Comparison of percentage of CD90-positive cells and osteogenic differentiation potential between mesenchymal stem cells grown on dish and nonwoven fabric

Abstract

Although nonwoven fabric (NWF) has been reported to be a candidate scaffold for the large-scale expansion of mesenchymal stem cells (MSCs), the quality of cells grown in NWF has not been well clarified. In this report, MSCs grown in an NWF disc for 3 weeks showed higher osteogenic differentiation potential and percentage of CD90 (+) cells than MSCs grown on the bottom surface of dish. The amount of the extracellular matrix (ECM) per unit surface area of fibers was larger than that on the bottom surface of the dish in the first 2 weeks of culture. The osteogenic differentiation potential of MSCs inoculated onto cell-free ECM increased with increasing amount of ECM. The higher percentage of CD90 (+) cells and osteogenic differentiation potential of cells grown in an NWF disc than of cells grown on a dish might, at least in part, be due to the higher amount of ECM.

Introduction

Human mesenchymal stem cells (MSCs) are an attractive candidate for cell-based therapies of disorders, such as cartilage impairments (Sato et al. 2013), graft-versus-host disease (Yamahara et al. 2014), and brain infarct (Leu et al. 2010), because they can be harvested by a minimally invasive procedure. Such therapies with an allograft system might require a large number of human MSCs in a single lot, and thus, large-scale growth cultivation. However, MSCs will gradually lose their stemness during in vitro expansion (Sekiya et al. 2002; Bonab et al. 2006; Røsland et al. 2009; Hoch and Leach 2014). This is not advantageous for the large-scale culture to obtain functional MSCs. Therefore, it is important to maintain the stemness of MSCs during in vitro large-scale culture.

NWF has three-dimensional fiber aggregates formed by heat bonding, which is different from woven fabrics and knitted fabrics. NWF has a long flat surface, on which a cell can elongate, in the longitudinal direction of a fiber. Moreover, NWF has sufficient space for a cell to crosslink between fibers. The protocol of MSC inoculation into a polyester NWF disc as a scaffold to obtain high inoculation efficiency and final cell density was previously studied (Fu et al. 2019). Thus, NWF was considered as a candidate material for the excellent porous scaffold. Besides the large-scale expansion of MSCs in NWF, it should be clarified whether MSCs grown in NWF can maintain a higher level of stemness, in terms of osteogenic differentiation potential and percentage of CD90 positive cells than those grown on a dish.

In this study, the differences in these qualities between cells grown on a dish and those grown in NWF were confirmed. The reason for the differences was examined in this paper in relation to Extracellular matrix (ECM).

Materials and methods

Isolation and cultivation of MSCs

MSCs were isolated from bone marrow aspirates obtained by routine iliac crest aspiration from human donors (76-year-old female) as reported previously (Sato et al. 2013). The donors gave their informed consent in this study, which was approved by our institutional committee on human research, as required by the study protocol.

The isolated cells, whose population doubling level (PDL) was defined to be zero, were inoculated at a concentration of 0.15 × 10⁴ cells/cm² on a dish (55 cm²; Corning, NY, USA) with a growth medium, Dulbecco’s modified Eagle’s medium–low glucose (DMEM-LG; Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% FBS (Thermo Fisher Scientific), 2500 U/l penicillin, and 2.5 mg/l streptomycin (MilliporeSigma, Burlington, MA, USA). The cells were cultivated at 37 °C in a humidified atmosphere containing 5% CO₂. When the density of cells reached near confluence, they were detached using trypsin–EDTA (MilliporeSigma) and inoculated onto dish or NWF (polyester, Y-15050, Asahi Chemical Co., Tokyo, Japan) as described below. GFP-rMSCs (Fu et al. 2019) used in this study were provided by Dr. Ryosuke Iwai of the National Cerebral and Cardiovascular Center Research Institute.

