IL-33 Promotes ST2-Dependent Fibroblast Maturation via P38 and TGF-β in a Mouse Model of Epidural Fibrosis

Abstract

BACKGROUND:

Recent evidence suggests that IL-33, a novel member of the IL-1β family, is involved in organ fibrosis. However, the roles of IL-33 and its receptor ST2 in epidural fibrosis post spine operation remain elusive.

METHODS:

A mouse model of epidural fibrosis was established after laminectomy. IL-33 in the wound tissues post laminectomy was measured with Western blotting, ELISA and immunoflurosence imaging. The fibroblast cell line NIH-3T3 and primary fibroblasts were treated with IL-33 and the mechanisms of maturation of fibroblasts into myofibroblasts were analyzed. To explore roles of IL-33 and its receptor ST2 in vivo, IL-33 knockout (KO) and ST2 KO mice were employed to construct the model of laminectomy. The epidural fibrosis was evaluated using H&E and Masson staining, western-blotting, ELISA and immunohistochemistry.

RESULTS:

As demonstrated in western blotting and ELISA, IL-33 was increased in epidural wound tissues post laminectomy. The immunoflurosence imaging revealed that endothelial cells (CD31⁺) and fibroblasts (α-SAM⁺) were major producers of IL-33 in the epidural wound tissues. In vitro, IL-33 promoted fibroblast maturation, which was blocked by ST2 neutralization antibody, suggesting that IL-33-promoted-fibroblasts maturation was ST2 dependent. Further, IL-33/ST2 activated MAPK p38 and TGF-β pathways. Either p38 inhibitor or TGF-β inhibitor decreased fibronectin and α-SAM production from IL-33-treated fibroblasts, suggesting that p38 and TGF-β were involved with IL-33/ST2 signal pathways in the fibroblasts maturation. In vivo, IL-33 KO or ST2 KO decreased fibronectin, α-SMA and collagen deposition in the wound tissues of mice that underwent spine surgery. In addition, TGF-β1 was decreased in IL-33 KO or ST2 KO epidural wound tissues.

CONCLUSION:

In summary, IL-33/ST2 promoted fibroblast differentiation into myofibroblasts via MAPK p38 and TGF-β in a mouse model of epidural fibrosis after laminectomy.

Introduction

The sedentary lifestyle is considered a risk factor for developing herniated discs [1], which require spine surgery (i.e., laminectomy) to relieve the compressed spinal cord. Unfortunately, up to 10–46% of patients may develop failed back surgery syndrome (FBSS) after laminectomy [2]. Epidural scarring is a process following spine surgery. Excessive scar tissues firmly attach to the dura matter and may reduce spine flexibility, leading to epidural fibrosis and FBSS. Although our team [3, 4] and other laboratories [5–7] have proposed many biological methods to relieve postoperative epidural scar tissue hyperplasia, the mechanisms of epidural fibrosis post laminectomy are still elusive.

It is well recognized that interleukin-1β (IL-1β), a typical inflammatory cytokine, is involved in acute inflammation after spine surgery [8]. As an IL-1-like family member, IL-33 is a multifunctional cytokine. IL-33 contributes to acute inflammation in the early stage and the repair process in the later stage. Once released, full-length IL-33 may be cleaved into mature IL-33 by neutrophil elastase [9]. ST2 is a receptor of mature IL-33 [10]. Either full-length IL-33 or mature IL-33 is involved in organ fibrosis via ST2-independent and ST2-dependent pathways [11]. Although the roles of IL-33 and ST2 in fibrosis have been documented in the lung [12], liver [13], kidney [14], heart [15] and other organs [16], the mechanisms of IL-33/ST2 in epidural fibrosis remain to be elucidated.

