Recombinant Human Amelogenin Protein Enhances Healing of Osteochondral Injury in a Goat Model

ABSTRACT

Osteochondral injuries are common, painful, and lack natural healing abilities. Previous studies demonstrated that recombinant human amelogenin promotes osteochondral repair in rat knee injuries, including improved defect filling and formation of hyaline cartilage. Here, we aimed to evaluate the efficacy and safety of Remelix®, a novel recombinant human amelogenin–based product, in a large animal goat model. Acute osteochondral injuries were created in the weight‐bearing region of the medial femoral condyle in thirteen goats. Seven goats received Remelix® at a protein concentration of 0.5 mg/mL, with the contralateral knee applied with PX 407 carrier alone. An additional six goats received Remelix® at 0.25 mg/mL, while contralateral knees were treated with normal saline. Potential adverse effects, along with evaluation of osteochondral healing were assessed. Healing was assessed 24 weeks post‐treatment using magnetic resonance imaging (MRI), macroscopic evaluation, and histological and immunohistochemical analyses. No adverse effects were observed. Both Remelix® concentrations resulted in substantial healing of cartilage and subchondral bone. Histological analyses demonstrated improved tidemark formation, predominantly hyaline cartilage filling in the deeper layers of the defect with extension toward the superficial zone, and reduced subchondral bone abnormalities. MRI and macroscopic evaluations revealed a continuous repair surface with good integration at the defect margins. In contrast, PX 407 and saline‐treated control knees exhibited inferior repair, characterized by mainly fibrous cartilage, heterogeneous MRI signal intensity, and lower histological scores. Overall, Remelix® substantially improved osteochondral healing in a goat model without observable adverse effects, supporting its potential for further evaluation in human clinical studies.

1. Introduction

Articular cartilage functions as both a shock absorber and a gliding surface within the joint. Injuries to articular cartilage and the underlying subchondral bone are common and often painful, primarily due to exposure of the highly innervated subchondral bone [1]. Since cartilage lacks blood vessels, lymphatic drainage, and neural components, it has minimal intrinsic healing capacity, making osteochondral lesions susceptible to deterioration and frequently progress to post‐traumatic osteoarthritis (PTOA). As a result, such injuries significantly impair the quality of life, with pain and restricted mobility affecting even routine daily activities [2]. A biochemical and mechanical crosstalk between the articular cartilage and subchondral bone enables the two tissues to function as an integrated osteochondral unit, which is essential for maintaining joint homeostasis and optimal function [3, 4]. Therefore in the context of osteochondral lesions, both tissues play a critical role in exerting a substantial influence on each other, actively modulating the joints' metabolism and mechanical environment [4].

The choice of treatment following osteochondral injuries (OCIs) depends on lesion characteristics, including size and location, as well as patient‐related variables such as age and activity level. Acute intra‐articular fractures and injuries are managed through stable anatomic reduction aimed at restoring the fragment. Whereas chronic injuries generally focus on repairing or replacing the damaged tissue [5]. The prevailing surgical approach for small lesions includes debridement of the injured cartilage, curettage and microfracture of the subchondral bone, transforming the chronic condition into an acute injury. This treatment usually achieves partial pain relief and yields fibrocartilage, not hyaline articular cartilage [6]. For the treatment of large chronic osteochondral injuries, traditional surgical interventions are based on the use of structural autologous or allogenous grafts. However, donor‐site morbidity limits autologous transplantation [6]. Allograft tissue availability is low, with a limited clinical success of osteochondral allograft transplantation, usually involving the development of osteoarthritic degenerative changes over time [7, 8].

The use of 3‐dimensional scaffolding biomaterials has been heavily investigated for regeneration of OCIs. In recent years, a shift towards designing multiphasic scaffolds has emerged, that better emulate the structural and mechanical heterogeneity of the native osteochondral unit. These studies were translated into a limited number of clinical trials using bi‐ or tri‐ multi‐layered scaffolds such as Agili‐C™ and MaiRegen® [9, 10], demonstrating improved clinical scores at short‐ and mid‐term follow up [9, 10, 11]. These scaffolds can further be supplemented with biological additives such as cells and/or growth factors, in order to facilitate cell adherence, motility and differentiation [12]. However, the design and fabrication of such scaffolds is complex and poses a great challenge. Several biological therapies‐ including mesenchymal stem cells (MSCs), platelet‐rich plasma (PRP), bone morphogenetic proteins (BMP), and autologous matrix‐induced chondrogenesis (AMIC)‐ have shown promise in both animal models and clinical settings [13, 14, 15]. However, variability in formulation, delivery, and durability remains a limitation across these approaches [16, 17].

