Angiogenic conditioning of peripheral blood mononuclear cells promotes fracture healing

Quality and quantity culture

Samples of human peripheral blood (20 mL) were collected from healthy volunteers, and PBMNCs were isolated by density gradient centrifugation using Lymphocyte Separation Solution (LSM) (d = 1.077, Histopaque-1077 sterile-filtered; Sigma-Aldrich, St Louis, Missouri), as previously reported.⁸ Informed consent was obtained from all volunteers. Isolated PBMNCs were treated by the QQ-culture method for seven days without any change of medium at a density of 2 × 10⁶ cells/2 mL/well in six-well culture plates (BD Biosciences, #353846, Primaria, BD Falcon, San Jose, California). The complete QQ-culture medium was prepared using serum-free medium (Stem Line; Sigma-Aldrich) supplemented with human stem cell factor (SCF) (#130-096-692, concentration 100 ng/mL; Miltenyi Biotec, Bergisch Gladbach, Germany), human thrombopoietin (TPO) (#130-094-011, concentration 20 ng/Ml; Miltenyi Biotec), human Fms-related tyrosin kinase (Flt)-3 ligand (#130-096-474, concentration 100 ng/Ml; Miltenyi Biotec), human vascular endothelial growth factor (VEGF) (#100-20, concentration 50 ng/mL; Peprotech EC Ltd, London, United Kingdom), and human interleukin (IL)-6 (#130-095-365, concentration 20 ng/mL; Miltenyi Biotec). Cell density in QQ culture corresponded to approximately 1 × 10⁶ PBMNCs in 1 mL as previously reported (Supplementary Fig. a).^12,13 After culture, QQMNCs were counted and used for further analysis as described below.

Cell characterisation by flow cytometry

To characterise PBMNCs and QQMNCs (n = 6 per group), fluorescence-activated cell sorting (FACS) analysis was performed using a BD LSRFortessa Cell Analyzer (BD Biosciences) and CellQuest software (BD Biosciences) after staining with mouse anti-human monoclonal antibodies against the following surface markers: Cluster of Differentiation (CD)34-Brilliant Violet (BV) 421 (clone 581; BD Biosciences); Vascular Endothelial Growth Factor Receptor(VEGFR)2-Phycoerythrin(PE) (clone 7D4-6; BioLegend, San Diego, California); Cluster of Differentiation (CD206)-fluorescein(FITC) (clone 15-2; BioLegend); and C-C chemokine receptor type 2 (CCR2)- Brilliant Violet (BV)605 (clone K036C2; BioLegend). Dead cells were excluded on the basis of 7-Amino-Actinomycin D (7-AAD) staining (BD Biosciences). Cells were stained with monoclonal antibodies for 20 minutes at 4°C following Fc receptors (FcR) blocking, washed twice using Hank’s buffered salt solution containing 2% fetal bovine serum (FBS), and then analysed. Relevant isotype controls (Immunoglobulin G(IgG)1-BV421 isotype control (BD Biosciences), IgG1-PE (BD Biosciences), IgG1-FITC (BD Biosciences), and IgG2a-BV605 (BioLegend)) were also used. In all samples, 20 000 events were acquired.

Cell characterisation by quantitative real-time polymerase chain reaction

Using TRIzol (Invitrogen, Thermo Fisher Scientific, Waltham, Massachusetts), total RNA was extracted from PBMNCs or QQMNCs (n = 6 per group). Complementary DNA (cDNA) was synthesised using 10 ng total RNA with the SuperScript VILO cDNA Synthesis Kit (Invitrogen). A real-time PCR assay was performed using TaqMan Gene Expression Assay (Applied Biosystems, Foster City, California). Relative expression was calculated using the ΔΔCt values and results are expressed as 2^-ΔΔCt. The value of each control sample was set at one and was used to calculate the fold difference in each target gene. All primers and probes used are listed in Table I.

