Supplementary Data Supplementary Methods: Release kinetics of from collagen sponges in vivo. 2μg of human recombinant 165 was labeled with lexafluor 555 (Microscale Protein Labeling Kit; Invitrogen) as described in manufacturer s instructions..5μg fluorescence labelled, loaded in collagen sponges, was inserted into the injury region when the cortical defect was generated. Tibial segments including the defect hole region were carefully isolated at various time points after surgery, and observed under confocal microscope. z-series file up to 2μm deep from the surface was acquired in area covering the hole region. The maximum intensity projection from the z-series images was chosen for the analysis of fluorescence density using Image J. The fluorescence intensities of samples at different time points represent the relative amount of remaining in the sponge. Isolation of murine articular chondrocytes. The primary joint chondrocytes were isolated from femoral head, femoral condyles and tibial plateau of 5-days old mice according to a protocol described previously (1). The primary chondrocytes were cultured in 1% DMEM with 1% FS. CCK-8 cell proliferation assay. 4, MSCs were seeded in each well of 96-well plate. The next day, cells were given αmem medium (1% or 2% FS) containing vehicle, 1-4ng/ml recombinant (R&D) and/or 2-4ng/ml recombinant PDGF- (R&D). fter two days of additional culture, 1µl cell counting kit 8 (CCK-8, Sigma-ldrich) was added to each well and incubated for 3 hours. bsorbance at 45 nm, representing relative cell numbers, was measured using a microplate reader. For chondrocyte proliferation, 2 primary chondrocytes were seeded in each well of 96-well plate. The next day, cells were given the complete culture medium with 1-4ng/ml recombinant (R&D). Cell numbers were monitored at day1, 3, 5, and 7 after seeding with medium change every other day. Complete blood analysis. Mice were placed under general anesthesia, and peripheral blood was collected from the heart. bout 2 µl of blood samples were analyzed by Hemavet 95 Chemistry nalyzer (Drew Scientific Inc.). Supplementary Reference: 1. Gosset M, erenbaum F, Thirion S, and Jacques C. Primary culture and phenotyping of murine chondrocytes. Nat Protoc. 28;3(8):1253-6.
Figure S1 Supplementary Figures PSD PSD1 PSD7 H&E PSD1 PSD14 PSD28 H&E PSD7 SO SO Figure S1. Repair of cortical defects recapitulates most steps in fracture healing. () Representative histological images of injury tibiae at various time points after surgery. Scale bar: 1mm. () Safranin O (SO)-stained sections from wild type mice shows absence of SO-positive cartilage in hole region and presence of cartilage in injured periosteum at PSD7. Scale bar: 1mm (upper image), 2µm (lower image). Representative image from 3 mice at each time point.
Figure S2 VE-cadherin-cre;tdTomato (Injury hole) VE-cadherin-cre;tdTomato (Growth plate) tdtomato C Vegfa fl/fl Vegfa fl/+ ;VE-cadherin-cre Vegfa fl/fl ;VE-cadherin-cre 3D Hole region V (mm 3 ).15±.42.13±.39.173±.2 V/TV (%) 1.74±2.47 11.5±2.99 15.33±1.51 T.N (/mm) 9.58±1.54 8.47±1.7 11.35±1.42 T.TH (/mm).21±.31.26±.2.27±.5 T.SP (/mm).98±.13.14±.38.95±.12 D 3D Structural parameters of newly formed bone in marrow space at PSD7 by µct Vegfa fl/fl Vegfa fl/+ ;VE-cadherin-cre Vegfa fl/fl ;VE-cadherin-cre V (mm 3 ).57±.13.79±.14.89±.13 V/TV (%) 13.82±2.74 17.6±3.37 21.5±2.75 T.N (/mm) 7.16±1.65 9.37±2.7 1.7±2.67 T.TH (mm).29±.2.3±.6.3±.1 T.SP (mm).23±.38.15±.49.13±.32 Figure S2. Deletion of Vegfa in endothelial cells does not affect intramembranous bone formation during bone repair. () Small amounts of anti- staining in tdtomato-positive endothelial cells in defects of VE-cadherin;tdTomato mice at PSD7. () Some tdtomato-positive endothelial cells in metaphyseal vessels express. GP: growth plate. MET: metaphysis. (C) 3D-reconstruction of mineralized bone formed in hole region of mice at PSD7 and 3D structural parameters by µct analysis. (D) 3D µct-based structural parameters of mineralized bone formed in wounded bone marrow space of mice at PSD7. Representative image from 3 mice (, ). N=5-6 for each genotype (C, D). Scale bars: 5µm (), 1µm (, C). NOV with Tukey s post-hoc test was used.
