The Immunoregulatory Effect of Mechanical Stress on Three-dimensional Cultured Mesenchymal Stem Cells After Extensive Expansion

Background: Mesenchymal stem cells (MSCs) have been used to treat immunopathy, and three-dimensional (3D) cultured MSCs show enhanced immunomodulatory property compared with those in two-dimensional (2D) culture. However, both the regulatory mechanisms remain unclear. The aim of the study was to investigate the role of mechanical stress in maintaining the immunomodulatory function of 2D and 3D cultured MSCs. Methods: Umbilical cord mesenchymal stem cells (UC-MSCs) were plated on tissue culture plastic (TCP) as 2D culture and 3D cultured UC-MSCs were seeded in matrigel. Surface markers, clonogenicity, proliferation and immunoregulatory property of UC-MSCs were evaluated. Meanwhile, we established the mouse models of colitis and type 1 diabetes mellitus (T1DM) to reveal the pharmacotherapeutic effects of 3D cultured MSCs in vivo. The effect of changing mechanical stress by modulating Yes-associated protein (YAP) on immunomodulatory function of 2D and 3D cultured UC-MSCs was evaluated by immunouorescent analysis, real-time quantitative polymerase chain reaction (qPCR) and western blot. Results: We veried early passage UC-MSCs in 2D and 3D cultures exhibited stemness, immunomodulatory property and therapeutic ecacy against immunopathy. However, these characteristics of 2D cultured UC-MSCs were impaired after extensive expansion, whereas 3D culture extended them for several passages by activating YAP. Moreover, prostaglandin E2 (PGE2) could up-regulate YAP to improve the immunomodulatory ability of 2D cultured UC-MSCs after extensive expansion. Conclusions: This work found for the rst time that the signicance of mechanical stress in maintaining immunoregulatory function of 2D and 3D cultured UC-MSCs, providing a new idea for improving the ecacy of MSCs-based immunotherapy.


Background
Mesenchymal stem cells (MSCs) are a subset of multipotent stem cells with the capacity of immunomodulation, which makes them promising candidates for stem-cell-based therapies [1].
Noteworthy is that although cell replacement plays an important role in the treatment of MSCs, the ultimate therapeutic effects are mainly the results of the immunomodulatory ability of MSCs, which reacts with the immune system [2,3]. It is well recognized that MSCs exert signi cant protective effects against immunopathy type 1 diabetes mellitus (T1DM) by inhibiting the expression of cytokines, such as interleukin (IL)-6, IL-1β, tumor necrosis factor (TNF)-α and IL-10, and the therapeutic effects persist during the long follow-up period [4,5]. Furthermore, MSCs can also modulate regulatory T (Treg)/ T helper (Th)17 cells in the spleen and improve endoscopic ndings, pathological ndings, Mayo scores and in ammatory bowel disease quality scores in colitis [6,7]. However, the mechanism by which MSCs treat immunopathy remains unclear, and discussing such content will help to further improve the therapeutic effects of MSCs on immunopathy.
Increasing evidences have indicated that MSCs cultured in 2D and 3D systems exhibited signi cant differences in cell morphology, proliferation, differentiation abilities and energy metabolism [8]. Under 2D culture conditions, replicative senescence can be induced by cell extensive expansion, which limit the e cacy of transplanted MSCs [9][10][11][12]. 3D culture is developed on the basis of 2D monolayer, but with the characteristics of microenvironments in vivo [13,14]. Thus, 3D cultured MSCs have been shown to enhance many therapeutic properties, including their immunosuppression and immunoregulation [1,15,16]. One study found several genes closely related to immune regulation, such as human leukocyte antigen (HLA)-G, indoleamine 2,3-dioxygenase (IDO)-1, prostaglandin synthase (PTGS)-2 and transforming growth factor (TGF)-β1 were more highly expressed after 3D culture than that after 2D culture. Further, 3D cultured MSCs could inhibit the proliferation of activated lymphocytes more effectively [17]. However, the mechanism of maintaining cellular immunomodulatory ability in 3D culture has not been reported, and exploring this mechanism provides an experimental basis for further elucidating the immunomodulatory characteristic of MSCs in 3D culture.
It has been known that Yes-associated protein (YAP) signaling is strongly activated during the culture of cells in a matrigel-encapsulated 3D culture system and maintains the proliferative capacity of the encapsulated cells [18]. Recent studies have shown that variation in the stiffness of 3D matrix changed mechanical stress, and was a potent regulator of MSCs expansion and differentiation. This may be due to the fact that YAP serves as a downstream nuclear mediator and effector of actin cytoskeletal contraction through β-catenin [19][20][21]. However, it is unknown whether activated YAP can improve the immunomodulatory ability of MSCs, the research in this regard may helpful to reveal the mechanism by which MSCs participate in immunoregulation through YAP in 3D culture, and to provide a research basis for exploring the way to enhance the immunomodulatory function of MSCs.
In the present study, we rstly veri ed that compared with 2D culture, 3D culture did extend the stemness and immunosuppression of UC-MSCs for several passages, as well as employed immune disorders such as colitis and T1DM mouse models to con rm the therapeutic effects of long-term passaged UC-MSCs in 3D culture. Furthermore, we explored the modulation of YAP pathway both in 2D and 3D cultured UC-MSCs and its therapeutic effects in colitis and T1DM mouse models, so as to prove the signi cance of mechanical stress in maintaining immunoregulatory function of 2D and 3D cultured UC-MSCs, and prostaglandin E2 (PGE2) could be used to up-regulate YAP and improve the immunomodulatory ability of 2D cultured UC-MSCs after extensive expansion. To the best of our knowledge, it was found for the rst time that the mechanical stress played a signi cant role in maintaining the immunomodulatory function and therapeutic potential both of 2D and 3D cultured UC-MSCs.