Amelogenin has emerged as a candidate for osteochondral regeneration based on our previous findings. It is expressed in bone marrow stromal cells and growth plate chondrocytes [18], and aged amelogenin‐null mice developed early osteoarthritic changes [19]. A single administration of a recombinant human amelogenin protein (rHAM⁺), produced in our laboratory using a eukaryotic expression system [20], was shown to induce healing of various musculoskeletal tissues. In a rat OCI model, rHAM⁺ promoted dose‐dependent repair of the articular cartilage and subchondral bone, demonstrating increased expression of proteoglycans and type II collagen and decreased expression of type I collagen at the newly formed cartilaginous tissue. It also facilitated healing of injured ligaments and tendons, including the medial collateral ligament of the knee, the calcaneofibular ligament of the ankle, and the infraspinatus tendon of the rotator cuff, restoring both mechanical strength and structural integrity [21, 22, 23]. Furthermore, rHAM⁺ attenuated the progression of osteoarthritis in both the rat OCI and experimental OA models by reducing cartilage degradation and preserving joint structure [19, 24]. Mechanistically, rHAM⁺ treatment increased mesenchymal stem cell (MSC) marker expression (CD105, STRO‐1) and promoted paracrine cell recruitment into the injury site, suggesting recruitment is an early event in amelogenin‐mediated regeneration [19, 22, 24].

These findings provide a strong proof‐of‐concept in a small animal rat model; validation in a large animal model is essential to better assess translational relevance. Therefore, in the present study, we extended these findings by evaluating potential adverse effects and animal welfare, as well as the regenerative potential of an amelogenin‐based product in a large animal goat model of acute OCI.

2. Materials and Methods

2.1. Production of Remelix®, a Recombinant Human Amelogenin‐Based Product

Remelix® is a proprietary, thermo‐sensitive aseptic composition (US20230082131A1) containing rAmelX (X‐linked human recombinant amelogenin, Merck, using an E. coli expression system), a thermo‐sensitive carrier (poloxamer 407; PX 407), and stabilizing excipients. The poloxamer carrier contained all ingredients except for the rAmelX protein. It is a triblock copolymer with a central hydrophobic block of polypropylene glycol flanked by two hydrophilic polyethylene glycol blocks. It has reversible gelation properties according to concentration and temperature, allowing it to reach the treatment site as a liquefied solution and undergo immediate in‐situ gelation [25]. The product has been tested for stability and biocompatibility and was verified to maintain protein integrity as well as its thermo‐reversible characteristics.

2.2. Animals

The experimental protocol (MD‐19‐15564‐4) was developed in collaboration with the veterinarian of the Authority for Biological and Biomedical Models, The Hebrew University, Jerusalem, Israel. The protocol was approved by the Authority for Biological and Biomedical Models committee. Animal welfare was continuously monitored by the authority's veterinary team and by the researchers, as detailed below. Twenty‐4‐month‐old female goats (n = 13) with a mean weight of 41.5 kg±1.6 were included in this study. The goats were housed in a large animal facility under standard conditions including controlled temperature (22°C ± 2), 12‐h light/dark cycles, humidity (30%–70%) ventilation, and ad libitum access to food and water. The goats were acclimated to the facility for a week upon arrival. The exclusion criteria were according to the ethical approval and included clinical and/or behavioral changes. We have adhered to the ARRIVE guidelines and have included the ARRIVE checklist.

2.3. Study Design

Thirteen goats were randomly allocated into two groups (n = 7 per treatment group A; n = 6 per treatment group B). Group sizes were determined on a priori power analysis using effect sizes derived from our previous rat OCI healing model, which was modified to a large animal model, and was sufficient to detect treatment effect [24]. Each goat underwent two separate procedures, spaced 7–8 days apart, to create an OCI in the medial femoral condyle (MFC) of each knee, followed by immediate application of treatment to each newly induced OCI. The procedures were performed on one knee at a time to reduce animal discomfort and allow sufficient recovery from the initial injury before the second intervention on the contralateral knee. In treatment group A, 0.283 mL of Remelix® at a protein concentration of 0.5 mg/mL was injected into the OCI of one knee, while to the contralateral OCI, 0.283 mL of carrier alone, PX 407. In treatment group B, one OCI was treated with 0.283 mL Remelix® at a protein concentration of 0.25 mg/mL, while the contralateral OCI received 0.283 mL of normal saline (sham treatment).