Fibroblast-to-myofibroblast transition, which is driven by TGF-β, is a key step in fibrosis [17]. Once activated, fibroblasts transdifferentiate into myofibroblasts, producing increased levels of fibronectin and alpha-smooth muscle actin (α-SMA) [18]. In the present study, we demonstrated that in vitro IL-33 promoted the activation of fibroblasts, and this effect was ST2-dependent; p38 and TGF-β were involved with IL-33/ST2 signal pathways in the fibroblasts maturation. In an established model of epidural fibrosis, IL-33 deficiency or ST2 deficiency decreased extracellular matrix proteins (fibronectin, α-SMA, and collagen) and TGF-β in scar tissues post spine surgery, suggesting that IL-33/ST2 contributed to epidural fibrosis. Our study provides evidence that IL-33 and ST2 may be novel targets in the prevention of epidural scar hyperplasia after laminectomy.

Materials and methods

Animals and model of laminectomy

Wild male C57BL/6J mice aged 6–8 weeks were obtained from Yangzhou University (Yangzhou, China) Experimental Animal Center and maintained at Nanjing Medical University under specific pathogen-free conditions. IL-33-knockout (KO) and ST2- KO mice with a C57BL/J background were obtained from Dr. Hong Zhou (Department of Immunology, Nanjing Medical University). All animal experimental protocols were examined and approved by the Animal Protection and Use Committee of Nanjing Medical University (1910018).

To construct a model of laminectomy, mice were anesthetized by an intraperitoneal injection of 10 mg/kg and 200 mg/kg silazane and ketamine hydrochloride, respectively. The limbs were fixed on the operating table in the prone position, and an incision of approximately 2 cm was made along the spine on the back of the mouse. The two sides of the incision were pulled apart by a retractor, and the peripheral muscles were separated along the spinal spinous process to fully expose the T12-L2 spinous process and the surrounding lamina tissue. After clamping the T12-L2 lamina with forceps, the spinous process was removed, and the dura mater was exposed. The fascia, muscles and skin of the spine were then closed.

Cell culture

The murine fibroblast cell line NIH-3T3 was obtained from the American Type Culture Collection (ATCC, CRL-1658, Manassas, VA, USA) and cultured with 10% fetal bovine serum (Lonsera) and 1% penicillin/streptomycin (Hyclone, GE, Logan, UT, USA) at 37 °C in 5% CO₂. NIH-3T3 cells (1*10⁵) were cultured in a 24-well plate. On the second day, the NIH-3T3 cells were treated with IL-33 or an equal volume of phosphate-buffered saline (PBS) (HyClone). ST2 neutralizing antibody (MAB10041, R&D Systems, Minneapolis, MN, USA), the p38 MAPK inhibitor RWJ 64809 (SB-203580, Selleck Chemicals, Houston, TX, USA) or the Smad3 inhibitor SIS3 (521984-48-5, MedChemExpress, Monmouth Junction, NJ, USA) was added to the medium when necessary. The working concentrations for the ST2 neutralizing antibodies, p38 MAPK inhibitor and Smad3 inhibitor were 5 μg/ml [19], 5 μM [20], and 10 μΜ [21], respectively. The cells were harvested after 24 h for further analysis.

Primary fibroblast cultures from mouse tissues were established as previously described [22]. Briefly, adult wild-type C57BL/6J mice were sacrificed by cervical dislocation. After sacrifice, the mice were soaked in 75% alcohol for 5 min and washed with PBS 3 times. The tail tip was minced and digested with collagenase (MilliporeSigma, Burlington, MA, USA). The tissue fragments were cultured in DMEM/F12 supplemented with 15% fetal bovine serum (Lonsera), 1% penicillin/streptomycin (HyClone, GE) and 1× antimycotic. The culture medium was changed every 2–3 days. On the 7th day of culture, a large number of cells swarmed out from the edge of the tissue mass, and pure murine primary fibroblasts were obtained by subculturing the cells for 2–3 generations.

