Development of a novel external fixation device for inducing knee arthrofibrosis in rats

Abstract

Arthrofibrosis (AF) is a common pathological condition characterized by joint dysfunction. However, traditional non-invasive external fixation methods are difficult to maintain on the small limbs of rats. This study aimed to develop a novel, non-invasive, and stable rat model of knee arthrofibrosis using a thermoplastic polymer resin. Sixty male Sprague-Dawley rats were randomly assigned into a Sham group and immobilization groups (1, 2, 4, and 6 weeks). Utilizing the material’s property of being malleable at high temperatures and rigid at room temperature, a custom-fitted “thigh-crus-trunk” external fixation device was fabricated to immobilize the knee at 135° of flexion. Total, arthrogenic, and myogenic contractures were assessed by measuring the range of motion (ROM). Histopathological changes were evaluated using H&E and Masson’s trichrome staining. The expression of fibrotic markers (α-smooth muscle actin, α-SMA and collagen type I alpha 1 chain, COL1A1) in synovial tissues was detected via immunohistochemistry, RT-qPCR, and Western blotting. Biosafety was assessed through histological and serum biochemical analyses of major organs. Prolonged immobilization resulted in a significant decrease in knee ROM, while joint capsule thickness, synovial hyperplasia, and collagen deposition increased, stabilizing after 4 weeks. Analysis revealed that myogenic contracture predominated in the first 2 weeks, whereas arthrogenic contracture became dominant in the later stage. Molecular analysis confirmed a time-dependent upregulation of α-SMA and COL1A1 in synovial tissues. Furthermore, no abnormalities were observed in major organs or serum biochemical indices, indicating favorable biosafety. A novel non-invasive rat model of knee arthrofibrosis was successfully established using thermoplastic polymer resin. This device is cost-effective, user-friendly, stable, and biocompatible. It effectively simulates immobilization-induced joint contracture without surgical trauma, serving as a valuable model for future arthrofibrosis research.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-46668-7.

Introduction

Arthrofibrosis (AF) is a pathological condition characterized by excessive fibrous proliferation of intra-articular or peri-articular tissues arising from various etiologies, resulting in the impairment of normal joint function^1–3. While the exact pathogenesis of AF remains elusive, it can be triggered by diverse factors, including trauma, chronic inflammation, immune stress, and metabolic diseases. This process involves complex biological mechanisms, such as cellular proliferation, the activation of fibrotic signaling pathways, and extracellular matrix (ECM) remodeling^4,5. AF is characterized by accelerated fibroblast proliferation, excessive ECM deposition, collagen overproduction, and adhesion formation⁶. Prolonged joint immobilization is a major etiology of AF. This immobilization can be involuntary, occurring in patients with muscle weakness due to neurological disorders or those under sedation in the intensive care unit, both of which can lead to fibrosis of the knee or ankle joints⁷. In addition, passive immobilization occurs in scenarios such as long-term casting following conservative or surgical treatment of fractures, or voluntary joint inactivity due to fear of pain, inducing fibrosis in the affected joints^8,9. The clinical manifestations of AF typically include joint pain, restricted ROM, swelling, and stiffness, which significantly compromise patients’ quality of life^10,11. Consequently, investigating the pathogenesis of AF and developing effective therapeutic strategies are of paramount importance.

Joint contractures secondary to arthrofibrosis are highly prevalent across multiple clinical disciplines, including orthopedics, rehabilitation medicine, neurology, and intensive care¹². Orthopedic surgeons have historically prioritized post-operative arthrofibrosis; consequently, invasive models involving surgical intervention predominate in the literature. However, these models often fail to recapitulate the progressive, non-traumatic onset of fibrosis observed clinically, such as that resulting from prolonged immobilization during the conservative management of ligament injuries or fractures. Non-invasive models induce fibrosis via extra-articular immobilization, thereby simulating clinical scenarios of external fixation in the absence of surgical intervention—a condition frequently encountered in rehabilitation and neurology. However, such non-invasive models remain significantly underrepresented in the literature. Currently, the majority of established external fixation models utilize large animal species, such as rabbits, dogs, or sheep, as their substantial joint dimensions facilitate reliable immobilization using splints, casts, or external fixators^13–17. Rat models offer distinct advantages in the study of arthrofibrosis, including cost-effectiveness and high genetic homology with humans, facilitating genetic manipulation studies such as knockout or knockin experiments. However, conventional external fixation methods, such as casting, are technically challenging to apply due to the small dimensions and irregular, conical morphology of the rat hindlimb.