Figure S3 C M C C V/TV (%) 2 15 1 5 ** * * Tb. N (/mm) 1 8 6 4 2 C Tb. Sp (mm).4.3.2.1 D Vegfa fl/+ ;Osx-cre Movat pentachrome Figure S3. Delayed intramembranous bone formation in defects of mice at PSD7. (-C) 3D µct-based structural parameters of mineralized bone formed in wounded marrow (yellow area). C: cortical bone, N=6. (D) Movat pentachrome staining shows histological details in hole region. Dense collagen fibers stained yellow and muscle and loose fibrous tissues stained red. Representative image from 3 mice for each genotype. Scale bar: 2µm. NOV with Tukey s post-hoc test was used *, P<.1; **, P<.5.
Figure S4 PS 4 ng/ml MSCs from Vegfa fl/fl mice Von kossa PS 4 ng/ml MSCs from Von kossa PS 1 ng/ml 4 ng/ml MC3T3-1E cells Von kossa Figure S4. dministration of recombinant does not affect the mineralization of MSCs and osteoblasts. () Von Kossa staining of MSCs from Vegfa fl/fl or Vegfa fl/fl ;Osxcre mice after 21 days of culture in mineralization medium (1nM dexamethasone) with PS or 4ng/ml. Scale bar: 1µm. ()Von Kossa staining of osteoblastic MC3T3-1E cells after 21 days of culture in mineralization medium with PS or recombinant. The data are representative of three independent experiments.
Figure S5 Density of ZsG-positive cells in hole region (/mm 2 ) 15 1 5 1 2 3 Density of anti-sp staining in hole region (%) 25 2 15 1 5 R 2 =.711 P =.4 5 1 15 2 25 Vessel area/total area in hole region (%) C Density of anti-ocn staining in hole region (%) 1.6 1.2.8.4 R 2 =.699 P =.13 1 2 3 Vessel area/total area in hole region (%) D MSCs Vegfa fl/fl E Osx-cre;ZsG ;ZsG FSP1 D TRP FSP1 D21 Density of osteoclasts in 11±78 72±42 hole region/mm 2 F Osx-cre;ZsG ;ZsG G Osx-cre;ZsG ;ZsG rdu ZsGreen TUNEL ZsGreen Density of proliferative cells in 133 ±26 116±14 hole region /mm 2 Density of dead cells in hole region /mm2 69±38 66±3 Figure S5. Reduced angiogenesis and osteoblast differentiation in ;ZsG mice at PSD7. () Similar density of ZsG-positive cells in defect hole region in different genotypes, N=5-6. (, C) Strong correlation between percentage of blood vessel areas and density of SP () orocn(c) stained areas in defect hole region. (D) Increased numbers of FSP+ fibroblasts in MSCs from mice after 21 days of culture in mineralization medium with 1nM dexamethasone. Representative images from three independent experiments. (E) No significant difference in density of TRP-positive osteoclasts in hole region of Osx-cre;ZsG and ;ZsG mice, N=3-4. (F, G) Similar density of rdu labeled cells (F) and TUNEL-positive cells (G) inholeregionofosx-cre;zsg and ;ZsG mice, N=5-6. Yellow stippled rectangles: hole region. C: cortical bone. Scale bars: 5µm (G), 1µm (D, F), 2µm (E). NOV with Tukey s post-hoc test (), spearman s correlation coefficient test (, C) and unpaired 2-tailed Student s t-test (E-G) were used.