Tube formation assay

PBMNCs and QQMNCs were evaluated by tube formation assay by co-culturing with human umbilical vein endothelial cells (HUVECs) on ECMatrix gel (Merck-Millipore, Darmstadt, Germany) to confirm their angiogenic potential. To solidify the matrix solution, 150 µL was incubated at 37°C for one hour in a 96-well plate. Then 1 × 10³ cells from each group (PBMNCs, QQMNCs) were mixed with 1.5 × 10⁴ HUVECs in 50 µL of Endothelial Basal Medium -2 (EBM-2) complete medium containing 2% FBS, seeded onto the surface of the polymerised ECMatrix and co-cultured. As a control, HUVECs were cultured alone. After incubation for 12 hours, photomicrographs of each well were taken under a light microscope (IX71N-22FL/PH; Olympus, Tokyo, Japan), and then the length of tube formation was calculated using ImageJ Software version 1.48 (National Institutes of Health, Bethesda, Maryland) (n = 16 in each group four donors, repeated four times ).

Osteogenic induction of MSCs in vitro

Commercially available primary human bone marrow MSCs (BMMSCs) from healthy donors (Lonza Ltd, Basel, Switzerland) were cultured and expanded in a specific medium, mesenchymal stem cell growth medium (MSCGM Bulletkit; Lonza Ltd). MSCs at passage four were seeded into 12-well plates with an initial seeding density of 10 000 cells/cm² and preconditioned in osteogenic induction medium (StemPro Osteogenesis Differentiation Kit; Thermo Fisher Scientific) at 37°C for one day. After preconditioning, the MSCs were co-cultured with PBMNCs or QQMNCs (1 × 10⁵ cells, n = 12 in each group; three donors, repeated four times) using cell culture inserts (Falcon Cell Culture Inserts: For Use With 12-Well Plates#353180 ) for seven days. As a control, MSCs alone were cultured in osteogenic induction medium. Total RNA was extracted from the MSCs at days 3 and 7 of co-culture using TRIzol, and cDNA was synthesised. A real-time PCR assay was performed in the same manner as described above. All primers and probes used in this analysis are listed in Table I. For evaluation of calcium deposition, the MSCs were stained with Alizarin Red S (CAS No. 130-22-3; Sigma-Aldrich) at day 14 of co-culture, and the stained area was measured using ImageJ Software.

Animal study

Animals and surgical procedure

Male athymic nude rats (F344/NJcl-rnu/rnu, nine to ten weeks old, (CLEA Japan, Inc, Tokyo, Japan) were used to evaluate the therapeutic potential of QQMNCs in vivo. All surgical procedures were performed under normal sterile conditions. The rats were anaesthetised using ketamine (100 mg/kg) and xylazine (12 mg/kg) by intraperitoneal administration. A transverse femoral fracture was created in the right femur of each rat and nonunion was induced by cauterising the periosteum to a distance of 2 mm on each side of the fracture, as previously described.^14,15 In order to avoid significant displacement of the fracture and achieve well-aligned stability, a 1.2 mm diameter Kirschner wire was inserted from the trochlear groove into the femoral canal in a retrograde manner. Finally, the wound was tightly closed in layers. Rats were then randomised to three groups: control; PBMNC; and QQMNC. In the PBMNC and QQMNC groups, 2 × 10⁵ cells were injected into the site immediately after surgery while the control group received phosphate-buffered saline (PBS) alone.

Evaluation of fracture healing

Rats were anaesthetised before imaging. Serial radiographs were taken at zero, two, four, six and eight weeks after surgery, and fracture healing was evaluated by Qui’s radiographic scoring system (n = 8 in each group and at each time point).¹⁶

For micro-CT imaging (µCT), rats were killed at two, four and eight weeks, and the right femurs were harvested. The femurs were fixed in 4% paraformaldehyde (Wako Pure Chemical Industries Ltd, Osaka, Japan) at 4°C for 24 hours, wires were removed, and µCT images were acquired (SkyScan 1176; Bruker, Kontich, Belgium). The imaging settings were as follows: acceleration voltage: 40 kV; beam current: 800  µA, no filter; resolution: 18  μm; and rotation: 360° in 1° steps. Acquired images were reconstructed using the manufacturer’s software (NRecon version 1.6.4.1; Skyscan), and total callus volume (TV), mineralised bone volume (BV), and bone volume fraction (BV/TV) were analysed on either side, 2 mm away from the fracture site using a CT Analyzer, version 1.11.8.0 (Skyscan) (two weeks: n = 6 per group; four weeks: n = 6 per group; and eight weeks: n = 8 per group).

