Genipin stabilized fibrin microbeads: Carrying cytokines to form niches for stem cell differentiation

Preparation of FMBs and gFMBs

FMBs were prepared from fibrinogen and thrombin based on oil emulsion according to a previous report with slight modifications. ²⁷ First, vegetable oil purchased from local grocery was preheated to 65 – 75°C with mechanical stirring. Second, Bovine fibrinogen (Bioing, Shenyang, China) solution (100 mg/mL) containing 2 mM Ca²⁺ was mixed with thrombin solution (2000 U/ml, Bioing) containing 2 U/ml factor XIII (MedChemExpress, Shanghai, China) with a v/v ratio 100:1. Third, the mixture was added into the oil with vertical flow drop by drop. After being stirred for 10 min, lots of FMBs could be spotted inside the oil. Last, the FMBs were collected, washed with hexane, rinsed in acetone and ethanol, sieved with a 20 mesh screen, and then stored in 75% ethanol. The preparation procedures of gFMBs were similar except that the mixture in the second step was prepared from solutions of fibrinogen (100 mg/mL), thrombin (2000 U/ml), and genipin of various concentrations (2, 4, and 8 mg/mL).

To obtain the gFMBs of appropriate diameters (around 300 – 700 μm), the impact of different stir velocities (300, 600, 900, and 1200 rpm) was studied. The gFMBs were washed and transferred to 6 well plates. Then the gross morphology of gFMBs was photographed with a digital camera and the diameters of gFMBs were measured with ImageJ.

Microstructures of FMBs and gFMBs

Hematoxylin and eosin (H&E) staining was applied to examine the inner microstructures of microbeads. Briefly, FMBs and gFMBs were prepared under the stir velocity of 900 rpm. These microbeads were fixed, dehydrated with gradient ethanol, embedded in paraffin, sectioned, deparaffinized, rehydrated and stained with H&E staining solutions. The images were captured with a microscope (Axio Observer, Zeiss, Germany).

Scanning electron microscopy (SEM) was applied to obtain the surface ultra-structures of FMBs and gFMBs. In short, microbeads were fixed with 2.5% glutaraldehyde, washed with phosphate buffered saline, and dehydrated with gradient ethanol. The dehydrated microbeads were then put in a CO₂ critical point dryer (CPD300, Leica) to dry for 2 h followed by being sputtered with gold. Finally the surfaces of FMBs and gFMBs were observed under an SEM (Joel, JAPAN).

Evaluation of crosslinking degrees of gFMBs

The crosslinking degrees were measured to optimize the concentration of genipin and duration of crosslinking. Ninhydrin assay was applied to evaluate the crosslinking index of gFMBs according to the previous reports. ²⁸ There were 12 groups with 3 levels of genipin concentrations (2, 4, and 8 mg/mL, designated as gFMBs-2, gFMBs-4, and gFMBs-8 respectively) and 4 crosslinking durations (6 h, 12 h, 24 h, and 36 h). Each group had 3 replicates and FMBs served as control. The crosslinking durations were defined as the period between the end of stirring and the harvest of microbeads from oil. During this period, the oil containing microbeads was put into an incubator at 37°C. Same amounts of FMBs and gFMBs were subjected to ninhydrin assay.

The ninhydrin assay was performed according to a previous report. Briefly, 50 mg FMBs or gFMBs were weighted and incubated in 1 mL water for 1 h. Then 450 μL ninhydrin solutions at 1 mg/mL (Sigma Aldrich, USA) were added. The mixture was vortexed and heated for 20 min at 90°C. After cooled down, the samples were centrifuged at 1000 g for 10 min to collect the supernatants. The optical absorbance at 570 nm of each supernatant was measured with a spectrophotometer (μQuant). The absorbance reading was positively correlated to the number of free amino groups which could react with ninhydrin to form Ruhemann’s purple. Additional, during crosslinking between genipin and FMBs, free amino groups were consumed. Therefore, the following formula can be used to represent crosslinking index:

Determination of water content in gFMBs

To assess the impact of genipin on the hydrophilic property of FMBs, water content (WC) of FMBs or gFMBs was measured. FMBs and gFMBs-2, gFMBs-4, gFMBs-8 (all cross-linked for 12 h) were subjected to WC analysis. Each sample was immersed in distilled water for 24 h to fully hydrate. Then the surface water was removed with filter papers and wet weight was measured. Finally, microbeads were lyophilized at 4°C for 24 h before the measurement of the dry weight. WC was calculated as: WC (%) = (wet weight – dry weight)/wet weight × 100%.

