Improving motor neuron-like cell differentiation of hEnSCs by the combination of epothilone B loaded PCL microspheres in optimized 3D collagen hydrogel

Spinal cord regeneration is limited due to various obstacles and complex pathophysiological events after injury. Combination therapy is one approach that recently garnered attention for spinal cord injury (SCI) recovery. A composite of three-dimensional (3D) collagen hydrogel containing epothilone B (EpoB)-loaded polycaprolactone (PCL) microspheres (2.5 ng/mg, 10 ng/mg, and 40 ng/mg EpoB/PCL) were fabricated and optimized to improve motor neuron (MN) differentiation efficacy of human endometrial stem cells (hEnSCs). The microspheres were characterized using liquid chromatography-mass/mass spectrometry (LC-mas/mas) to assess the drug release and scanning electron microscope (SEM) for morphological assessment. hEnSCs were isolated, then characterized by flow cytometry, and seeded on the optimized 3D composite. Based on cell morphology and proliferation, cross-linked collagen hydrogels with and without 2.5 ng/mg EpoB loaded PCL microspheres were selected as the optimized formulations to compare the effect of EpoB release on MN differentiation. After differentiation, the expression of MN markers was estimated by real-time PCR and immunofluorescence (IF). The collagen hydrogel containing the EpoB group had the highest HB9 and ISL-1 expression and the longest neurite elongation. Providing a 3D permissive environment with EpoB, significantly improves MN-like cell differentiation and maturation of hEnSCs and is a promising approach to replace lost neurons after SCI.


Materials and methods
Isolation and culture of human endometrial stem cell (hEnSC). hEnSCs were isolated according to a previous study 56 . The collection of discarded endometrium biopsies from patients was achieved via the informed consent from the patients/legal guardian. All methods were performed in accordance with the relevant guidelines and regulations of Tehran University of Medical Sciences and approved by the university's ethical committee (code: IR.TUMS.REC.1394.1137). Briefly, biopsies of endometrium were washed with phosphatebuffered saline (PBS; Sigma, P4417) containing 2-3% amphotericin B and penicillin/streptomycin, were cut into small pieces and digested with collagenase type I (1 mg/mL, Sigma-Aldrich) in Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 (DMEM-F12; Invitrogen) for 2 h at 37 °C. The enzyme was then neutralized by adding a complete medium composed of DMEM-F12 with 10% fetal bovine serum (FBS; Gibco, 10270-106). Subsequently, the samples were passed through 70 μm and 40 μm filters and were centrifuged at 1200 rpm for 5 min. After the supernatant was removed, the pellet of cells was resuspended in a complete medium containing 10% FBS and 1% penicillin/streptomycin. The cell suspension was transferred to 6-well cell culture plates and incubated in a humidified chamber at 37 °C with 5% CO 2 . The medium was exchanged every 3 days. Three or four passages of hEnSCs were used for all of our experiments.

Flow cytometry.
To confirm the identity of the isolated cells, passage three cells were used for flow cytometry BD FACSCalibur (BD Biosciences, USA). The evaluated markers were as follows: endometrial stem cell Microsphere fabrication and sterilization. The oil in water (o/w) single emulsion method was used to synthesize the EpoB-loaded PCL microspheres. First, 250 mg of polycaprolactone (PCL; Sigma, 363170) was dissolved in 2 mL of dichloromethane (DCM; Sigma, 32222) using a magnet stirrer (Heidolph, Hei-Tec) for 25 min at 35 °C. Three groups of EpoB-loaded PCL microspheres were fabricated by adding 0.625 µg, 2.5 µg, and 10 µg of EpoB (Abcam, 152044-54-7) to the PCL solution for an EpoB loading of 2.5 ng/mg, 10 ng/mg, and 40 ng/mg (EpoB/PCL), respectively. Unloaded microspheres were synthesized through the elimination of EpoB from the formulation. 2 mL of pure ethanol (Merck, 111727) was added to the solution while stirring, and 2 mL of 2% (w/v) polyvinyl alcohol (PVA) (Sigma, 363170) solution was slowly added drop by drop without disrupting the boundary layer. Then, the solution was mixed and emulsified for 12 s using a vortexer (Heidolph) and instantly added to a 100 mL beaker containing 50 mL of 0.3% (w/v) PVA solution at 35 °C. To evaporate the organic solvent, the solution was stirred with a 3 cm × 0.5 cm round magnetic stirring bar at 500 revolutions per minute (rpm) for 4 h in a 100 mL beaker. Since EpoB is light sensitive, the entire fabrication process was performed in a dark room. It is worth noting, due to the importance of mentioning parameters such as magnetic bar's length, diameter, and the size of cylindrical container for having a reproducible particle synthesis method, these specifications were mentioned above 58 .
