Mycophenolate mofetil reduces epidural brosis by regulating TGF-β1/Smad2/3 axis to repress the proliferation, migration and differentiation of broblasts

Excessive broblasts proliferation is believed as a major reason in the process of epidural brosis, which is known as a troublesome complication of lumbar laminectomy clinically. Mycophenolate mofetil (MMF) is a kind of immunosuppressant, previous studies had shown that it has the function of preventing brosis, but the role and mechanism are still unclear. In this study, we determined the repressed effect of MMF on broblasts proliferation, migration and differentiation in surgical-induced epidural brosis and the underlying mechanism.

One of the major complications after lumbar laminectomy is epidural brosis, which can seriously affect the surgical outcome [1]. An important factor contributes to this complication is the massive proliferation of broblasts, which form a matrix of scar adhesions outside the epidural, causing many related symptoms such as low back pain [2]. As the number of patients undergoing laminectomy continues to increase, the incidence of epidural brosis is increasing year by year. The main way to solve this complication is re-surgical, which not only reduces the patient's quality of life, but also increases the patient's nancial burden [3]. This troublesome postoperative complication is an urgent problem for contemporary spine surgeons.
Many researchers have explored various ways to prevent epidural brosis, such as the use of antimetabolites, injecting drugs at the surgical site, and locally implanting synthetic materials to ll the postoperative defect area [4][5][6]. These measures have achieved some curative effects more or less, but there are many side effects and limitations [7].
Related reports in recent years indicate that the main cause of epidural brosis after the operation is broblasts undue proliferation [8][9][10]. Many articles have concluded that inhibiting broblast proliferation can effectively reduce epidural brosis and reduce scar tissues hyperplasia. Some articles show that the TGF-β1 / Smad2/3 axis is extensively participated in the formation of brosis, and repression of the channel can inhibit the proliferation of broblasts [11,12]. In addition, many studies have indicated that inhibition of broblasts migration and differentiation can also achieve better anti-brotic effects [13,14].
Mycophenolate mofetil (MMF) is a kind of immunosuppressant that is mainly used to prevent rejection and treat refractory rejection in patients with allogeneic kidney transplantation [15]. Recently, many experiments have reported that MMF has a therapeutic effect on a variety of brotic diseases [16][17][18].
This study aims to explore whether MMF is effective in inhibiting epidural brosis after laminectomy, and to explore its possible mechanism of action: Whether MMF could inhibit broblast proliferation via TGF-β1 / Smad2/3 axis. To nd a more effective and safe treatment to help prevent this surgical complication.

Cell cycle analysis
To explored the principle of the MMF which induced proliferation of the broblasts, we used Cell Cycle Testing Kit (KeyGEN BioTECH, Jiangsu, China) to analysis the cell cycle distribution of various concentration groups. We cultured broblasts in 10 cm dishes, and treated cells with different concentrations of MMF (0, 0.1, 1, 10 µM) or TGF-β1 (5 ng/ml) in the logarithmic growth phase. Collected cells after the treatment then centrifuged at 1000 g/min for 5 min, throwed away the supernatant. Used the cold phosphate buffered saline to wash the wells once and used 70% ice-cold ethanol to xed them. In dark room, resuspended the broblasts in 0.5 ml propidium iodide (PI) dye solution with 37℃ water bath about 30 min. We detected the cells by ow cytometry (MoFloTM XDP, German). Used ModFit software to analyze cell cycle distribution.

Cell wound scratch assay
Inoculated the cells in 6 well plates, and grown until 85-90% con uence. Then wounded the cells with white pipette tips and washed three times with phosphate buffered saline. Then hatched the cells with MMF (0, 0.1, 1, 10 µM) or TGF-β1(5 ng/ml) in medium supplemented that with 2% serum for 24 h. We took the photo of the scratch area under an inverted microscopy. We measured the wounded area from treatment wells at three randomly selected wound sites for statistical analysis, and we calculated three independent scratch-wound experiments.

Transwell assay
We prepared transwell chambers to conduct the experiment. The cells were collected in serum-free medium, its nal concentration were 1 × 10 5 cell/ml. Then treated with MMF (0, 0.

