Establishment of an ex vivo cartilage damage model by combined collagenase treatment and mechanical loading

Abstract

Background

There is a substantial need for ex vivo cartilage damage models to assess new emerging cartilage repair strategies. Ex vivo cartilage explant models have the advantages of achieving standardized and reproducible experimental conditions while maintaining the cells in their native tissue environment. This study aimed to establish a bovine cartilage damage model to evaluate the safety and efficacy of novel cartilage repair therapies. We hypothesized that combining transient exposure to matrix-degrading enzymes with mechanical loading on bovine cartilage would simulate cartilage damage.

Methods

Prior to mechanical load, bovine osteochondral plugs underwent a brief 5-minutes treatment with collagenase to induce mild cartilage damage by disrupting the collagen network. To induce a moderate cartilage damage, aggrecanase 1 and aggrecanase 2 were additionally applied to the cartilage for 40 min post-collagenase treatment to degrade aggrecan. Data was analyzed using ANOVA or the Friedman test.

Results

Observations revealed a statistically significant loss of sulphated glycosaminoglycan (sGAG) using both Collagenase Treatment (CT) and Collagenase and Aggrecanase Treatment (CAT), while chondrocytes viability was maintained. Both treatments resulted in a significantly elevated release of inflammation markers during the initial two days, including IL6 and nitric oxide. Collagenase treatment also significantly increased neo-epitopes of aggrecan compared to the untreated plugs at day 7, suggesting endogenous aggrecanase activation upon collagen network disruption. The additional effect of mechanical loading on cartilage degeneration was also explored in the CT group. Mildly damaged cartilage treated solely with collagenase could withstand 1 h per day of cyclical load, at 10-20% compression of cartilage thickness combined with interfacial shear at 25 degrees. However, higher compression levels (20-40% of cartilage thickness) with the same shear stress regimen led to a significant increase in surface chondrocyte death, with no evidence of TUNEL staining.

Conclusions

This study establishes a promising model for evaluating cartilage repair strategies, and screening anti-catabolic drugs, particularly overload-related cartilage damage.

Introduction

Osteoarthritis (OA) represents a prevalent chronic musculoskeletal disorder that significantly impacts individuals’ quality of life globally [1]. Due to the incomplete understanding of the molecular mechanisms driving OA and the lack of effective interventions to slow its progression, total joint replacement surgery remains the sole option for severe OA [2]. Therefore, there is a significant demand for a reliable OA model capable of assessing targeted treatment strategies. However, OA is a complex disease of the whole joint, including cartilage, bone, muscles, tendons, ligaments and synovium [3]. Furthermore, OA is an end result, with multiple potential initiators and should be considered as multiple diseases that lead to a similar outcome.

The degradation of the cartilage extracellular matrix (ECM) is a key feature of OA pathogenesis, contributing to joint damage and structural failure. Breakdown of type II collagen and aggrecan, the two most abundant cartilage matrix biomolecules, is particularly significant, as their loss diminishes tensile strength and resistance to compression [4]. As the disease progresses, cell clusters [5], surface fibrillation and ultimate tissue destruction occurs. Intra-articular injection of collagenase in the knee is commonly applied in mice [6] or rats [7] to induce osteoarthritis in vivo by weakening ligaments and degrading the cartilage collagen network. However, the thinner cartilage of mice and rats, and greater potential for intrinsic healing than humans, limit the translational value of rodent studies [8]. With the incorporation of the principles of the 3Rs (Replacement, Reduction, and Refinement) into national and international animal use regulations, new methods that minimize the utilization of animals in research is crucial [9]. Bovine osteochondral explants present a translational and easily accessible method for assessing cartilage repair [10]. To mimic the in vivo environment, an ex vivo platform capable of separating cartilage and bone, can be utilized to culture explants with tissue-specific medium [11].

