Size-controlled human adipose-derived stem cell spheroids hybridized with single-segmented nanofibers and their effect on viability and stem cell differentiation

Background Fabrication of three-dimensional stem cell spheroids have been studied to improve stem cell function, but the hypoxic core and limited penetration of nutrients and signaling cues to the interior of the spheroid were challenges. The incorporation of polymers such as silica and gelatin in spheroids resulted in relatively relaxed assembly of composite spheroids, and enhancing transport of nutrient and biological gas. However, because of the low surface area between cells and since the polymers were heterogeneously distributed throughout the spheroid, these polymers cannot increase the cell to extracellular matrix interactions needed to support differentiation. Methods We developed the stem cell spheroids that incorporate poly(ι-lactic acid) single-segmented fibers synthesized by electrospinning and physical and chemical fragmentation. The proper mixing ratio was 2000 cells/μg fibers (average length of the fibers was 50 μm - 100 μm). The SFs were coated with polydopamine to increase cell binding affinity and to synthesize various-sized spheroids. The function of spheroids was investigated by in vitro analysis depending on their sizes. For statistical analysis, Graphpad Prism 5 software (San Diego, CA, USA) was used to perform one-way analysis of variance ANOVA with Tukey’s honest significant difference test and a Student’s t-test (for two variables) (P < 0.05). Results Spheroids of different sizes were created by modulating the amount of cells and fibers (0.063 mm2–0.322 mm2). The fibers in the spheroid were homogenously distributed and increased cell viability, while cell-only spheroids showed a loss of DNA contents, internal degradation, and many apoptotic signals. Furthermore, we investigated stemness and various functions of various-sized fiber-incorporated spheroids. In conclusion, the spheroid with the largest size showed the greatest release of angiogenic factors (released VEGF: 0.111 ± 0.004 pg/ng DNA), while the smallest size showed greater effects of osteogenic differentiation (mineralized calcium: 18.099 ± 0.271 ng/ng DNA). Conclusion The spheroids incorporating polydopamine coated single-segmented fibers showed enhanced viability regardless of sizes and increased their functionality by regulating the size of spheroids which may be used for various tissue reconstruction and therapeutic applications. Supplementary Information The online version contains supplementary material available at 10.1186/s40824-021-00215-9.