To address these limitations, we developed a novel, non-invasive rat model of arthrofibrosis utilizing a thermoplastic polymer resin. This cost-effective and commercially available material is solid at room temperature but becomes semi-solid and malleable when heated above 80 ℃, allowing it to be custom-molded to the specific anatomy of the rat handlimb. Upon cooling to room temperature, the material solidifies, thereby completing the immobilization process. The thermoplastic polymer resin exhibits high rigidity, resistance to gnawing, and biocompatibility, ensuring it poses no hazard to the animals. In contrast to previous methods, our technique overcomes the challenges associated with the small dimensions of rat joints and the poor conformability of plaster casts by providing a close-fitting, adjustable fixation. This approach effectively minimizes slippage and ensures stable, long-term immobilization. In the present study, a series of physiological and pathological assessments were conducted to validate the reliability of our model.

Methods

Animals

Following an a priori power analysis (G*Power 3.1) based on pilot study data, sixty male Sprague-Dawley (SD) rats (8 weeks of age, 180–220 g) were obtained from the Laboratory Animal Center of North Sichuan Medical College. The animals were randomly assigned to five groups (n = 12) using a random number table: a Sham group and four immobilization groups (1, 2, 4, and 6 weeks, respectively). All rats were housed in standard cages under controlled environmental conditions (20–25 ℃, 12/12-h light/dark cycle) with ad libitum access to standard laboratory chow and water. At the designated time points (0, 1, 2, 4, and 6 weeks), animals in the corresponding groups were euthanized via an intraperitoneal injection of sodium pentobarbital. All experimental procedures were approved by the Ethics Committee of North Sichuan Medical College (Approval No. 2025090).

Joint immobilization

Rats (including those in the Sham group) were anesthetized with sodium pentobarbital (20 mg/kg i.p.). Upon achieving satisfactory anesthesia, the thermoplastic polymer resin (No. TN-DIY-2603-01; Guangdong Tiannuo, China) was immersed in hot water (> 80 °C) for several minutes until it reached a semi-solid, malleable state (Fig. 1A-C). Initially, the thigh and crus were immobilized in a flexed position at an angle of 135° (Fig. 1D, green circle). At this stage, the fixation was applied relatively tightly, given that this was an external fixation, and the soft tissues between the thigh and crus provided a protective cushion, thereby mitigating the risk of direct compression injury. Subsequently, to prevent anterior displacement of the device, a connecting strut was fashioned posterior to the knee joint to secure the crus (Fig. 1D, red circle). Importantly, this distal fixation was applied without excessive constriction to avoid impairing venous return, which could otherwise lead to distal limb edema or tissue necrosis. Finally, a circumferential band was molded around the rat trunk to connect the proximal components, preventing the entire apparatus from being dislodged (Fig. 1D, yellow circle). Notably, the melted thermoplastic polymer resin maintained a malleable state for approximately 10 min; the immobilization process was completed within this duration. The gross appearance of the rat following immobilization is shown in Fig. 1F. For the first 3 days post-immobilization, the lower limbs were inspected twice daily (morning and evening) to detect any signs of ischemia or necrosis. Once the rats had adapted to the immobilization (after 3 days), monitoring was conducted every other day. All fixation procedures were performed by the same investigator to ensure consistency.

Fig. 1.

Establishment of the non-invasive rat knee arthrofibrosis model using thermoplastic polymer resin. (A–C) Preparation of the thermoplastic polymer resin. The material transitions from a rigid solid state to a malleable semi-solid form upon immersion in hot water (> 80℃); (D, E) Gross appearance of the customized external fixation device and the sequential immobilization protocol (green circle to red circle to yellow circle); (F) Macroscopic view of the rat following immobilization. The device maintains the knee in flexion while sparing the ankle joint.