Histological evaluation

After µCT scanning, the right femurs were decalcified with 0.5 M ethylenediaminetetraacetic acid (EDTA) solution for three weeks. The samples were embedded in paraffin, and cut into 5 µm thick sections through the site of the fracture. The sections were stained with Toluidine blue and the process of endochondral ossification was evaluated at two, four and eight weeks after surgery. The degree of fracture healing was evaluated using the five-point scale (grades 0 to 4) proposed by Allen, Wase and Bear.¹⁷ The assessment was blinded. In addition, the sections at two weeks after surgery were stained for smooth muscle actin to assess angiogenesis. Researchers blinded to the identity of each group counted the number of capillary structures in five random areas around the fracture sites using a light microscope (×200 magnification). The capillary density was then calculated as the number of vessels/mm² as previously described.¹⁸

Gene expression analysis (around the fracture site)

From the two-week old rat model, the tissue around the fracture site was harvested after death, and RNA was isolated (n = 6 per group). Real-time PCR was performed to evaluate fracture healing-related gene levels as described above. All primers and probes used are listed in Table I.

Statistical analysis

All values are expressed as means and standard error. Student’s t-test (for two groups), analysis of variance (ANOVA) and Tukey’s post hoc test (more than two groups) were used. Probability (p) values of less than 0.05 were considered statistically significant.

Cell number and population transition after QQ culture

The fold decrease of QQMNCs to PBMNCs per well in all subjects was an average of 0.358-fold (Fig. 1a).

Fig.

Cell characterisation of quality and quantity control-cultured mononuclear cells (QQMNCs). (a) Total cell counts of peripheral blood mononuclear cells (PBMNCs) and QQMNCs and their images: 2 × 10⁶ cells/2 mL/well were cultured in Quality and Quantity(QQ) culture using five wells in a six well-plate. After culture, QQMNCs were counted. The graph shows total counts of PBMNCs and QQMNCs in five wells. This graphs show that total cell number decreases after culture (**p < 0.01, n = 12). Phase contrast image of PBMNCs and QQMNCs (×40 magnification); b) gene expression of angiogenic markers relative to glyceraldehyde 3-phosphate dehydrogenase: quantitative real-time polymerase chain reaction (qRT-PCR) assay of PBMNCs and QQMNCs. Each value indicates a mean sd standard deviation (*p < 0.05, **p < 0.01; n = 6 volunteers). The expression of both vascular endothelial growth factor (VEGF) and angiopoietin (Ang) 2 were significantly greater in QQMNCs than in PBMNCs; c) enhanced in vitro tube formation by human umbilical vein endothelial cells(HUVECs) when co-cultured with QQMNCs. Representative images of each group are shown (×40 magnification). The total length of tube formation was calculated by ImageJ Software version 1.48 (National Institutes of Health, Bethesda, Maryland). The overall length in the QQMNC group was significantly longer than in other groups (**p < 0.01).

Based on fluorescence-activated cell sorting, QQMNCs exhibited enrichment of CD34+ and CD206+ cells compared with PBMNCs (10.01-fold in CD34+ cells, 26.04-fold in CD206+ cells, Table II). In contrast, the population of Vascular Endothelial Growth Factor Receptor(VEGFR)-2+ and C-C chemokine receptor type 2 (CCR2+ cells) was significantly lower in QQMNCs than in PBMNCs (0.33-fold in VEGFR-2+ cells, 0.50-fold in CCR2+ cells) (Supplementary Fig. b).