Background
Stem cells have been evaluated in tissue engineering due to their multipotent ability to differentiate into diverse tissue lineages, for example osteoblasts, neuronal cells, hepatocytes, and adipocytes [1,2]. Adult stem cells (ASCs) can be isolated from various organs and tissues including bone marrow, turbinate, periosteum, and synovial membranes, and have been widely utilized for cellbased therapeutic approaches [3][4][5]. In particular, human adipose-derived stem cells (hADSCs) are an attractive source due to their ability to secrete diverse paracrine factors and their rapid doubling time of less than 70 h [6][7][8]. hADSCs have been used with direct injection [9,10], in combination with scaffolds [11,12] and as cell sheets [13,14] to target tissues, such as skin and muscles. However, directly delivering stem cells without any scaffolds or cell related interactions are easily cleared or trapped in capillaries, limiting their therapeutic efficacy [15,16]. Also, use of scaffolds can often activate potential inflammation and may reduce cell-cell interactions, leading to limited stemness [17][18][19]. Furthermore, mono-layered cell sheets have shown to be difficult to handle because of weak mechanical properties, and limit control of spontaneous and direct stem cell differentiation [20][21][22]. In addition, thick, multilayered cell sheets limit oxygen and nutrient diffusion and the stem cells are unable to survive in vivo without blood vessels [23].
Alternatively, fabrication methods for three-dimensional (3D) cell spheroids have been studied as 3D microenvironment for cells. For instance, methods such as hanging drop, round micro-well, centrifugation, and a spinner flask with a bioreactor were mainly used for preparing spheroids [24][25][26][27]. Although these spheroids had improved stem cell function, there are still challenges in tissue engineering due to the hypoxic core and limited penetration of nutrients and signaling cues to the interior of the spheroid. By the reason, most of the spheroid diameters were less than 200 μm because larger-sized spheroids showed severe hypoxia from diffusion limitations [26,28]. Severe hypoxia induces apoptosis, and it may limit the spheroid viability and survival time. So it was hard to expect the long survival for differentiation such as osteogenic or chondrogenic lineage which were needing at least 14 or 21 days of culturing under differentiation medium. The dynamic cultures in bioreactors could attenuate the hypoxia in spheroids and improve their viability, however, the fabricated spheroids showed heterogeneous sizes [29]. Also, spheroids composed only of cells could maintain their stemness because of high cell-cell interactions, but less cell-extracellular matrix (ECM) interactions may limit some differentiation, especially osteogenic differentiation in which cell-ECM interactions are an important factor [30].
The incorporation of polymers such as silica and gelatin in spheroids resulted in relatively the larger and relaxed assembly of composite spheroids because the polymer improved the viability of the spheroids by enhancing transport of nutrient and biological gas. However, because of the low surface area between cells and since the polymers were heterogeneously distributed throughout the spheroid, these polymers cannot increase the cell-ECM interactions needed to support differentiation [31][32][33]. Our previous study dealt with these issues by using homogeneously distributed fragmented nanofibers in spheroids and was successful in preventing hypoxia-related cell death and increased cellular function [26]. As the volume of spheroids increased, hypoxia was increased due to oxygen diffusion limitations and result in reducing osteogenic differentiation while increasing the release of angiogenic factors. These phenomenon have been demonstrated only in 2D cultures and not with a 3D culture system [34,35]. Previous spheroids that had standardized sizes have been investigated regardless of the relation between the size and hypoxia induced stem cell reaction. Therefore, spheroids were showed limited sizes and functionally limited to one therapeutic outcome or differentiation; the spheroids could not be used in multi-functional approaches [36][37][38]. To successfully mimic the 3D microenvironment of the body, we needed to understand the different biological reaction of cells on spheroid from the change of sizes causing different hypoxia condition and customize the spheroids for diverse therapeutic approaches.
In this study, we synthesized single fragmented fibers (SFs) from an electrospun nanofiber sheet, hypothesizing that this method could increase spheroid viability, and prevent severe hypoxia, apoptosis, and diffusion limitations. To prove our hypothesis, we (1) evaluated the differences in spheroid morphology and viability and (2) investigated the difference of growth factor secretion and differentiation comparing variably-sized spheroids.

Preparation and characterization of single-segmented fibers (SFs)
Thin nanofiber sheets were prepared by electrospinning using 10 ml of 4% PLLA in dichloromethane (DCM) and trifluoroethanol (TFE) (8,2, v/v). The solution was injected into a rotating collector (200 rpm) through a 23-G blunt end needle under 11-12 kV at 5 ml/hr. The electrospun sheet was then dried at room temperature overnight. To prepare the SFs, the sheet was cut into 5 × 5 mm pieces, which were subsequently treated with 10% (v/v) ethylenediamine (EDA) solution in isopropyl alcohol (IPA) and mixed for 30 min at 200 rpm on a parallel rotator. Fibers were then collected by centrifugation at 3000 rpm for 5 min and washed three times in IPA, once in 70% ethanol (EtOH), and three times in distilled water. The washed SFs were freeze-dried for 24 h. Fieldemission scanning electron microscopy (FE-SEM) (AUR-IGA, Carl Zeiss, Germany) was used to observe the morphology of the SFs. In order to improve cell adhesion, SFs were rinsed once in 70% EtOH, three times with DW, and then immersed in a dopamine hydrochloride solution (2 mg/ml, 10 mM Tris-HCl buffer, pH 8.5) under mild shaking at 50 rpm for 10 min. Then, polydopamine-coated SFs (P-SFs) were washed three times with DW at 70 rpm for 10 min and lyophilized. The surface atomic compositions of SFs and P-SFs were analyzed with X-ray photoelectron spectroscopy (XPS) (Theta Probe base system; Thermo Fisher Scientific, Waltham, MA, USA).