One goat from treatment group A was euthanized 12 weeks post‐treatment due to a significant weight loss caused by a gastrointestinal viral infection. To avoid the need of killing all animals and to allow their rehabilitation at the end of the study, one goat from group A was sacrificed 24 weeks post‐treatment. Analysis of both femoral condyles demonstrated that harvesting the OCI site along with surrounding tissues provided sufficient data. Therefore, the remaining 11 goats underwent two additional procedures, spaced 7–8 days apart, to harvest the OCI area for assessment. For animal welfare, following the harvesting of the OCIs during these additional procedures, all OCIs were filled with rHAM⁺ dissolved in propylene glycol alginate (PGA) carrier, previously demonstrated to induce tissue regeneration following induction of OCIs [24]. The goats were kept under close veterinary supervision for several days following surgery and were released to a rehabilitation farm only after demonstrating pain‐free ambulation, including running and jumping.

The primary outcomes were assessment of potential adverse effect and animal welfare, evaluated through gait and weight‐bearing assessment, feeding behavior, wound status, and body weight. Secondary outcomes included evaluation of the overall anatomy and structural integrity of the repaired cartilage and subchondral bone, as well as and the adjacent tissues. These secondary outcomes were assessed by 1) MRI scans conducted 2 weeks prior to the third operation; 2) macroscopic evaluation; and 3) histological and immunohistochemical analyses. A schematic diagram of the study design is presented in Figure 1.

Figure 1.

Schematic diagram of the study design.

2.4. The Surgical Procedure

The surgeries were performed at the Surgical Innovation and Technology Center of Hadassah University Medical Center. Prior to the procedure, the goats underwent a 16‐h fasting period. For general anesthesia, the goats were premedicated with an intramuscular injection of 0.025 mg/kg medetomidine (Zoetis) and 1 mg/kg ketamine (Vetoquinol), intravenously induced with 2% Propofol (Fresenius Kabi) to effect, followed by the use of 3% isoflurane inhalation to maintain general anesthesia throughout the operation. The goats were intubated, and a stomach tube was inserted. The goats received 1 mg/mL of Tramadol (Grünenthal) intramuscularly for pain management.

Under aseptic conditions, the MFC was exposed via a minimal medial parapatellar approach. An Osteochondral Autologous Transplantation System device (OATS; Arthrex) was used to create a 6 mm in diameter and 10 mm in depth OCI in the weight‐bearing area of the MFC, verified by fluoroscopy (GE Healthcare) (Figure Supporting Information 1). The defect area was then rinsed with normal saline, and 0.283 mL of the desired treatment or control was applied. The joint capsule and the subcutaneous tissue were separately sutured by a 2‐0 vicryl stitch and the skin by a 2‐0 nylon stitch, followed by an injection of 10 mL of 0.5% Marcain (Baxter) around the wound. Opsite spray (Smith+Nephew) was sprayed for protection over the wound site. The goats were then intramuscularly injected with Antisedan® (Orion Pharma) reversal agent and monitored until they were able to stand and bear weight on the operated leg. Post‐operation, 0.1 mg/kg Buprenorphine (Richter Pharma) analgesics were administered intramuscularly twice a day for 4 days and monitoring was conducted no less than twice a day in the 4 days post‐operation, and once a day for additional 10 days. Until the end of the experiment, the animals were monitored twice a week, including weighting once a week. All monitoring were performed by the veterinarians, technicians and the researchers. By the end of the experiment, 11 goats were transferred to a rehabilitation farm.

2.5. MRI Scans

Following general anesthesia as described above, each knee was injected with 10 mL of contrast media containing 0.5 mmol/mL Cyclolux (SANO CHEMIA Pharmazeutika GmbH) diluted in normal saline. MRI scans were performed using the GE 3 T Premier scanner using a knee coil to capture proton density fat‐saturated (PDFS) and T1 sequences with a slice thickness of 3 mm and a slice distance of 0.3 mm. For coronal PDFS sequence, the average TR was 2488 ms, TE 28.73 ms, resolution 220 ×300. For the coronal T1 sequence, the average TR was 578 ms, TE 6.63 ms, resolution 280 ×320. Imaging parameters were selected to balance spatial resolution with total acquisition time while the animals were maintained under general anesthesia. The scan duration was approximately 3 minutes per sequence and 30 minutes per goat. The scans were evaluated and scored using the MOCART 2.0 scale [26] to assess cartilage repair based on parameters including volume filling, integration, surface, structure, and signal intensity of the cartilage, bony defects or overgrowths, and subchondral changes. The scoring scale ranges from 0 (indicating no cartilage repair) to 100 (representing excellent cartilage repair). The scans were analyzed by a musculoskeletal radiologist and an orthopedic surgeon, both of whom were blinded to the specific treatment applied.