Cultivation in NWF

An NWF disc (diameter: 15.1 mm, depth: 0.1 mm) was placed in a 24-well suspension culture plate (Sumitomo Bakelite Co., Ltd., Tokyo, Japan). A 10 μl suspension of MSCs was inoculated at the center of the NWF discs at a predetermined inoculation density and incubated at 37 °C in a 5% CO₂ atmosphere. Then, 10 μl of growth medium was added to the center of each disc 1 h after inoculation. Eight hours after inoculation, 1 ml of the medium was added into the wells and the culture was incubated further for a specified time.

Cell number analysis

The cell numbers in the dish and discs were counted by the nucleus staining method (Sanford et al. 1950). Briefly, after the culture supernatant was removed from the wells, an NWF disc was transferred to a 1.5 ml microtube. The NWF disc was mixed with 0.5 ml of nuclei staining solution (500 µl: crystal violet, 1 g/l; citrate, 21 g/l), and incubated for 3 days at 25 °C. Then, the mixture was vigorously mixed, and the number of cells with their nuclei stained was counted using hemocytometer.

Flow cytometry

Percentages of CD90 and CD166 positive cells were determined by flow cytometry. Both of them are one of MSC markers and CD90 is also proved to be related to the undifferentiated status of MSC (Moraes et al. 2016). Briefly, The NWF disc in a 24-well plate was transferred to a 100Ø dish and washed two times with 20 ml of PBS. Then, 5 ml of trypsin–EDTA was added, and the disc was incubated in 5% CO₂ at 37 °C for 10 min. Thereafter, 15 ml of the growth medium was added, and the suspended cells were recovered. The cells cultured in the dish were recovered by the similarly same way.

An antibody was diluted with PBS containing 1% BSA. The cell suspensions (1 × 10⁶ cells/ml) recovered from the NWF and dish were incubated with a mouse anti-human CD90 IgG1 antibody (MilliporeSigma) at 0 °C for 45 min. After washing with PBS(-), an FITC-conjugated goat anti-mouse IgG1 antibody (MilliporeSigma) was added as the secondary antibody and the cells were further incubated at 0 °C for 45 min. The cells were blocked using mouse IgG1 (Beckman Coulter, Inc., Brea, CA, USA) for 45 min at 0 °C. Then the cells were stained with a PE-conjugated mouse anti-human CD166 antibody (Beckman Coulter, Inc.) for 45 min at 0 °C. As the isotype control, mouse IgG1 (Beckman Coulter, Inc.) and PE-conjugated mouse IgG1 (Beckman Coulter, Inc.) were used. The cells were analyzed by flow cytometry (EPICS XL, Beckman Coulter, Inc.).

Osteogenic differentiation

MSCs isolated from the NWF and dish were inoculated on 6-well plates at a density of 0.8 × 10⁴ cells/cm² using the growth medium. They were cultured until they reached confluence and cultured for another 12 days using osteogenic induction medium, during which the medium was changed with a fresh one every 3 days. The osteogenic induction medium was DMEM-HG medium (Thermo Fisher Scientific) containing 55.6 μM ascorbic acid (Wako, Osaka, Japan), 11.1 mM glycerol 2-phosphate (MilliporeSigma), 0.11 μM dexamethasone (MP Biomedicals, Santa Ana, CA, USA), 10% FBS (Thermo Fisher Scientific), 2500 U/l penicillin, and 2.5 mg/l streptomycin. As the negative and positive control cells, uninduced MSCs and normal human osteoblasts (Lonza, Basel, Switzerland) were used, respectively.

Osteogenic differentiation was evaluated by Alizarin Red S staining. Samples were fixed with 95% ethanol for 30 min, washed two times with PBS, and stained for 5 min with Alizarin Red S solution (Wako) at a pH in the range of 4.2–5.1. Samples were washed with distilled water until the supernatant became clear and then observed under an inverted microscope. The percentage of the stained area was calculated as the average of 7 photos using ImageJ.