Range of motion measurement

Knee joint ROM was measured according to previously established protocols with minor modifications. Following the induction of anesthesia, a protective pad was placed between the skin and the fixation device to prevent thermal injury. The device was then removed by partially melting the thermoplastic polymer resin with heated forceps, and the entire removal process was completed within 1 min. The rat was positioned in a neutral trunk alignment, and skin incisions were made over the thigh and crus to expose the underlying structures. Subsequently, the ankle was connected to a customized pulley system via a fine suture. To standardize the measurement of passive knee extension, a force gauge (NK/ZP; AIGU, China) was utilized to apply a constant torque of 14.6 N·mm to the knee joint. This torque was sufficient to achieve maximum physiological extension in rats without inducing tissue damage. A goniometer was subsequently employed to record the maximum extension angle, with its pivot centered on the lateral femoral epicondyle, the stationary arm aligned with the long axis of the femur, and the moving arm aligned with the long axis of the tibia¹⁸. The procedures described above were performed independently by two investigators. Each investigator conducted three measurements, and the mean values were calculated for analysis. The investigators were blinded to the group allocation. To isolate arthrogenic contracture and eliminate the confounding influence of myogenic factors, all muscles crossing the knee joint were surgically transected (myotomy). The maximum passive extension angle measured under these conditions was recorded as arthrogenic passive extension ROM. The measurement procedure was repeated to determine arthrogenic contracture using the parameters described above. (1) Degree of total contracture = ROM before myotomy (knee joint in the Sham group) - ROM before myotomy (knee joint in the immobilization group); (2) Degree of arthrogenic contracture = ROM after myotomy (knee joint in the Sham group) - ROM after myotomy (knee joint in the immobilization group); (3) Degree of myogenic contracture = Degree of total contracture - degree of arthrogenic contracture.

Histological preparation and assessment

Following the completion of ROM measurements, blood samples, major internal organs, knee joint specimens, and synovial tissues were harvested sequentially from the rats. Blood samples were collected via retro-orbital puncture under general anesthesia and then centrifuged at 3,000 × g for 10 min. The serum was separated and analyzed using the ADVIA 2400 Chemistry System (Siemens Healthcare Diagnostics, Tarrytown, NY, USA). The major organs of all rats and the knee joints of six rats were harvested for histopathological analysis. For hematoxylin and eosin (H&E), Masson’s trichrome staining, and immunohistochemistry (IHC), tissue specimens were processed according to standardized protocols, including fixation, decalcification (joints only), paraffin embedding, sectioning, and staining. Knee joint tissues were sectioned in the sagittal plane, oriented parallel to the long axes of the femur and tibia. Care was taken to obtain sections as close to the mid-sagittal plane as possible. To ensure consistency across all samples and groups, the sectioning level was standardized by identifying the posterior horn of the meniscus and the midpoint of the femoral condyle as reference landmarks. Histological evaluations were then performed under light microscopy. For the remaining six rats, the synovial tissue and the posterior capsule were isolated from the popliteal space between the gastrocnemius heads. Specimens were snap-frozen in liquid nitrogen and stored at −80 ℃ for RT-qPCR and Western blotting analyses.

The thickness of the posterior joint capsule was measured according to previously established protocols¹⁹. Briefly, H&E-stained sections were examined at 4× magnification to visualize the posterior aspect of the knee joint, and the distance between the posterior margin of the meniscus and the joint capsule was measured using ImageJ software (1.8.0). This measurement was defined as the posterior capsule thickness. For major organs, after staining of the sections, histological images were obtained using an optical microscope, and the sections were thoroughly examined at 10× magnification to identify any potential pathological alterations. For fibrosis quantification, three non-overlapping representative fields were selected from the posterior joint capsule of each Masson-stained section. Quantification was conducted on images captured at 10× magnification to ensure clear visualization of collagen fibers and accurate area ratio calculation. The extent of fibrosis was quantified as the Collagen Volume Fraction, calculated as the percentage of blue-stained collagen area relative to the total tissue area using the Color Threshold tool in ImageJ software. Immunohistochemical (IHC) staining was semi-quantitatively evaluated using the H-score method, processed via ImageJ software. For each tissue section, at least five random high-power fields (10× magnification) were analyzed. The H-score was determined using the following formula: H-score = ∑ (pi × i) = (percentage of weak intensity × 1) + (percentage of moderate intensity × 2) + (percentage of strong intensity × 3). To ensure objectivity and reproducibility, all measurements were conducted by two independent observers blinded to the group allocation, and the mean values were utilized for statistical analysis.