The synthesized microspheres were collected through centrifugation at 4000 rpm (Eppendorf, 5810R) for 10 min and were washed four times with distilled water and centrifugation at 4000 rpm for 5 min each time to remove residual PVA. Then, the microspheres were frozen at − 80 °C, lyophilized for 24-48 h, and were stored at − 20 °C. For cell culture experiments, microspheres were sterilized by air-plasma for 30 s on low power (Harrick, PDC-32G).
Microsphere characterization. Scanning electron microscopy (SEM). To evaluate the size and morphology of microspheres, a ZEISS DSM 960A Oberkochen (Germany) scanning electron microscope (SEM) was used. Microspheres were sputter-coated with gold (20 kV for 4 min), and after imaging, their diameters were quantified using ImageJ software. The distribution of microsphere diameters was plotted as a histogram. Each sample was displayed from smallest to largest diameter.
Determination of encapsulation efficiency (EE). After synthesizing 2.5 ng/mg, 10 ng/mg, and 40 ng/mg (w/w, EpoB/PCL) EpoB microspheres, the suspension of each group was centrifuged at 4000 rpm for 10 min at room temperature to separate the microparticles and supernatant. The amount of EpoB in the supernatant was evaluated as the extent of unencapsulated EpoB using liquid chromatography-mass/mass spectrometry (LC-Mas/ Mas; HPLC Alliance 269S Waters, Quattro Micro API micromass micromass) with a C18 column (4.8 × 150 mm, 5 µm, Agilent Zorbax XDB). The standard solutions of EpoB and each microsphere sample were loaded to LC-MS/MS and first eluted with methanol: water (40:60) for 5 min at the speed of 0.5 mL/min and then www.nature.com/scientificreports/ methanol:water (95:5). The retention time of EpoB was 9 min. The samples were quantified based on the standard curve of EpoB (all samples were run in triplicate). The encapsulation efficiency (EE) of EpoB in the PCL microparticle was determined as the following formula: EpoB release measurement. 10 mg of 2.5 ng/mg, 10 ng/mg, and 40 ng/mg (w/w, EpoB/PCL) EpoB-loaded microspheres were suspended in 1.5 mL of PBS in a Costar ® Spin-X ® centrifuge tube filter (0.22 µm pore CA membrane; Corning, 8160). The microspheres were placed in the above chamber containing the filter, filled with 0.5 mL PBS, and the tube filled with 1 mL PBS. These tubes were used to prevent microspheres from being mistakenly collected at the time of sampling. The Costar ® Spin-X ® centrifuge tubes were then placed on a shaker (Lab Tech, LSI-3016R) at 100 rpm and incubated at 37 °C. For the in vitro study, to simulate the conditions of exchanging the culture medium, in the release test, the PBS was exchanged every 2 days. Three samples were considered for each time point (7-time points: 1 and 18 h and 1, 3, 6, 14, and 21 days). The PBS of each sample was collected entirely at each time point. First, the PBS of the tubes was collected (not the above chamber that contained the filter), then each chamber was centrifuged with its tube at 4000 rpm for 5 min until the PBS in the chamber inter the tube and could be collected. The PBS was then stored at − 80 °C and protected from light until all the samples were collected for further analysis. The amount of EpoB was determined by using LC-Mas/Mas. In addition, to better identify the drug amount, the samples were first concentrated six times. For this purpose, each 1.5 mL samples were freeze-dried for 24-48 h and then dissolved in 200 µL deionized water and 50 µL methanol. For cell culture application, the candidate microspheres were then sterilized by air plasma (Harrick, PDC-32G) on low power for 30 s.