2.6. Tissue Harvesting

Twenty‐4 weeks after the creation of OCIs and treatment, the goats underwent two additional surgeries to harvest the area of interest from the MFC for macroscopic and microscopic analyses. The operative procedure mirrored the described process, with the exception that the MFC was exposed and photographed. A 10 mm OATS device (Arthrex) was then used to extract a cylindrical section enclosing the OCI and the surrounding tissue at a depth of 10 mm. All OCIs were filled with rHAM⁺ dissolved in a propylene glycol alginate (PGA) carrier. Monitoring was conducted as described above.

Macroscopic evaluation of the repaired tissue was conducted on the harvested osteochondral cylinders and the images captured during the surgeries. A semi‐quantitative scoring system developed by Goebel et al. [27] was used. The scoring criteria included color, presence of blood vessels, surface characteristics of repaired tissue, OCI filling, and degeneration of adjacent cartilage. The scoring scale ranges from 0 (indicating excellent cartilage repair) to 20 (indicating no repair and extension of the defect into adjacent cartilage). Image analyses were performed by an orthopedic surgeon and a researcher, both of whom were blinded to the treatments applied.

2.7. Tissue Processing, Histological, and Immunohistochemical Analyses

The harvested cylinders were processed and stained for histological (H&E and toluidine blue, safranin‐O/fast green) and immunohistochemical analyses for type I and II collagen cartilaginous components (ab34712 Abcam, and 1310‐08 SouthernBiotech, respectively), according to Helwa‐Shalom O. et al. [24]. Modifications to the original protocol included a fixation duration of 48 h and the use of a secondary anti‐goat antibody (Vectorlabs) for type I collagen staining. The degree of hyaline cartilage repair at the OCI site was assessed using the International Cartilage Research Society histological scoring system; ICRS‐II [28]. We present the thirteen parameters of the ICRS‐II scoring system, using a scale ranging from 0 (indicating fibrous tissue) to 100 (indicating hyaline cartilage). The parameters include: subchondral bone abnormalities, basal integration, surface architecture, mid/deep assessment, tidemark formation, matrix staining, tissue and cell morphology, vascularization, chondrocyte clustering, inflammation, abnormal ossification, and overall assessment. The quantification was performed by an orthopedic surgeon and a researcher, both of whom were blinded to the specific treatments applied.

2.8. Statistics

Statistical analyses were conducted using GraphPad Prism 9 software (GraphPad Software). Normality was assessed using the Shapiro‐Wilk test. For assessment of efficacy, both treatments of Remelix®, and the placebo (PX 407 carrier), were compared to sham‐ treatment with normal saline alone. One‐way analyses of variance (ANOVA) with Dunnett's correction for multiple comparisons was employed. The data are presented as the mean value and the standard deviation (SD). Statistical significance was defined as p‐values ≤ 0.05.

3. Results

3.1. Clinical Observations

Post‐operatively, the goats underwent routine clinical monitoring to assess for adverse effects and evaluate their overall welfare. As expected, the goats exhibited an antalgic gait during the initial 3–4 days following surgery, which resolved promptly. Joint swelling was minimal. Throughout the experiment, no signs of local or systemic infection or allergic reactions were observed, with the exception of one goat that experienced a severe gastrointestinal viral infection. In addition, the goats received pain management for 4 days after the surgery, with no requirement for chronic treatment.

3.2. Evaluation of the OCIs 24 Weeks Post‐Treatment

3.2.1. Macroscopic Observation

Macroscopic evaluation revealed that OCIs treated with Remelix® at protein concentrations of 0.5 or 0.25 mg/mL exhibited substantial, though incomplete, cartilage healing. Twenty‐4 weeks post‐treatment, defects treated with Remelix® at either protein concentration were predominantly filled with repair tissue exhibiting a more continuous surface morphology, minimal superficial fissures, and improved integration with the adjacent native cartilage. (Figure 2A–B). In contrast, OCIs treated with PX 407 carrier only or normal saline (sham) appeared more fibrillated and incompletely repaired toward the defect center, with increased vascularization (Figure 2C–D).

Figure 2.