Enhanced angiogenic gene expression in QQMNCs

Expression of genes encoding angiogenic factors, including VEGFA and angiopoietin-2 (Ang2), was significantly upregulated in QQMNCs compared with PBMNCs (3.58-fold for VEGF-A, 25.6-fold for Ang2). In contrast, there was no significant difference in Ang1 expression between QQMNCs and PBMNCs (Fig. 1b).

QQMNCs promote tube formation in vitro

In an in vitro tube formation assay involving co-culture with HUVECs, QQMNCs showed greater angiogenic potential at 12 hours. Total tube length in the QQMNC group was significantly longer than in the PBMNC and control groups (tube length 44.5 mm, sd 3.4 for HUVECs and QQMNCs; 39.4 mm, sd 4.7 for HUVECs and PBMNCs; and 38.5 mm, sd 5.1 for HUVECs alone (Fig. 1c).

QQMNCs enhance osteo-induction of MSCs in vitro

At day 3 of co-culture, real-time PCR revealed that the expressions of RUNX2 and Type 1 collagen were significantly upregulated in the QQMNC group compared with the PBMNC and control groups. At day 7, the expression of osteocalcin was significantly higher in the QQMNC group compared with the PBMNC and control groups (Fig. 2a). Alizarin Red staining of co-cultures at day 14 showed a significantly larger stained area in the QQMNC group compared with the control group. However, there was no significant difference between the PBMNC and QQMNC groups (Fig. 2b).

Fig.

Real-time polymerase chain reaction (RT-PCR) assay of osteogenic induction of mesenchymal stem cells (MSCs) co-cultured with peripheral blood mononuclear cells (PBMNCs) or quality and quantity control-cultured mononuclear cells (QQMNCs): a) the graphs show the relative gene expression levels of osteogenic factors (Type 1 collagen, RUNX2, and osteocalcin) (**p < 0.01, Tukey’s post hoc test). Expression of osteocalcin was undetectable at day 3. The expression levels were compared between the QQMNC group and the control and PBMNC groups. Each graph column represents a mean standard deviation (SD); n = 12 per group; b) quantification of Alizarin red S staining: MSCs were stained with alizarin red at day 14 of co-culture. Microscopic image of each well of 12-well plate (upper panel: × 40 magnification in 12 wells and lower panel: × 200 magnification) showed enhanced alizarin red staining of MSCs in co-culture with QQMNCs compared with control.

QQMNCs accelerate fracture healing

Fracture healing was evaluated by radiographic, µCT, and histological examinations. Rats which received QQMNCs demonstrated the highest rate of fracture healing, with bridging callus formation observed radiographically (50% (3/6) at four weeks, 62.5% (5/8) at eight weeks). In the group of rats receiving PBMNCs, only one achieved radiographic fracture healing at eight weeks (12.5% (1/8)), while none of the rats in the control group achieved radiographic fracture healing (Fig. 3a, Table III). Qi’s radiographic fracture healing score was significantly higher in the QQMNC group than in the other groups at eight weeks (QQMNCs: 5.88, sd 0.64; PBMNCs: 3.0, sd 0.33; control: 2.63, sd 0.182) (Fig. 3a). The µCT analysis confirmed the same rates of fracture healing at eight weeks (QQMNCs: 62.5%; PBMNCs: 12.5%; control: 0%) (Fig. 3b). In the quantitative evaluation of bone formation, the value of BV/TV was significantly higher in the QQMNC group compared with the other groups at eight weeks (QQMNCs: 48.81%, sd 0.90%; PBMNCs: 37.04%, sd 9.13%; control: 34.61%, sd 6.21%) (Fig. 3b).

Fig.