Determining the effect of microspheres on cell viability and proliferation. According to experimental release results, the microsphere containing 2.5 ng/mg (EpoB/PCL) was selected as the candidate microsphere for this study, referred to as Mic-2.5-EpoB in this text. To determine the biocompatible concentration of microparticles for hEnSCs culture, first, different concentrations of unloaded PCL microspheres (5, 10, 15, and 40 mg/mL) were used by indirect MTT assay for 24, 48, and 72 h. After evaluating their results, a second indirect MTT assay was performed for Mic-2.5-EpoB for 24, 48, and 72 h to assess the effect of the loaded microsphere on viability and proliferation of hEnScs. For both MTT assays, 2 × 10 5 hEnSCs were seeded in each 24-well cell culture plate containing 500 µL of complete medium and incubated at 37 °C, 5% CO 2 for 24 h. Then, the cell culture was replaced with extract solutions of the microspheres and incubated at 37 °C, 5% CO 2 for 24, 48, and 72 h. The microsphere extract solutions of the (5, 10, 15, and 40 mg/mL unloaded PCL microspheres and Mic-2.5-EpoB) in complete medium (DMEM-F12 + 10% FBS) were collected every 24 h (each group was performed in triplicate). For each time point, the medium was removed and replaced with 5 mg/mL tetrazolium (Sigma) in PBS (MTT solution) and incubated for 3 h at 37 °C, 5% CO 2 . The optical density (OD) of formazan dye dissolved in dimethyl sulfoxide solution (DMSO) was obtained using an ELISA microplate reader (Gen5, Power Wave XS2, BioTek, USA) at 590 nm. The following formula calculated the cell viability percent of hEnSCs: The MTT test was also used to determine the permissible particle limit in the cell culture for hEnSCs.
Fabrication of collagen hydrogels. Collagen type I was extracted from rat tails following a previously developed method 59 . Collagen was sterilized through dialysis with 1% chloroform (Merck) using dialysis bags with a molecular cut of 6-8 kDa (SpectraPor, 132660) for 1 h at 4 °C.
Collagen hydrogels were fabricated using a ratio of 8:1:1 of sterile collagen solution:10 × DMEM/F12:HSS buffer. The HSS buffer was prepared by dissolving 4.77 g HEPES and 2.2 g sodium bicarbonate in 100 mL of 0.5 M sodium hydroxide. For gelation, 500 µL of the prepared mixture was poured into 24-well cell culture plates and incubated at 37 °C for 1 h. Two final concentrations of collagen were used to fabricate the hydrogels: 2 mg/ mL and 4 mg/mL.
To improve stability of the hydrogel, collagen hydrogels were also cross-linked using different concentrations of proanthocyanidin (Grape Seed, Proanthocyanidins, Shari ® , Iran) in PBS, including 0.25% PA, 0.5% PA, and also 0.5% PA with 0.05 M Ca(OH) 2 (PA-Ca(OH) 2 ). In the last formulation, we used Ca(OH) 2 based on evidence indicating that Ca(OH) 2 facilitates the penetration of PA and collagen cross-linking 52 . After one-hour incubation of collagen solutions (2 mg/mL and 4 mg/mL) at 37 °C, 100 µL of PA solution was added, and the cell culture plates were further incubated for 48 h at room temperature in a sterile condition. Then, the hydrogels were washed with PBS three times to remove the excess crosslinkers. The specifications of hydrogel preparation are summarized by code in Table 2.
Fabrication of collagen hydrogel containing EpoB-loaded microspheres. To mix EpoB microspheres with collagen hydrogel for 3D culture application, candidate microspheres were coated with a dilute sterilized collagen solution (0.5 mg/mL) overnight at 4 °C. Before coating, the collagen solution was neutralized using a 1 N solution of filter-sterilized NaOH in water (syringe-filter 0.22-µm membrane), bringing the pH to around 7.4. Hydrogel biodegradation degree. We assessed the degradation degree of collagen, 2-Col/PA-CH, and 2-Col/PA-CH/Mic EpoB hydrogels based on their weight loss in DMEM-F12 medium at 37 °C. The DMEM-F12 medium was exchanged every 3 days. Hydrogel weight alterations were evaluated over 14 days (all time points were run in triplicate). The degradation degree was calculated using the following equation:

EE%
where, W 0 and W t indicate the weight of hydrogels before and after immersing in medium, respectively.
Swelling degree. The swelling degree of the hydrogel scaffolds, including collagen, 2-Col/PA-CH, and 2-Col/PA-CH/Mic EpoB were measured for 96 h. The hydrogels were allowed to swell in DMEM-F12 at 37 °C. Each time point was performed in triplicate. The weight of the wet samples was measured, and the swelling degree was obtained via the following formula: where, W 0 and W t are the weight of the primitive hydrogel and the weight of the swollen hydrogel, respectively.