Western blotting

After sacrifice, the spine of the mouse was exposed, and the surrounding scar tissue was carefully peeled off along the original laminectomy. The total protein was isolated from cells or scar tissue by lysis on ice with RIPA buffer (89900, Thermo Fisher Scientific, Waltham, MA, USA) containing protease and phosphatase inhibitors (78443, Thermo Fisher Scientific) for 20 min. Then, the sample was centrifuged for 10 min, and the supernatant was collected and transferred to a new EP tube. The sample was then treated with ultrasonication on ice for 10 s. Then, the sample was centrifuged for 10 min, and the supernatant was collected, transferred to a new EP tube and stored at − 80 °C for later use. The protein concentration was determined by the BCA method (P0012S, Beijing, China). Proteins were separated by polyacrylamide gel electrophoresis and transferred to poly (vinylidene fluoride) (PVDF) membranes at a current of 300 mA. The PVDF membrane was sealed with 5% skimmed milk powder for 1 h at room temperature and then incubated overnight at 4 °C with primary antibodies (Table 1). The PVDF membrane was washed with Tris-buffered saline and 0.1% Tween (TBST) 4 time for 5 min each and then incubated with goat anti-rabbit horseradish peroxidase (HRP)-conjugated antibodies (EarthOx Life Science) or goat anti-mouse HRP-conjugate antibodies (EarthOx Life Science, Millbrae, CA, USA) for 1 h at room temperature. The PVDF membrane was washed with TBST 4 times for 7 min each. The signals were captured using Immobilon Western chemiluminescence HRP substrate (Ripoli, MA, USA) and a G:BOX system (Tanon 5200, Shanghai, China). All experiments were carried out at least three times.

Table 1.

Primary antibodies for western blot analysis

Antibody	Brand	Code	Source	Dilution
Anti-IL-33	Abcam	ab187060	Cambridge, England	1:1000
Anti-α-SMA	Abcam	ab7817	Cambridge, England	1:200
Anti-TGF-β1	Abcam	ab170874	Cambridge, England	1:1000
Anti-fibronectin	Proteintech	15613-1-AP	Wuhan, China	1:1000
Anti-β-actin	CST	4970	Shanghai, China	1:1000
Anti-GAPDH	CST	5174	Shanghai, China	1:1000
Anti-P38	CST	8690	Shanghai, China	1:1000
Anti-p-P38 (D3F9)	CST	4511	Shanghai, China	1:1000

Hematoxylin and eosin (HE) staining

At 30 days after the laminectomy, the tissue in the surgical area was collected and fixed in 4% buffered formaldehyde solution for 48 h. Then, the tissue was decalcified in 10% neutral EDTA for approximately 4 weeks, and finally, the tissue in the surgical area was embedded in paraffin. Paraffin sections were cut into 5-μm-thick pieces with a Leica 2165 rotating microtome (Leica, Wetzlar, Germany) and placed on glass slides. The dewaxed tissue was dyed with hematoxylin for 5–10 min and then washed with distilled water for 1–2 min. Then, the tissue was immersed in 1% hydrochloric acid–ethanol for 5 min to differentiate it rapidly, washed back to blue with distilled water, and stained with eosin for 5 min. After being dehydrated, the slices were sealed with neutral resin. Images were collected under an optical microscope (BX-53, Olympus Optical Company, Tokyo, Japan) and saved for subsequent analysis.

Masson staining

Masson staining as used to evaluate collagen deposition in the model of epidural fibrosis [4]. Briefly, the deparaffinized tissue slides were soaked in potassium dichromate overnight. Next, the slides were rinsed with distilled water and stained with ferric hematoxylin solution A and solution B (1:1) for 3 min. Furthermore, the slides were soaked in liprune acid magenta for 5 min. After being impregnated with the molybdophosphate solution for 2 min and aniline blue medium for 5 min, the slides were differentiated with 1% glacial acetic acid and dehydrated with anhydrous ethanol. Images were recorded by an optical microscope (BX-53, Olympus Optical Company) for subsequent analysis.