In vitro degradation

Sterilized microbeads of similar weights (W₀) were soaked in cell culture medium (Dulbecco’s modified Eagle’s medium/F12, with addition of 10% fetal bovine serum and antibiotics) in tubes which were rotated at the speed of 20 rpm for different durations (7 d, 14 d, 21 d, and 28 d). During this period, half of the culture media were changed every 3 days. At each time point, the final weights (W₁) of microbeads were measured. The degradation rate was calculated as following: Degradation rate (%) = (W₀ − W₁) / W₀ × 100%. Another batch of FMBs and gFMBs were also soaked in the culture media for 10 days, and the media were changed every 3 days. Then these microbeads were fixed to prepare H&E staining sections.

Preparation of gFMB@SHH

First, an SHH stock solution (Applied Biological Materials Inc., Canada) was prepared and mixed with various concentrations of genipin solutions. The final concentrations of SHH was 500 ng/mL in each mixture, while those of genipin were 0, 1, 2 or 4 mg/mL respectively. The gFMBs weighted about 50 mg were mixed with the SHH and genipin solution at 37°C for 24 h using a shaker incubator. Finally, gFMBs were collected and their surface liquid was removed with paper filters. According to the concentrations of genipin, there were four kinds of gFMB@SHH, designated as gFMB@SHH/gp0 (without genipin), gFMB@SHH/gp1 (genipin 1 mg/mL), gFMB@SHH/gp2 (genipin 2 mg/mL), gFMB@SHH/gp4 (genipin 4 mg/mL).

Crosslinking efficiency and cumulative release of SHH

The crosslinking efficiency of SHH was calculated with the following formula: Crosslinking efficiency (%) = (1-quantities of SHH remained in solution after crosslinking/ whole amount of SHH used to crosslink) × 100%. The gFMBs were mixed with 1 mL solutions containing 500 ng SHH and different concentrations of genipin. After being shaken for 24 h, the supernatants were collected to quantify the amounts of SHH remained in solutions by using an enzyme-linked immunosorbent assay (ELISA) kit according to the manufacturer’s instructions.

The cumulative releasing of SHH in each type of gFMB@SHH was measured as previously reported. Briefly, each type of sterilized gFMB@SHH/gp0, 1, 2, 4 was soaked in 1 mL PBS in a tube and incubated at 37°C for different days (1, 7, 14, 21, and 28 d). At each time point, 0.5 mL PBS was collected to determine the concentration of SHH by ELISA assay. Then, 0.5 mL fresh PBS was added to the tube.

Culture and characterization of ectomesenchymal stem cells

Ectomesenchymal stem cells (EMSCs) were harvested from two C57/B6L mice as previously reported. ²⁹ The animal experiments were conducted following the national regulations and approved by the Jiangsu University Animal Ethics and Experimentation Committee. These cells were cultured with DMEM/F-12 (DF12) with 10% fetal bovine serum in an incubator containing 5% CO₂ at 37°C. The culture media were changed every 2 days and the cells were passaged when cells become confluent.

To characterize EMSCs, several markers including cluster of differentiation 44 (CD44), cluster of differentiation 133 (CD133), neurotrophin receptor p75 (p75), Nestin and SRY-box transcription factor 2(Sox2) were selected to perform immunofluorescence. In short, the cultured cells were fixed with 4% paraformaldehyde, washed with phosphate buffered saline (PBS), and incubated in 10% bovine serum albumin (BSA) with 0.3% triton X-100. Then these cells were incubated with primary antibodies overnight at 4°C. The primary antibodies include anti-CD44 (Boster, 1:100), anti-CD133 (Boster, 1:100), anti-p75 (Boster, 1:100), anti-Nestin (Abcam, 1:200), and anti-Sox2 (proteintech, 1:100). After several washes, the cells were incubated with corresponding Cy3-conjugated secondary antibodies (proteintech, 1:300) at 37°C for 2 h. Then the cells were counter stained with 4′,6-diamidino-2-phenylindole (DAPI) before the capture of images by using a multifunctional microscope (Axio Observer, Zeiss).