In this study, we aimed to establish a bovine cartilage damage model to evaluate the safety and efficacy of novel cartilage repair therapies. We hypothesized that minor damage induced by short term exposure to matrix-degrading enzymes would be exacerbated by mechanical loading. As patients are physically active, this combination of mild damage and load would partially simulate in vivo cartilage damage.

The cartilage ECM consists primarily of type II collagen and the sulfated proteoglycan aggrecan [12]. Apart from collagenases, aggrecanases which belongs to the ‘A Disintegrin and Metalloproteinase with ThromboSpondin motifs’ (ADAMTS) family also plays a significant role in osteoarthritic cartilage [13]. Specifically, aggrecanase 1 (ADAMTS-4) and aggrecanase 2 (ADAMTS-5) are key enzymes in osteoarthritis progression [14].

Cyclical mechanical load is required to maintain a healthy cartilage tissue. Mechanical loading of cartilage induces complex changes, including matrix and cell deformation, fluid flow, and alterations in matrix water content, ion concentration, and fixed charge density. Excessive mechanical loading affects chondrocyte metabolism, causing an imbalance between anabolic and catabolic activities, which depletes matrix components [15]. Specifically, cartilage responds positively to physiological dynamic compression (~ 10–20%) by enhancing ECM synthesis. Excessive loading (> 20%) impairs matrix production, while hyper-physiologic levels (> 50%) cause cellular damage and promote catabolism [16]. Therefore, we also investigated the effect of mechanical load including compression and shear on cartilage damage model.

In this study, we aimed to establish an ex vivo cartilage damage model by combining enzyme digestion and mechanical loading using bovine osteochondral explants. The dose and treatment duration of collagenase and aggrecanases were investigated. In the case of collagenase treatment, the role of additional mechanical stimulation was also assessed. The cell viability and tissue histology of explants were determined by viability staining and immunohistology. The expression of genes related to cartilage anabolism and inflammation was evaluated by quantitative real-time polymerase chain reaction (qRT-PCR). In addition, the sulphated glycosaminoglycan (sGAG), nitric oxide (NO), and inflammatory cytokine releases into the conditioned medium were quantified. Finally, the effect of compression and shear force on collagenase damaged cartilage was explored.

Materials and methods

Materials and reagents

Dulbecco’s Modified Eagle’s Medium (DMEM) high glucose, penicillin-streptomycin (Pen/Strep), HEPES, and non-essential amino acids were purchased from Gibco. ITS supplement and fetal bovine serum (FBS) were supplied by Corning. Collagenase II, HBSS, and SuperScript Vilo RT Kit were supplied by ThermoFisher. Recombinant human ADAMTS4 protein (aggrecanase 1) and recombinant human ADAMTS5 protein (aggrecanase 2) were purchased from R&D Systems. Keratanase II was purchased from Glycofinechem. Biotinylated Anti-Mouse IgG [H + L] made in horse, ABC‐complex, and DAB (3,3′‐diaminobenzidine) was supplied by Vector Laboratories. Bovine IL6, IL-1β and TNF-α Elisa kit was supplied by KINGFISHER. Griess Reagent System was supplied by Promega. TUNEL Assay Kit - HRP-DAB was purchased from Abcam. Quant-iT PicoGreen dsDNA Assay Kits were purchased from Invitrogen. TRI reagent and PolyAcryl carrier were supplied by Molecular Research Center. RNeasy Mini Kit was purchased from Qiagen. TaqMan Gene Expression Master Mix was supplied by Applied Biosystems. Sircol™ Insoluble Collagen Assay kit was purchased from Biocolor. All other materials and reagents were purchased from Sigma-Aldrich.