Immunohistochemical (IHC) analysis

The tissue was fixed in paraffin blocks, cut into 5-μm-thick sections and deparaffinized with xylene. Then, the slides were fully rehydrated with gradient alcohol and water, placed in an autoclave for antigen recovery with citrate buffer (pH 6.0) for 15 min, and then cooled on ice for 20 min. Once at room temperature, the activity of endogenous peroxidase was inhibited with a 3% hydrogen peroxide solution for 15 min, after which the slides were washed with PBS on a shaker 3 times for 5 min each and then blocked with 3% bovine serum albumin BSA (G5001, Servicebio, Hubei, China) for 30 min. Then, the glass slides were incubated with rabbit anti-mouse fibronectin antibodies (1:1000, GB13091, Servicebio) at room temperature for 1 h. Next, the slides ere further stained with goat anti-rabbit antibodies coupled with horseradish peroxidase (1:200, GB23303, Servicebio) for 50 min at room temperature. Then, newly prepared DAB chromogenic reagent was added to stain the tissue. Finally, the sections were counterstained with hematoxylin staining solution and captured with a microscope (BX-53, Olympus Optical Company) at a magnification of ×400.

Panoramic histocyte immunofluorescence analysis

The slides were prepared as described for IHC analysis. After being blocked with 3% BSA (G5001, Servicebio), the slides were incubated with anti-IL-33 (1:5000, ab187060, Abcam, Cambridge, UK), anti-CD31 (1:5000, gb13428, Google Biology), and anti-α-SMA (1:1000, gb13044, Servicebio) overnight at 4 °C. Next, the slides were incubated with the corresponding fluorescence-conjugated secondary antibodies for 2 h. Finally, the slides were stained with DAPI (36308ES20, Yeasen, China) and observed with an Olympus IX73 fluorescence microscope.

Enzyme-linked immunosorbent assay (ELISA)

To evaluate the type I collagen level, 20 mg of tissue from the surgical area was homogenized in 200 µL of PBS and centrifuged at 12,000 × g for 10 min at 4 °C. Type I collagen levels in the supernatant were determined using an ELISA kit (E-EL-M0325c, Elabscience, Houston, TX, USA) [23]. To measure IL-33, the homogenized samples were analyzed using a commercial ELISA kit (DY3626, R&D Systems). The absorbance at 540 nm was measured. IL-33 levels were calculated according to the standard curve.

Statistical analysis

GraphPad Prism 8.0 software was used for statistical analyses. The values are expressed as the mean ± SD. Differences between groups were evaluated by one-way ANOVA and the Student–Newman–Keuls test. Statistical significance was defined as follows: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; no significance (NS).

Results

Elevated IL33 in a mouse model of epidural fibrosis

According to our previous reports, we established a mouse model of laminectomy. As shown in Fig. 1A, H and E staining indicated scar development around the dura matter 28 days post spine surgery. Masson staining further showed the excessive deposition of collagen in the wound tissues of mice that underwent spine surgery (Fig. 1B). Accordingly, in contrast with that in control mice that underwent a sham operation, fibronectin and α-SMA levels were elevated in the surgical tissues from mice post laminectomy (Fig. 1C). Immunohistochemical analysis showed that in the scar tissues, fibronectin was mainly located around the dura matter (Fig. 1D). Collectively, epidural fibrosis developed in the mouse model of laminectomy.

Fig. 1.

IL-33 was increased in scar tissues 28 days post laminectomy. A Epidural scarring as shown by HE staining; bar = 500 μm. B Masson staining revealed collagen deposition in the epidural scar; bar = 500 μm. C Western blot analysis of fibronectin and α-SMA expression in scar tissues. D Immunohistochemical analysis of fibronectin levels in scar tissues; bar = 100 μm. E Western blot analysis of IL33 expression in scar tissues. F IL33 levels in scar tissues were measured using ELISA. G Immunofluorescence imaging of IL33 in α-SMA⁺ fibroblasts in scar tissues. H Immunofluorescence imaging of IL33 in CD31⁺ endothelial cells in scar tissues; bar = 200. The black arrow indicates the epidural scar, the red arrow indicates the dura matter, and ‘SC’ indicates the spinal cord. ***p < 0.001

To explore whether IL-33 was involved in postepidural scarring, we quantified IL-33 in wound tissues. As shown in Fig. 1E, IL-33, especially full-length IL-33, was increased in the wound tissues from mice post laminectomy. Compared with control mice that underwent sham operation, mice that underwent spine surgery produced more IL-33, which was further validated by ELISA (Fig. 1F). Immunofluorescence imaging revealed that IL-33 was colocalized with α-SMA (Fig. 1G) or CD31 (Fig. 1H), suggesting that fibroblasts (α-SMA⁺) and vascular endothelial cells (CD31⁺) may produce IL-33 in scar tissues. In summary, laminectomy caused epidural fibrosis, which was accompanied by elevated IL-33 production in the surgical tissues.