Cytotoxicity of gFMB@SHH

Four types of gFMB@SHH (/gp0, /gp1, /gp2, and /gp4) were sterilized and rinsed. Then, each type of microbead weighting around 10 mg was transferred into the wells of 96-well plates. After that, 100 µl EMSCs suspension of 5 × 10⁴ cells per ml culture media was added into each well. The culture media were changed every 3 days. The wells containing only EMSCs served as control. Four time points (1, 3, 5, 7 d) and three replicates were set up in this experiment.

When each time point arrived, the microbeads in the well were carefully removed, and the cells were washed before the addition of 100 µl DF12 containing 10 µl CCK-8 reagent (Biosharp, Anhui, China). The plates were put back to incubator for 1 h before the absorbance at 450 nm was measured by a spectrophotometer (µQuant, Biotek). EMSCs suspensions of known concentrations were tested with CCK-8 to establish a standard curve. The OD values were transformed to cell numbers which were then normalized to the results of control group.

Relative growth rates of EMSCs on gFMB@SHH

Four types of gFMB@SHH (/gp0, /gp1, /gp2, and /gp4) and gFMB were sterilized, rinsed and weighted before being mixed with 4 mL culture media containing 1 × 10⁴ EMSCs in a centrifuge tube. The tubes were rotated at the rate of 20 rpm in a mini-bioreactor (AmProtein, Hangzhou, China). Several hours later, some EMSCs attached to gFMB@SHH or gFMB and began to proliferate. There were four time points (1, 4, 7, 10 d) and three replicates in each group. At each time point, the microbeads in the corresponding tubes were washed, collected and transferred to 96 well plates. Then 100 µl DF12 containing 10 µl CCK-8 reagent was added. The plates were shaken slowly for 1 h at 37°C. To avoid the interference of microbeads, the supernatants were transferred to other blank wells to determine the absorbance at 450 nm. The relative growth rate was calculated by using: relative growth rate = (OD450 of gFMB@SHH)/(OD450 of gFMB) × 100%.

Induced differentiations of EMSCs on microbeads

The EMSCs were seeded on microbeads and cultured in tubes which were rotated at the rate of 20 rpm. There were three culture conditions, that is gFMBs (control group), gFMBs with addition of SHH in culture media (gFMBs + SHH), gFMBs crosslinked with SHH by 2 mg/mL genipin (gFMBs@SHH). Both of the total amounts of SHH used in gFMBs + SHH and gFMBs@SHH equaled 2 µg. The inducement media was composed of DF12 supplemented with 10% FBS, 2% B27 (Gibco), 1 μg/mL ATRA (Sigma Aldrich, USA), 20 ng/mL TSA (Sigma Aldrich, USA) and 100 ng/mL NGF. Every 3 days, two thirds of the media were changed, and the whole process lasted for 28 days.

At the end of differentiation process, the cells on the microbeads were fixed, washed, and blocked. Then, The samples were incubated with primary antibodies including anti-Gap43 (Boster, 1:100), anti-Tuj1 (proteintech, 1:100), anti-ChAT (proteintech,1:100), and anti-MBP (Boster, 1:100) overnight at 4°C. After several washes, the samples were incubated with secondary antibodies Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:500, ImmunoWay) at 37°C for 1 h. All samples were stained with DAPI. At last, the images were captured with a confocal microscope (DeltaVision Elite, GE healthcare).

Quantitative real-time PCR

The cells growing on gFMB or gFMB@SHH were collected through digestion with Accutase (Stemcell) followed by centrifugation at 200 rpm to collect the supernatant, which was subjected to centrifugation at 2000 rpm to harvest cells. Then, total RNA was extracted with TRIzol reagent (Invitrogen) and reversely transcribed by using HiScript III 1st Strand cDNA synthesis kit (Vazyme). qPCR was performed by using AceQ qPCR SYBR Green Master Mix (Vazyme) according to its instructions. The cDNA were amplified and detected with CFX Opus Real-Time PCR System (Biorad). The 2^−ΔΔCT method was applied to gauge the relative mRNA level by using β-actin as housekeeping gene and gFMB as control group. The assay was repeated for three times. The primer pairs were listed in the following table.

