Fig 2: Analysis of staining intensity per lesion of CD31- and Collagen IV-stained vessels (A–F). Error bars indicate the mean and standard deviation for each lesion. p-values: Not statistically significant (ns) p value > 0.05; for statistical significance, accepted p value = 0.05. ** p value = 0.01, *** p value = 0.001, and **** p value = 0.0001.

Fig 3: Identification of TSP1 cleavage sites for MT1-MMP AWestern blot of TSP1 expression (developed with an antibody against an epitope nearby the N-terminus; Lee et al, 2006) in lysates from siCtrl and siMT1-silenced HUVEC. MT1-MMP and tubulin are included as silencing and loading controls, respectively. The arrowhead marks full-length TSP1 and the asterisk the N-terminal TSP1 fragment.BWestern blot of in vitro digested TSP1 (developed with the same antibody as in A) incubated with increasing amounts of recombinant human MT1-MMP catalytic domain. rhMT1-MMP catalytic domain is also included. The arrowhead marks full-length TSP1 and the asterisk the N-terminal TSP1 fragment generated by MT1-MMP cleavage.C In silico model of the membrane-anchored MT1-MMP dimer (blue/orange) and TSP1 type 1 repeat domains 2 and 3 (green). Yellow marks the catalytic pocket in the MT1-MMP protease, and red indicates the two selected cleavage sites in TSP1.DScheme depicting the TSP1 domain structure with the binding sequences to CD36, CD47, and avß3 integrin, as well as the positions of the identified cleavage sites for MT1-MMP.EDAF-FM mean fluorescence intensity (MFI) in HUVEC expressing MT1-MMP siRNA and left untreated or treated with 200 ng of full-length TSP1 or the E123CaG-1 fragment; n = 128–184 cells analyzed per condition in two independent experiments. Data are shown as individual cell values and mean ± SEM and were tested by the Kruskal–Wallis test; *P < 0.05, ***P < 0.001.FRepresentative maximum-intensity projection images of CD31 staining (green) in whole-mount colon mucosal plexus of MT1f/f mice injected intraperitoneally with 1 µg full-length TSP1 or the E123CaG-1 fragment at day 0 and day 2 during 1% DSS treatment for 3 days. Scale bar, 40 µm. Arrows, arrowheads, and asterisks indicate duplications, loops, and pillars, respectively; n = 6 and 7 mice per condition in two independent experiments.GQuantification of IA events in mice treated as in (F). Data are shown as mean ± SEM and were tested by t-test; *P < 0.05.Data information: Please see Appendix Table S1 for exact P-values.Source data are available online for this figure.

Fig 4: Loss of NOX2 does not affect the cellular composition of islets. (A) Representative immunofluorescence stainings of insulin/glucagon, insulin/somatostatin and insulin/CD31 in WT and Nox2-/- islets within the pancreas. Cell nuclei were stained with Hoechst 33,342 (blue). Scale bar: 50 µm. (B-E) Quantitative analysis of insulin- (B), glucagon- (C), somatostatin- (D) and CD31-positive cells (E) in WT and Nox2-/- islets within the pancreas in % of all islet cells (n = 15 each). Mean ± SEM. (F) Representative immunofluorescence stainings of insulin/glucagon, insulin/somatostatin and insulin/CD31 in isolated WT and Nox2-/- islets. Cell nuclei were stained with Hoechst 33,342 (blue). Scale bar: 50 µm. (G-J) Quantitative analysis of insulin- (G), glucagon- (H), somatostatin- (I) and CD31-positive cells (J) in isolated WT and Nox2-/- islets in % of all islet cells (n = 20 each). Mean ± SEM. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig 5: Choroidal neovascularization (CNV) detection by CD31 immunofluorescence. Mice that received only the vehicle (upper left, group V) had larger areas of CNV than mice that received the supplement (upper right) and/or anti-vascular endothelial growth factor (VEGF) treatment. Treatment groups included the nutritional supplement (S) group, the supplement and anti-vascular endothelial growth factor agent (S + aVEGF) group, and the vehicle and anti-VEGF agent (V + aVEGF) group.