Figure S6 SPG+ unlabeled SPG+ lexafluor 555 labeled PSD PSD PSD1 PSD3 PSD7 PSD1 Relative levels compared to PSD (%) 1 21.7 8.7 1.3.7 C PS 1µg PS 1µg 3D Hole region niline lue V/TV (%) D 4.46±1.27.4±.14** PS 1µg Density of niline blue staining (%) Mineralization to collagen ratio (%) E 26.7±4.4 3.9±1.9* 25.6±1. 14.5±5.5 PS 1µg SP ZsG CD31 ZsG Density of anti-sp staining (%) Density of anti-sp staining per 1 ZsG+ cells (%) F Cell density in hole region (/mm 2 ) 25 2 15 1 5 6.66±1.27.81±.63* 8.35±1.9 2.3±.95** PS 1µg * ** ZsG-positive cells Total cells Vessel density in hole region /mm 2 G TRP PS 163±12 11±18** 1µg Osteoclast density in 43±6 54±8 hole region/mm 2 C
Figure S6. Treatment with 1µg in collagens sponges inhibits mineralized bone formation in cortical defects of Osx-cre;ZsG mice at PSD7. () Confocal image of the hole region containing absorbable collagen sponge (SPG) with lexafluor 555-labelled at various time points after surgery. The relative amount of remaining in SPG, based on fluorescence density, was calculated. Data represent two independent experiments. () ased on µct, less mineralized bone formed in hole region in Osx-cre;ZsG mice treated with 1µg (V/TV.4 ±.14 %) compared with PS (4.46 ± 1.27%). (C) 1µg reduces density of aniline blue staining in hole region of Osx-cre;ZsG mice compared with PS, but does not alter mineralization to collagen ratio. (D) 1µg reduces density of anti-sp staining with or without normalization to total number of ZsGpositive cells in hole region of Osx-cre;ZsG mice. (E) Low density (11 ± 18 /mm 2 ) of blood vessels (determined by anti-cd31 staining) in hole region of Osx-cre;ZsG mice treated with 1µg compared with PS treatment (163 ± 12/mm 2 ). (F) Reduced density of ZsG-positive cells and total cells in hole region of Osx-cre;ZsG mice treated with compared with PS treatment. (G) No difference in low density of TRP-positive osteoclasts in hole region of mice treated with or PS. Yellow stippled rectangles: hole regions (C, G). C: cortical bone. Scale bars: 1 µm (, ), 2µm (C, G), 5 µm (D, E), N=4-5 (-G). Unpaired 2-tailed Student s t-test was used.*, P<.1; **, P<.5.
Figure S7 Osx-cre;ZsG PS ;ZsG PS.1µg 1µg CD31 Vessel density in hole region /mm 2 182±15 7±8 123±13**, # 84±13 Osx-cre;ZsG PS ;ZsG PS.1µg 1µg TRP Density of osteoclast s/mm 2 25 2 15 1 5 Hole region adjacent marrow * # * # * # *## WT+PS CKO+PS CKO+.1µg CKO+1µg Figure S7. lood vessel formation and osteoclast recruitment in cortical defects with delivery of PS or recombinant at PSD1. () Enhanced density (123 ± 13/mm 2 ) of blood vessels (based on anti-cd31 staining) in defect hole region of ;ZsG mice treated with.1µg compared with PS (7 ± 8/mm 2 )or1µg(84±13/mm 2 ). N=5-6 mice for each treatment group. () Density of TRP-positive osteoclasts in hole region and adjacent area of injured M. Scale bars: 1µm (), 2µm (), N=5-6. Yellow stippled rectangles: hole region (, ); blue stippled rectangles: adjacent bone marrow regions (). C: cortical bone. WT: Osxcre;ZsG mice. CKO: mice. Unpaired 2-tailed Student s t-test was used. *, #, P<.1; **, ##, P<.5.*vs. Osx-cre;ZsG+PS. # vs. ;ZsG+PS.