RT-qPCR

Total RNA was extracted from synovial tissues using an E.Z.N.A.^® Total RNA Kit I (Omega Bio-TEK, R6834-02, Georgia, USA) according to the manufacturer’s instructions. 1 µg of total RNA was then reverse-transcribed using the HiScript^® II 1st Strand cDNA Synthesis Kit (Vazyme, R211-02, Nanjing, China) in accordance with the provided protocol. Quantitative real-time PCR (RT-qPCR) was performed utilizing the ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711-02-AA, Nanjing, China) alongside a Roche 480 II system (Roche, Basel, Switzerland). The mRNA expression levels were normalized to Gapdh via the 2^−ΔΔCT method. All primers were synthesized by Sangon Biotech (Shanghai, China), and their sequences were as follows: α-SMA forward 5´-TACTACTGCTGAGCGTGAGATTG-3´ and reverse 5´-TGATGCTGTTGTAGGTGGTTTCA-3´, COL1A1 forward 5´-AATGAAGGGACACAGAGGTTTCA-3´ and reverse 5´-AGCACCAGTAGCACCATCATTTC-3´. GAPDH forward 5´-ACAACTTTGGTATCGTGGAAGG-3´ and reverse 5´-GCCATCACGCCACAGTTTC-3´.

Western blotting

Tissues were homogenized in radio-immunoprecipitation assay (RIPA) buffer (Beyotime, P0013, Shanghai, China) supplemented with 1 mM PMSF (phenylmethanesulfonyl fluoride; Calbiochem, 539134, Germany). Protein concentrations were determined using a bicinchoninic acid (BCA) protein quantification kit (Thermo Fisher Scientific, Waltham, MA, USA). Subsequently, the protein samples were electrophoresed and then transferred to nitrocellulose membranes (Pall, 66485, Port Washington, NY) at 4 ℃. The membranes were blocked with 5% non-fat milk (BIOFROXX, 1172GR500, Germany) for 1 h and then incubated overnight at 4 ℃ with the specified primary antibody, and subsequently incubated with an HRP-conjugated secondary antibody (Goat anti-rabbit, 1:5000; Boster Biological Technology Co., Ltd., BA1054, California, USA) for one hour at room temperature. Target proteins COL1A1 (139 kDa) and α-SMA (42 kDa) were separated using 8% and 10% SDS-PAGE gels, respectively, while the loading control GAPDH (36 kDa) was separated using 12% SDS-PAGE gels to ensure optimal resolution. All gels were processed in parallel using the same biological samples. Protein bands were visualized through chemiluminescence using an ECL kit (Biosharp, BL520B, China) and exposed to X-ray film (Fusion FX VILBER LOURMAT, Vilber Lourmat, Colmar, France). α-SMA (Rabbit pAb, 1:1000; Proteintech, 80008-1-RR, China), COL1A1 (Rabbit pAb, 1:1000; Proteintech, 86093-1-RR, China), and GAPDH (Rabbit pAb, 1:5000; HUABIO, HO1224, China) were used for Western blotting.

Statistical analysis

Statistical analysis was conducted using GraphPad Prism 9.5. Differences among multiple groups were analyzed using one-way analysis of variance (ANOVA). Data normality was assessed using the Shapiro-Wilk test, and homogeneity of variances was evaluated using Brown-Forsythe or Bartlett’s test. For variables with equal variances (P > 0.05), Tukey’s post-hoc test was employed. For variables with unequal variances (P < 0.05), Welch’s ANOVA followed by the Games-Howell post-hoc test was applied. Quantitative results are expressed as mean ± standard deviation (SD). Significance levels were established at ****P < 0.0001, ***P < 0.001, **P < 0.01, and *P < 0.05.

Results

Joint range of motion and capsule thickness

With prolonged immobilization, total passive extension ROM and arthrogenic passive extension ROM progressively declined (Fig. 2A and B). Myogenic contracture served as the primary contributor during the initial 2 weeks of immobilization, whereas arthrogenic contracture became predominant in the later stages (Fig. 2C). Notably, the severity of contracture tended to stabilize after 4 weeks. Detailed quantitative data regarding total, arthrogenic, and myogenic contracture degrees are presented in Table 1. Correspondingly, consistent with these functional changes, the thickness of the joint capsule also increased with the duration of immobilization, reaching a plateau after 4 weeks (Fig. 3B).