Western blotting analysis
After treatment with MMF (0, 0.1, 1, 10 µM) or TGF-β1 (5 ng/ml), we lysed the cells with equal volume lysis buffer (CST, American) on ice. Then used ultrasonic wave to treat cell lysate. Centrifugated at 15000r, 4℃ for 10 minutes. Measured the protein concentrations by BCA Protein Assay Kit (Thermo, USA). We used 10% sodium dodecyl-sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) (Beyotime, Jiangsu, China) to separate equal amount protein extracts, then transferred onto polyvinylidene di uoride membranes (Millipore, America). After transferred, at indoor temperature, immersed the polyvinylidene uoride membrane in 5% skim milk in 1x TBST about 2 h. Then hatched with diluted primary antibodies at 4℃ all night according to the instruction books. Washed the membranes 3 times with 1x TBST for 10 minutes and then hatched with secondary antibody (anti-mouse or anti-rabbit IgG, CST, America) for 2 hours at indoor temperature. Membranes were washed for 3 times with 1x TBST as well after incubated. Finally, added ECL reagent (enhanced chemiluminescence detection) on the membranes then used the ChemiDoc XRS + system (Bio Rad, America) to explore the membranes.

Animals
We bought SD male rats weighing between 300 and 350 g from the experimental animal center of Yangzhou University (Yangzhou, China). Animal Ethics Committee of Yangzhou University rati ed this study about animals, and we cared each rat carefully. We divided the rats into four groups (12 rats per group) randomly: control group (saline) and MMF groups (low, medium and high dose groups).

Rats laminectomy model and topical application of MMF
Established laminectomy models according to the methods in previous study. Brie y, we anesthetized the rats through intraperitoneal injection with 1% sodium pentobarbital (40 mg/kg), then located the lumbar 1 position by pinching the rib pedicle. Revealed fascia and isolated the paravertebral muscles bluntly. Then we used mini-biting forceps to remove the L1 vertebral plate and expose the 5 × 3 mm defect in the dura mater district. After the surgical area got adequate hemostasis, we washed the area with saline. Then, the gauzes thorough soaked with 0.2 mg/ml saline and MMF in different concentrations were placed on the operation area for 15 min. Then we washed the surgical area and sutured the wounds layer by layer. For the sake of preventing notch infection, each rat had intramuscular injection with penicillin (Lukang, Shandong, China) in once a day for three days, and injected 50 mg of penicillin per kilogram of body weight each mouse.

Histological analysis
1 month later, killed the rats with fetal dose of sodium pentobarbital after laminectomy. Then we perfused the rats in cardiovascular with 4% paraformaldehyde. After that, we removed the whole L1 spinal column and immerged in 4% paraformaldehyde. We decalci ed the samples in ethylenediamine tetraacetic acid (EDTA) about 1 month and then embedded in para n. Continuous 4 µm samples were made of para n lumps which selected from these rats randomly. Then we used hematoxylin-eosin (H&E) and Masson's trichrome to stain the samples respectively. We used light microscope to take photos of epidural brotic tissues under 40 and 200fold microscope. The study calculated the broblasts in three counting elds (100 × 100 mm each) of the surgical areas by using Image J, the magni cation was × 200.

Immunohistochemical staining
We used immunohistochemical staining to explore the impacts of MMF on broblasts proliferation, migration and differentiation in epidural brosis. PCNA can be a sign of recent cell proliferation, tubulin can be a sign of cell migration, and α-SMA can be a sign of cell differentiation. Therefore, we used immunohistochemical staining to detect the content of PCNA, tubulin and α-SMA in brotic tissues. After depara nization and alcohol dehydration, we used citrate buffer to acting on the samples to repair antigen, and then exposed the samples in 3% H 2 O 2 to interdict endogenous peroxidase activity. We washed these slices with PBS three times, and hatched them with anti-PCNA, anti-tubulin and anti-α-SMA antibodies all night at 4℃. Followed by incubation with anti-rabbit or mouse IgG at RT for 2 h. Then we used DAB kit to analysis the slices and detect the binding of antibodies. In the end, we counterstained the slices with hematoxylin and used light microscope to observe them. For the study, we used Image J to calculate the proportion of positively cells per ve 200 × magni cation elds.

Statistical analysis
We used SPSS 19.0 statistical tool to analyzed the results in our experiments, and used mean ± standard deviation (S.D.) to indicate all results. The one-way analysis of variance and independent-samples T test was used to compare the data of each group. P value < 0.05 is considered with statistical signi cance.

MMF suppress broblasts proliferation
As shown in Fig. 1B, the results of CCK-8 assay demonstrated that MMF repressed broblasts proliferation with the increased concentration and longer effect time. The results of EdU incorporation assay proved the inhibition effect of MMF on the proliferation of broblasts (Fig. 1C). To further con rmed how MMF regulated the broblasts proliferation, we used ow cytometry to explore the cell cycle. As the results in Fig. 1D, MMF prevented the cell cycle course in the G0/G1 phase. The analysis about Western blotting indicated that the synthesis of proliferation proteins, for example, PCNA and Cyclin D1, decreased as MMF concentration increased (Fig. 1E). The results of the above experiments proved that MMF could inhibit the proliferation of broblasts.