Adipogenic differentiation

MSCs isolated from the NWF and dish were inoculated on 6-well plates at a density of 0.15 × 10⁴ cells/cm² using the growth medium. After reaching confluence, the cells were cultivated for another 14 days using an adipogenic induction medium, during which the medium was changed with a fresh one every 3 days. The adipogenic induction medium was DMEM-HG medium containing 168 μM indomethacin (Wako), 0.55 mM 3-isobutyl-1-methylxanthine (MilliporeSigma), 11 μg/ml insulin (Wako), 1.1 μM dexamethasone, 10% FBS (Thermo Fisher Scientific), 2500 U/l penicillin, and 2.5 mg/l streptomycin. Uninduced MSCs and human subcutaneous preadipocyte cells (Lonza) were used as the negative and positive control cells, respectively. The samples were fixed in 4% PFA/PBS for 30 min, rinsed with PBS two times, and stained with Oil Red O (Wako) for 15 min, and washed with distilled water until the supernatant became clear. The samples were observed under an inverted microscope and photographed. The percentage of the stained cells was calculated as the average of 7 photos using ImageJ.

Chondrogenic differentiation

MSCs isolated from the NWF and dish were resuspended at a density of 5 × 10⁵ cells/ml in the chondrogenic induction medium, the DMEM-LG medium containing 3.5 g/l glucose (Wako), 50 μg/ml l-ascorbic acid phosphate magnesium salt n-hydrate (Wako), 100 μg/ml pyruvic acid sodium salt (ICN Biomedicals, Irvine, CA, USA), 40 μg/ml proline (ICN Biomedicals), 100 nM dexamethasone (MP Biomedicals, Santa Ana, CA, USA), 1% ITS + premix (BD Biosciences, San Jose, CA, USA), 10 ng/ml TGF-β3 (R&D Systems, Minneapolis, MN, USA), and 100 ng/ml IGF-1 (PeproTech, Rocky Hill, NJ, USA). The cell suspension (0.5 ml) was added to a 15 ml centrifuge tube and centrifuged at 1000 rpm for 5 min, and the collected cells were cultivated at 37 °C in 5% CO₂ for 14 days, during which the medium was replaced with a fresh medium every 3 days.

The formed cell pellets were washed with PBS, fixed in 4% PFA/PBS, and embedded in paraffin blocks. The blocks were cut into 4 μm sections, deparaffinized with xylene, and rehydrated with ethanol. The sections were washed with 3% acetic acid solution (Wako) three times before stained with alcian blue (Wako) solution. Then, pellet sections were washed with 3% acetic acid solution 3 times and washed with distilled water three times. The resulting samples were observed under an inverted microscope.

The distribution of type II collagen in the above-mentioned sections was analyzed by immunohistochemical staining. The sections were treated with 50 μl of 0.2% Triton X-100 (Wako)/PBS for 5 min. After washing with PBS three times, the sections were blocked using 1% BSA/PBS for 30 min and washed again with PBS three times. Then, the sections were incubated in 30 μl of a mouse anti-human type II collagen antibody (MilliporeSigma) for 1 h at room temperature. After washing, 30 μl of an FITC-conjugated goat anti-mouse antibody (MilliporeSigma) was added as the secondary antibody and incubated for 30 min. The sections were washed with PBS followed by Elix water. After air drying, the sections were observed under a fluorescence microscope (IX70, Olympus, Tokyo, Japan).

Quantification of ECM around cells

After the cells were cultured in the dish and NWF for various specified periods, the dish and NWF were washed with PBS(K⁺-) two times. Then, the samples were treated with 0.5% Triton X-100 containing 20 mM NH₄OH in PBS(K⁺-) for 30 s to remove cellular components from the dish and NWF (Lai et al. 2010; Chen et al. 2007), the cell-free scaffolds were then washed with PBS(K⁺-) two times and stored at − 30 °C.

The thawed samples were mixed with 0.3 ml of 0.001% trypsin in PBS(K⁺-) and digested in 5% CO₂ atmosphere at 37 °C for 40 min. Then, 0.1 ml of 4% SDS was added and the resulting solution containing trypsin and decomposed ECM was recovered into a tube. The protein concentration in the recovered solution was determined using a Micro BCA Protein assay kit (Thermo Fisher Scientific). Finally, the amount of ECM was calculated by subtracting the amount of trypsin from total amount of protein.