OPEN IN VIEWER

Gene symbol	Forward	Reverse
Gli1	CCAAGCCAACTTTATGTCAGGG	AGCCCGCTTCTTTGTTAATTTGA
Ascl1	GCAACCGGGTCAAGTTGGT	GTCGTTGGAGTAGTTGGGGG
Ccnd1 (CyclinD1)	GCGTACCCTGACACCAATCTC	CTCCTCTTCGCACTTCTGCTC
Bcl2	GTCGCTACCGTCGTGACTTC	CAGACATGCACCTACCCAGC
Neurog2 (Ngn2)	AACTCCACGTCCCCATACAG	GAGGCGCATAACGATGCTTCT
Neurog3 (Ngn3)	CCAAGAGCGAGTTGGCACT	CGGGCCATAGAAGCTGTGG
NeuroD1	ACAGACGCTCTGCAAAGGTTT	GGACTGGTAGGAGTAGGGATG
NeuroD2	AAGCCAGTGTCTCTTCGTGG	GCCTTGGTCATCTTGCGTTT
NeuroD6	ACGACTGGAAAGGGTCAAGTT	GACCACTTTTCGCAAATTGTCC
Mnx1	TCTATGGACACCCGGTCTACA	CCCCAAGAGGTTCGACTGC
NeuN	AACCACGAACTCCACCCTTC	GACCTCAATTTTCCGTCCCTC
Actb (β-actin)	GGCTGTATTCCCCTCCATCG	CCAGTTGGTAACAATGCCATGT

Western blotting

The cells growing on gFMB or gFMB@SHH were harvested as being described in the previous section. The cell pellets were then lysed by using RIPA containing protease inhibitor cocktail on ice. After ultra-sonication and centrifugation, the supernatants were collected to measure the protein concentrations. The samples were mixed with loading buffer and boiled for 5 min.

The samples were loaded to the wells of PAGE gels, separated and transferred to polyvinylidene difluoride (PVDF) membranes. After being blocked with bovine serum albumin, the membranes were incubated with antibodies overnight including anti-Gap43, anti-Tuj1, anti-ChAT, and anti-GAPDH. The next day, these membranes were washed thoroughly before being incubated with HRP-conjugated secondary antibodies. The bands were developed using an ECL substrate (Millipore) and were captured with an imaging system (Vilber Fusion Solo, France).

Statistical analysis

All data were shown as Means ± standard errors (SEM). To compare the means between different groups, analysis of variance (ANOVA) or repeated measurements ANOVA (RM-ANOVA) were first performed, and then a post-hoc test (Student–Newman–Keuls test) was done. The normality and homogeneity of variance were tested before ANOVA. The statistical analyses were performed with Graphpad prism 8.0 (La Jolla, USA). When p < .05, the difference was considered statistically significant.

The impact of stirring speed on the size of microbeads

Basically, the diameters of gFMBs were negatively correlated to the stirring speed used to produce them, as shown in Figure 1(a)–(d). When the stirring speeds were set at 300, 600, 900 and 1200 rpm, the corresponding median diameters of gFMBs were around 2100, 1300, 550, and 280 μm respectively (Figure 1(e)). Considering the cell carrying capability, the gFMBs of diameters of around 550 μm (stirring speed = 900 rpm) were chosen to perform the following experiments. The homogeneity of gFMBs was increased after being sieved with a mesh 20 screen (Figure 1(f)). Notably, the color of fibrin microbeads became dark blue to black when crosslinked by 4 mg/mL genipin.

OPEN IN VIEWER

Figure 1.

The effects of stirring speed on the diameters of genipin modified fibrin microbeads (gFMBs). (a, b, c, d), the gross morphology of gFMBs formed under the stirring speeds of 300, 600, 900, and 1200 rpm respectively, Bar = 5 mm; e, the violin plot of diameters of gFMBs produced under various stirring speeds; f, the microscopic appearance of gFMBs which were produced under the speed of 900 rpm and sieved thereafter, Bar = 1 mm.