Figure S8 F4/8 NIMP-R14 FSP1 Vegfa fl/fl PSD3 Vegfa fl/fl CD45R Density of CD45-positive 26±8 47±9 lymphocytes in hole region /mm 2 C Vegfa fl/fl Osterix Density of Osterix-positive cells in hole region /mm 2 Density of nuclei in hole region /mm 2 11±16 53±5** 1522±144 653±79* D Vegfa fl/fl CD31+ Endothelial cells F4/8+ Macrophages Osx+ Osteoblast precursors CD45+ Lymphocytes Other cell types Figure S8. promotes cell infiltration at PSD3. () Co-staining with antibodies against, macrophage marker F4/8, neutrophil marker NIMP-R14 or fibroblast marker FSP1 in injury sites of Vegfa fl/fl mice at PSD3. Representative images from three independent experiments. () Similar density of CD45R-positive lymphocytes in hole region of and Vegfa fl/fl mice, N=4-6. (C) Low density of osterix-positive osteoblast precursor cells (53 ± 5 /mm 2 ) and total cells (653 ± 79/mm 2 ), the latter determined by counting total numbers of nuclei, in hole region of compared with Vegfa fl/fl mice (11 ± 16/mm 2 and 1522 ± 21/mm 2 ), N=5-6. (D) Percentage of different cell types in hole region of and Vegfa fl/fl mice. Yellow stippled rectangles: hole region. C: cortical bone. Scale bars: 1µm (, C), 2µm (). Unpaired 2-tailed Student s t-test was used.*, P<.1; **, P<.5.
Figure S9 PSD PSD1 Vegfa fl/fl Vegfa CKO Vegfa fl/fl Vegfa CKO Normal range WC (K/µL) 3.6 ±.3 3.5 ±.8 3.5 ±.4 4.3 ±.4 1.8-1.7 NE (K/µL).4 ±.1.6 ±.1 1.2 ±.1*.9 ±.1*.1-2.4 LY (K/µL) 3.1 ±.3 2.9 ±.7 2.2 ±.3 3.3 ±.3.9-9.3 NE % 11.1 ± 1.6 17.8 ± 2.7 34.5 ± 3.1* 2.3 ± 1.2 6.6-38.9 LY % 85.4 ± 2.1 79.1 ± 2.8 61.1 ± 3.5* 76.7 ± 1.6 55.8-91.6 RC (M/µL) 8.9 ±.4 8.7 ±.7 9. ±.4 8.8 ±.4 6.4-9.4 Hb (g/dl) 1.9 ±.4 11.1 ±.9 11.5 ±.4 1.8 ±.4 11.-15.1 HCT % 44.9 ± 2.4 43.4 ± 3. 44.6±2.2 43.7 ± 2.5 35.1-45.4 PLT (K/µL) 589.3 ± 45 417.3 ± 76.5 581.5 ± 36.6 578.9 ± 22.1 592-2972 Vegfa fl/fl NIMP-R14 Density of neutrophils in hole region/mm2 51±85 331±66 C Vegfa fl/fl F4/8 Density of macrophages in hole region/mm 2 31±13 4±11 Figure S9. lteration of peripheral blood cells and local infiltration of inflammatory cells during acute inflammation. () Complete analysis of peripheral blood from mice before and 24 hours after surgery, N=6-7. (, C) Similar density of NIMP-R14 positive neutrophils () and F4/8-positive macrophages (C) in defect hole region of and Vegfa fl/fl mice at PSD1, N=4-7. Scale bar: 2µm (, C). Yellow stippled rectangles: hole region. C: cortical bone. Unpaired 2-tailed Student s t-test was used. *, P<.1.*,PSD1 vs. PSD.