Culture of human adipose-derived stem cells
The hADSCs were purchased from Invitrogen (StemPro; Carlsbad, CA, USA). The cells were cultured under standard culture conditions (5% CO 2 and 37°C) in minimum essential medium, MesenPRO RS™ basal medium (Carlsbad, CA, USA) with 2% growth supplements (Carlsbad, CA, USA), 1% penicillin/streptomycin (P/S), and 1% L-glutamine. The cells were incubated in a 175 T culture flask and the medium was refreshed every 3 days. For the study, hADSCs that had attached to the bottom of the culture dish were detached by using 4 ml of 0.05% trypsin-ethylenediaminetetraacetic acid (EDTA). Cells surviving after the fourth passage were used for all experiments.

Viability of cells within the spheroids
Live and dead assay kit was used to investigate the viability of cells within the spheroids at days 1 and 7. A working solution containing calcein AM (11000) and ethidium homodimer (1500) in DPBS were added to the spheroids and incubated for 20 min. Next, images were captured via fluorescence microscopy (TE2000) (Tokyo, Japan). The sizes of spheroids were measured by Photo-Shop CS6 software (Adobe) (San Jose, CA, USA) from the captured images. DNA assays was performed to assess cell proliferation in the spheroids. The spheroids were collected at each time point, lysed in 100 μL of radioimmune precipitation assay (RIPA) lysis buffer (150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 150 mM Tris, pH 7.2), and detected with the Quant-iT Picogreen dsDNA Assay kit according to the manufacturer's instructions. Fluorescent intensity was measured using a spectrometer (VARIOSKAN LUX, Thermo Fisher Scientific) (Vantaa, Finland) with excitation and emission wavelengths of 480 to 520 nm.

Histological analysis of the spheroids
The microscale images of the spheroid surface and interior were visualized using the sectioned spheroids. Transmission electron microscopy (TEM; LIBRA 120, Carl Zeiss, Oberkochen, Germany) was performed to visualize cell-fiber interactions within the spheroid. Briefly, spheroids cultured for 2 and 7 days were fixed in modified Karnovky's fixation solution (2% paraformaldehyde and 2% glutaraldehyde in 0.05 M sodium cacodylate buffer, pH 7.2) and 1% osmium tetroxide (in 0.05 M sodium cacodylate buffer, pH 7.2) for 2 h each. The fixed spheroids were then treated with 2% uranyl acetate for membrane staining. Serial dehydration of fixed spheroids was carried out by incubating in increasing concentrations of EtOH (30 to 100%) followed by a transition step consisting of a 15 min treatment with 100% propylene oxide at room temperature. Samples were then embedded in Spurr's resin at 70°C for 24 h. Polymerized blocks of Spurr's resin were sectioned with an ultra-microtome (MT-X; RMC, Tucson, AZ, USA). The spheroids were analyzed using TEM.