Animals
Four-week-old male C57BL/6J mice were purchased from the Laboratory Animal Center of the Fourth Military Medical University. The mice were housed with a 12 h/12 h light/dark cycle at an ambient temperature of 22-25℃ with free access to food and water. All animal experimental procedures were accordance with National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by Animal Care Committee of Xi'an Jiaotong University.

Isolation and Culture of UC-MSCs
UC-MSCs were isolated from human umbilical cords freshly obtained from women who gave birth in Xi'an No. 4 Hospital. All the donors signed informed consent to this study. The experimental procedures of human samples were approved by the Institutional Review Board for Human Subjects Research of Xi'an No. 4 Hospital. The umbilical cord was careful removal of blood vessels, Wharton's jelly tissue was chopped into small pieces of 1 mm 3 and seeded in minimum essential medium-α (α-MEM, Gibco, USA, 12000063) with 10% fetal bovine serum (FBS, Gibco, USA, 16140071), as described in previous reports [22]. 5-7 days after seeding, the cells reached 80% con uency, then passaged and cultured in α-MEM supplemented with 10% FBS, streptomycin (100 IU/ml, Gibco, USA, 15140122) and penicillin (100 µg/ml, Gibco, USA, 15140122), and maintained in a humidi ed 5% CO 2 incubator at 37℃.
In the 2D method, UC-MSCs at P3 were plated directly on tissue culture plastic (TCP) and passaged using 2.5% trypsin (Gibco, USA, 15050057). In the 3D method, UC-MSCs at P3 were seeded in matrigel (Corning, USA, 356231) and passaged with cell recovery solution (Corning, USA, 354253). MSCs were continuously passaged until P15. Cells at P5 and P15 were used in the following studies.

Cell Proliferation Assay
P5 and P15 UC-MSCs (5×10 2 cells/well) were seeded on 96-well plates. Cell proliferation was analyzed with cell counting kit-8 (CCK-8, Yeasen Biotech, China, 40203ES60) according to the standard protocol. First, 10 µL CCK-8 solution per 100 µL complete medium was added into intraday wells, and the plates were incubated in 37℃ and 5% CO 2 for 3h after being blended. The optical density (OD) value was recorded by a microplate reader (Bio-TEK Instruments, USA) at 450 nm.

Isolation of Human Peripheral Blood Mononuclear Cells (PBMCs) and CD3 + T Cells
PBMCs were isolated from fresh whole blood from Xi'an No. 4 Hospital. PBMCs were prepared from peripheral blood diluted in PBS, slowly added to Lymphoprep (MP biomedicals, USA, 0916922-CF), and then centrifuged. The separated PBMCs were collected and washed twice with PBS. CD3 + T cells were then isolated from PBMCs using anti-human CD3 antibody (Biolegend, USA, 300308) and FACS cell sorting system (Miltenyi Biotec, Germany).