Figure S1 Vegfa fl/fl Vegfa fl/fl niline lue rdu Density of niline bluestained area(%) 29.3±3. 2.±.7* Density of rdu incorporated cells in 211±63 511±49* hole region/mm 2 C D.22 E MSCs from Vegfa CKO mice 4 PSD3 PSD7 1% FS.22 MSCs from Vegfa fl/fl mice 1% FS Total cell density/mm 2 3 2 1 OD (45 nm).18.14 # # OD (45 nm).18.14 ## Vegfa fl/fl Osx-cre;fl/fl Fold increase 1.8 4.5.1 VEH 1 ng/ml 4 ng/ml.1 VEH 1 ng/ml 4 ng/ml F 1.4 1.2 MSCs from Vegfa CKO mice 2% FS ** * OD (45nm) 1.8.6.4.2 VEH 2 ng/ml PDGF- 1 ng/ml PDGF- 4 ng/ml PDGF- 1 ng/ml PDGF- + 4 ng/ml 4 ng/ml PDGF- + 4 ng/ml Figure S1. Increased cell proliferation and decreased collagen accumulation in Vegfa fl/fl ;Osxcre mice at PSD5. () Reduced density (2. ±.7%) of aniline blue-stained areas in hole region of mice compared with Vegfa fl/fl mice (29.3 ± 3.%), N=4-5. Scale bar: 2µm. () Increased density (511 ± 49/mm 2 ) of rdu incorporated cells in hole region of mice compared with Vegfa fl/fl mice (211 ± 63/mm 2 ), N=3-4. Scale bar: 1µm. (C) Fold increase in total cell density in hole region of and Vegfa fl/fl micefrompsd3topsd7.(d-f) Proliferation of MSCs cultured in medium for 48 hour monitored by CCK8. OD 45nm represents relative cell number. t least 4 wells used for each group; data are representative of three independent experiments. Yellow stippled rectangles: hole region. C: cortical bone. Unpaired 2- tailed Student s t-test was used. *, #, P<.1; **, ##, P<.5. # vs. VEH (D, E)
Figure S11 Vegfa fl/fl Osx-cre;ZsG ;ZsG PSD7 SO PSD1 PSD1 Percent of cartilage (%) 1.±.5 26.±6.7** C Osx-cre;ZsG ;ZsG PSD14 ZsG SOX9 D rdu ZsGreen Osx-cre;ZsG ;ZsG E OD Value (45nm).3.25.2.15.1.5 PS 1 ng/ml 4 ng/ml Density of proliferative cells in 389±3 264±66 injured PO /mm 2 1 3 5 7 Day(s) Figure S11. Delayed replacement of cartilaginous callus by bony callus in mice. () niline blue staining on sections from mice at PSD7, PSD1 and PSD14 shows bony callus in periosteum. Representative images from 4-8 mice for each genotype. () Increased percentage (26. ± 6.7%) of safranin O (SO)-stained cartilage in injured periosteum of ;ZsG mice compared with Osx-cre;ZsG mice (1. ±.5%), N=3-5. (C) Increased number of SOX9-positive chondrocytes in injured periosteum of ;ZsG mice compared with Osx-cre;ZsG mice; few ZsG-positive osteoblast lineage cells express SOX9. Representative images from 3-4 mice for each genotype. (D) Similar density of rdu labeled cells in injured PO of Osx-cre;ZsG and ;ZsG mice, N=4-5. (E) dministration of recombinant does not affect proliferation of primary murine chondrocytes; The data are representative of three independent experiments. C: cortical bone (while stippled area); PO: periosteum (blue stippled area). Scale bars: 5µm (D), 1µm (C), 2µm (, ). Unpaired 2- tailed Student s t-test was used. **, P<.5.