ATR-FTIR spectroscopy. Proanthocyanidin (PA; GrapeSeed, Shari ® , Iran) blended tablets were first dissolved in water. Then the precipitated additives of the tablet were removed and filtered through a 0.22 μm membrane. The samples (PA, Collagen, 2-Col/PA, and 2-Col/PA-CH) were freeze-dried for 48 h and then analyzed using an ATR accessory (Nicolet Avatar, Thermo Fisher Scientific, Waltham, MA, USA).

Differentiation of hEnSCs into motor neuron-like cells within 3D collagen hydrogels containing EpoB-loaded microspheres. Differentiation of hEnSCs was conducted in a three-step approach 26
using 24-well cell culture plates with different experimental groups, including collagen hydrogel without Mic EpoB (2-Col/PA-CH), collagen hydrogel with Mic EpoB (2-Col/PA-CH/Mic EpoB), tissue culture plate (TCP; hEnSCs differentiation without Mic EpoB and collagen hydrogel), and control (undifferentiated hEnSCs). 2 × 10 5 hEnSCs in complete medium (DMEM/F12 and 10% FBS) at passage three were seeded on different substrates and incubated at 37 °C and 5% CO 2 . After 24 h, in step one, the complete medium was replaced with a preinduction (PrI) medium, and cells were incubated for 24 h. In step two, the PrI medium was exchanged with the induction medium, and cells were incubated for 7 days (the medium was exchanged after 3 days). In step three, www.nature.com/scientificreports/ the induction medium was replaced with a maturation medium, and incubation was continued for an additional 7 days (the medium was exchanged after 3 days). All procedures and reagents are summarized in Table 3.
Immunofluorescence (IF) staining. All samples were fixed using 4% paraformaldehyde in PBS for 30 min at room temperature and then permeabilized using 0.2% Triton X-100/PBS. The cells were blocked using 5% BSA/PBS for 45 min at room temperature and incubated overnight with primary antibodies at 4 °C (Table 4). Then, each sample was washed three times with PBS/Tween 20 (0.1%, Sigma). Secondary antibodies were added, and samples were kept in the dark for one hour at room temperature. Cells were again washed with PBS/Tween 20 (0.1%) three times. Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, D8417). For the negative control group, only secondary antibodies were used.
RNA extraction and cDNA synthesis. Total RNA extraction was conducted using RNX Plus™ low copy RNA isolation (CinnaGen, Iran, EX6101), chloroform, isopropyl alcohol, and 75% ethanol. The purity and quantity of each extracted RNA were assessed using a Nanodrop spectrophotometer (Thermo Scientific, USA). cDNA synthesizing was conducted using PrimeScript™ RT Reagent Kit (Takara, RR037A) following the manufacturer's instruction (Thermocycler schedule: 15 min at 37 °C, 5 s at 85 °C).
Real-time RT-PCR for gene expression analysis. All primers were obtained from Qiagen (Germany) ( Table 5). RealQ RNA quantitative measurement was evaluated using Plus 2 × Master Mix Green (AMPLIQON, 5000850-1250) on a real-time thermocycler (Rotor-Ge4ne Q, Corbett Life Science, USA) ( Table 6). PCR product specificity was evaluated by confirming a single peak in the melting curve analysis. All experimental samples were performed three times in duplicate. RPL13A was selected as a housekeeping gene, and each gene fold- Table 3. The procedure and reagents used for differentiation of hEnSCs into motor neuron-like cells on 3D collagen hydrogels with or without EpoB-loaded PCL microspheres or cell culture plates.
Step Differentiation medium Ingredients  www.nature.com/scientificreports/ change was calculated relative to RPL13A based on relative quantitation evaluation via the ΔΔCT method, using the 2 −ΔΔCT relative expression equation. A summary of the different steps of the study is shown in Fig. 1.

Analysis and quantification of neurite extension. Analysis of neurite extension was investigated
using an inverted phase-contrast microscope (Olympus, Japan) attached to a digital camera (DP50). For each experimental group, five random representative areas were selected for neurite length measurement using ImageJ software.
Statistical data analysis. Data
Cell GFP transduction. Two types of cells, HEK293-T, and hEnSCs were labeled. As shown in Fig. 2b, cells were effectively transduced by GFP-encoding lentiviral particles. The GFP-positive cells were observed by fluorescent phase-contrast microscopy and used to assess cell adhesion and proliferation within hydrogels.