Characterization of genipin modified fibrin microbeads

As shown in Figure 2(A), the results of H&E staining of microbeads’ sections indicated that the outlines of fibrin microbeads (FMBs) were irregular while those of the gFMBs-4 and gFMBs-8 (crosslinked with 4 mg/mL and 8 mg/mL genipin respectively) were approximately round. In contrast, the silhouette of gFMBs-2 was more regular than FMBs but less round than gFMBs-4 or gFMBs-8. Furthermore, the inside structure of gFMBs was more homogenous than that of FMBs.

OPEN IN VIEWER

Figure 2.

Morphology of microbeads. (A) The microstructure of microbeads after H&E staining. Bar = 100 μm. (B) The SEM micrographs of microbeads. Bar = 500 nm. (a) FMB, (b) gFMB-2, (c) gFMB-4, (d) g FMB-8.

The results of SEM demonstrated the effects of genipin treatment on the surface ultrastructure of FMBs. Bundles of fibrils could easily be identified from the surface micrograph of FMBs (Figure 2(B)(a)), whereas large bundles of curved fibers were discernible on the surface of gFMBs-2 (Figure 2(B)(b)). However, no apparent fibril or fiber could be found on the surface of gFMBs-4 and gFMBs-8 (Figure 2(B)(c) and (d)), which might be owing to the crosslinking effects of genipin. Figure 3(A) also demonstrated that the shapes of gFMBs-4 and gFMBs-8 were more regular than those of FMBs and gFMBs-2. Interestingly, after being treated by genipin, gFMBs eventually became red fluorescent beads. Additionally, the fluorescence intensity of gFMBs-4 was similar to gFMBs-8 but higher than that of gFMBs-2.

OPEN IN VIEWER

Figure 3.

Characterization of microbeads. (A) The microscopic morphology of FMB and gFMB, left column, white field, right column, red fluorescent field; (B), the impact of crosslinking time and genipin concentrations on crosslinking levels of gFMBs; (C), the effects of genipin concentrations on water contents of FMBs or gFMBs; (D), the H&E staining results of FMBs or gFMBs soaked in culture media for 10 days; (E), the impact of genipin concentrations on the in vitro degradation rate of microbeads. (a) FMBs, (b) gFMBs-2, (c) gFMBs-4, (d) g FMBs8. *, p < .05, **, p < .01, bar = 100 μm.

Figure 3(B) showed the time-dependent and concentration-dependent (2, 4, 8 mg/mL) effects of genipin on crosslinking index of microbeads. After 12 h, the crosslinking index arrived to the peak. Meanwhile, the crosslinking index of gFMBs-4 and gFMBs-8 were similar but higher than that of gFMBs-2.

The FMBs showed the highest water content among the four groups (Figure 3(C)). After treatment of various concentrations of genipin, the water contents of microbeads decreased significantly. Specifically, the water content of gFMBs-4 was similar to that of gFMBs-8, but was significantly lower than that of gFMBs-2 (Figure 3(C)).

Genipin altered the degradation rate of FMBs as well as water content. As demonstrated by Figure 3(D), there were lots of large cavities emerged inside FMBs after being soaked in culture media for 10 days. In contrast, there were only several small cavities lying inside gFMBs-4 and gFMBs-8. More importantly, the outlines of gFMBs-4 and gFMBs-8 retained their original round shape while those of FMBs and gFMBs-2 had already become irregular. If these microbeads were submerged in culture media for 28 days, almost all FMBs and more than 70% of gFMBs-2 would degrade, whereas only less than one third of gFMBs-4 and gFMBs-8 would degrade (Figure 3(E)). In addition, there was no significant difference between the the degradation rate of gFMBs-4 and gFMBs-8 as indicated by Figure 3(E).

It was noteworthy that the crosslinking indexes of gFMBs-4 and gFMBs-8 were higher than that of gFMBs-2, and there were no significant differences between gFMBs-4 and gFMBs-8 as to the inner and surface structures, the crosslinking index, and degradation rates. Based on these results, gFMBs-4 was chosen to finish the following experiments.