ELISA for Cytokine Production
The supernatant of co-cultured MSCs and activated T cells were collected. Cytokine measurements were done for IL-6 (EHC007), IL-10 (EHC009) and TGF-β (EHC107b), using a commercially available enzymelinked immunosorbent assay (ELISA, Neobioscience, China) according to the manufacturer's protocol.
The serum of experimental T1DM mice was collected. Cytokine measurements were done for Glycated hemoglobin (GHb) and C-peptide (Fankewei, China, F2159-A, F2580-A) by ELISA kit.

Western Blot Analysis
After collection, protein concentration measurements were carried out using a bicinchoninic acid (BCA) kit (Beyotime Biotechnology, China, P0012). Equal protein amount from each sample was taken for western blot analysis. The protein bands were visualized and captured using Amersham Imager 600 (GE Healthcare Life Sciences, China). Anti-YAP1 antibody (ab56701) was obtained from Abcam (USA).

RNA Extraction and Real-time Quantitative Polymerase Chain Reaction (qPCR)
RNA was isolated from UC-MSCs using the TRIzol Reagent (Invitrogen,USA, 15596026). The cDNA was transcribed using a PrimeScript RT reagent kit (TaKaRa, Japan, RR036A) and random primers following the manufacturer's instructions. The qPCR analysis was performed using the TB Green Premix Ex Taq II kit (TaKaRa, Japan, RR820L) and detected by CFX96™ Real-time RT-PCR System (Bio-Rad, USA).

Experimental Colitis and Treatment by UC-MSCs
Colitis was induced in C57/BL6J mice (8 weeks old female) by administration of 2.5% dextran sulfate sodium salt (DSS, MP Biomedicals, USA, 02160110-CF) in drinking water for 5 days followed by normal water. These mice were randomly and evenly divided into 6 groups: and P15 2D or 3D UC-MSCs (1×10 6 cells/ dose) respectively. Mice in normal and DSS groups were administrated with equal volume of PBS. The body weights were recorded every day. Disease activity index (DAI) calculated according to changes of body weight, diarrhea and hematochezia as previously described [23]. Four days after UC-MSCs injection, colons were entirely resected for length measurement and pathological analysis. For pathological analysis, the colon segments were xed with 4% paraformaldehyde, embedded in para n and subjected to H&E staining. The conditions of in ammation and edema of colons were evaluated under a light and polarized microscopy.

Experimental T1DM and Treatment by UC-MSCs
T1DM has been induced by a single intraperitoneal injection of streptozotocin (STZ, 50 mg/kg, MP Biomedicals, USA, 02100557-CF) in female C57/BL6J mice (8 weeks old) for 5 days. The blood glucose was measured on the 8th and 11th day. The experimental T1DM mice are those the levels of blood glucose were than 11.1 mmol/L. For 3 consecutive weeks, diabetic mice were divided into 6 groups and intravenously administrated with P5 and P15 2D or 3D UC-MSCs twice (day1 and day14). At the end of the experiment, the changes in glucose metabolism, C-peptide and GHb levels were evaluated. Furthermore, histological examination of pancreatic tissues was performed.

Statistical Analysis
Statistical analysis was carried out with GraphPad Prism version 7.0 software. T-test and one-way analysis of variance (ANOVA) were used for differences between two groups and more than two groups respectively. P-value < 0.05 was considered statistically signi cant. The data are presented here as Mean ± Standard Error of Mean (SEM).