Osteogenic differentiation on the dish coated with ECM

MSCs were inoculated at several densities (0, 0.25, 0.5, 1 × 10⁴ cells/cm²) in 6-well plates and cultured using the growth medium containing 50 mM ascorbic acid for 3 days (Rakian et al. 2015). After removing the cells using 0.5% Triton X-100, the amount of ECM was determined, or by washing the wells with PBS(K⁺-) two times for inoculate the flesh cells. For a parallel experiment, 0.3 ml of type 1 collagen solution (0, 17, 34, and 170 μg/ml) was added to other 6 well plates and the plates were dried in a clean bench overnight resulting in 0, 0.53, 1.06, and 5.3 μg/cm² collagen coating the plates, respectively.

MSCs were inoculated at a density of 0.01 × 10⁴ cells/cm² on the above decellularized ECM wells and collagen-coated wells and cultivated using the growth medium for 21 days. The medium was changed with a fresh one every 7 days. After 21 days, the cells were harvested using trypsin and inoculated to new 6-well plates at a density of 0.8 × 10⁴ cells/cm². After 1 day of incubation for cell adhesion, the medium was changed with the osteogenic medium. The cells were further incubated for 12 days, after which the osteogenic differentiation potential of MSCs was analyzed.

Results

Percentages of CD90 (+) and CD166 (+) cells

The percentages of CD90 (+) and CD166 (+) cells among the inoculated cells (Fig. 1a) and cells grown for 21 days on the dish (Fig. 1b) and in NWF disc (Fig. 1c) were determined (Table 1). The percentages of CD90 (+), CD166 (+), and CD90/CD166 (+) cells among the inoculated cells were high (98.2, 99.6, and 98.3%, respectively). The percentage of CD166 (+) cells among the cells cultivated for 21 days on a dish was still high (99.1%), and the percentages of CD90 (+), and CD90/CD166 (+) cells were low (67.9 and 67.9%, respectively). On the other hand, percentages of CD90 (+), CD166 (+), and CD90/CD166 (+) cells among cells cultivated for 21 days in NWF were all high (97.9, 98.3, and 97.2%, respectively) and almost the same as those among the inoculum cells. Thus, the percentage of CD90 (+) cells was maintained at a high level during the culture in NWF, whereas that among cells cultivated on the dish markedly decreased.

Fig. 1.

Flow cytometry analysis of CD90 (+) and CD166 (+) of MSCs. Percentages of CD90 (+) and CD166 (+) cells after preculture on the dish (a), after main culture on dish for 21 days (b), and after main culture in the NWF for 21 days (c)

Table 1.

Percentages of CD90 (+) and CD166 (+) cells

	CD90 (+) (%)	CD166 (+) (%)	CD90/CD166 (+) (%)
Preculture
Dish	98.2	99.6	98.3
Main culture
Dish	67.9	98.1	67.9
NWF	97.9	98.3	97.2

Comparison of differentiation potential between cells grown in NWF and on dish

To compare the differentiation potentials of MSCs grown on the dish and in NWF, cells cultured on the dish and in NWF for 21 days were recovered by trypsinization and induced with osteogenic, adipogenesis, and chondrogenic induction medium. The induced cultures were stained with Alizarin Red S (Fig. 2a–d), Oil Red O (Fig. 2e–h), and Alcian Blue (Fig. 2i–k) and by immunohistochemical staining of type II collagen (Fig. 2l–n). The staining results were analyzed using ImageJ and shown in Fig. 2o, p.

Fig. 2.

Differentiation of MSCs grown in dish and NWF for 21 days. Cells were grown on dish (b, f, i, l) and NWF (c, g, j, m). Uninduced MSCs were used as the negative control (a, e). Osteoblasts (d), preadipocytes (h) and chondrocytes (k, n) were used as the positive control. The osteogenic (a–d, o), adipogenic (e–h, p), and chondrogenic (i–n) differentiation potentials of MSCs were tested by Alizarin Red, Oil Red O, Alcian Blue, and immunohistochemical staining against Type II collagen. The percentages of stained areas (o) and stained cells (p) were measured using ImageJ (n = 3, average ± SD, *P < 0.05)