Diffusion of fluorescent dye into the spheroids
For staining of the spheroid nuclei, Hoechst pentahydrate (bis-benzimide) (Invitrogen, Oregon, USA) nuclear staining dye was diluted in PBS to a 1 to 200 ratio. The solution (100 ml) was used to treat each spheroid for 2.5 h. The stained spheroids were cross-sectioned and images of the stained area were captured by confocal microscope (TCS SP5, Leica) (Hessen, Germany). For quantification, images of the whole area and unstained area of sectioned spheroids were analyzed by Photoshop CS6 to calculate the ratio of stained area.
Analysis of the stemness of the hADSCs within the spheroids Immunofluorescence (IF) staining and quantitative polymerase chain reaction (qPCR) were performed to assess expression of stemness markers, octamer-binding transcription factor 4 (OCT4), sex determining region Y-box 2 (SOX2), and NANOG. The spheroids were evaluated on day 7. For IF staining, cryo-sectioned samples were fixed in 4% paraformaldehyde for 15 min followed by a 1 h treatment with blocking buffer (5% FBS, 0.1% Tween-20 in PBS). Subsequently, samples were incubated at 37°C with primary antibody (1:100) overnight and then incubated with anti-mouse IgG biotin conjugate (1:200) for 1 h. This was followed by treatment with FITCconjugated streptavidin (1:200) for 1 h. IF stained samples were then mounted using DAPI-containing mounting media and observed under a fluorescence microscope. For gene expression analysis, mRNA was extracted using a QIAGEN RNA extraction kit (QIAG EN, Hilden, Germany) following manufacturer's instructions. RNA concentrations were estimated by measuring the absorbance at 260 nm using a nanospectrometer (Nanodrop 2000; Thermo Scientific, Wilmington, DE, USA). The cDNA was synthesized from 1 μg of RNA using a Maxime RT PreMix kit from Intron Biotechnology (Seoul, Korea) and was processed in a Bio-Rad Thermocycler (Bio-Rad Laboratories, Hercules, CA, USA). Before RT-PCR was performed, 20 μl of cDNA solution, which was diluted with Ambion® DEPC-treated water (Life Technologies, Carlsbad, CA, USA), was mixed with 10 μl of SYBR® Premix Ex Taq™ (2X) Tli RNaseH Plus (Takara, Kusatsu, Japan), 0.4 μl of 10 pmol primers, 0.4 μl of ROX reference dye (50X), and 6.8 μl of DEPC-treated water. The sample solutions were processed in an RT-PCR StepOnePlus™ instrument (Life Technologies, Carlsbad, CA, USA). The amplification reaction was performed with denaturation at 95°C for 10 min, followed by 40 cycles of annealing at 95°C for 15 s and extension at 60°C for 1 min. A melting curve stage was performed from 60.0°C to 95.0°C in increments of 0.5°C per 5 s. To investigate fold changes of the spheroids, the hADSCs were cultured on tissue culture plate as a monolayer in parallel and then used as a control group. All reactions were conducted in triplicate. Primer sequences are listed in supporting file.

LOX-1 staining and VEGF secretion
For LOX-1 staining, the working reagent (2 μmol LOX-1/L in growth media) was prepared and the spheroids were treated for overnight under standard culture conditions. To examine VEGF secretion, spheroids was cultured for 1, 4, and 7 days. Supernatants from each time point were collected and stored at − 20°C. After collection, an enzyme-linked immunosorbent assay (ELISA) was performed using a VEGF ELISA development kit (PeproTech, Rocky Hill, NJ, USA) according to manufacturer's instructions. RT-PCR for apoptosis-related genes (B-cell lymphoma 2 (BCL-2) and Bcl-2-associated X protein (BAX)) was performed following the same procedure as described previously. All primer sequences are listed in supporting file.

Osteogenic differentiation
Spheroids were cultured under osteogenic media for 14 days to assess differentiation. For visualization of the deposited calcium minerals, the spheroids were fixed with 4% paraformaldehyde, treated with Alizarin Red S (Sigma-Aldrich) (2% in DW with pH 4.2), and the images were captured by the microscope. The calcium ions were extracted by incubating the spheroids for overnight in 0.6 N HCL, and then the calcium assay was then performed according to the QuantiChrom Calcium Assay kit's instructions. Quantitative PCR analysis of the osteogenic gene markers; runt-related transcription factor 2 (Runx2), osterix (OSX), osteocalcin (OCN), and osteopontin (OPN) was performed as following the same PCR protocols used for investigation of the stemness markers, as described in the previous section. Primers are listed in Table 1. To investigate fold changes of the spheroids, the hADSC monolayer was used as a control group.

Statistical analysis
Quantitative data are expressed as means ± standard deviation. Graphpad Prism 5 software (San Diego, CA, USA) was used to perform one-way analysis of variance (ANOVA) with Tukey's honest significant difference (HSD) test and a Student's t-test (for two variables). Statistical significance was denoted as P < 0.05. All quantitative values were calculated from at least three samples (n ≥ 3). Experiments were independently replicated at least three times. We found that the trend of all results was the same each time.