Fig. 2.

Biomechanical and macroscopic assessment of joint contracture with prolonged immobilization . (A) Total passive extension range of motion; (B) Arthrogenic passive extension ROM; (C) Contribution of myogenic and arthrogenic components to total contracture. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Table 1.

Quantification of knee joint contracture degrees at different immobilization time points. The data represent the calculated degree of total, arthrogenic, and myogenic contractures. All values are expressed as mean ± standard deviation (mean ± SD).

Immobilizatio time	Total passive extension ROM (°)	Arhrogenic passive extension ROM (°)	Myogenic passive extension ROM (°)
1 W	18.89 ± 4.88	7.88 ± 2.18	11.01 ± 4.79
2 W	39.74 ± 4.56	18.06 ± 4.16	21.68 ± 5.49
4 W	77.45 ± 2.46	48.30 ± 2.84	29.15 ± 2.87
6 W	78.30 ± 2.50	49.89 ± 2.97	28.41 ± 3.29

Fig. 3.

Histopathological evaluation of the posterior joint capsule with prolonged immobilization. (A) Representative H&E and Masson’s trichrome staining images of the posterior joint capsule; (B) Measurement of posterior joint capsule thickness; (C) Quantitative analysis of the collagen volume fraction. Significance is noted in the figures with a standard asterisk convention (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Scale bar = 250  µm and 25 µm.

H&E and Masson’s trichrome staining

H&E staining revealed a comprehensive overview of the histological alterations within the entire knee joint. In the Sham group, abundant adipocytes and loose collagen fibers were observed, along with minimal vasculature and virtually no inflammatory cells. As the immobilization duration progressed, marked fibroblast proliferation was noted, accompanied by a reduction in synovial adipose tissue within the fibrotic capsule from 1 to 4 weeks; however, a trend of adipocyte recovery was observed at 6 weeks. Furthermore, mild neovascularization and scattered mononuclear inflammatory cell infiltration were identified, with these pathological changes peaking at 2 weeks post-immobilization. These findings confirmed an early-stage inflammatory and angiogenic response in our model, which gradually diminished as mature fibrosis developed by 4–6 weeks (Fig. 3A). Masson’s trichrome staining distinctly visualized changes in collagen deposition. Quantitative analysis of the collagen volume fraction revealed that collagen fibers within the posterior joint capsule began to gradually increase from the first week post-immobilization, and reached a plateau by the fourth week (Fig. 3A, C). Both H&E and Masson’s trichrome staining revealed that the thickness of the posterior joint capsule and the extent of synovial hyperplasia progressively increased with the duration of immobilization, tending to plateau after 4 weeks. These histological findings were consistent with the observed changes in joint ROM. Furthermore, H&E staining demonstrated that the cartilage in all rat groups remained intact without damage (Fig. S2).

Immunohistochemical analysis

The expression profiles of α-SMA and COL1A1 in the synovial tissues were remarkably similar. The Sham group showed minimal baseline expression for both markers. However, their expression levels were significantly upregulated as early as the first week post-immobilization. Subsequently, these pro-fibrotic markers displayed a continuous, time-dependent increase as the duration of immobilization was prolonged, eventually plateauing after 4 weeks (Fig. 4A-C).

Fig. 4.

IHC evaluation of α-SMA and COL1A1 expression in synovial tissues. (A) Representative IHC images for α-SMA; (B) Representative IHC images for COL1A1; (C) Quantitative analysis of α-SMA expression (H-score); (D) Quantitative analysis of COL1A1 expression (H-score). Positive expression is indicated by brownish-yellow staining. Significance is noted in the figures with a standard asterisk convention (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Scale bar = 50 μm.

RT-qPCR and western blotting analysis

RT-qPCR and Western blotting were performed to quantify the mRNA and protein levels of α-SMA and COL1A1 in the synovial tissues. The results demonstrated consistent expression profiles for both markers at the mRNA and protein levels. Overall, the Sham group exhibited low baseline expression, whereas expression levels were significantly upregulated after 1 week of immobilization. Both fibrotic markers displayed a continuous, time-dependent increase as the duration of immobilization progressed, eventually reaching a plateau after 4 weeks. Notably, a marginal downward trend was observed at both mRNA and protein levels after 6 weeks of immobilization, although this trend was not statistically significant (Fig. 5A-E).

