Bone tissue engineering using adipose‐derived stem cells and endothelial cells: Effects of the cell ratio

Abstract The microvascular endothelial network is essential for bone formation and regeneration. In this context, endothelial cells not only support vascularization but also influence bone physiology via cell contact‐dependent mechanisms. In order to improve vascularization and osteogenesis in tissue engineering applications, several strategies have been developed. One promising approach is the coapplication of endothelial and adipose derived stem cells (ADSCs). In this study, we aimed at investigating the best ratio of human umbilical vein endothelial cells (HUVECs) and osteogenic differentiated ADSCs with regard to proliferation, apoptosis, osteogenesis and angiogenesis. For this purpose, cocultures of ADSCs and HUVECs with ratios of 25%:75%, 50%:50% and 75%:25% were performed. We were able to prove that cocultivation supports proliferation whereas apoptosis was unidirectional decreased in cocultured HUVECs mediated by a p‐BAD‐dependent mechanism. Moreover, coculturing ADSCs and HUVECs stimulated matrix mineralization and the activity of alkaline phosphatase (ALP). Increased gene expression of the proangiogenic markers eNOS, Flt, Ang2 and MMP3 as well as sprouting phenomena in matrigel assays proved the angiogenic potential of the coculture. In summary, coculturing ADSCs and HUVECs stimulates proliferation, cell survival, osteogenesis and angiogenesis particularly in the 50%:50% coculture.


| INTRODUC TI ON
The reconstruction of critical sized bone defects can be challenging in clinical practice. Critical bone defects can be caused by malformation, cancer, trauma or infection. Regardless of the entity, the current gold standard is autologous tissue transfer, which can be associated with significant donor side morbidity and limited tissue availability. One way to circumvent these problems is the generation of bioartificial bone tissue.
For bone formation and regeneration, a sufficient vascularization providing oxygen and nutrition supply is indispensable. [1][2][3] Strategies to improve vascularization in bone tissue engineering applications include the use of angiogenic growth factors, endothelial cells (ECs) and the surgical induced angiogenesis by means of arteriovenous loops. [4][5][6] In many studies, the cocultivation of ECs and mesenchymal stem cells (MSCs) has already proven to be beneficial for proliferation and osteogenic differentiation. [7][8][9][10][11] With regard to clinical practice, the isolation of MSCs from the bone marrow can be limited in terms of quantity and donor side morbidity. Stem cells from fat tissue are an interesting alternative to MSCs derived from bone marrow. Isolation, characterization and multiple differentiation potential have already been described in the literature, and it has been shown that adipose-derived stem cells (ADSCs) are also suitable for bone defect healing in animals. 12,13 In addition to that, it has been shown that ADSCs and bone marrow MSCs have both osteogenic differentiation potential. [14][15][16] Furthermore, ADSCs are even superior to MSCs in terms of immunomodulatory capabilities and secretion of proangiogenic factors and extracellular matrix components. [17][18][19][20][21] Especially in coculture with endothelial cells, the osteogenic differentiation of ADSCs can be further increased. 22 Moreover, human umbilical vein endothelial cells (HUVECs) have a pronounced vascularization capacity. 5,21,23,24 The coapplication of ADSCs and HUVECs in terms of tissue engineering applications seems to be a promising approach to increase vascularization and bone formation.
The aim of this study was to investigate the optimal ratio of HUVECs and osteogenic differentiated ADSCs in the two-dimensional cell culture and the effects on proliferation, cell survival, osteogenesis and angiogenesis. Using negative immunoselection, we tried to enlighten the cell type specific effects regarding apoptosis, angiogenesis and osteogenic differentiation more in detail.