Evaluation of bone union by radiography and µCT: a) representative radiographs of each group at 0, two, four, six, and eight weeks post-operatively and the radiographic scores by Qui in each group at 0, four, and eight weeks. The score in the quality and quantity control-cultured mononuclear cell (QQMNC) group was significantly higher than in the other groups (*p < 0.05, ** p < 0.01, Tukey’s post hoc test; n = 8); b) representative µCT images from each group at four and eight weeks post-operatively. The union rates reflected the radiographic images. Mineralised callus volume (BV)/total callus volume (TV) (%) was calculated using µCT. TV is shown as the red area in the left picture. BV is shown as the pink area in the right picture. The BV/TV ratio was calculated from the fracture site to 2 mm in each case. BV/TV in the QQMNC group was significantly higher compared with in the peripheral blood mononuclear cells (PBMNC) and control groups. The graph represents the mean sd standard deviation (*p < 0.05, **p < 0.01, Tukey’s post hoc test; n = 6).

Histological evaluation demonstrated that nonunion sites in the PBMNC and QQMNC groups showed enhanced endochondral ossification consisting of numerous chondrocytes and newly formed trabecular bone at two and four weeks after surgery. In five rats of the QQMNC group and one in the PBMNC group, bridging callus formation was observed at eight weeks. The histological score of fracture healing was significantly greater in the QQMNC group than in the other groups at eight weeks (QQMNCs: 3.50, sd 0.75; PBMNCs: 2.12, sd 0.99; control: 1.87, sd 0.83) (Fig. 4a).

Fig.

Histological evaluation of bone healing after cell transplantation: a) representative histological image (× 40 magnification) of each group stained with toluidine blue at two, four, and eight weeks post-operatively (bc, bridging callus; cb, cortical bone; ca cartilage; ft, fibrous tissue; tb trabecular bone). The nonunion sites of the peripheral blood mononuclear cell (PBMNC) group and the quality and quantity control-cultured mononuclear cell (QQMNC) group demonstrated enhanced endochondral ossification consisting of numerous chondrocytes and newly-formed trabecular bone. The degree of bone healing evaluated by Allen’s scoring scale was significantly higher in the QQMNC group than in the other groups at eight weeks post-operatively. The graph represents the mean sd standard deviation (*p < 0.05, **p < 0.01;, Tukey’s post hoc test n = 6 at two, and four weeks, n = 8 at eight weeks); b) immunohistochemical evaluation of neovascularisation by α smooth muscle actin staining at two weeks post-operatively: representative images of each group are shown (top images; ×40 magnification, bottom images; ×200 magnification). The density of capillaries was calculated and can be seen in the graph. The density was significantly higher in the QQMNC group than in the control or PBMNC groups. Each column represents the mean sd standard deviation (**p < 0.01, Tukey’s post hoc test; n = 6) (PBS, phosphate-buffered saline).

QQMNCs enhance angiogenesis and osteogenesis in vivo

Capillary density was significantly greater in the QQMNC group than in the PBMNC and control groups in the tissue harvested at two weeks after transplantation (QQMNCs: 21.67 vessels/mm², sd 4.96; PBMNCs: 10.95 vessels/mm², sd 2.56; control: 10.62 vessels/mm², sd 3.02) (Fig. 4b). In addition, in the QQMNC group, several bone healing-related genes were significantly upregulated (Fig. 5). There were no significant differences in expression of type 1 collagen and osteocalcin among the three groups (Fig. 5).

Fig. 5

quantitative real-time polymerase chain reaction (qRTPCR) of callus around the fracture site at two weeks after implantation. The relative gene expression level relating to angiogenic factors (vascular endothelial growth factor (VEGF), fibroblast growth factor 2 (FGF2)). The expression of VEGF was significantly greater in the quality and quantity control-cultured mononuclear cell (QQMNC) group than in the control group and the expression of FGF2 was significantly higher in the QQMNC group than in the peripheral blood mononuclear cell (PBMNC) and control groups. The relative gene expression levels of osteogenic factors (Type 1 collagen, RUNX2, osteocalcin) are shown in the second row of graphs. The expression of RUNX2 was significantly greater in both experimental groups than in the control group (*p < 0.05, **p < 0.01, Tukey’s post hoc test; n = 6 per group). Each graph column represents a mean sd standard deviation.