Osteochondral plugs (OC plugs) harvest and culture

Osteochondral plugs were isolated from fresh bovine femoral condyles (343, 356, 363, 401, 512, 521, 723, 942 days old) on the day of slaughter for human consumption. The biopsies were isolated using an 8 mm trephine drill. Biopsies were cut to a bone length of 5 mm, then cultured in an osteochondral culture bioreactor platform, which was supplied by LifeTec Group (Eindhoven, Netherlands). The osteochondral explants were mounted into the inserts using an O-ring positioned exactly at the interface between cartilage and bone, creating two isolated culture compartments: the upper compartment containing the cartilage and the lower compartment containing the bone, each with respective media [11]. Cartilage was cultured with 1.5 mL cartilage medium consisting of DMEM high glucose supplemented with 1% Pen/Strep, 50 µg/mL ascorbic acid, 1% ITS + Premix, 1% non-essential amino acids, and 25mM HEPES. Bone medium consisted of 1.5 mL DMEM high glucose supplemented with 1% Pen/Strep, 50 µg/mL ascorbic acid, 10% FBS, 5mM β-glycerophosphate and 25mM HEPES. Each experiment was performed with at least 3 biological repeats.

Induction of cartilage damage

To establish a mild cartilage damage, 1 mL 0.1% Collagenase II (285 U/mL) in HBSS was added to the cartilage compartment for 5 min at 37 °C, then washed three times with PBS followed by fresh cartilage medium. For a moderate damage, the cartilage was further treated with 800 µL 2 µg/mL aggrecanase 1 and 2 µg/mL aggrecanase 2 in HBSS for 40 min at 37 °C after Collagenase II treatment, followed by a PBS wash. The OC plugs were then cultured for 2 days or 7 days, after which the cartilage and subchondral bone were separated. The cartilage and medium were collected for further analysis.

Mechanical loading of osteochondral plugs

A previously described cartilage bioreactor system [17] was utilized to load the osteochondral plugs. A 32 mm ceramic hip ball was pressed onto the osteochondral plug. Interfacial motion was induced by oscillating the ball ± 25° at 1 Hz around an axis perpendicular to the explant axis. Simultaneously, dynamic compression was applied at 0.2 mm to 0.4 mm (around 10–20%) or at 0.4 mm to 0.8 mm (around 20–40%) sinusoidal strain related to the cartilage thickness in the center of the explant. Mechanical load of the OC plugs was performed for 1 h per day for 3 and 7 days of culture (10–20% compression and shear group) or 1, 2 and 3 days of culture (20–40% compression and shear group), then the cartilage part was collected for histology.

sGAG, DNA, and collagen measurement

Cartilage was digested overnight at 56 °C in 0.5 mg/mL proteinase K. The total sulfated glycosaminoglycan (sGAG) content in cartilage and in the medium were measured using the 1,9-dimethyl-methylene blue (DMMB) colorimetric method [18]. Briefly, 20 µL of digested cartilage samples or 50 µL of conditioned medium were added to each well, followed by 200 µL DMMB reagent. Absorbance at 535 nm was immediately read using a Tecan microplate reader. DNA content in the proteinase K-digested samples was measured using PicoGreen reagent, following the manufacturer’s instructions. Briefly, the samples were added in duplicates (100 µL) into a 96-well white plate. 100 µL PicoGreen working solution was added and the fluorescence of the samples was measured at an excitation wavelength of 485 nm and emission of 535 nm by using a Tecan reader after incubation for 3 min at room temperature. On the same digested samples, the collagen content was evaluated with Sircol™ Insoluble Collagen Assay kit, according to the manufacturer’s instructions. After adding the acid step (Fragmentation Reagent) directly to the harvested media, soluble and insoluble collagen fragments are both measured. Briefly, 1 ml Sircol dye reagent was added to 200 µl of sample or standard and gently mixed on a shaker for 30 min. The tubes were centrifuged at 13,000 x g. for 10 min and the supernatant removed. After removing the unbound dye with an ice-cold Acid-Salt Wash reagent, the pellet was dissolved in Alkali reagent. The absorbance was recorded at 550 nm.