Figure S12 Flk1fl/fl Flk1fl/+;Osx-cre Flk1fl/fl;Osx-cre SP Density of anti-sp stained areas (%) 7.6±.9 Flk1fl/fl 7.1±.8## 3.1±1** Flk1fl/+;Osx-cre Flk1fl/fl;Osx-cre CD31 Vessel Density in hole region /mm2 137±12 15±13 Flk1fl/fl 137±28 Flk1fl/+;Osx-cre Flk1fl/fl;Osx-cre C Density of staining in hole region (%) Flk1fl/fl MSCs D 4.3±.5 4.5±.9 4.1±.8 Flk1fl/fl;Osx-cre E Flk1fl/fl Flk1fl/fl;Osx-cre 82±17 98±13 lizarin Red TUNEL Von Kossa Density of TUNEL+ cells in hole region /mm2 Figure S12. Increased osteoblast maturation and mineralization in defects of Flk1 fl/fl;osx-cre mice at PSD7. () Low density of anti-sp stained areas in hole region of Flk1fl/+;Osx-cre mice compared with Flk1fl/fl;Osx-cre and Flk1fl/fl mice at PSD7, N=4-7. (, C) Similar density of blood vessels (based on CD31 staining) () and levels (C) in hole region of different genotypes, N=4-7. (D) lizarin red and Von Kossa staining of MSCs from Flk1fl/fl or Flk1fl/fl;Osx-cre mice after 21 days of culture in mineralization medium with 1nM dexamethasone. Representative images from three independent experiments. (E) Similar density of TUNEL+ cells in hole region of Flk1fl/fl and Flk1fl/fl;Osx-cre mice, N=7. Scale bars: 5µm (), 1µm (, C, E). Yellow stippled rectangles: hole region. C: cortical bone. NOV with Tukey s post-hoc test (-C) and unpaired 2-tailed Student s t-test (E) were used. **, P<.5. ** vs. Flk1 fl/fl. ## vs Flk1 fl/+;osx-cre. ##
Figure S13 Flk1 fl/fl Flk1 fl/fl ;Osx-cre SO PSD1 Callus thickness (µm) 213.7±29 229.9±2.2 Flk1 fl/fl Flk1 fl/fl ;Osx-cre SO PSD14 Callus thickness (µm) 345.7±6.4 38.8±11.8 Figure S13. Periosteal endochondral ossification is not altered when Vegfr2 is removed from osteoblastic cells. () Safranin O-stained cartilage in periosteum distal to hole region in Flk1 fl/fl and Flk1 fl/fl ;Osx-cre mice at PSD1. Similar percentage of stained cartilage and callus thickness, represented by average distance from edge of periosteal callus to cortical bone, in at least 3 tissue sections, N=3-5 for each genotype. () lmost no safranin O-stained cartilage in either Flk1 fl/fl and Flk1 fl/fl ;Osx-cre mice at PSD14. No difference in periosteal callus thickness, N=3-5. C: cortical bone. Scale bar: 2µm. Unpaired 2-tailed Student s t-test was used.
Figure S14 Figure S14. Diagrams summarizing roles of osteoblast-derived in bone defect healing. In response to hypoxia during inflammation, osteoblasts release angiogenic factors, including, which induce proliferation and migration of endothelial cells. Increased vasculogenesis and angiogenesis bring bone-forming progenitors as well as nutrition, oxygen and minerals. In addition, vascular cells release osteogenic factors, such as MP2, further promoting osteoblast differentiation and mineralization. In the bone healing process, physiological levels of are important. Reduced levels in osteoblasts result in interrupted communication between bone and blood vessels, and compromised bone healing. Levels of beyond physiological levels have detrimental effects on bone repair, likely due to inhibition of osteoblastic maturation and mineralization via stimulated R2 signaling.