3D-culture Preserves the Stemness and
Immunosuppression of UC-MSCs during Long-term Passaging MSCs are promising candidates for disease treatment and tissue regeneration. However, the stemness and immunosuppression of UC-MSCs is decreased after long-term passaging ( Fig. S1A-C), resulting in failure of MSCs therapy. Here, we explored the impact of 3D-culture-based strategy on the preserve of stemness of MSCs. UC-MSCs were cultured to P3 on TCP, then were cultured by 2D and 3D method respectively. UC-MSCs were continuously passaged until to P15 (Fig. 1A). After P15, 3D-culture was superior to 2D method in the maintenance of UC-MSCs surface speci c antigens (Fig. 1D, S2A). Proliferation assay revealed that 3D method was signi cantly improved colony formation of P15 UC-MSCs (Fig. 1B, C). Similarly, the 3D environment promoted the proliferation of P15 UC-MSCs (Fig. 1E).
Additionally, the P5 and P15 UC-MSCs were co-cultured with activated T cells respectively (Fig. 1F). The proliferation of T cells and ELISA assay for IL-6, IL-10 and TGF-β of the supernatant showed that immunosuppression of 3D UC-MSCs was enhanced compared with 2D UC-MSCs both on P5 (Fig. 1I) and P15 (Fig. 1J). Furthermore, ow cytometry analysis showed that 3D UC-MSCs inhibited the differentiation of Th1 and Th17 and promoted the differentiation of Th2 and Treg. The ratios of Th1/Th2 and Th17/Treg in T cells co-cultured with 3D UC-MSCs were signi cantly decreased (Fig. 1G, H, S2B). The results showed that 3D UC-MSCs can better maintain the balance of T cell subsets, and thus maintaining the immunity stability of the body.

3D-culture Preserves the Therapeutic Effects of Longterm Passaged UC-MSCs on Immune Disorders
We adopted experimental colitis mice model and T1DM, two widely used models of immune disorders.
After DSS feeding and UC-MSCs transplantation, we recorded the data on mice, including loss of weight, diarrhea and bloody stool. Transplantation of P15 3D cultured MSCs was still showed the therapeutic effects on colitis, while transplantation of P15 2D cultured MSCs was less effective ( Fig. 2A, B, D, S3A). Histological analysis con rmed that DSS feeding caused severe damage in colon mucosa. Tissue section displayed P5 2D and 3D MSCs had similar therapeutic effects. However, infusion of P15 3D cultured MSCs was more effective than 2D cultured MSCs on prevent damage in colon mucosa (Fig. 2C).
To test the therapeutic effects on immune disorders of long-term passaged UC-MSCs, we established a T1DM mice model by STZ. T1DM mice showed blood glucose levels higher than 11.1mmol/L, while the normal group was lower than 10mmol/L, regardless if the mice were fasted or re-fed. The results of continuous blood glucose tests showed that 3D-culture MSCs treatment provided better performance. The 3D cultured MSCs provided therapeutic effects till P15 while the 2D not (Fig. 2E). Reduced insulin sensitivity and decreased islet size were indicated by the results of oral glucose tolerance tests (OGTTs) (Fig. 2F) and pancreas histology (Fig. 2G). The 3D cultured MSCs were more effective on islet protection than 2D cultured MSCs both on P5 and P15. The histology showed that P15 3D cultured MSCs maintained their therapeutic effects (Fig. 2G). GHb assay can re ect the overall control of blood glucose in T1DM mice (Fig. 2I). The results showed that the 3D cultured MSCs was more effective on blood glucose control than 2D cultured MSCs. The P15 2D cultured MSCs were similar to the PBS group. In contrast, the P15 3D cultured MSCs were still effective (Fig. 2H). The C-peptide levels were proportional to serum insulin and can accurately re ect the secretion function of islet β cells. After 7 days infusion of MSCs, C-peptide levels of 3D cultured MSCs treated T1DM mice were maintained better, which did show signi cant differences compared to that of the 2D cultured MSCs (Fig. 2J).

3D Culture Supports Long-Term Expansion through YAP Activation
We then explored the mechanism of 3D culture to prevent stemness loss of long-term passaged UC-MSCs. The immuno uorescent analyses showed that 3D culture system signi cantly promoted the proliferation of UC-MSCs both on P5 and P15 (Fig. 3A, B). The staining against anti-active YAP1 revealed that YAP1 expressions in the nucleus were signi cantly enhanced by 3 times of P5 and 7 times of P15 respectively (Fig. 3C). Notably, 3D culture system increased the active YAP1 expression in the nucleus by 25 times through qPCR assay of P5 UC-MSCs (Fig. 3D). Western blot analysis also supported that 3D culture enhanced YAP activation (Fig. 3E).