Safranin O / Fast green staining

Cartilage was fixed in 10% buffered formalin for two days and dehydrated in an ascending ethanol series. Then the samples were embedded in paraffin and sectioned at 6 μm thickness. Sections were dewaxed and then rehydrated in ethanol. To visualize the nuclei, slides were first stained with Weigert’s iron Haematoxylin for 10 min. Sections were then stained with 0.02% Fast Green in ddH₂O for 6 min to reveal collagen deposition, followed by 0.1% Safranin O for 12 min to show proteoglycan deposition, then rinsed in ddH₂O, and dehydrated using ascending concentrations of ethanol (50%, 70%, 96%, 100%) before mounting in Eukitt mounting medium. Images were acquired on an Olympus BX63 microscope.

Lactate dehydrogenase (LDH) and ethidium homodimer (EthD) staining

Cartilage was snap frozen in liquid nitrogen and stored at -20 °C. Sections with thickness of 10 μm were made using a cryosection microtome (Leica, Wetzlar, Germany). An LDH assay was performed on cryostat sections after drying at room temperature. Sections were washed twice with PBS for 5 min, then stained with 1 µg/mL ethidium homodimer for 45 min at room temperature. Sections were then incubated with freshly prepared LDH solution containing 40% Polypep, 2 mM Gly-Gly, 5.4 mg/mL lactic acid, 1.75 mg/mL ß-nicotinamide adenine and 3 mg/mL Nitroblue Tetrazolium pH 8.0, for 3 h at 37 °C. Slides underwent 2 washes with 50 °C tap water, followed by 4% formalin. Slides were washed with PBS, followed by washing with water. Coverslips were placed on the slides with aqueous Faramount mounting medium. Images were acquired on an Olympus BX63 microscope. Images were analyzed and quantified using ImageJ software. Cell viability was calculated as the ratio of live cells (number of blue signals) to total cells (number of EthD signals) by Image J in the full-thickness section.

Immunohistochemistry of COL2A1, ACAN, neo-epitopes of aggrecan

Paraffin sections were dewaxed and rehydrated in ethanol. Endogenous peroxidase activity was blocked by incubation with 0.3% hydrogen peroxide for 30 min. Then cartilage sections underwent different antigen retrieval protocols for the separate antibodies, which are shown in Table 1. Afterwards, unspecific antigen binding was blocked with horse serum for 1 h at room temperature. Samples were then incubated with primary antibody at 4 °C overnight. The sections were washed with PBST, and then incubated with anti-mouse secondary antibody at room temperature for 30 min. After washing with PBST, the sections were incubated with ABC‐complex at room temperature for 30 min. Following thorough washing with PBST, the sections were incubated with DAB for 4 min, followed by counterstaining with hematoxylin. The sections were then washed and dehydrated. Finally, cartilage sections were mounted with Eukitt. Images were acquired on an Olympus BX63 microscope. Images were analyzed and quantified using ImageJ software.

Table 1 Antibodies used for immunohistochemistry staining

Interleukin-6, Interleukin-1 beta, tumor necrosis factor alpha (TNF-α) and nitric oxide release

Bovine interleukin-6 (IL 6), Interleukin-1 beta (IL-1β) and Tumor Necrosis Factor alpha (TNF-α) released by cartilage was measured by using a commercial enzyme-linked immunosorbent assay (ELISA) kit according to manufactures’ instructions. Nitrite, a stable product of nitric oxide in aqueous solutions, was measured using the Griess reaction system following the manufacturer’s instructions.

TUNEL staining

TUNEL staining was used for detection of potential cell apoptosis in cartilage sections. Briefly, after dewaxing and dehydration, paraffin sections were stained with a TUNEL assay kit according to the manufacturer’s instructions. Images were acquired on an Olympus BX63 microscope.