As shown in Fig. 2b, c, the MSCs on the NWF showed a significantly higher amount of mineralized matrix, than those cells cultured on the dish, and there was a statistically significant difference in the mean area stained with Alizarin Red S (Fig. 2o). On the other hand, the intensities of Oil Red O staining of cultures on the dish (Fig. 2f) and in NWF (Fig. 2g) were similar and very much lower than that of the control (Fig. 2h). The intensity of Alcian Blue staining of cells grown in NWF (Fig. 2j) seemed higher than that of cells grown on the dish (Fig. 2i). However, there was no such difference in the intensity of immunohistochemical staining of type II collagen between those cells (Fig. 2l, m). Thus, as shown in Fig. 2 the cells grown in NWF showed a higher osteogenic differentiation capacity than the cells grown on the dish, and there were no marked differences in adipogenic and chondrogenic differentiation capacities between them.

Validity of quantification method for ECM in culture

To confirm the possibility that the reason for the difference in the osteogenic differentiation and percentage of CD90 (+) cells mentioned above was the difference in the amount of ECM, the validity of the method of ECM quantification in culture was studied. GFP-MSCs were inoculated onto the dish and NWF at densities of 0.15 and 0.123 × 10⁴ cells/cm², respectively, and cultured for 3 days. After treatment with 0.5% Triton X-100 for 30 s, both types of scaffold were observed under a fluorescence microscope. Although a markedly large number of green fluorescent cells were observed on the dish (Fig. 3a) and NWF (Fig. 3d) before TritonX-100 treatment, the green fluorescence was invisible on the dish and NWF after treatment with Triton (Fig. 3b, e) similarly to the noninoculated dish and NWF (Fig. 3c, f). Thus, the successful removal of cells from the dish and NWF using TritonX-100 was confirmed.

Fig. 3.

Removal of cells from dish and NWF using TritonX-100. After GFP-MSCs were cultured on the dish (a, b) and NWF (d, e) for 3 days, the cells were observed under a fluorescence microscope before treatment with TritonX-100 (a, d), and after the treatment (b, e). Noninoculated dish (c) and NWF (f) were also observed

In the next step, the method of quantification of ECM remaining on the dish and NWF after the cells were removed was studied. Type I collagen solution was added to both the 6-well plate (9.9 μg/well, 0.3 ml) and NWF disc (8.2 μg/NWF, 0.01 ml), and the plate and disc were dried in the clean bench overnight. Type I collagen coating both types of scaffold was recovered using trypsin and SDS and the amount of protein in the recovered solution was determined by the BCA method. As shown in Fig. 4, the amounts of protein in the solution recovered from the well and NWF were 11.8 and 9.3 μg, indicating the recovery rates (determined amount/calculated amount of coating × 100) of 90.6 and 102.1%, respectively.

Fig. 4.

Recovery of type 1 collagen coated on dish and NWF. The left white and black columns are the calculated total amounts of collagen coating the dish and NWF, respectively, and added trypsin. The right white and black columns are the were determined amounts of protein recovered

Comparison of amount of ECM between cultures on dish and in NWF

The MSCs were inoculated in NWF (0.0245 × 10⁴ cells/cm²) and 6-well plates (0.01 × 10⁴ cells/cm²) and incubated for 21 days, during which the medium was changed with a fresh one every 7 days. At same time, before each medium change every 7 days, three wells and NWF discs were treated with Triton X-100 and stored at − 30 °C. Cell density was also determined every 7 days (Fig. 5a). Although there was a lag phase in the initial (0–7 days) culture in NWF, a logarithmic growth at almost the same rate was observed in the cultures on the dish (0–14 days) and NWF (7–14 days) (Fig. 5a). On the other hand, the growth rates in later (14–21 days) cultures (decline phase) on the dish and NWF markedly decreased compared with those in the log phase.

Fig. 5.