Results
Multi-cellular spheroids from stem cells have been actively investigated in tissue engineering, however, the relatively hypoxic environment within the core may be detrimental to cell survival, which often limits their universal application. We prepared various sizes of hADSC spheroids by incorporating an artificial ECM from celladhesive single-segmented fibers (SF) and examined stem cell behavior within the spheroid. The SFs were coated with cell adhesive polydopamine (P-SFs) and co-assembled with hADSCs to form spheroids (Fig. 1). The objectives of this study were (1) to examine the effect of differing amounts of hADSCs and P-SFs on spheroid size, (2) to investigate the effect of P-SF on the viability of the spheroids, and (3) to investigate the effect of the spheroid size on secretion of pro-angiogenic factors and osteogenic differentiation.
Preparation and characterization of the single-segmented fibers SEM images showed the morphology of individual SFs (Fig. 2a) and the aggregated network (Fig. 2b). The length with 50 to 100 μm was selectively chosen by sieving and the average length of fiber was 68 ± 2 μm (Supplementary Figure 1). of the surface chemical composition using XPS revealed that the high-resolution carbon spectrum analysis revealed that the peak for the C-N bond (286.0 eV) on the surface of the fibers was 0.0096% in the SF group and 6.66% in the P-SF group ( Fig. 2c and d), which indicated that the SFs were successfully coated with polydopamine.

Preparation of size-controlled hADSCs spheroids
After synthesizing the P-SFs, we prepared the cellonly spheroids (CS-1, CS-2, and CS-4) and the fiber incorporated spheroids (FS-1, FS-2, and FS-4). The phase contrast images of the spheroids after 24 h from the centrifugation of cells and fibers showed that all spheroids were spherical shape and all the fibers and cells were well assembled to form the spheroid (Supplementary Figure 2). The live and dead assay at day 7 showed dead cell signals in the CS groups, particularly CS-4 while all FS groups showed comparatively minimal dead cells (Fig. 3a). The size of spheroids cultured for 24 h was well-controlled by the cell number within the spheroids. The area of spheroids prepared from 40,000 cells was the largest (0.256 ± 0.027 and 0.322 ± 0.025 mm 2 for CS-4 and FS-4, respectively) while those of 10,000 cells showed the smallest area (0.078 ± 0.004 and 0.065 ± 0.002 mm 2 for CS-1 and FS-1, respectively) (Fig. 3b). The DNA contents in the CS groups were decreased while the FS groups retained their DNA content during 7 days (Fig. 3c). DNA content on day 7 normalized to day 1 was 66.31 ± 6.43% (CS-1), 72.12 ± 7.05% (CS-2), 63.28 ± 2.52% (CS-4), 93.30 ± 3.80% (FS-1), 86.16 ± 3.94% (FS-2), and 94.38 ± 2.00% (FS-4).

Histological analysis and diffusion of fluorescent dye
TEM images showed that CS-2 exhibited disintegrated membranes (DM) and empty spaces (E) (Fig. 4a). In contrast, the cell membranes (CM) in FS-2 were tightly bound with each other or with P-SFs (F), and disconnected or empty regions were minimal (Fig. 4b).
The high magnification image clearly shows polydopamine  particles (black arrows) on the surface of P-SFs and the fibers bound with adjacent cell membrane (Fig. 4c). To quantitate the difference in small molecule diffusion into the spheroid center (Fig. 4d), CS-2 and FS-2 were treated with Hoechst nuclear staining dye. The dye infiltration was much higher in FS-2, in which 65.13 ± 9.86% of the area was stained. The CS group showed only 46.63 ± 4.40% staining (Fig. 4e).

Spheroid size and stemness
To investigate stemness maintenance of viable spheroids, all FS were cultured for 7 days, and the expression of the stemness markers NANOG, OCT4, and SOX2 were evaluated. PCR analysis of the markers revealed significantly higher expression in the FS groups as compared to the cells which were seeded as mono-layered on tissue cul- , and SOX2, respectively (Fig. 5a, b and c). Despite the significant increase in stemness markers in spheroids over the group M, no significant difference in expression was found between the different spheroid sizes. In addition, high immunofluorescence signals for all markers were observed in FS-1, FS-2, and FS-4, while the group M showed less intense signal (Fig. 5d).