Gene expression analysis

Cartilage samples were precooled with liquid nitrogen and pulverized before transferring into 1 mL TRI Reagent and 5 µl PolyAcryl carrier, then homogenized with a Retsch tissue lyser. After centrifugation, the supernatant was collected. 0.1 mL 1-Bromo-3-chloropropane (BCP) was added to get the aqueous phase containing RNA, which was purified with RNeasy Mini Kit. Reverse transcription was conducted with Vilo RT Kit. qRT-PCR was performed with TaqMan gene expression master mix on a QuantStudio™ 6 Pro Real-Time PCR system. Primer and probes sequences or gene expression assay on demand numbers are presented in Table 2. Target mRNA was quantified using the 2^−ΔΔCt method, with RPLP0 as the endogenous reference. Gene expression levels were normalized to the mean values of bovine cartilage at day 0.

Table 2 Primers and probes used for PCR

Primers and probes presented with sequences were custom designed; primers and probes presented with the RPLP0 catalogue number from Applied Biosystems.

Statistical analysis

Data are presented as the mean ± standard deviation (SD) of at least three independent experiments. Shapiro–Wilk normality test was used to define whether data were normally distributed in GraphPad Prism 8.1.0 software. Analysis of variance (ANOVA) was conducted to determine differences with normal distribution. Friedman test was conducted to evaluate the differences for non-normally distributed data. Multiple comparison was corrected by Tukey test. P value < 0.05 was considered significant.

Results

Mild and moderate cartilage damage

An ex vivo culture platform was applied for osteochondral plugs (Fig. 1A), which aimed at simulating the in vivo environment and enabled supply of tissue specific factors to meet the different needs of bone and cartilage [11]. Mild damage was induced by treating cartilage with collagenase II for a duration of 5 min (CT group), whereas moderate damage was induced by additional digestion using aggrecanase1 and aggrecanase2 for a period of 40 min (CAT group). The collagenase treatment time was chosen based on a pilot series where tissue was exposed to collagenase II for up to 30 min (Supplementary Fig. 1).

Compared to the untreated samples, the release of sGAG in the CT and CAT groups primarily occurred within the initial two-day period (Fig. 1B). The ratio of tissue sGAG to DNA was decreased in damaged cartilage both at day 2 and day 7 comparing to untreated cartilage (Fig. 1C). The sGAG/DNA ratio remained unchanged between days 2 and 7 for all treatments, consistent with the measured sGAG release slope. Collagen release was measured during cartilage damage. Interestingly, the CT and CAT groups showed lower collagen release compared to untreated cartilage on day 2 (Fig. 1D). However, increased collagen release was observed in the CT group on days 6 and 7 compared to untreated cartilage.

Fig. 1

sGAG loss of cartilage after enzyme treatment were assessed. (A) Schematic illustration of the platform for ex vivo culture of bovine osteochondral explants and procedures of enzyme treatment. (B) sGAG release. On Day 0 within the CAT group, the initial sGAG content did not originate from zero due to the collection of Aggrecanase buffer for sGAG measurement. (C) Normalization of sGAG retained in cartilage to DNA content. (D) Collagen release (µg). ∗P < 0.05. ∗∗P < 0.01. Triangles, dots or square of different color represent the 4 biological repeats. Abbreviations: CT: Collagenase II treated cartilage, CAT: Collagenase II and Aggrecanase treated cartilage

Histological evidence of mild and moderate cartilage damage under static conditions

SO-FG staining revealed sGAG loss in the CT and CAT groups compared to untreated cartilage (Fig. 2A). As CAT resulted in complete loss of safranin O staining in the upper zone, with an intact surface, we considered this as moderate damage. Notably, chondrocytes within cartilage exhibited high viability in both the CT and CAT groups according to LDH staining (Fig. 2A and B). In the CT cartilage explants, IHC analysis revealed increased COL2A1 compared to the untreated group, although this finding was not statistically significant. In the CAT group, COL2A1 staining varied inconsistently among the four bovines. Elevated ACAN staining was observed in the CT and CAT groups in three of the four bovines. In addition to the CAT explants, the CT group also exhibited the presence of neo-epitopes of aggrecan (Fig. 2A and B), indicating that upon treatment with collagenase II, cartilage undergoes a process wherein endogenous aggrecanases were activated to facilitate the breakdown of aggrecan spontaneously.