Macroscopic evaluation of the joints and defect surface 24‐ weeks after creation of the OCIs and treatment. (A–D) Pictures were taken during the surgery in which a cylinder surrounding the OCI was removed. Surgery was performed 24 weeks after treatment with (A) Remelix® at amelogenin concentration of 0.5 mg/mL, (B) Remelix® at amelogenin concentration of 0.25 mg/mL, (C) PX 407, or (D) normal saline (sham control). Scale bar: 1 cm; Round images: the excised 10 mm in diameter and depth osteochondral cylinders. (E) Semi‐quantitative analyses, according to Goebel et al. [27], after treatment with Remelix® at amelogenin concentrations of 0.5 mg/mL (n = 6), or 0.25 mg/mL (n = 6), PX 407 (n = 6) and normal saline (n = 6). Data are presented as mean (standard deviation). The images were taken from the same goats presented in Figures 3, 4 and 5. These goats received the overall assessment scores of 2.5, 3, 6.5, and 9.5 for the treatments with Remelix® at amelogenin concentration of 0.5 mg/mL, 0.25 mg/mL, PX 407, and saline, respectively.

Semi‐quantitative macroscopic scoring [27] yielded lower (improved) gross scores in the experimental groups treated with Remelix® at protein concentrations of 0.5 and 0.25 mg/mL (mean 4.8 (SD 2.1) and mean 4.5 (SD 2.7), respectively), although non‐significant, compared to the saline‐treated (sham) group (mean 8.9 (SD 3.7); p = 0.100 and p = 0.070, respectively). The control, PX 407‐ treated group, yielded a similar gross score to the saline group (mean 8.3 (SD 4.0); p = 0.969 compared to saline), indicating inadequate healing at both control groups (Figure 2E).

3.2.2. MRI scans

MRI scans were performed 22 weeks after creation of the OCIs and treatment. At both Remelix® protein concentrations of 0.5 and 0.25 mg/mL, the defects exhibited advanced but incomplete repair (Figure 3A–D). The OCIs treated with Remelix® revealed almost complete integration of the treated area with the adjacent tissue. The tissue filling the defect exhibited a signal intensity similar to that of the adjacent cartilage. However, minor superficial fissures were observed.

Figure 3.

MRI images of the knee joints 22 weeks after creation of the OCIs and treatment. Coronal images were taken from MRI scans performed using T1 (A, C, E, G) or PDFS (B, D, F, H) sequences. OCIs were treated with (A, B) Remelix® at amelogenin concentration of 0.5 mg/mL, (C, D) Remelix® at amelogenin concentration of 0.25 mg/mL, (E, F) PX 407, and (G, H) normal saline (sham control). Scale bar: 1 cm. (I) Grading of cartilage repair at the OCIs according to the MOCART 2.0 scale [26] after treatment with Remelix® at a protein concentration of 0.5 mg/ml (n = 6), or 0.25 mg/ml (n = 6), PX 407 (n = 6) and normal saline (n = 6). Data are presented as mean (standard deviation). The images were taken from the same goats presented in Figures 3, 4 and 5. These goats received the overall assessment scores of 75, 50, 20, and 10 for the treatments with Remelix® at amelogenin concentration of 0.5 mg/mL, 0.25 mg/mL, PX 407, and saline, respectively.

In contrast, OCIs treated with controls showed poorer restoration. Carrier‐only control, PX 407, displayed irregular cartilage surfaces, incomplete defect fill, and heterogenous signal intensity. (Figure 3E–F). Most saline‐treated (sham) OCIs exhibited fluid‐like hyperintense signal with disorganized tissue alterations, indicating a heterogeneous structure of the tissue at the defect site (Figure 3G–H). These defects often lacked a discernible cartilaginous layer.

A subset of Remelix®‐treated knees exhibited high intensity signal in the PDFS sequence at the subchondral bone, suggesting edema within the subchondral bone at the defect center (Figure 3A–D). However, the surrounding bone architecture remained intact, without evidence of cyst formation.

Quantification assessment using the MOCART 2.0 scoring scale [26] further supported the qualitative observations. Analyses of individual parameters, as well as the total score, revealed that defects treated with Remelix® at both protein concentrations yielded higher gross scores, indicating substantially improved repair (total scores: 0.5 mg/mL, mean 57.5 (SD 23.4); 0.25 mg/mL, mean 60.0 (SD 15.2)) compared to the saline‐treated (sham) group (mean 31.7 (SD 19.2); p = 0.070 and p = 0.044, respectively). The PX 407 treated group yielded lower scores (mean 37.5 (SD 16.7); p = 0.909 compared to saline) similar to saline (Sham), suggesting poor healing (Figure 3I, Supporting Information Table 2).