Pro-angiogenic factor secretion of the spheroids
To examine the oxygen supply in the spheroids, the FS groups were cultured for 7 days and treated with LOX-1. FS-4 showed more intense red signal as compared to those from the other groups; the signal intensity decreased as spheroid size decreased (Fig. 6a). VEGF secretion had a similar trend to LOX-1 staining. VEGF secretion from FS-4 was 2.43 ± 0.06 pg/ng DNA on day 7, but it was significantly lower for the group M (0.08 ± 0.06 pg/ng DNA), FS-1 (0.81 ± 0.05 pg/ng DNA), and FS-2 (0.98 ± 0.08 pg/ng DNA) groups (Fig. 6b). We then analyzed the expression of the BCL-2 and BAX genes. FS-1 showed the highest expression of BCL-2 (12.8 ± 1.09 times greater than BCL-2 expression in group M), and FS-4 showed the lowest expression of BCL-2 (4.35 ± 1.13 times greater than BCL-2 expression in group M). In contrast, BAX expression was reversed. With the exception of the control group M, FS-4 showed the highest BAX expression (0.44 ± 0.19 times lower than BAX expression in group M), and FS-1 showed the lowest BAX expression (0.010 ± 0.0029 times lower than BAX expression in group M) ( Fig. 6c and d).

Osteogenic differentiation of the spheroids
Finally, we investigated the osteogenic differentiation of FS via PCR analysis, a calcium assay, and staining. Alizarin red S staining resulted in mineralized calcium appearing as a reddish color and was more intense in FS-1 than in FS-4 (Fig. 7a). The quantification of mineralized calcium content revealed that the smaller spheroid contained significantly more calcium minerals (18.1 ± 0.27, 14.3 ± 0.28 and 10.6 ± 1.35 ng/ng DNA for FS-1, FS-2, and FS-4, respectively) (Fig. 7b). Furthermore, the expression of the initial and intermediate differentiation markers RUNX2 and OSX were upregulated in the FS groups as compared to the control group (Fig. 7c) (Fig. 7d).