| Cell culture
Human ADSCs were isolated from five patients undergoing autologous breast reconstruction, according to an established protocol. 25  ADSCs underwent osteogenic differentiation for 14 days, before coculturing with HUVECs. To induce osteogenic differentiation, ADSCs were cultured in ECGM medium, modified with 50 µg/mL L-ascorbic acid, 10 mmol/L glycerophosphate, 1 × 10 −8 mol/L dexamethasone, 0.01 µmol/L 1,25-dihydroxyvitamine D3 (all supplements purchased from Sigma), 10% FCS superior, 100 U/mL penicillin and 100 µg/mL streptomycin, according to an established protocol. 26 To determine the ideal concentration of ADSCs and HUVECs, five groups were formed, the monoculture of ADSCs and HUVECs and cocultures with different ratios (75%:25%, 50%:50% and 25%:75%). All experiments were performed with the ADSCs from the 5 donors in a concentration of 5000 cells/cm 2 with osteogenic modified ECGM. The cells were grown on plastic cell culture plates as two-dimensional cell cultures.

| Cell viability assay
After 3, 7 and 14 days, cell viability was assessed using a WST-8 assay.
After refreshing medium and adding CCVK-I-Solution (PromoCell), all five groups were incubated for 2 hours at 37°C. The resulting supernatant was analyzed photometrically by using an ELISA Reader at 450 nm.

| Negative immunoselection
In order to describe celltype specific effects, negative immunoselection was carried out after 3, 7 and 14 days. Briefly, cells were de-

| Cell death detection Assay
To quantify the DNA fragmentation in apoptotic cells, a cell death detection ELISA (Sigma) was performed after 7 and 14 days.

| Alizarin red assay
Matrix mineralization was measured with an Alizarin Red-based assay after 14 days, according to manufacturer recommendation.
Briefly, samples were fixed with 4% paraformaldehyde. After fixation, the samples were washed with phosphate-buffered saline (PBS) (Sigma). Then, 1 mL Alizarin Red staining solution (ScienCell) was added. After incubation for 30 minutes, the dye was removed, and the samples washed and 800 µL acetic acid (ScienCell) added.
Afterwards, the cells were collected using a cell scraper and the samples heated at 85°C for 10 minutes. After cooling and centrifugation, the supernatant was collected, neutralized using 10% ammonium hydroxide (ScienCell), and absorbance measured using an ELISA Reader at 405 nm.

| Quantification ALP activity
Osteogenic differentiation was assessed by alkaline phosphatase (ALP) activity. The ALP assay (Abcam) was performed after 3 and 7 days according to manufacturer information. Briefly, the cells were detached by using Accutase ® solution (Sigma), washed with ice-cold PBS, centrifuged at 300 × g for 4 minutes, resuspended in 200 µL assay buffer and centrifuged again at 18 800 × g for 15 minutes.
Afterwards, the resulting supernatant was transferred into microtitre plates and the pNPP solution added. After 60 minutes, a stop solution was added and absorbance measured at 405 nm.

| Matrigel assay
To investigate the angiogenic potential of cocultures containing ADSCs and HUVECs, a matrigel assay was performed. 10 µL matrigel (Corning) were pipetted into each well of a µ-Slice (ibidi). After polymerisation for 30 minutes at 37°C, 10 000 cells per well were added and incubated for 4 hours, at 37°C, 5% CO 2 . Vital cells were visualized using calcein staining (Sigma). The number of branches and the length of vessel network were analyzed by Angiogenesis Analyzer (ImageJ version 2 NIH).

| Proliferation
A WST-8 assay was performed as a surrogate assay for cell prolifera- After 14 days, we were able to prove statistically significant more vital cells in all coculture groups compared to the monocultures.

| Apoptosis
Proportional to the HUVEC ratio, we measured an increasing apoptosis rate in ADSCs with the highest values in the coculture group with 75% ADSCs after 7 and 14 days ( Figure 3A). Conversely, HUVECs displayed an unidirectional reduction of apoptosis upon coculture with ADSCs. Interestingly, apoptosis was even more reduced in cocultures with an ADSC ratio ≥50% in the first week. After 2 weeks, we were not able to detect any significant influence of the ADSC ratio on apoptosis in HUVECs ( Figure 3B). To enlighten a putative mechanism, we performed a phospho-BAD ELISA confirming a lower proportion of phosphorylated BAD which might explain the higher apoptosis rate in cocultured ADSCs. On the other hand, we found increasing levels of phosphorylated BAD in cocultured HUVECs leading to decreased apoptosis ( Figure 3C,D).