Overall, MRI analysis demonstrated that Remelix® treatment at both protein concentrations resulted in enhanced cartilage repair, better defect filling, improved cartilage‐like signal characteristics, and better tissue integration compared with both carrier‐ only PX407 and saline controls.

3.2.3. Histological evaluation

Histological analyses of OCIs treated with Remelix® at either protein concentration of 0.5 or 0.25 mg/mL demonstrated an improved organization of the cartilage layers compared with the PX 407 carrier only or sham controls. The cartilage in most areas of the Remelix®‐treated OCIs progressed toward hyaline cartilage maturation, with distinct zonal layering becoming apparent, whereas the central region still exhibited immature, unorganized tissue. However, in control OCIs, the majority of the defect exhibited unstructured fibrous cartilage (Figures 4, 5).

Figure 4.

Histological analyses of the OCIs 24 weeks after creation and treatment. Hematoxylin and Eosin staining of thin sections obtained from OCIs treated with (A) Remelix® at amelogenin concentration of 0.5 mg/mL, (B) Remelix® at amelogenin concentration of 0.25 mg/mL, (C) PX 407, or (D) normal saline (sham control). Enlarged pictures were taken from the regions marked with black and yellow frames. Line represents the size of the 6 mm in diameter OCI. C, Cartilage; SB, Subchondral bone; SZ, Superficial zone; MD, Middle zone; DZ, Deep zone. Scale bar: 1 cm; scale bar for enlarged images: 100 µm. The images were taken from the same goats presented in Figures 3, 4 and 5.

Figure 5.

Histological and immunohistochemical analyses of the OCIs 24 weeks after creation and treatment. Treatment with (A, E, I, M) Remelix® at amelogenin concentration of 0.5 mg/mL, (B, F, J, N) Remelix® at amelogenin concentration of 0.25 mg/mL, (C, G, K, O) PX 407, or (D, H, L, P) normal saline (sham control). Scale bar: 1 cm. Histological staining of (A–D) Toluidine blue, (E–H) Safranin O/Fast Green. Immunohistochemical staining (I–L) for type II collagen and (M–P) type I collagen. (Q) Histological grading of the OCI according to ICRS‐II [28], after treatment with Remelix® at amelogenin concentrations of 0.5 mg/mL (n = 6), or 0.25 mg/mL (n = 6), PX 407 (n = 6) and normal saline (n = 6). Data are presented as mean (standard deviation). *p ≤ 0.05, **p ≤ 0.01. The images were taken from the same goats presented in Figures 3, 4 and 5. These goats received the overall assessment scores of 85, 40, 40, and 30 for the treatments with Remelix® at amelogenin concentration of 0.5 mg/mL, 0.25 mg/mL, PX 407, and saline, respectively.

Quantification of the histological analyses [28] revealed that treatment with Remelix®, resulted substantially reduced bone abnormalities compared with saline control (0.5 and 0.25 mg/mL Remelix®: mean 68.3 (SD 25.6) and 75 (SD 21.9) versus saline: mean 22 (SD 11.3); p = 0.006 and p = 0.002, respectively) and improved integration of cartilage with the subjacent bone (0.25 mg/mL Remelix®: mean 85.8 (SD 22.4) vs. saline mean 45.8 (SD 23.3); p = 0.012). Moreover, the pathological features of inflammation and abnormal ossification were absent or minimal in the Remelix®‐treated and the PX 407 carrier‐treated groups, in accordance with healthy cartilage [29], but were present following treatment with the saline control (Supporting Information Table 3).

Administration of Remelix® at either concentration also enhanced cartilage repair across all cartilaginous layers. Substantial healing was demonstrated in the middle and deep zones compared with saline control, indicated by improved formation of hyaline cartilage (0.25 mg/mL Remelix®: mean 72.5 (SD 22.7) vs. saline: mean 35.0 (SD 16.4); p = 0.021) and tidemark formation (0.25 mg/ml Remelix®: mean 85 (SD 21.9) vs. saline: mean 45.33 (SD 25.7); p = 0.025) (Figure 5Q). Matrix staining for proteoglycans, type II and type I collagen only showed minimal differences among the various groups, consistent with the incomplete nature of osteochondral repair at 6 months in large animal model [30, 31] (Figure 5, Supporting Information Table 3). Similarly, examination of cell and tissue morphologies revealed no statistically significant differences between the groups (Supporting Information Table 3). Consequently, the overall assessment demonstrated substantial healing following treatment with Remelix® at both protein concentrations of 0.5 and 0.25 mg/mL (mean 72.5 (SD 23.4) and mean 74.2 (SD 22.2), respectively), compared to treatment with either carrier only PX‐407 or saline, sham control (mean 48.33 (SD 23.2) and mean 46.2 (SD 15.4), respectively) (Figure 5Q). Together, these findings indicate that Remelix® promotes superior osteochondral repair characterized by improved cartilage quality, enhanced integration, and reduced subchondral bone pathology.