Time courses of cell density and ECM on dish and NWF. MSCs were cultured on the dish or NWF for 21 days. Cell density (a), ECM per area (b), and ECM per cell (c) on the dish (white) and NWF (black) were determined every 7 days. Asterisks indicate a significant difference between NWF and dish (n = 3, average ± SD, *P < 0.05)

On the 21st day, all wells and NWF discs were thawed. ECM was recovered using trypsin and SDS and quantified by the BCA method. The amounts of ECM are shown per area (Fig. 5b) and per cell (Fig. 5c).

In the first week (7 day), the amounts of ECM per area and per cell on NWF were significantly higher than those on the dish. Although ECM on the dish and NWF per area continued to accumulate in the second week, its amount was still markedly higher in the NWF. Thus, during the lag and log phases (0–14 days), cells in NWF grew markedly surrounded by a higher amount of ECM than those on the dish. During the third week of the decline phase, the increase in the amount of ECM in the NWF stopped, and the amount of ECM on the dish further increased. In the first and second weeks, the ECM amount per cell on NWF was significantly higher than that on the dish.

Amount of protein in medium adsorbed to dish and NWF

To determine the amount of protein in the medium adsorbed to the scaffold, the growth medium was added to a 6-well plate (2 ml/well; area: 9.6 cm²) and NWF discs in 24 wells of a suspension culture plate (1 ml/well; Area: 12.24 cm²), and the plates were incubated in 5% CO₂ at 37 °C for 21 days. The medium was changed with a fresh one every 7 days.

On the 7th day, only a small amount of protein adsorbed to both the dish and NWF (Fig. 6). On the 14th day, the amount of protein adsorbed to NWF was statistically significantly higher than that adsorbed to the dish. On the 21st day, protein accumulation in the NWF stopped whereas that on the dish continued to increase.

Fig. 6.

Time courses of protein density on dish and NWF without MSCs. The growth medium was added to the dish and NWF and incubated for 21 days, during which the medium was changed with a fresh one every 7 days. Protein amounts per area on dish (white) and NWF (black) were determined every 7 days. Asterisks indicate a significant difference (n = 3, average ± SD, *P < 0.05)

The difference between the amounts of protein shown in Figs. 5 and 6 may be due to the amount of ECM secreted by the cells (Fig. 7). The amount of ECM secreted by each cell in NWF was always higher than that on the dish and decreased over time (Fig. 7).

Fig. 7.

Time course of protein amount per cell. Calculated difference between protein amounts in Figs. 5 and 6. White columns, dish; black columns, NWF

Effect of amount of ECM on osteogenic differentiation potential of MSCs grown on dish

To investigate the effect of the amount of ECM on the MSC osteogenic potential, MSCs of various densities (0, 0.25, 0.5, 1 × 10⁴ cells/cm²) were cultivated and then removed with TritonX-100, which resulted in various densities of ECM coating the dishes (A, 0; B, 1.48; C, 3.41; D, 4.89 μg/cm²). Other wells were coated with collagen at various density (E, 0; F, 0.53; G, 1.06; H, 5.3 μg/cm²). MSCs were inoculated onto these 6-well plates and cultured for 21 days. After 21 days, the osteogenic differentiation of the cells was examined (Fig. 8). The stained area significantly increased as the amount of ECM and collagen increased (Fig. 8).

Fig. 8.

Effect of the amount of ECM and collagen coating on wells on MSC osteogenic potential. hMSCs cultured on wells coated with ECM (a–d, 0, 1.48, 3.41, 4.89 μg/cm²) and type I collagen (e–h, 0, 0.53, 1.06, 5.3 μg/cm²) for 21 days were induced to form osteoblasts and then stained with Alizarin Red. The percentage of stained area was measured using ImageJ (i, j). Asterisks indicate a significant difference (n = 3, average ± SD, *P < 0.05)

Discussion

It is important to maintain the high quality of MSCs during large scale culture. Important quality points of MSCs should include CD90 expression and multilineage differentiation potential such as osteogenic differentiation. CD90 expression which is one of the stem cell markers, and multi lineage differentiation potential might be some of the important characteristics determining the high quality of MSCs. MSCs grown on a dish (CD90 (+), 98.2%; CD166 (+), 99.6%) were inoculated to dishes and NWF and cultivated for 21 days.