MMF inhibits migration of broblasts in vitro
Cell wound scratch assay demonstrated that structure and quantitative damage of the cytoskeleton led to the reduction of migration capacity ( Fig. 2A), the percentage of scratch reduction area decreases with increasing MMF concentration. In the transwell assay, we proved the inhibited impact of MMF on the migration of broblasts as well. The amount of cells went through the transwell was reduced along with the increase concentration of MMF (Fig. 2B). Immuno uorescence demonstrated that MMF repressed the synthesis of migration proteins and changed structure of the cytoskeleton (Fig. 2C). As the increase of MMF concentration, the direction of the migrating proteins changed from uniform to chaotic and diffused.
Western blot was used to detect the synthesis levels of the migration proteins like vinculin and tubulin (Fig. 2D). MMF repressed the synthesis of vinculin and tubulin obviously. Our results indicated that MMF repressed migration of broblasts in vitro.

MMF inhibits differentiation of broblasts in vitro
We explored the impacts of MMF on the broblasts differentiation. The outcomes displayed that the quantitative decrease synthesis of differentiation protein like α-SMA was indicated by immuno uorescence (Fig. 3A). Besides, the synthesis quantity of α-SMA was decreased with concentrations of the drug increased (Fig. 3B). These results indicated that MMF inhibits differentiation of broblasts in vitro.

MMF regulates TGF-β1/Smad2/3 axis expression
To explore whether the MMF regulated TGF-β1/Smad2/3 axis expression, we treated broblasts with MMF of 0, 0.1, 1, 10 µM, then used Western blot to detect channel correlative protein expression. As the results in Fig. 4A, along with the treatment of various concentration of MMF, the synthesis quantities of TGF-β1, p-Smad2, p-Smad3 were decreased with concentrations of the drug increased. TGF-β1, which was known as an activator of TGF-β1/Smad2/3 axis, partially increased the low under-expression quantities of p-Smad2, p-Smad3 induced by MMF (Fig. 4B). The above results showed that MMF could regulate TGF-β1/Smad2/3 axis expression.

MMF repressed broblasts proliferation, migration, differentiation, and reduced epidural brosis in rats
The results of H&E-staining showed that applied topically of MMF was able to reduce brotic tissues in epidural. Compared with the saline group, the topical application of MMF could signi cantly arrest the epidural brosis, and its action trend was basically consistent with the concentration. The brotic tissues in 0.16 mg/ml group were obviously less than that in medium, low concentration groups and saline group (Fig. 8A). We calculated the amount of broblasts in brotic tissues taken from surgical areas demonstrated that MMF could obviously repressing the broblasts proliferation (Fig. 8B, Fig. 8C). Meanwhile, we used Masson's trichrome staining to con rm collagen content in brotic tissues from the surgical areas. Optical density analysis proved that the content of collagen in the experimental groups was signi cantly reduced compared with which in the saline group (Fig. 8D). The PCNA immunohistochemistry staining in broblasts also re ected that the proliferation of cells in treatment group was weaker than the proliferation of cells in control group (Fig. 9A). Moreover, the tubulin immunohistochemistry staining suggested that MMF could inhibit the migration of broblasts (Fig. 9B). The α-SMA immunohistochemistry staining proved that MMF could inhibit the differentiation of broblasts (Fig. 9C).