IL-33 induced the fibroblast-to-myofibroblast transition in vitro

Fibrosis initiates the differentiation of fibroblasts into myofibroblasts, which produce more fibronectin and α-SMA [24]. As α-SMA-positive cells were IL-33 producers (Fig. 1G), we wondered whether IL-33 played a role in the transition of fibroblasts into myofibroblasts. NIH-3T3 cells that were treated with IL-33 produced more fibronectin and α-SMA than untreated cells (Fig. 2A), suggesting that IL-33 may directly promote the fibroblast-to-myofibroblast transition. Within 48 h (Fig. 2B), the production of fibronectin and α-SMA in NIH-3T3 cells gradually increased. Similarly, IL-33 increased the expression of fibronectin and α-SMA in primary fibroblasts (Fig. 2C). ST2 is an IL-33 receptor expressed by fibroblasts [25]. As expected, the ST2 neutralizing antibody rescued the effects of IL-33 in the increased expression of fibronectin and α-SMA in fibroblasts (Fig. 2D), suggesting that IL-33 may promote the ST2-dependent fibroblast-to-myofibroblast transition.

Fig. 2.

IL-33/ST2 promoted the differentiation of fibroblasts into myofibroblasts. A, B Western blot analysis of fibronectin and α-SMA expression in NIH-3T3 cells treated with IL-33. C Western blot analysis of fibronectin and α-SMA expression in primary fibroblasts treated with IL-33. D Western blot analysis of fibronectin and α-SMA expression in NIH-3T3 cells treated with IL33 and ST2 neutralizing antibodies. *p < 0.05; **p < 0.01

p38 mediated IL-33-induced fibroblast maturation

Mitogen-activated protein kinase (MAPK) p38 plays vital roles in the differentiation of fibroblasts into myofibroblasts; p38 deficiency blocks fibroblast differentiation into myofibroblasts [26]. As shown in Fig. 3A, IL-33 increased the level of phosphorylated p38, which is an indicator of p38 activation. In contrast, the ST2 neutralizing antibody inhibited p38 activation in IL-33-treated fibroblasts (Fig. 3B). Moreover, the p38 inhibitor decreased fibronectin production in IL-33-treated fibroblasts (Fig. 3C), suggesting that the p38 signaling pathway mediated the induction of the fibroblast-to-myofibroblast transition via IL-33/ST2.

Fig. 3.

P38 was required for IL-33-mediated promotion of the fibroblast-to-myofibroblast transition. A Western blot analysis of p-38 activation in NIH-3T3 cells treated with 1 ng/ml IL33. B Western blot analysis of p-38 activation in NIH-3T3 cells treated with IL33 and ST2 antibodies for 24 h. C Western blot analysis of fibronectin expression in NIH-3T3 cells treated with IL33 and a p38 MAPK inhibitor (SB-203580) for 24 h. *p < 0.05; **p < 0.01; ****p < 0.0001

TGF-β mediated IL-33-induced fibroblast maturation.

TGF-β is a master cytokine associated with fibroblast activation [27]. Consistent with this finding, IL-33 directly increased TGF-β levels in NIH-3T3 fibroblasts (Fig. 4A, B) and primary fibroblasts (Fig. 4C). Similarly, the ST2 neutralizing antibody decreased TGF-β production in IL-33-treated fibroblasts (Fig. 4D), and the TGF-β inhibitor reduced fibronectin and α-SMA levels in IL-33-treated fibroblasts (Fig. 4E), suggesting that TGF-β was involved in the IL-33/ST2-induced fibroblast-to-myofibroblast transition. Collectively, IL-33 promoted ST2-dependent fibroblast maturation into myofibroblasts via the p38 and TGF-β signaling pathways.