Fig. 5.

Gene and protein expression of fibrotic markers in synovial tissues. (A) Relative mRNA expression levels of α-SMA determined by RT-qPCR; (B) Relative mRNA expression levels of COL1A1 determined by RT-qPCR; (C) Representative Western blot images showing protein expression of α-SMA, COL1A1, and GAPDH; (D) Quantitative analysis of Western blot results for α-SMA (normalized to GAPDH); (E) Quantitative analysis of Western blot results for COL1A1 (normalized to GAPDH). The samples were derived from the same experiment and gels/blots were processed in parallel. Significance is noted in the figures with a standard asterisk convention (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Hematological indices and histopathology of major organs

To evaluate the systemic biosafety of the thermoplastic polymer resin used in this study, histological examinations were performed on major organs, including the heart, liver, spleen, lung, and kidney. H&E staining revealed no pathological abnormalities or signs of systemic toxicity in any of the harvested tissues (Fig. 6A). Furthermore, serum biochemical analyses were conducted to assess hepatic and renal functions (Fig. 6B-E). The results demonstrated that all measured parameters were within normal physiological ranges.

Fig. 6.

In vivo biosafety evaluation of the thermoplastic polymer resin. (A) Representative H&E staining images of major organs, including the heart, liver, spleen, lung, and kidney; (B) Serum levels of alanine aminotransferase(ALT); (C) Serum levels of aspartate aminotransferase (AST); (D) Serum levels of blood urea nitrogen (BUN); (E) Serum levels of creatinine (Cre). Scale bar = 25 μm.

Discussion

In this study, we developed a novel immobilization method utilizing a thermoplastic polymer resin to establish a rat model of knee arthrofibrosis. This material is commercially available, cost-effective, and allows for customized molding to the specific dimensions of the rat knee through simple adjustments. Our findings showed that immobilization for up to 6 weeks led to a progressive decline in knee ROM, which became significant by week 4. Histopathological analysis revealed concomitant synovial thickening and collagen deposition. Furthermore, the markedly elevated expression levels of α-SMA and COL1A1 in synovial tissues further validated the model at the molecular level. Given that α-SMA and COL1A1 are well-established markers of fibrosis, these findings confirm the successful induction of arthrofibrosis. Specifically, the upregulated expression of α-SMA and COL1A1, accompanied by a slight decrease in ROM, was detected as early as 1 week post-immobilization, indicating an early onset of the fibrotic process. By 2 weeks post-immobilization, significant elevations in all fibrotic markers were observed compared to the 1-week mark. By 4 weeks, the thickness of the posterior joint capsule increased substantially, and histological examination showed synovial hyperplasia and disorganized collagen fiber arrangement; these observations suggest that synovial fibrosis may be a primary driver of joint contracture. During the initial 2 weeks of immobilization, myogenic contracture accounted for a higher proportion of the total contracture. However, after 4 weeks, arthrogenic contracture became the dominant component. This shift may be attributed to the aggravation of capsular contracture and the concurrent onset of significant muscle atrophy, which diminishes the relative contribution of myogenic factors²⁰. In our study, changes in joint ROM, capsule thickness, and the expression of synovial α-SMA and COL1A1 were evident after 1 week and tended to stabilize after 4 weeks. These temporal trends are highly consistent with previous reports, demonstrating that our model is as stable and reliable as established models^21–23. We noted a partial increase in adipocytes in some rats after 6 weeks of immobilization, which represents an intriguing phenomenon. Additionally, the Collagen Volume Fraction statistics revealed that the collagen proportion in the 6-week immobilization group showed a slight decrease compared with the 4-week group, and the expression levels of α-SMA and COL1A1 at 6 weeks also showed a slight decline. Similar results were reported, showing that fibrosis peaked at 4 weeks of immobilization, followed by a decrease in collagen area at 6 and 8 weeks, accompanied by a declining trend in proteins such as TGF-β₁ and p-Smad2 in the joint capsule tissue²⁴. We propose two potential mechanisms for this phenomenon: (1) the transition from acute to chronic inflammation may lead to reduced expression of pro-fibrotic factors; (2) as active fibroblast proliferation plateaus and progresses toward chronic scar formation, collagen remodeling becomes prominent. With the resolution of active fibrosis and some matrix degradation, relative retention or mild reappearance of adipose tissue may occur. This may be attributed to the transition to a chronic fibrotic response. Notably, our study focused on developing a novel immobilization device and examined changes only up to 6 weeks, without long-term follow-up of fibrotic progression—which represents a limitation of the present work. Future studies should extend the immobilization duration to better characterize the late-stage fibrotic process in the knee joint.