Discussion
Epidural brosis is a deformed healing of the incision site after laminectomy. In this process, broblasts accumulate at the incision after a large number of proliferation [19]. As a pathological process, epidural brosis can cause a large amount of brotic tissue to form in the surgical site after laminectomy, which seriously affects the surgical effect of patients [20]. The mechanism of the proliferation of broblasts is relatively complicated, and there are still many related problems that need to be solved. However, it is generally recognized that excessive proliferation of broblasts is the main factor causing tissue brosis [21][22][23].
TGF-β1 has extensive biological functions, including regulating cell growth, differentiation, migration, matrix formation, and damage repair [24][25][26]. It is the most in-depth and most powerful pro brotic cytokine currently studied [27][28][29]. TGF-β1 is a powerful chemokine for monocytes and broblasts [30]. As well as, it can increase the proliferation ability of broblasts, promote their secretion of extracellular matrix such as collagen and bronectin [31]. TGF-β1 can induce broblasts phenotype Transformation to myo broblasts phenotype, characterized by high synthesis of α-SMA, causing matrix contraction and participating in accelerating the extracellular matrix deposition [32]. It also has the function of repressing extracellular matrix degradation and promoting apoptosis of epithelial cells [33]. At the same time, it is an important cell that induces epithelial-mesenchymal transition, making epithelial cells lose their polarity, adhesion molecules such as cadherin, and obtain characteristics of mesenchymal cells, such as the synthesis of bronectin, α-SMA, and exhibit migration [34]. It can be seen that it has a signi cantly impacts on the mechanism of epidural brosis formation.
The Smad protein is the main signal transduction factor of TGF-β1, which transfers the signal of TGF-β1 from the cell to the nucleus regulates target gene transcription [35]. Co-regulates the transcription of target genes with other transcription factors [36]. Many articles indicate that repress the TGF-β1 channel can inhibit broblasts functions [37][38][39].
MMF is a derivative of mycophenolic acid (MPA), and it is a new immunosuppressant. Rapid hydrolysis of MMF after oral administration into the active metabolite MPA in the body. MPA inhibits the synthesis of guanine nucleotides by repressing the key rate-limiting enzyme inosine monophosphate dehydrogenase (IMPDH), a new channel for the synthesis of purine nucleotide [40]. Due to its speci city, MMF has recently been used in the treatment of proliferative brotic diseases and has achieved good results. Badid et al. [13]found that 72 hours after MMF treatment, MMF can dose-dependently inhibit rat broblasts proliferation, reduce in ltration of interstitial myo broblasts and deposition of type III collagen. Liu et al. [41] found that MMF improves renal function and reduces renal interstitial brosis in a rat renal interstitial brosis model with 5/6 resection of the kidney. Nina [42] and other studies found that MMF can inhibit the expression of collagen genes, the contraction of extracellular matrix, and the migration of broblasts suggest that it may have an anti-brotic effect. Our experimental results also proved the anti-brosis effect of MMF. Dubus et al. [43] found that MMF can reduce TGF-β1-induced human epithelial cell proliferation and extracellular matrix production (such as type I collagen and laminin deposition). It can be seen that MMF can not only inhibit lymphocyte proliferation and recruitment, but also have a direct anti-brosis effect, but there is no related research on epidural brosis. We speculate that MMF can terminate the signal transduction of TGF-β1-related channel, thereby inhibiting the proliferation, migration and differentiation of broblasts. Our study showed that MMF could also inhibit the differentiation of broblasts, which was barely involved in the above studies, and our study on cell differentiation helped deepen the understanding of the mechanism of MMF anti-brosis.
Therefore, we speculated that MMF may repress the functions of broblasts by regulating the TGF-β1 channel. In this research, we used Western blot to research uctuation in related protein synthesis. The results showed that the synthesis quantities of TGF-β1, p-Smad2, p-Smad3, PCNA and Cyclin D1 decreased after MMF treatment. These results indicate that MMF can regulate TGF-β1 channel and its channel proteins to repress broblasts proliferation.
In vitro experiments, we conducted a number of cell experiments to investigate whether MMF can inhibit the proliferation, migration, and differentiation of broblasts. Our results indicate that MMF can play these roles. At the same time, we also found that MMF signi cantly repressed the TGF-β1 / Smad2/3 axis. Therefore, we used a potent activator, TGF-β1, to activate the channel. According to the results of Western blot analysis, we found that TGF-β1 effectively activated the TGF-β1 / Smad2/3 axis. The activated channel attenuated impacts of MMF on the proliferation, migration, and differentiation of broblasts.
In vivo experiments, we established a laminectomy model in rats. After laminectomy, different concentrations of MMF were applied locally to rat incisions. The selection of the MMF concentrations was based on previous studies in rats. H&E-staining histological analysis of epidural brotic tissues demonstrated that brosis in the brotic tissues from the surgical areas in the MMF treatment group was signi cantly weaker, and the number of broblasts was signi cantly reduced. We used Masson's trichrome staining to evaluate collagen content in brotic tissues. Our results showed that MMF can signi cantly inhibit collagen content. Further research on histochemical staining demonstrated that the synthesis of PCNA, tubulin and α-SMA in brotic tissues in the group treat with MMF was obviously less than in the saline group. These studies indicate that MMF could reduce epidural brosis caused by operation.
In conclusion, based on above results, we can preliminary explain the role of MMF on inhibiting epidural brosis. In short, when applying MMF to the laminectomy site, the TGF-β1 / Smad2/3 axis in broblasts is repressed, and nally the proliferation, migration and differentiation of broblasts are inhibited, thereby reducing the ber formation.
The mechanism of epidural brosis formation is intricacy, there are still some limitations in our research. Firstly, MMF was administered topically. Although all animals survived until the study nished, a few side effects may happen on other regions. Secondly, we only studied the effect of TGF-β1 / Smad2/3 axis in inducing broblasts proliferation, migration, and differentiation, without further exploring MMF may affect broblasts proliferation through other signaling pathways. These are the shortcomings of this research, and we will start to conduct more in-depth research in these areas in the future.