Encapsulation efficiency. The encapsulation efficiencies (EE) for different loading concentrations of EpoB are shown in Table 7. A greater EE was obtained in PCL microspheres with higher EpoB loading levels, indicating that the initial concentration of EpoB can potentially affect EE of EpoB in PCL microspheres using the single emulsion fabrication method.  www.nature.com/scientificreports/ EpoB release. As shown in Fig. 3i, there was a burst release in the initial first 24 h for all groups, followed by a slower gradual release for 2.5 and 10 ng/mg EpoB/PCL microspheres, which reached a plateau. However, compared to 2.5 and 10 ng/mg EpoB/PCL microspheres, the 40 ng/mg EpoB/PCL microspheres maintained a higher release rate over 21 days, indicating that the release rate is influenced by the amount of EpoB encapsulated in the microspheres. According to the EpoB release results, 2.5 ng/mg was selected due to its appropriate sustained release dosage for neuronal differentiation.
Proliferation and viability assessment of microspheres. To find the permissible limit, the effect of four different concentrations of unloaded PCL microspheres (unloaded Mic: 5, 10, 15, and 40 mg/mL) were assessed for viability and proliferation of hEnSCs during 24, 48, and 72 h using the MTT assay. The effect of the candidate microsphere 2.5 ng/mg (Mic EpoB) on hEnSCs was also evaluated by indirect MTT assay for 24, 48, and 72 h (Fig. 3j) Cell morphology, attachment, and proliferation. Fluorescent microscopy. Collagen hydrogels improved adhesion and proliferation of HEK293-T (as a typical well-transected GFP cell) and hEnSCs-GFP. As shown in Figs. 4 and 5, the effect of hydrogel fabricated using 2 mg/mL collagen was more prominent on cell proliferation compared with hydrogel fabrication using 4 mg/mL collagen. Furthermore, among different Figure 1. A schematic presentation of different steps of the study, including; EpoB microspheres synthesis using single emulsion, collagen extraction from rat tail, synthesis of collagen hydrogel and its crosslinking using proanthocyanidin from grape seed and calcium hydroxide (PA-CH), cell isolation, and culturing from human endometrium, and a three-step MN-differentiation (during 15 days). www.nature.com/scientificreports/ groups, the 0.5% PA-CH hydrogel had the greatest cell proliferation compared to other PA-crosslinked collagen hydrogels in both 2 mg/mL and 4 mg/mL collagen concentrations (i.e., 2-Col/PA-CH and 4-Col/PA-CH). Since the 2-Col/PA-CH hydrogel had a more prominent effect on cell proliferation compared to 4-Col/PA-CH, this formulation was selected as the optimized hydrogel for further experiments.
Scanning electron microscope (SEM). SEM images of hEnSCs cultured on collagen, 2-Col/PA-CH, and 2-Col/ PA-CH/Mic EpoB hydrogels indicated cell-scaffold interactions and cell attachment on hydrogel scaffolds 48 h after cell seeding (Fig. 6a-c,e). The interconnected porosity and surface morphology of the hydrogel scaffolds was revealed by SEM. In the 2-Col/PA-CH/Mic EpoB hydrogel, the EpoB microspheres were visible and detected (Fig. 6d,e). The 2-Col/PA and 2-Col/PA-CH hydrogels showed more fibrous and regular structure compared to collagen (Fig. 6f-h), and the 2-Col/PA-CH hydrogel indicated a smooth surface morphology.

The effect of 3D collagen hydrogel and EpoB release on differentiation of hEnSCs into motor neuron-like cells. Immunofluorescence staining. After 15 days of MN-induction, the morphology of hEn-
SCs changed from a fibroblast-like shape to a bipolar and round cell body shape with neurites. The neuron-like morphology was observed using an inverted phase-contrast microscope (Olympus, Japan). IF staining results showed up-regulation of the MN markers ISL-1, CHAT, and HB9 in all three experimental groups, but no expression in the control group (Fig. 8a).
Differentiation rate was investigated by counting each positive cell marker (beta-tubulin III, ISL-1, CHAT, NES, HB9) as a percentage of the total number of DAPI-stained cells. The level of positive MN markers expression in cells seeded on 2-Col/PA-CH/Mic EpoB group was significantly higher (P < 0.001) than the other two groups (i.e., TCP, and 2-Col/PA-CH), except for NES. The positive marker expression of different evaluated markers in TCP, 2-Col/PA-CH, and 2-Col/PA-CH/Mic EpoB is shown below, respectively. Quantitative results are also shown in Fig. 8b.