Historically, the establishment of rat arthrofibrosis models has predominantly fallen into two categories. The first involves intra-articular trauma, such as anterior cruciate ligament (ACL) reconstruction, which restricts joint mobility to induce fibrosis²⁵. Invasive methods inflict direct injury within the joint cavity, potentially damaging multiple structures including the joint capsule, meniscus, cruciate ligaments, and articular cartilage. While effective, the fibrotic response is likely primarily driven by the healing of surgical trauma rather than immobilization itself. The second category involves extra-articular surgical trauma, where skin and muscle incisions are made to immobilize the knee using stainless steel wires or sutures wrapped around the femur and tibia^18,21,26,27. These minimally invasive models also inevitably introduce confounding factors like intramedullary hemorrhage, bone-derived inflammation, and potential skin infections due to the foreign body (suture) penetrating the skin. There are also other methods, such as the intra-articular injection of agents like TGF-β₁ to stimulate collagen deposition^28,29. Regardless of the approach, these invasive techniques inevitably cause surgical trauma. The resulting hemorrhage, inflammation, and stress responses may introduce significant bias into experimental outcomes. Consequently, these models fail to adequately recapitulate the clinical features of patients suffering from non-traumatic contractures.

Although previous studies have reported non-invasive knee flexion immobilization models in rats^30–32, there are distinct technical and practical differences between their models and ours. Previous studies used traditional materials such as plaster bandages or aluminum splints combined with tape. We attempted to replicate these methods in our pilot experiment. The results showed that a large number of rats developed limb swelling or device slippage by the second day of immobilization. For rats fixed with plaster, skin ulceration occurred within 7 days even without swelling due to the poor breathability of the material. Furthermore, some reports³² suggest that the device needs to be replaced every 2–3 days; however, the lack of cooperation from the rats requires repeated anesthesia for each replacement, which not only compromises the continuity and scientific rigor of the experiment but also poses a concern for animal welfare. Regarding splints and tape, rats are inherently covered in fur, which makes tape ineffective at securing a firm fixation; additionally, splints lack moldability, rendering effective immobilization unattainable. We also acknowledge that some studies have employed Velcro, custom-made splints, or tape for immobilization in mouse models^33,34; however, this approach is difficult to replicate in rats.

In contrast, our model represents a completely non-invasive approach with high feasibility, effectively eliminating the confounding factors associated with surgical trauma. It achieves stable and rigid immobilization in rats without compromising the intergrity of articular cartilage (Fig. S2), thus providing a reliable protocol for investigating immobilization-induced knee contractures. Furthermore, compared to previously reported external immobilization devices, our design offers another distinct advantage: it spares the ankle joint²⁴; consequently, ankle range of motion remains unaffected, thereby avoiding additional confounding variables related to multi-joint immobilization (Video S2). Unlike traditional plaster casts that fully encase the limb, our device features a semi-open design. This is critical because the rat hindlimb exhibits an irregular, conical morphology with a significant discrepancy between the thigh and crural circumferences. With circumferential casting, a loose application inevitably leads to device slippage, whereas a tight application risks compromising circulation, resulting in distal limb ischemia and edema. A distinct advantage of our device is that the majority of the limb remains exposed. This design not only minimizes dermatological complications but also mitigates the risk of ischemic necrosis associated with full circumferential pressure. Furthermore, it facilitates unimpeded daily observation of the limb’s condition. Our device demonstrated a high success rate, typically requiring only a single anesthesia session and device application. In this experiment, only three rats needed re-fixation during the entire experiment (two with mild swelling and one with device slippage); notably, no animal mortality was observed, and no further abnormalities occurred following re-fixation. We have summarized several practical considerations to ensure optimal immobilization: (1) Anesthesia dosage: A minimal dose is required. Since the immobilization procedure is painless and can be completed within approximately 10 minutes, rats can be maintained in a state of light sedation. This approach minimizes the potential systemic toxicity associated with high-dose anesthetics. (2) Sequential application: To facilitate a more intuitive understanding and replication of our device, the design can be conceptually simplified into three elliptical components. Strict adherence to the sequential steps outlined in the Methods section is crucial. Through repeated trials, we determined that the optimal and most time-efficient sequence is the ‘thigh-crus-trunk’ order (indicated by the green, red, and yellow circles in Fig. 1D). (3) Rapid cooling: A hair dryer (on a cool setting) can be utilized to accelerate the cooling and hardening of the thermoplastic polymer resin. (4) Post-immobilization monitoring: Ischemic swelling can develop rapidly within hours. Consequently, twice-daily inspections (morning and evening) are mandatory during the initial 72 h to prevent tissue necrosis.