4. Discussion

Our previous studies have shown that the recombinant human amelogenin protein, rHAM⁺, induced healing of acute OCIs in a rat model [24]. Here, we evaluated the efficacy of Remelix®, a recombinant human amelogenin composition, to heal OCIs in a large‐animal model. The goat was selected as the translational model due to its joint size, subchondral bone‐to‐cartilage ratio, and loading characteristics which closely resemble those of the human knee [32, 33, 34]. For potential clinical translation, rHAM⁺ was replaced with the prokaryotic‐manufactured Remelix®, previously shown to promote periodontal wound regeneration in minipig models [35, 36]. The chosen concentration of 0.5 mg/mL Remelix® was based on our prior studies utilizing both amelogenin proteins rHAM⁺ [19, 21, 22, 23, 24] and Remelix® [35, 36] for regeneration. An additional lower dose (0.25 mg/mL) was included to account for future manufacturing considerations.

Throughout the experiment no adverse effects related to the treatments were detected in any of the goats. Twenty‐four weeks post‐treatment, healing was demonstrated, although analyses indicated that complete repair had not yet been achieved, consistent with the prolonged osteochondral matrix maturation in a large animal model [30]. In Remelix®‐treated groups, the central portion of the defect had not fully healed with normal hyaline cartilage and subchondral bone, while the defect margins more closely resembled a near normal joint surface, demonstrating early hyaline cartilage formation. Significant subchondral bone healing supported advanced repair of deeper cartilage layers, characterized by the onset of hyaline cartilage differentiation and improved integration. Since the mid and deep cartilage zones are primarily responsible for resistance to compressive loading, improved repair in these regions is considered critical for durable osteochondral regeneration [37].

This pattern of healing, progressing from the subchondral bone toward the superficial layers, aligns with the proposed paradigm drawn from in silico and in vivo analyses of osteochondral regenerative processes reported in various models [38, 39]. It has been proposed that migration of MSCs into the OCI follows this pattern; however, in empty OCIs this process results in fibrous tissue formation [39], as was observed in the current study in the control OCIs. The ability of amelogenin to act as a chemoattractant for MSCs has been demonstrated in vivo following induction of OCI in a rat model. rHAM⁺ treatment increased recruitment of cells expressing MSC markers from adjacent bone marrow into the injury site, and thereby contribute to osteochondral repair [24]. Nevertheless, the possibility of Remelix®‐induced recruitment of MSCs into the osteochondral niche in the goat model remains to be clarified.

A subset of Remelix®‐treated knees exhibited edema‐like signals. Such signals are frequently observed after cartilage repair procedures [40] and may reflect a normal healing response, typically resolving over time [41]. However, in humans, large bone marrow edema following osteochondral autograft transplantation has been linked to implant failure [42]. In the present study, no graft transplantation was performed, and, together with our other findings, we interpret the observed edema‐like signal as indicative of ongoing healing, although confirmation at later time point(s) is required.

Critical‐sized full‐thickness OCIs increase mechanical stress at the defect rim and disrupt bone‐cartilage interactions [43, 44], contributing to post‐traumatic osteoarthritis (OA) [45, 46]. Therefore, by promoting coordinated healing of subchondral bone and cartilage, Remelix® may reduce this risk. This interpretation is consistent with our previous study showing that rHAM⁺ inhibited progression and even partially healed OA in an induced‐OA rat model [19]. Given that intra‐articular fractures are typically treated days after injury, administration of Remelix® may offer a clinically relevant therapeutic strategy to attenuate the progression toward OA [47].

Remelix® at both protein concentrations exhibited similar healing abilities. However, only Remelix® at a protein concentration of 0.25 mg/mL resulted in significantly improved MRI scores and some histological parameters compared to the saline treated (sham). This suggests a potential dose‐dependent effect, with the lower concentration yielding slightly better healing in this goat model. Interestingly, in our previous rat OCI model, dose dependent effect was demonstrated, with 0.5 mg/mL rHAM⁺ yielding the best results [24].