Discussion
In recent years, increasing evidence suggests that small molecules are effective chemical agents for cell survival and differentiation, and may be candidates for various interventional therapies 60  Many studies have proposed tissue engineering as a practical approach to repairing various damaged tissues, and nerve tissue engineering is no exception to this rule, especially given the increasing rate of SCI 23,63 . Many combinational approaches have proven advantageous compared with using each method separately. However, these approaches require further elucidation to ensure a meaningful effect compared to their individual usage 12 .
It is known that ECMs have active roles in cell survival, neuronal differentiation, and neurite development 12 . Embedding hydrogels alone without any treatment may slightly improve injury outcomes after SCI 8 . However, for the hydrogel approach, it has been shown drug incorporation into hydrogels is necessary for effective healing of the injured environment 8,40 .
In this study, we synthesized EpoB/PCL microspheres to have a sustained release during the MN differentiation process. The results indicated that PCL microspheres mediated the sustained release of EpoB as a microtubule-stabilizing small molecule that enhances differentiation of hEnSCs to MN-like cells and increases axonal elongation. The addition of released EpoB, along with the differentiation protocol, promoted neurite outgrowth www.nature.com/scientificreports/ and neuronal morphology. It also significantly increased the efficiency of motor neuron generation (˃ 96%), and compared to most recent studies, it gained a higher percentage of motor neurons 64,65 . PCL was chosen for microsphere synthesis because of its good biocompatibility, low biodegradation, and cost-effectiveness 39 . All four groups of EpoB/PCL microspheres (0, 2.5, 10, and 40 ng/mg) were synthesized using a single emulsion technique, and their round shape microspheres with a smooth surface were demonstrated by SEM. Based on the EpoB release test results, the 2.5 ng/mg EpoB/PCL microsphere group was selected for MN differentiation given a more appropriate dose and drug release and normal size distribution. The indirect MTT test also confirmed the suitability of 2.5 ng/mg EpoB/PCL microsphere for our purposes with cell viability ˃ 77% after 72 h.
Recent studies indicate that culturing stem cells in 3D hydrogels improves cell maintenance, expansion, and differentiation by mimicking the native ECM 66,67 . For neuronal replacement therapies, given that neurons are not capable of dividing, stem cell expansion with hydrogel assistance is essential to achieve a suitable cell number before neuronal differentiation 68 .
Cell attachment to the ECM is crucial for cell differentiation and proliferation via cell anchoring and activation of signaling pathways that can lead to tissue regeneration and development [69][70][71] . Collagen is a key component of the ECM and is extensively used as a scaffold for its effective properties of improving cell attachment, survival, proliferation, differentiation, and migration 72 . One study also reported that collagen might promote higher survival, adhesion, and proliferation of mesenchymal stem cells 73 . These cell processes are regulated by the interaction of collagen with integrins in cell surface receptors 74 . Moreover, collagen has garnered attention as a delivery system for the sustained release of different drugs and chemical agents 75,76 .
GFP cell culture in the two collagen hydrogel concentrations (2 and 4 mg/mL) presented a dose-dependent proliferation of hEnSC. The lower concentration, 2 mg/mL, was selected because of its greater potential for cell proliferation and cell connection, which are important for neural differentiation. In line with our study, a previous report indicated that lower concentrations of collagen are more appropriate for neuronal differentiation and neurite outgrowth 77 .
Moreover, adding either PA-CH or PA to the collagen hydrogel increased its stability during the 15-day differentiation process by decreasing the hydrogel's degradation degree. However, hydrogels cross-linked with PA-CH showed a greater cell proliferation than PA alone. In addition, cross-linking the collage hydrogel also decreased the swelling degree of the hydrogel. Interestingly, the 2-Col/PA-CH/Mic EpoB had a slower degradation degree than the collagen and 2-Col/PA-CH groups, which can be attributed to the interactions between PCL microspheres and the collagen hydrogel. The 2-Col/PA-CH/Mic EpoB had a higher swelling ratio than the collagen and 2-Col/PA-CH hydrogels. All three groups showed interconnected porous, proper cell-scaffold integration and cell attachment.