Discussion
The aims of this study were to prepare various sizes of hADSC spheroids with ECM-mimicking SFs and to investigate the functional differences of stem cells. To prepare the spheroids, segmented fibers were selected ranging from 50 to 100 μm in length. When the fiber length was too long, the spheroid was not compact and loosely assembled, which resulted in inefficient spherical shape formation. As shown in Fig. 4b and c, P-SFs were homogenously distributed with cells and tightly incorporated with cells. Furthermore, FS could maintain its initial sphere shape through day 7, while previously studied polymer-containing spheroids [39,40] could not maintain their initial shapes. These observations demonstrate that the P-SFs were appropriate for maintaining the overall spheroid shape and cell binding, akin to the role of the ECM in natural tissues [41,42].
To investigate various cellular functions, we used hADSCs, which have the potential to differentiate into various cell types including osteocytes [43], adipocytes [44], chondrocytes [45], and neuronal cells [46]. In addition, hADSCs have a quicker doubling time than other adult stem cells. To stably assemble hADSCs with the fibers, the SFs were coated with polydopamine in order to enhance cell binding affinity. Previous reports have demonstrated that dopamine spontaneously polymerizes in a weak basic solution on the surface of diverse scaffolds via hydrophobic interaction and π-π stacking. This polydopamine layer allowed cell adhesion and absorption of serum proteins on the coated surface [47,48]. When fibers exhibited insufficient cell adhesion properties, the spheroids would easily dissociate or take on an irregular shape during long term culture. For example, one previous study incorporating poly(lactic-coglycolic acid) (PLGA) fragments with human embryonic kidney 293 cells which were cultured for 24 h showed many unmixed fibers outside of the spheroids [49]. The weak interactions between the fibers and cells within the spheroid may limit cell functions, for example mineralization, myofibroblast differentiation, cell migration, and proliferation [50,51]. The P-SFs succeeded in binding with cells and providing space for oxygen and nutrient diffusion into the spheroids.
As demonstrated from Live/Dead and DNA assays (Fig. 3a and c), the CS groups showed several dead cell signals, and their DNA content was decreased after 7 days, while the FS groups successfully maintained the stem cell viability. A previous report had demonstrated that the internal cell death from apoptosis resulted in disconnected vacant space, and thus, spheroid size decreased and its morphology was changed [52], which was also confirmed on the TEM image of CS (Fig. 4a). Furthermore, the CS surface was arranged more compactly than that of FS (Fig. 4d), and this compact arrangement may limit penetration of biological gases and nutrients, leading to cell death [53]. Awadhesh N. Jha et al. had previously evaluated oxygen distribution within the cell spheroid, suggesting that there are discrete zones composed of outside, middle, and inside spheroid regions along the oxygen supply from high to low, exhibiting proliferating, quiescent viable and apoptotic core property, respectively. When a spheroid size is increased, the hypoxic zone would increase dramatically. For example, the spheroid with 350 ± 117 μm radius showed a 273 μm hypoxic zone and only a 77 μm viable zone [54]. Similarly, previous studies have reported that hypoxic conditions could cause loose of cell-cell binding, and the cell membrane degradation provides evidence supporting apoptosis [55][56][57][58]. Collectively, our results suggested that the presence of P-SFs within the spheroid improved the stem cell viability by mitigating the diffusion limitation of crucial survival factors including oxygen. To be useful for stem cells, stemness should be wellmaintained during the expansion period, and differentiation should be precisely controlled upon application [59][60][61]. As shown in Fig. 5, the 3D cultured FS spheroids showed the higher expression of stemness marker than the 2D cultured stem cells. This difference may be attributed to the complex microenvironment of cell aggregates recreated by the 3D cultured spheroid, which enhanced cell-to-cell interactions and induced cytoskeletal cellular reorganization, maintaining their multipotent stemness [37]. Similar results were prevalent for other types of stem cells. For instance, spheroids composed not only with mesenchymal stem cells (MSCs) increased gene expression of diverse differentiation markers (angiopoietin, bone morphogenetic proteins, and transforming growth factor beta-3), but also spheroids composed of human umbilical cord stem cells showed up-regulation of stem cell markers, and adipogenic or osteogenic differentiation makers than those of 2D cultured cells [37,62]. Similarly, our results demonstrated that FS-1, FS-2, and FS-4 significantly preserved stemness as compared to 2D cultured cells over the 7 days, regardless of their sizes. All FS groups had the same ratio of hADSCs to the amount of P-SFs along with homogeneous assembly which caused similar levels of cellular interactions and that may have led to similar stemness gene expression.
The spheroid cores of larger size may be subjected to low oxygen supply due to diffusion limitation (Fig. 6a). Our results showed that spheroids released greater amounts of pro-angiogenic growth factors as size increased (Fig. 6b). BAX expression showed a similar trend. These are consistent with previous studies that adipose-, bone marrow-, and umbilical cord-derived stem cells show greater secretion of pro-angiogenic factors and apoptotic gene expression under hypoxic conditions [63,64]. Additionally, all spheroids revealed greater osteogenic marker expression than was seen in 2D cultured hADSCs (Fig. 7c and d). These results may be related to differing levels of oxygen supply in 2D cultured cells and spheroids. According to previous studies, the presence of stem cells under low oxygen availability stimulated expression of diverse growth factors such as VEGF and TGF-beta, which could increase osteogenic differentiation [35]. The factors significantly enhanced osteogenic differentiation in the spheroid groups as Fig. 6 The secretion of angiogenic factor from the FS groups. a LOX-1 staining from FS-1, FS-2, and FS-4 (white arrow indicates a highly stained region in FS-4, scale bar: 200 μm). b The secreted amounts of VEGF from the all groups were quantified by ELISA ("*", "#", and " §" indicate significant differences between group M, FS-1, and FS-2, respectively, for each time point). PCR analysis of the expression of (c) BCL2 and (d) BAX genes in all groups compared to the 2D cultured cells. Among the spheroid groups, there were no differences in the initial and intermediate osteogenic differentiation markers RUNX2 and OSX, regardless of spheroid size; whereas late osteogenic markers OCN and OPN, which are related to mineralization and mineralized calcium content, showed significantly greater expression in the smallest spheroid (FS-1). This phenomenon may be related to different hypoxia levels of the spheroids. Previous studies have demonstrated that more severe hypoxic condition (2% oxygen) increases the expression of insulin growth factor binding protein (IGBP) in ADSCs and blocks insulin growth factor-1 (IGF-1) binding to its receptor (IGFR) on the cell membrane. This IGF-1 and IGFR binding has been shown to have important role in stem cell mineralization [34,65]. The core of large sized spheroids (> 350 μm of radius) underwent similar level of hypoxia because of limited oxygen supply [54] despite the presence of engineered fibers (Fig. 6a) comparing with smaller sized one, so the smallest spheroid (FS-1) which were under the lowest hypoxia could show the highest capacity of osteogenic differentiation specifically amount of mineralized calcium contents and expression of OCN and OPN which were related with mineralization signaling pathway because FS-1 was more favored the IGF-1 and IGFR binding than the larger spheroids.
The spheroids described here that incorporate P-SFs address limitations of previous spheroids, specifically limited oxygen diffusion, viability, and cell-to-ECM interactions. Due to this approach, the spheroids could be dramatically increased in size (up to a radius of 700 μm). The variety of the sizes make it possible to support different stem cell functions, such as angiogenic factor secretion and osteogenic differentiation. Taken together, these spheroids may have the potential to be used for customized treatments.