Following the initial success of our model, we explored the potential of 3D printing to further streamline the fabrication process, drawing inspiration from a previous study³⁵. That study utilized an ‘S-shaped’ 3D-printed splint to immobilize mouse hindlimbs, securing both the knee and ankle joints. However, we noted that the device described in that study involved full circumferential wrapping of the limb, which can lead to dermatological complications and limb edema. Nevertheless, guided by this concept, we prototyped 3D-printed molds for our rat model (Fig. S1, Video S1). We encountered significant challenges due to individual anatomical variations: even among rats of identical body weight, there were substantial discrepancies in limb circumference and length. Consequently, a standardized 3D-printed mold failed to achieve the secure and close-fitting immobilization required. Furthermore, identifying a 3D-printing material that matched the superior hardness, gnaw resistance, and biosafety profile of the thermoplastic polymer resin proved difficult within a reasonable timeframe. Nevertheless, with the rapid evolution of 3D printing and material science, it is foreseeable that researchers will eventually be able to mass-produce standardized yet adaptable immobilization templates. Such devices could potentially achieve secure and stable fixation with only minor individualized adjustments, thereby significantly streamlining the experimental workflow and enhancing overall efficiency.

Limitations

This study has several limitations. First, our primary objective was to establish and validate a novel rat knee arthrofibrosis model; consequently, a detailed investigation into the dynamic biological processes of contracture development was not fully conducted. Specifically, early-stage inflammatory markers such as IL-1β and IL-6 were not assessed. Second, since our results indicated that the joint contracture plateaued after 4 weeks and our primary focus was on validating the successful induction of arthrofibrosis rather than investigating the multifaceted pathophysiological changes during the chronic fibrotic process, the maximum immobilization period was set at 6 weeks. Therefore, the effects of extended immobilization beyond this timeframe were not evaluated. Future studies should consider extending the observation period to characterize the long-term pathological evolution of arthrofibrosis.

Conclusion

In conclusion, the novel immobilization method presented herein offers simplicity, ease of operation, and high accessibility of materials. It yields a stable and reproducible model with a low incidence of complications. Consequently, this non-invasive approach represents a reliable strategy for future research into the pathophysiology and therapeutic interventions of arthrofibrosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Jiameng Jia: Conceptualization, Data curation, Validation, Visualization, Writing–original draft, Writing–review & editing. Weiwei Li: Data curation, Validation, Visualization, Writing–review & editing. Yu Pan: Validation, Writing–review & editing. This work currently described has not been published, is not being considered for publication elsewhere, and its publication was approved by all authors.

Funding

This study was funded by Beijing High-level Innovation and Entrepreneurship Talent Support Program (“Dengfeng” Project Grant No. G202511047) and National Key Clinical Specialty Construction Project of China (Grant No. XKB2023A2001).

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Ethics approval

All animal experiments were conducted in strict accordance with the institutional guidelines for the care and use of laboratory animals. The experimental protocol was reviewed and approved by the Ethics Committee of North Sichuan Medical College (Approval No. 2025090). All surgical procedures were performed under sodium pentobarbital anesthesia, and animals were euthanized via intraperitoneal injection of sodium pentobarbital at the study endpoints.

Competing interests

The authors declare no competing interests.