In contrast, control OCIs treated with PX 407 carrier or saline (sham) exhibited poor healing with aberrant phenotypes of the subchondral bone and cartilage. Although comparable results were observed in most of the evaluated parameters, PX 407 showed a slightly reduced pathological features of inflammation and abnormal ossification as compared to saline (sham). Poloxamer polymers have been widely investigated as injectable scaffolds, often combined with cells and growth factors, for tissue repair and reconstruction [48]. It is compelling to suggest that filling the defect with any scaffolding material may be sufficient for marginal improvement of some clinical parameters. Nevertheless, the effect of PX 407 on healing osteochondral tissues was limited, while the combination with amelogenin led to superior healing throughout the various layers. The observation that a subset of control OCIs treated with PX 407 carrier or saline (sham) exhibited partial repair aligns with previous reports showing that some 6 mm goat defects can be spontaneously filled with fibrocartilage over time [49, 50, 51]. Such spontaneous repair may stem from the goats' skeletal immaturity at the start of the study [52] and, nonetheless, challenges the assumption that 6 mm diameter osteochondral injuries represent critical‐sized defects in goats [53].

When combined with the thermo‐sensitive carrier PX 407, the formulation conforms to the defect morphology by gelling at body temperature. This mechanism functionally mimics scaffold‐based biomaterials commonly applied in goat OCI repair studies [54]. Both scaffolds and Remelix® have demonstrated the capacity to promote osteochondral tissue regeneration in large‐animal models [49, 55]. However, the scaffolds have occasionally been associated with irregular matrix deposition [55, 56], which was not observed with Remelix®, where more consistent tissue integration was achieved. Furthermore, Remelix® eliminates the need for complex scaffold design and fabrication, as well as intraoperative trimming or fitting to patient‐specific defects [57]. Consequently, it may provide a simpler and more adaptable treatment approach applicable to OCIs of varying shapes and sizes.

This study has several limitations. The quadrupedal locomotion of the goats does not fully recapitulate human joint biomechanics. In addition, the use of a bilateral model may introduce confounding effects, though it was employed to reduce the total number of animals required for the study. The model only partially mimics the “acutization” of chronic osteochondral lesions that occurs during surgical debridement or microfracture procedures [58]. Enlargement of the sample size and evaluation of treatment efficacy over longer durations would have been beneficial. In this work, Remelix®, was produced in E. coli, which lacks post‐translational modifications. Human amelogenin carries at least one post‐translational modification, phosphorylation at serine 16 [59], which regulates enamel mineralization in vitro [59, 60]. Its role in musculoskeletal regeneration is unknown, and the use of eukaryotically produced amelogenin may enhance healing and clarify the importance of this modification. The impact of repeated dosing of Remelix® also remains to be determined.

5. Conclusion

Our findings demonstrate that treatment with Remelix® induces the healing of both hyaline cartilage and subchondral bone 6 months following acute osteochondral injury in a large animal goat model. These results suggest that Remelix® is a promising candidate for clinical trials as a biological adjunctive treatment.

Author Contributions

Research design: H.N., H.S., A.B., A.H.; Data acquisition: O.H.S., H.N., T.N., K.B.D., S.I.M., H.S., A.B., A.H.; Data analysis: O.H.S., H.N., M.G., A.Y., H.S., A.B., A.H.; Data interpretation: O.H.S., H.N., M.G., A.Y., Y.N.G., S.B., H.S., A.B., A.H.; Drafting the manuscript: O.H.S., H.N., H.S., A.B., A.H.; Revising the manuscript: O.H.S., H.N., Y.N.G., S.B., H.S., A.B., A.H. All authors have read and approved the final submitted manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Figure 1: The surgical procedure for creating a large osteochondral injury (OCI) in the goat medial femoral condyle (MFC).

Supporting Table 1: Scoring of the macroscopic evaluation of the OCI according to Goebel et al. [27]. Data are presented as mean (standard deviation). *p ≤ 0.05, **p ≤ 0.01.

Supporting Table 2: Scoring of the MRI scans of the OCI according to MOCART 2.0 scale [26]. Data are presented as mean (standard deviation).

Supporting Table 3: Scoring of the histological grading of the OCI according to ICRS‐II [28]. Data are presented as mean (standard deviation). *p ≤ 0.05, **p ≤ 0.01.