Conclusion
In this study, we prepared stem cell spheroids of various sizes by incorporating P-SFs and studied stem cell function within those spheroids. We demonstrated that the surface of fiber successfully coated with polydopamine, and analysis of the viability and histological experiments showed that the incorporation of fibers in a spheroid Fig. 7 Osteogenic differentiation of the FS groups. a Alizarin red S staining (scale bar: 100 μm) and (b) a calcium assay ("*", "#", and " §" indicate significant differences between groups M, FS-4 and FS-2, respectively) were performed to investigate the amounts of mineralized calcium. c PCR analysis of osteogenic differentiation markers (c) RUNX2, OSX, (d) OCN, and OPN on day 7 ("*", "#", and " §" indicate significant differences between groups M, FS-2, and FS-4, respectively) significantly enhanced stem cell viability by partially solving the diffusion limitation (FS showed 65.13 ± 9.86% dye infiltration, while CS showed 46.63 ± 4.40%). Also, we synthesized three different-sized spheroids using differing amounts of P-SFs and cells. The spheroids demonstrated similar expression of stemness markers, however, their differentiation capacities varied due to differences in hypoxia levels. The smallest FS showed 18.10 ± 0.27 times higher calcium content than the control groups; and higher osteogenic differentiation marker expression than larger-sized spheroids while released the lower amount of angiogenic factor. In contrast, the largest FS showed the lowest osteogenic differentiation capacity, but released the greatest amounts of VEGF (2.43 ± 0.06 pg/ng DNA) as compared to the control group (0.08 ± 0.06 pg/ng DNA). Taken together, our results suggest that biocompatible P-SFs can modulate the size of spheroids, make long-term spheroid culture possible, and may be important for tissue regeneration and therapeutic applications.
Additional file 1: Supplementary figure 1. The initially synthesized single-segmented nanofibers were sieved to selectively choose the fibers with 50 to 100 mm, and the length of selected fibers were measured (n= 10 to each range). Supplementary figure 2. The phase contrast images of spheroids after 24 hr from the centrifugation of cells and fibers.