Regulation of YAP is Effective on Immunosuppression of MSCs
To con rm whether the activation of YAP affect the therapeutic effects of the cells, we further investigated the function of YAP1 on UC-MSCs. We pretreated UC-MSCs with PGE2 to promoted YAP1 and vertepor n (VP), which is a YAP speci c inhibitor to inhibit YAP1 [24,25]. The immuno uorescent and western blot analysis showed that PGE2 treated UC-MSCs have higher level of YAP1, while VP treated UC-MSCs have lower level of YAP1 (Fig. 4A-D). Furthermore, the 2D UC-MSCs, PGE2-treated 2D UC-MSCs, 3D UC-MSCs and VP-treated 3D UC-MSCs were co-cultured with activated T cells respectively. Flow cytometry analysis showed that UC-MSCs inhibited the differentiation of Th1 and Th17 and promoted the differentiation of Th2 and Treg (Fig. 5A, B, S4A-D). This immunosuppression was enhanced after PGE2treated and declined after VP-treated. The ratios of Th1/Th2 and Th17/Treg in T cells co-cultured with PGE2-treated UC-MSCs were signi cantly decreased while the VP-treated UC-MSCs increased. It is suggesting YAP activation preserved the therapeutic effects of MSCs.

Regulation of YAP is Effective on the Therapeutic Effects of MSCs on Immune Disorders
We next con rmed the function of YAP1 activation by experimental colitis and T1DM Model. The experimental colitis mice were divided into 6 groups (normal, DSS + PBS, DSS + 2D UC-MSCs, DSS + PGE-2D UC-MSCs, DSS + 3D UC-MSCs and DSS + VP-3D UC-MSCs). By comparing loss of weight (Fig. 6A), diarrhea, bloody stoolt (Fig. 6B), colon length (Fig. 6D, S5A) and histological analysis (Fig. 6C), PGE2treated UC-MSCs were more effective on colitis than no-treated, while VP-treated UC-MSCs were less effective than no-treated. The experimental T1DM mice were also divided into 6 groups. The results of therapy showed that PGE2-treatment improved the therapeutic effects of UC-MSCs on blood glucose control (Fig. 6E, F), insulin sensitivity and islet protection (Fig. 6G-J). On the contrary, VP-treatment reduced the therapeutic effects of UC-MSCs (Fig. 6E-J). In this study, we demonstrated that PGE2-treated UC-MSCs could improve the therapeutic effects of UC-MSCs. On the contrary, VP-treatment decrease the therapeutic effects. It is suggested that YAP activation improves the therapeutic effects of UC-MSCs, while YAP inhibition decreases it.