After 21 days, the cells in NWF remained higher positive rate of CD90 (97.9%), which was proved to be related to the undifferentiated status of MSC (Moraes et al. 2016), whereas that of cells on the dish decreased to 67.9% (Fig. 1, Table 1). In addition, MSCs grown in NWF for 21 days showed markedly higher osteogenic differentiation potential than those grown on the dish (Fig. 2). Consequently, MSCs grown in NWF appeared to maintain higher osteogenic differentiation potential and CD90 expression than those grown on the dish.

The reason why the MSCs cultured in the NWF have both high osteogenic differentiation potential and percentage of CD90 (+) cells was considered. Stromal cells cultured in a NWF had 1.74-fold expression of laminin α5 mRNA compared with cells in a dish (Sasaki et al. 2003). A large amount of fibrous network like ECM was observed in MSC cultured in 3D non-woven polyvinylidene fluoride scaffolds (Schellenberg et al. 2014). These hints made us suppose that there seems to be more ECM in NWF. ECM plays vital roles in the regulation of cell proliferation and differentiation, and the maintenance of MSCs feature (Engler et al. 2006; Pandolfi et al. 2017; Anasiz et al. 2017;). In fact, MSCs cultured on ECM showed better stemness than those grown on standard cell culture dish (Lai et al. 2010; Sudhakararao et al. 2014; Rakian et al. 2015). Therefore, we developed an ECM quantification method to compare the amount of ECM between NWF and dish. The cells were lysed with Triton X-100, and cellular debris was washed away with PBS (Fig. 3). After the cells were removed, ECM was extracted consistent with references (Chen et al. 2007). However, some types of ECM extraction buffer (Turoverova et al. 2009) containing strong reducing agents such as DDT were incompatible with most protein assay kits. Therefore, we decomposed macromolecular ECM with trypsin before extraction using SDS and quantified it by the BCA method. Indeed, macromolecular ECM such as collagenI could be quantified well by this method (Fig. 4).

The amounts of ECM around MSCs cultured on the dish and NWF were determined by this method. In the lag and log phases (0–14 days), both the amounts of ECM per area (Fig. 5b) and per cell (Fig. 5c), in NWF were higher than those on the dish. In contrast, in the decline phase (14–21 days), the amount of ECM was higher on the dish than in NWF. Because those trends were the same as those observed for the amount of ECM adsorbed on the dish and NWF (Fig. 6), the differences in ECM amount between the dish and the NWF (Fig. 5b, c) might be due to adsorption rather than the cell activity for ECM production.

The effect of the amount of ECM on osteogenic potential was investigated (Fig. 8). The results strongly suggested that the higher the amounts of ECM and typeIcollagen, the higher the osteogenic differentiation potential of MSCs (Fig. 8i, j). Therefore, the MSCs have higher osteogenic differentiation potential when the amount of ECM is increased. Thus, the MSCs grown on the dish gradually lost their stemness in the low-ECM environment in the first two weeks of culture. Although the amount of ECM was increased in the third week, the lost stemness could not be recovered (Bonab et al. 2006). On the other hand, the MSCs in the NWF were always grown in a high-ECM environment. This may be one of the reasons why MSCs in the NWF maintained their higher osteogenic differentiation potential and CD90 expression levels.

The difference in the ECM amount between the medium with MSCs culture (Fig. 5) and the medium incubated without MSCs (Fig. 6) was considered to be due to the ECM produced by cells (Fig. 7). The ECM (protein) productivity by cells was higher in the NWF than on the dish throughout the entire culture period, which may be due to that the 3D porous structure of the NWF being similar to the environment in the marrow (Hoch and Leach 2014; Xu et al. 2016).

Consequently, MSCs grown in the NWF scaffold maintained a markedly higher percentage of CD90 positive cells and osteogenic differentiation potential than cells grown on the dish. This might be due to the higher amount of ECM around cells in the NWF than that on the dish. This difference might be true even using the larger size of dish. NWF culture may provide a great amount of high-quality MSCs with improved efficiency for clinical applications.