Fig. 2

sGAG loss, viability, distribution of cartilage matrix components COL2A1 and ACAN after enzyme treatment were assessed. (A) Representative images of SO-FG, LDH/EthD, immunohistochemistry staining of COL2A1, ACAN, neo-epitopes of aggrecan which indicated activity of Aggrecanase. Scale bar = 100 μm. (B) Quantification of cell viability, COL2A1, ACAN, neo-epitope of aggrecan under static conditions on day 7. ∗P < 0.05. Triangles, dots or square of different color represent individual donors (n = 4). Abbreviations: CT: Collagenase II treated cartilage, CAT: Collagenase II and Aggrecanase treated cartilage, SO-FG: Safranin O and Fast Green

Expression of inflammatory and cartilage related markers

The release of IL6 (Fig. 3A) and the generation of nitric oxide (NO) (Fig. 3C) was significantly higher in both the CT and CAT groups compared to untreated explants during the initial two days of observation. In the first two days, IL-1β release was elevated in the CT group compared to untreated cartilage (Fig. 3B), while levels in the CAT groups remained unchanged. These findings suggest that cartilage underwent an inflammatory response initially, with normal levels of these parameters thereafter. Notably, on day 2, the CT group exhibited a higher production of IL-1β and nitric oxide (NO) compared to the CAT group, indicating a more pronounced release of inflammatory markers within the CT group. However, the release of IL-6, IL-1β, and NO in both CT and CAT groups did not show any significant changes on days 4, 6, and 7 when compared to the untreated cartilage. No differences in TNF-α release were observed (Supplementary Fig. 2). We also did not observe changes in the mRNA expression levels of ACAN, COL2A1, and SOX9 in the damage groups at both day 2 and day 7, indicating a lack of a detectable cartilage repair response at the transcriptional level in 7 days (Fig. 3D). Additionally, there were no apparent alterations in the mRNA expression levels of IL6, ADAMTS4 and ADAMTS5 (Fig. 3D).

Fig. 3

External collagenase II and aggrecanase led to release of inflammatory markers, but mRNA expression of ACAN, COL2A1, SOX9, IL6, ADAMTS4, and ADAMTS5 did not change. (A, B) IL6 and IL-1β release in the medium were detected by ELISA (n = 4). (C) Release of Nitric oxide was detected by Griess reaction (n = 4). (D) RT-qPCR analysis of ACAN, COL2A1, SOX9, IL6, ADAMTS4, and ADAMTS5 (n = 5). Relative quantification of target mRNA was performed according to the 2^−ΔΔCt method with RPLP0 as the endogenous control. Gene expression values were normalized to the average values of bovine cartilage at day 0. Triangles, dots or square of different color represent individual donors

Impact of compressive forces ranging from 10 to 20% of cartilage thickness and shear stress at 25 degrees on mild cartilage damage

As the damage to the CAT group already led to a total loss of safranin O stain, mechanical load was only applied to the CT group to assess whether mild histological changes under static conditions would be exacerbated during kinematic load. Chondrocytes within cartilage treated with Collagenase II remained alive when exposed to compression of 10–20% of cartilage thickness and ball rotation of +/- 25 degrees, as assessed by LDH staining (Fig. 4A and B). COL2A1 mRNA levels remained unchanged in the CT loading groups at both day 3 and day 7 comparing untreated loading group. However, increased ACAN mRNA levels were observed in the CT loading groups compared to untreated cartilage at both time points, although this was not statistically significant (Fig. 4B). Additionally, higher levels of neo-epitopes of aggrecan were detected in the CT loading groups at day 7 but not at day 3 (Fig. 4B).