Fig. 4.

TGF-β was required for IL-33-mediated promotion of the fibroblast-to-myofibroblast transition. A Western blot analysis of TGF-β1 expression in NIH-3T3 cells treated with IL-33 for 24 h. B Western blot analysis of TGF-β1 expression in NIH-3T3 cells treated with 1 ng/ml IL-33. C Western blot analysis of TGF-β1 expression in primary fibroblasts treated with 1 ng/ml IL-33 for 24 h. D Western blot analysis of TGF-β1 expression in NIH-3T3 cells treated with IL33 and ST2 antibodies for 24 h. E Western blot analysis of fibronectin and α-SMA expression in NIH-3T3 cells treated with IL33 and a TGF-β inhibitor. *p < 0.05; **p < 0.01; ***p < 0.001

IL33 deficiency abrogated epidural fibrosis

Thus far, we have demonstrated that IL-33 promoted the ST2-dependent fibroblast-to-myofibroblast transition via the p38 and TGF-β signaling pathways. Considering that IL-33 was increased in the wound tissues of mice that underwent spine surgery, we asked whether IL-33 contributed to epidural fibrosis in a mouse model of laminectomy. In contrast with wild-type mice, IL-33-KO mice developed less epidural scarring, as evidenced by H&E and Masson staining (Fig. 5A, B). In line with the decreased deposition of collagen fibers, IL-33 deficiency also reduced fibronectin and α-SMA levels in scar tissues (Fig. 5C, D). In addition, ELISA provided quantitative results showing that IL-33 deficiency decreased type 1 collage in the scar tissues of mice that underwent spine surgery (Fig. 5E). As TGF-β was involved in IL-33-induced fibroblast activation, IL-33 deficiency caused a reduction in TGF-β in the wound tissues post spine surgery (Fig. 5F, G). In summary, IL-33 deficiency caused a reduction in TGF-β and alleviated epidural fibrosis.

Fig. 5.

IL-33 deficiency mitigated epidural fibrosis 28 days post spine operation. A Representative images of HE staining of wound tissues bar = 500 μm. B Representative images of Masson staining of wound tissues bar = 500 μm. C Western blot analysis of fibronectin and α-SMA expression in wound tissues. D Representative immunohistochemical images showing fibronectin expression in wound tissues; bar = 100 μm. E Type I collagen in wound tissues was measured by ELISA. F, G Western blot analysis of TGF-β1 expression in wound tissues. The black arrow indicates the epidural scar, and ‘SC’ indicates the spinal cord. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001

ST2 deficiency alleviated epidural fibrosis

Although the IL-33/ST2 axis has been widely discussed in organ fibrosis [11], full-length IL-33 is active in an ST2-independent manner in vivo [28]. Because the major form of elevated IL-33 in scar tissues was full-length IL-33 (Fig. 1E), we examined the roles of ST2 in epidural fibrosis. Surprisingly, mice that were defective in ST2 were less prone to epidural scarring than WT mice, as evidenced by H&E or Masson staining (Fig. 6A, B). ST2 deficiency also decreased the expression of fibronectin and α-SMA in epidural scar tissues (Fig. 6C, D). Similarly, type I collagen in scar tissues was significantly reduced in the ST2-defective mice (Fig. 6E). In addition, ST2 deficiency decreased TGF-β expression in wound tissues post spine surgery (Fig. 6F). In summary, ST2 deficiency reduced TGF-β production and mitigated epidural fibrosis in a mouse model of laminectomy.

Fig. 6.

