) or control peptide. Subsequently, platelets were coincubated with human umbilical vein endothelial cells (HUVECs) on Matrigel. Endothelial tube formation was significantly enhanced by preincubation with PMX53 compared with the control. Data are shown as the mean ± SEM (n = 4 independent experiments with separate donors) of the total tube length after 6.5 h. The cumulative length of endothelial tubes after incubation with platelets preincubated with control peptide represents 100%. *p < 0.05. m There was no significant difference in adhesion to endothelial cells under static conditions between WT and C5aR1-deficient platelets under normoxic conditions. Data are shown as the mean ± SEM (n = 4–5 independent experiments) and as the percentage of control. Adherent WT platelets expressed as the area fraction of platelet-specific staining divided by the number of endothelial cells expressed as the DAPI count per area represents 100%. *p < 0.05 compared to control-stimulated platelets. n Representative images of static platelet adhesion of WT and C5ar1-/- platelets to endothelial cells. Scale bars represent 200 µm. Two-sided Student’s t test in b, f–h, j, l. ANOVA in c–e, m.

Fig 5: C5aR1 deficiency promotes ischemia-induced revascularization in vivo.WT or C5ar1-/- mice were subjected to hindlimb ischemia and analyzed after 2 weeks. a Revascularization of the hindlimbs after femoral artery ligation was visualized by laser Doppler fluximetry (LDI). We found increased revascularization in C5ar1-/- animals. Data are shown as the mean ± SEM (n = 7 animals per group) and are displayed as the percentage of the perfusion in the contralateral control limb. *p < 0.05. b Representative LDI images of mouse hindlimbs after femoral artery ligation illustrate increased revascularization in C5ar1-/- animals compared with WT controls. c Vessel density in the gastrocnemius muscle of the ischemic limbs was quantified by immunofluorescence staining. Vessels were visualized by isolectin B4 (IB4 in green, nuclei in blue), and images of whole-muscle sections were acquired as tile scans and analyzed. At 14 days after the induction of ischemia, C5ar1-/- mice exhibited a significantly higher vessel density than WT controls. ×200 magnification, scale bars represent 200 µm. d Quantification of the IB4-positive area fraction in the muscle sections reveals higher vessel abundance in ischemic C5ar1-/- hindlimbs. Data are shown as the mean ± SEM (n = 10 whole-muscle sections per group) and are displayed as the percentage of control. The IB4-positive area fraction in WT control hindlimbs represents 100%. *p < 0.001. e Furthermore, WT or C5ar1-/- mice were subjected to hindlimb ischemia and platelets were depleted systemically by injection of platelet depleting serum starting on the first day post induction of ischemia. We could not detect significant differences (p < 0.05) in revascularization. Data are shown as the mean ± SEM (n = 7 animals per group) and are displayed as the percentage of the perfusion in the contralateral control limb. f Vessel density in the gastrocnemius muscle of the ischemic limbs was quantified by immunofluorescence staining as described in c. At 14 days after the induction of ischemia, WT and C5ar1-/- mice with platelet depletion did not exhibit significant differences in vessel density. ×200 magnification, scale bars represent 200 µm. Image is representative of 10 whole-muscle sections per group analyzed. Two-way ANOVA with Bonferroni’s post hoc test in a, e. Two-sided Student’s t test in d.