Immunofluorescence staining indicated that the beta III-tubulin marker was increased in all three differentiation groups (i.e., TCP, 2-Col/PA-CH, and 2-Col/PA-CH/Mic EpoB). In agreement with previous reports, undifferentiated hEnSCs and other sources of human MSCs have a spontaneous expression of this marker 78 , demonstrating that it is possible for MSCs such as hEnSCs to undergo neuronal differentiation. Furthermore, other studies have shown that hEnSCs can induce spinal motor neuron generation during development by upregulating dopamine 79 . However, compared to the differentiation groups, the undifferentiated hEnSCs in our study had a lower beta III-tubulin expression. hEnSCs also spontaneously express NES. Thus, both NES and beta III-tubulin markers show hEnSCs to be a suitable source for neuronal differentiation and spinal cord restoration.
Real-time RT-PCR determined significant gene up-regulation of NEFH, NES, ISL-1, and HB9 after 15 days of induction. Higher gene expression of ISL-1 and HB9 in the 2-col/PA-CH/MicEpoB group compared to the other two differentiation groups showed that incorporation of EpoB microspheres in the collagen hydrogels may significantly increase the MN differentiation efficacy of hEnSCs.
The results show that the release of EpoB from microspheres, in addition to RA and SHH, increase HB9 gene expression 80,81 , and can significantly enhance the gene expression of HB9 and also ISL-1, which are two crucial regulators of mature motor neurons 82,83 . Moreover, among the other two groups, the 2-Col/PA-CH group demonstrated higher ISL-1 gene expression.
Furthermore, the neurite length of MN-like cells in the 2-Col/PA-CH/Mic EpoB group increased remarkably to 146 µm, which is more than three times longer compared to the other two groups without Epo B microspheres (2-Col/PA-CH, and collagen groups). Recently, it was reported that acellularized spinal cord scaffolds incorporating bpV(pic)/PLGA microspheres promote axonal regeneration by about 45 µm 88 .
Another study indicated that adding 3 nM of paclitaxel as an MSA can enhance neurite length of mature rat retinal ganglion cells (RGCs) by 40 µm while using 10 nM paclitaxel significantly decreases neurite outgrowth 89 . This shows that paclitaxel has less of an effect on neurite outgrowth than EpoB. Additionally, compared with paclitaxel, EpoB has a smaller molecular size, higher solubility, and the capability to pass the blood-brain barrier (BBB) 90,91 . It is worth noting that EpoB can be effective for axon elongation and MN-differentiation, despite the presence of the neurite outgrowth inhibitor A, which is known to be an axonal inhibitor in SCI 23,26 .
One proposed method of increasing neurite outgrowth in recent years is electrical stimulation [92][93][94] . A study reported that the neurite length of PC12 cells via gold-nanoparticle after electrical stimulation (ES) was 120 µm 95  www.nature.com/scientificreports/ Another study reported, the final neurite length of the primary prefrontal cortex (PFC) in collagen 3D cultures with ES to be ~ 110 µm, compared to the control wild-type where only ~ 10 µm of elongation was observed 96 . Thus, our results indicate that the fabricated 2-Col/PA-CH/Mic EpoB hydrogel achieved a higher neurite elongation than the mentioned studies. While understanding the exact mechanism of EpoB and adapting it to the complex mechanism of SCI yet need to achieve a clear vision, today its role in stem cell differentiation and SCI has garnered attention, and our study demonstrated the enhanced MN differentiation efficiency of hEnSCs and neurite growth by 3D crosslinked collagen hydrogel containing EpoB-loaded microspheres which can be used as a combinatorial approach for MN-diseases modeling and a future aid in SCI repair.

Conclusion
To our knowledge, this is the first time that ultra-low doses of EpoB loaded microspheres have been synthesized for the purpose of MN differentiation from hEnSCs. Our study demonstrates that incorporation of EpoB microspheres with collagen hydrogel can significantly increase the MN differentiation efficiency of hEnSCs by up-regulating ISL-1 and HB9, known as mature motor neuron key markers. Moreover, our results showed that the addition of EpoB microspheres significantly enhances neurite growth during MN differentiation. Utilizing proanthocyanidin and calcium hydroxide as cross-linkers significantly reduces the degradation degree of the 3D collagen hydrogel and provides an appropriate stable ECM for supporting cell attachment, proliferation, and differentiation. Our findings emphasize the active role that 3D cross-linked collagen hydrogels containing EpoB-loaded microspheres for nerve tissue engineering and the potential of using combinatorial approaches to overcome the limitations of individual methods for SCI repair.

Data availability
The datasets used and/or analyzed during the current study are available on reasonable request.