| Osteogenic differentiation
Coculturing ADSCs and HUVECs increased matrix mineralization as proved by alizarin red assay ( Figure 4A). Moreover, matrix mineralization increased with higher HUVEC ratios in the coculture groups.
Alkaline phosphatase (ALP) activity is another surrogate parameter for osteogenic differentiation. On day 3, we observed an induction of ALP activity in the coculture groups with ≥25% HUVECs. After 7 days, this effect was even more pronounced in all cocultures (twofold induction) ( Figure 4B).
After negative immunoselection and PCR, we were able to prove that the induction of ALP gene expression is limited to ADSCs.
Additionally, we measured the highest ALP mRNA levels in the coculture group with 50% ADSCs ( Figure 5A). Moreover, we found increasing levels of the osteogenic transcription factor RUNX2 in ADSCs after coculturing. The coculture with 50% ADSCs displayed the highest RUNX2 upregulation ( Figure 5B).

| Angiogenesis
To investigate the angiogenic potential of ADSC/HUVEC cocultures, matrigel assays were performed. On the contrary to osteogenic F I G U R E 2 Coculturing ADSCs and HUVECs increased proliferation after 7 and 14 days compared to the monocultures. Statistically significant differences between the experimental groups are indicated for *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001

TA B L E 1
The primer sequences used for polymerase chain reaction (PCR) differentiated ADSCs, undifferentiated ADSCs formed tubes in matrigel. Proportionally to the amount of HUVECs, the number and length of tubes increased in the cocultures ( Figure 6A-H). Using VEGF ELISA, we were able to prove the highest VEGF production in ADSCs under monoculture conditions on days 3 and 7. Moreover, the production of VEGF was even more pronounced in cocultures containing ≥50% HUVECs on day 3. No statistically significant differences were detected between the cocultures on day 7 ( Figure 6I). (A, B). The amount of phosphorylated protein BAD was assessed in ADSCs and HUVECs upon coculture demonstrating higher phosphorylated BAD amounts in cocultured HUVECs (C, D). Statistically significant differences are indicated for *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001

F I G U R E 4
Alizarin red assay measuring matrix mineralization (A). Alkaline phosphatase (ALP) activity was measured as a surrogate parameter for osteogenic differentiation. Higher ALP activity was found in the cocultures upon 3 and 7 days (B). Statistically significant differences are indicated for *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001 Gene expression of angiogenic molecules, such as endothelial nitric oxidase (eNOS) or matrix metalloprotease 3 (MMP3) increased in the cocultured HUVECs proportionally to higher ADSC ratios ( Figure 7B,D).

F I G U R E 5 PCR analysis concerning
ALP and RUNX2 gene expression after negative immunoseparation. In the coculture group containing 50% ADSCs, the expression of ALP increased significantly (A). A same trend towards higher RUNX2 expression was also observed in cocultured ADSCs (B). Statistically significant differences are indicated for *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001

F I G U R E 6
Matrigel assay demonstrating that the number of tubes (G) as well as the total tube length (H) increased with the HUVEC ratio. In addition to that, undifferentiated ADSCs (F) formed tubes. VEGF ELISA depicts the highest VEGF production in ADSCs under monoculture conditions (I). Statistically significant differences are indicated for *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001 F I G U R E 7 PCR analysis of cocultured ADSCs and HUVECs after negative immunoseparation. The gene expression of eNOS (A), Ang2 (E) and Flt (G) is higher in cocultured ADSCs. eNOS (B) and MMP3 (D) are higher expressed in cocultured HUVECs, whereas a lower MMP3 (C) gene expression was observed in cocultured ADSCs. Statistically significant differences are indicated for *P ≤ 0.05 and **P ≤ 0.01