Crosslinking SHH to gFMBs by genipin

As shown by Figure 4(A), the addition of genipin significantly increased the crosslinking efficiency of SHH to gFMBs. Although the crosslinking efficiencies in gFMB@SHH/gp2 and gFMB@SHH/gp4 groups were similar, they were higher than that in gFMB@SHH/gp1 group. Consistently, the cumulative releasing curve also indicated the increased bound of SHH to gFMBs when crosslinked by genipin. In gFMB@SHH/gp0 group, half of the attached SHH was released into PBS in 1 day. If genipin of 2 mg/mL or 4 mg/mL was used to crosslink SHH to gFMBs, more than 75% of SHH bound to gFMBs would still remain on these gFMBs for nearly 1 month (Figure 4(B)).

OPEN IN VIEWER

Figure 4.

Crosslinking SHH to gFMB. (A) The impact of genipin concentration on the crosslinking efficiency of SHH to gFMB. (B) The effects of genipin concentration on the cumulative release of SHH from gFMB. gFMB@SHH/gp0, /gp1, /gp2, and gp4 designated the concentrations of genipin used to crosslinking SHH (0, 1, 2, and 4 mg/mL respectively). **, p < .01.

Biocompatibility of gFMBs@SHH evaluated by using EMSCs

The EMSCs displayed a spindle-like shape, and positively expressed the markers of mesenchymal and neural stem cells, such as CD44, CD133, p75, nestin, and sox-2 (Figure 5(A)). When different kinds of gFMBs were added into the culture media to co-culture with EMSCs for 7 days, no obvious toxicity of gFMBs could be found (Figure 5(B)). Moreover, when EMSCs were cultured on the surface of gFMBs, the cells on gFMBs@SHH/gp2 proliferated faster than those on gFMBs in the first week, but slower than the latter from the 7th to the 10th day (Figure 5(C)).

OPEN IN VIEWER

Figure 5.

(A) The morphology and characterization of EMSCs, (a) white field, (b) CD44, (c) CD133, (d) p75, (e) nestin, (f) sox2. (B) The cytotoxicity of gFMB on EMSCs, (C) the growth of EMSCs on gFMB at different time. Bar = 10 μm.

gFMB@SHH served as platform to induce neural differentiation of EMSCs

After been cultured on gFMB, gFMB + SHH, and gFMB@SHH/gp2 for 28 days, the EMSCs positively expressed GAP43, Tuj1, and ChAT, but negatively expressed MBP as demonstrated by the results of immunofluorescence staining and western blotting (Figures 6(A)–(D) and 7(A) and (B)). In particular, the expression levels of GAP43, Tuj1, and ChAT in EMSCs grown on gFMB@SHH/gp2 were significantly higher than those in EMSCs grown on gFMBs with addition of SHH to the culture media (Figure 7(B)). In line with this, the results of qPCR indicated that EMSCs grown on gFMB@SHH expressed higher levels of SHH downstream molecules including Gli1, CyclinD1, and Bcl2 than the cells on gFMBs (Figure 7(C)). On the basis of other inducing reagents, the activated SHH pathway stimulated the increased expression of NeuroD1, NeuroD6 and Ngn2 in EMSCs on gFMB@SHH compared with those on gFMB. However, the expression of NeuroD2 was very low, and the expression of Ngn3 was undetectable in in the EMSCs after induction. Furthermore, the mRNA levels of neuron-related markers including Asc1, Mnx1, and NeuN were significantly elevated in induced EMSCs on gFMB@SHH than those on gFMB (Figure 7(C)).

OPEN IN VIEWER

Figure 6.

The induced differentiation of EMSCs into neuron-like cells. (A, B, C, D) The immunofluorescence staining of GAP43, Tuj1, ChAT, and MBP in the differentiated EMSCs growing on the gFMBs respectively. Bar = 100 μm.

OPEN IN VIEWER

Figure 7.

The induced differentiation of EMSCs into neuron-like cells. (A) and (B) The representative and quantification results of western blot demonstrating the expression of GAP43, Tuj1, and ChAT in the differentiated EMSCs growing on the gFMBs. (C) The results of RT-qPCR showing the relative levels of molecules related to SHH signaling or neural differentiation pathways.