Conclusion
In summary, we have initially veri ed that MMF can inhibit broblasts proliferation, migration, and differentiation. Then we propose a new concept: MMF may affect these three broblasts activities by repressing the transduction of the TGF-β1 / Smad2/3 axis. To this point, further exploration is needed in the future.

Ethics approval and consent to participate
The research was conducted in accordance with the guidance of the Animal Ethics Committee of Yangzhou University, China, and pathology laboratory of Yangzhou University.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was supported by Jiangsu Provincial Medical Innovation Team (grants#CXTDB2017004).
Authors' contributions DZ and JZ performed the whole experiments and were responsible for the data and drafting of the article. YS and LY designed the study and contributed to the preparation of the manuscript. XZ, XL, and YL helped in the performance of animal surgeries and the interpretation of data. All authors read and approved the nal manuscript.     proliferation of broblasts by EdU incorporation assay. All nuclei were blue and the nuclei of EdU-positive cells were green. Five different elds were randomly selected under a uorescence microscope to calculate the percentage of EDU-positive cells. Each group counted more than 50 broblasts for EDU analysis. After treatment with the activator TGF-β1, MMF-induced broblasts proliferation inhibition was reversed to a certain extent. MMF group compared with MMF mixed TGF-β1 group, *P < 0.05. (B) Cell cycle analysis showed that after treatment with the activator TGF-β1, the proportion of cells in G1 phase was decreased. The proliferation inhibition of broblasts induced by MMF was reversed by the activator TGF-β1. MMF group compared with MMF mixed TGF-β1 group, *P < 0.05.  showed that after treatment with the activator TGF-β1, the percentage of scratch reduction area were higher than the treatment group which only treated with MMF (10μM). The migration inhibition of broblasts induced by MMF was reversed by the activator TGF-β1. MMF group compared with MMF mixed TGF-β1 group, *P < 0.05. (B) Transwell assay showed that after treatment with the activator TGF-β1, the amount of cells went through the transwell was increased. MMF group compared with MMF mixed TGF-β1 group, *P < 0.05. (C) Immuno uorescence staining of tubulin and vinculin showed that after treatment with the activator TGF-β1, the amount of these protein which expressed in the cytoplasm of broblasts were increased. MMF inhibits broblasts differentiation by inhibiting TGF-β1 / Smad2/3 axis. Immuno uorescence staining of α-SMA showed that after treatment with the activator TGF-β1, the amount of α-SMA in the cytoplasm of broblasts were increased. MMF group compared with MMF mixed TGF-β1 group, P < 0.05.

Figure 8
Histological analysis of the laminectomy areas. (A) In the HE images, we found dense or thick brosis tissues (*) with extensive or rm adherence (arrow) to dura mater in control group (0 mg/kg). We observed medium brosis tissues (*) with medium adherence (arrow) to dura mater in laminectomy defect areas of 0.04 mg/ml MMF-treated group. Then we found sparse brosis tissues (*) with few adhesions to dura mater (arrow) in laminectomy defect sites of 0.08 mg/ml MMF-treated group. Looser brosis tissues (*) with fewer adhesions to dura mater (arrow) were observed in 0.16 mg/ml MMF-treated group. The laminectomy defect site was marked by "L", and spinal cord was marker by "S". All sections were stained by the mean of HE, and the magni cation was ×40. (B) The images of broblasts in epidural brosis tissues of each group showed that the number of broblasts was decreased along with the increase of MMF concentration. The magni cation was ×200. (C) Fibroblasts number was calculated by randomly selecting three counting areas from every section. * Compared with the control group, P < 0.05.
(D) After the slices were stained with Masson's trichrome, the images showed that collagen content in epidural brosis tissues decreased in MMF treatment group. The magni cation was ×200. Optical density analysis of Masson staining images showed that MMF could reduce collagen production. * Compared with the control group, P < 0.05.