| D ISCUSS I ON
Osteogenesis and angiogenesis are two directly related processes.
A stable microvascular network, providing an adequate supply of oxygen and nutrients, is essential for bone formation and regeneration. 9,28 For this reason, the interaction between mesenchymal stem cells and endothelial cells has been widely studied demonstrating auspicious effects concerning cell growth, survival and osteogenic differentiation. 10,11,29,30 Although the molecular mechanisms are not fully understood, heterotypic cell contacts between mesenchymal stem cells and endothelial cells seem to play an important role. 10,11,[30][31][32][33] So far, most studies used MSCs or osteoprogenitor cells isolated from bone marrow. Considerably fewer studies used MSCs isolated from fat tissue. However, ADSCs can be isolated in sufficient quantity without significant donor side morbidity and expanded in vitro. [34][35][36] In several in vivo studies, ADSCs were used to reconstruct critical sized bone defects. [37][38][39][40] Considering the fact that most studies used rodent cells and/or focused on global effects of coculturing ADSCs and endothelial cells, we pursued a translational approach using human primary cells and described cell type specific effects upon coculturing using negative immunoselection. In addition to that, we tried to determine the ideal cell ratio of ADSCs and HUVECs with regard to proliferation, apoptosis, osteogenic differentiation and angiogenesis.
According to previous studies, our results indicate that cell proliferation increased in all groups over 14 days. Moreover, the proliferation rate was significantly higher in the cocultures after 7 and 14 days, especially if the cell ratio constituted >50% ADSCs. 30, 33 We also investigated apoptosis in the cocultures after 7 and 14 days.
Using negative immunoseparation, we were able to analyze apoptosis of ADSCs and HUVECs separately. Our results indicate an anti-apoptotic effect of ADSCs on HUVECs. This effect was even more pronounced with an increasing ADSC ratio. Contrary to a previous study using HUVECs and MSCs, HUVECs did not reduce apoptosis in ADSCs. 30 In fact, apoptosis increased in cocultured ADSCs.
A possible explanation for this unexpected phenomenon would be that ADSCs are more vulnerable than HUVECs to increasing cell density. In this respect, Kim  Bearing in mind that angiogenesis and osteogenesis are two directly linked processes in terms of bone regeneration, we wanted to enlighten the angiogenic potential of ADSC/HUVEC cocultures.
In the pertinent literature, the angiogenic potential of ADSCs on endothelial cells is controversially discussed. [46][47][48] In our experiments, we were able to prove proangiogenic effects upon coculturing ADSCs and HUVECs correlating with the amount of HUVECs.
In this regard, no sprouting was observed in the osteogenic differ- ising growth factor is angiopoetin 2 (Ang2), which is significantly upregulated in cocultured ADSCs. Ang 2 plays a critical role in angiogenesis and supports bone healing in rabbits. 53 Angiogenesis is a well-orchestrated process supported by extracellular matrix degradation and migration mediated by matrix metalloproteases. 54,55 In our experiments, MMP-3 gene expression in HUVECs correlated with increasing ADSC ratios of ≥50% but not vice versa.
Consistent with the results from the matrigel assay, one gets the impression that a minimum of 50% HUVECs is necessary to stimulate proangiogenic effects of cocultured ADSCs.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
DS, REH and AA made substantial contributions to conception and design of the study. HM, DS, SW and VW made substantial contributions to acquisition, analysis and interpretation of data. DS, HM, SW, REH, AA and VW were involved in drafting the manuscript or revising it critically for important intellectual content. All authors have given final approval of the manuscript.

D I SCLOS U R E
The present work was performed in fulfilment of the requirements for obtaining the degree "Dr. med." of Hilkea Mutschall.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.