Discussion
Due to their immunomodulatory property, MSCs have been increasingly used to treat a range of immunopathy [26,27], and 3D cultured MSCs show enhanced immunomodulatory property compared with those in 2D culture [1]. However, both the regulatory mechanisms remain unclear. Meaningfully, we modulated YAP pathway both of 2D and 3D cultured UC-MSCs and validated its role in colitis and T1DM mouse models, which proved that mechanical stress played an important role in maintaining the immunomodulatory function and therapeutic potential of 2D and 3D cultured UC-MSCs, and PGE2 could be used to up-regulate YAP and improve the immunomodulatory ability of 2D cultured UC-MSCs after extensive expansion.
As a kind of adult stem cells, MSCs possess two characteristics that are distinct from other adult cells. First, they have stemness, as embodied by their self-renewal and tri-lineage differentiation into osteoblasts, chondrocytes and adipocytes. Second, MSCs own immunomodulatory property [28]. Studies have shown that although cell replacement plays an important role in MSCs therapy for speci c diseases, the nal therapeutic effects are mainly the results of MSCs-derived immunomodulatory ability of inhibiting the proliferation and functions of various immunocytes, including T lymphocytes, B lymphocytes, dendritic cells, macrophages, natural killer cells and neutrophils, suppressing the differentiation of naive CD4 + T cells into proin ammatory Th1 and Th17 cell lineages and promoting the generation of Treg cells [3,29,30]. Here, the P5 UC-MSCs did exhibit both features, which were consistent with previous ndings.
MSCs release cytokines into the microenvironment through autocrine and paracrine to maintain living conditions. The microenvironment, in turn, conducts variation of the environment in the whole body through changes in metabolism, secretion, immunity and functions, limiting and affecting the occurrence and development of MSCs [31]. However, traditional 2D culture cannot mimic the in vivo architecture and microenvironment well. Therefore, 2D cultured MSCs display many characteristics of replicative senescence compared with in vivo MSCs [1,32]. Furthermore, senescent cells secrete IL-6 and IL-8, which act in an autocrine and a paracrine way to reinforce senescence [33]. In this study, compared with P5 UC-MSCs, when P15 UC-MSCs were co-cultured with activated T cells, it showed enhanced T cell proliferation, increased IL-6 and decreased IL-10 and TGF-β, which may be the reasons for the decrescent therapeutic effects of 2D cultured UC-MSCs.
Increasing studies have explored drugs that can reverse the replicative senescence of MSCs. For example, reduced glutathione (GSH) and melatonin displayed an anti-senescent effect in MSCs and preserved stem cell functions including cell migration, stemness, and multidirectional differentiation potential through reducing reactive oxygen species (ROS) generation during long-term in vitro expansion [34]. Imperfection of 2D culture has also inspired the emergence of 3D cell culture systems. Up to now, various types of architectures and biomaterials that support 3D cell proliferation have been reported. For one instance, TableTrix™ (Cytoniche) is a commercially available microcarrier based on macroporous and elastic gelatin [35]. Hydrogels, a common material form, have also been successfully used to support 3D cell growth in vitro, such as Extracell™ (Dextran Biological System) [36]. However, existing drugs or products have mostly focused on retarding the senescence phenotype of MSCs and promoting the proliferation and differentiation in the process of long-term expansion, while the improvement of immunomodulatory ability, a hinge of boosting the therapeutic effects of MSCs, has not received much attention.
In a 3D culture system, cells can better communicate with neighboring cells, and the cell-to-cell and cellto-matrix connections are re ected in the culture environment, which are consistent with the in vivo microenvironment [37,38]. Moreover, 3D cultured MSCs signi cantly maintain elevated expression levels of stemness genes, yield high frequencies of colony-forming units and show enhanced osteogenic and adipogenic differentiation e ciency [39]. One research also indicated that the expressions of HLA-G, IDO1, PTGS2 and TGF-β1, which were closely related to immune regulation, were higher in 3D culture than those in 2D culture. Further, 3D cultured MSCs could inhibit the proliferation of activated lymphocytes more effectively [17]. Here, we compared the stemness and immunomodulatory ability between 2D and 3D UC-MSCs and found that 3D culture did preserve the stemness, T cell proliferative inhibition and the expression of anti-in ammatory cytokines such as IL-10 and TGF-β of long-passaged UC-MSCs.
Moreover, we observed that 3D cultured UC-MSCs at P5 or P15 was more effective in treating T1DM and colitis mice than those in 2D culture, adding to the current knowledge of 3D cultured MSCs for the treatment of immune disorders.
Most notably, the morphology of MSCs in the 3D cultured systems markedly differ from those on 2D surfaces throughout the cultivation process [40], and the morphology of MSCs can be affected by a variety of mechanical properties, including elasticity, geometry and adhesion, which will ultimately affect the tension of the cytoskeleton [41]. It has been widely accepted that the transcriptional effector of the hippo pathway-the YAP/transcriptional coactivator with PDZ-binding motif (TAZ) acts as a sensor and mediator of mechanical signals in response to extracellular matrix cues [42,43] Meanwhile, YAP signaling is strongly activated during the culture of cells in the matrigel-encapsulated 3D culture system and maintains the proliferative capacity of the encapsulated cells [18]. In line with these phenomena, we found the YAP expression both of P5 and P15 UC-MSCs in 3D culture was signi cantly higher than those in 2D culture. Furthermore, we demonstrated that mechanical stress played a signi cant role in maintaining immunoregulatory function of 2D and 3D cultured UC-MSCs by modulating YAP pathway, and PGE2 could be used to up-regulate YAP and improve the immunomodulatory ability of long-passaged 2D cultured UC-MSCs, providing theoretical supports for the regulation of cellular immunity and therapeutic e cacy by adjusting mechanical stress.

Conclusions
Our study indicated that 3D culture extended the stemness, immunomodulatory ability and therapeutic  The ratios of Th1/Th2 and Th17/Treg were shown in the graphs. (I, J) Cell proliferation and cytokines in supernatant of UC-MSCs co-cultured with activated CD3+ T cells were detected by CCK-8 and ELISA.
(H) Area of pancreatic islets in six groups sections were quanti ed. (I-J) GHb and C-peptide levels in serum were evaluated by ELISA at day 21 (scale bar=250 μm, values were shown as mean ± SD, n = 4 per group).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryData.docx