ST2 deficiency alleviated epidural fibrosis 28 days post spine operation. A Representative images of HE staining of wound tissues bar = 500 μm. B Representative images of Masson staining of wound tissues bar = 500 μm. C Western blot analysis of fibronectin and α-SMA expression in wound tissues. D Representative immunohistochemical images showing fibronectin expression in wound tissues; bar = 100 μm. E Type I collagen in wound tissues was measured by ELISA. F Western blot analysis of TGF-β1 expression in wound tissues. The black arrow indicates the epidural scar, and ‘SC’ indicates the spinal cord. *p < 0.05; **p < 0.01; ****p < 0.0001

Discussion

In recent years, the roles of IL-33 in the pathogenesis of various organ fibrosis diseases have aroused widespread concern [11]. The increase in IL-33 in disease tissues is a predictable index for the fibrosis and disease severity in lung [29, 30], liver [31], kidney [32] and other organs [33]. In the present study, IL-33 was increased in wound tissue post laminectomy, highlighting the importance of IL-33 in the pathogenesis of epidural fibrosis. Once the cells are damaged or activated, full-length IL-33 is released from the nucleus and cleaved into its mature form, which binds to ST2 and initiates the type 2 immune response [34]. Both full-length IL-33 and mature IL-33 contribute to organ fibrosis [11]. Unlike mature IL-33, full-length IL-33 mainly enhances the inflammatory response independent of ST2 [35]. Elastase from neutrophils cleaved full-length IL-33 into mature IL-33, which induced fibrosis in a ST2-depedent manner. As neutrophils infiltrated into the wound tissues following spine operation released elastase [36], we speculated that full-length IL-33 elevated in the wound tissues maybe be cleaved into mature IL-33 and initiated fibrosis process.

Cells expressing α-SMA or CD31 produced IL-33 in the wound tissue post laminectomy. Spine surgery triggered the secretion of alarmin molecules (i.e., HMGB1) [36] and HMGB1 from wound tissues may boost the production of IL-33 [37], which was in line with the previous report that inflammation induced the production of IL-33 by murine endothelial cells [38]. In addition to CD31⁺ cells, α-SMA⁺ cells also produced IL-33. CD31 was mainly expressed by endothelial cells; α-SMA was constitutively produced by fibroblasts. IL-33 in the fibroblasts facilitated type 2 inflammation [39]. Therefore, IL-33, from either endothelial cells or fibroblasts, may promote epidural fibrosis following spine operation.

In the present study, IL-33 promoted the transdifferentiation of fibroblasts into myofibroblasts in an ST2-dependent manner in vitro. IL-33 deficiency or ST2 deficiency mitigated epidural fibrosis, suggesting that IL-33 and ST2 were involved in fibroblast activation. TGF-β promotes the differentiation of fibroblasts into myofibroblasts [40], which is an important process associated with the structural modification of epidural scars. Therefore, TGF-β is considered one of the most important factors in the pathogenesis of epidural fibrosis. In our study, a large amount of TGF-β was detected in postoperative epidural scars. After treatment with IL-33, the fibroblast cell line NIH-3T3 or primary fibroblasts produced more TGF-β, which was at least partially dependent on MAPK p38. ST2 neutralizing antibodies blocked p38 activation, suggesting that ST2 was involved with IL-33 and p38 signaling pathway in the fibroblasts maturation. As expected, in both the IL-33-deficient mice and ST2-deficient mice, TGF-β was drastically decreased in wound tissues after spine surgery, confirming that ST2 was required for IL-33-induced epidural fibrosis. The absence of IL-33 and ST2 reduced hepatic fibrosis, accompanied with decreased TGF-β [41]. However, IL-33 but not ST2 was involved with skin wound fibrosis [42], suggesting that roles of IL-33/ST2 in the organ fibrosis may vary in different cells, i.e., fibroblasts, hepatic stellate cells [41] and keratinocytes [42].

Overall, we demonstrated that the IL33/ST2 signaling pathway may be related to the development of epidural fibrosis. IL-33 in scar tissue activated MAPK p38 signaling pathway through ST2, thereby increasing the expression of TGFβ1 and inducing fibrosis. Blocking the IL-33/ST2 signaling pathway, such as with an anti-IL-33 or anti-ST2 neutralizing antibody may be a potential alternative strategy for the alleviation of epidural fibrosis.

Declarations

Conflict of interest

The authors report no declarations of interest.

Ethical Statement

All animal experimental protocols were examined and approved by the Animal Protection and Use Committee of Nanjing Medical University (1910018).