Fig. 4

Cartilage treated with collagenase II could resist 10–20% compression of cartilage thickness and 25 degrees shear. (A) LDH/EthD staining, SO-FG, IHC-stained with anti-COL2A1, anti-ACAN, and anti-neo-epitopes of aggrecan (n = 3). (B) Quantification of cell viability, COL2A1, ACAN, neo epitope of aggrecan under loading conditions (10–20% compression and 25 degrees shear) on day 3 and day 7. Triangles, dots or square of different color represent individual donors (n = 3). Scale bar = 100 μm

Influence of compressive forces within the range of 20–40% of cartilage thickness and shear stress at 25 degrees

Cartilage was exposed to compressive forces ranging from 20 to 40% of cartilage thickness and ball rotation of +/- 25 degrees. In a preliminary study, cell death was observed on the cartilage surface in the CT group after 7 days of loading. We then focused on determining the timing of cell death. LDH staining revealed cell death on the cartilage surface in the CT loaded groups from day 2 (Fig. 5A and B) while TUNEL staining was negative. The loss of viability indicates the cartilage’s incapacity to withstand mechanical stresses of 20–40% after collagenase treatment. SO-FG staining additionally revealed a greater loss of sGAG when compared to compression within the range of 10–20%. Furthermore, following mechanical loading, damaged cartilage exhibited reduced COL2A1 staining compared to untreated loading samples (Fig. 5A and B). In the CT loading group, ACAN and neo-epitopes of aggrecan increased on day 1 and day2 comparing to untreated loading group.

Fig. 5

Cartilage surface cells did not survive under 20–40% compression of cartilage thickness and 25 degrees shear. While there were reduced numbers of LDH positive viable cells, TUNEL staining was negative. (A) LDH/EthD staining, TUNEL staining, SO-FG, IHC-stained with anti-COL2A1, anti-ACAN, and anti-neo-epitopes of aggrecan (n = 3). (B) Quantification of cell viability, COL2A1, ACAN, neo-epitope of aggrecan under loading conditions (20–40% compression and 25 degrees shear). Scale bar = 100 μm

Discussion

Osteoarthritis (OA) is the most common degenerative joint disorder and coupled with an incomplete understanding of its pathophysiology, OA contributes to the challenges associated with cartilage repair [19]. There is a great need for ex vivo models that allow for assessment of cartilage therapeutics, particularly at early stages of degeneration, to reduce the number of animal experiments [20]. The objective of this study is to establish a convenient ex vivo cartilage damage model capable of assessing the safety and efficacy of novel cartilage regeneration therapies. Use of a media separation insert allows for separate media to be used for cartilage and bone. Within this study, serum free media was used for cartilage, due to the detrimental effect of FBS on cartilage explant mechanical properties [21]. Future studies could add serum, depending on the research question. This proof-of-concept study used tissue from a wide age range of animals (343–942 days old). As the bovine growth plate starts closing around the age of 2 years [22], when using this model in future interventional studies the age of the animals, as well as the sex, should be taken into consideration.

Cartilage extracellular matrix (ECM) destruction and disrupted homeostasis are hallmarks of osteoarthritis and key therapeutic targets for cartilage regeneration [23]. The cartilage ECM is predominantly composed of type II collagen and the sulfated proteoglycan aggrecan [12]. In our study, transient exposure to enzymes such as collagenase and aggrecanase were used to induce cartilage damage by targeting these matrix components.

Collagenase has previously been employed to induce osteoarthritis in mice or rats by joint instability leading to altered loading [8]. In a preliminary study, treatment durations of 5, 15, and 30 min with collagenase were tested. No significant changes were observed between 5 and 15 min. However, 30 min resulted in excessive sGAG loss, as indicated by SO FG staining in Fig S1. In this study, a short treatment of 5 min with collagenase was applied to the cartilage part of bovine osteochondral plugs to induce a mild cartilage damage most likely by breaking the superficial collagen network. Osteochondral plugs were then cultured in a platform that enabled the separation of cartilage and bone, thereby more accurately mimicking the in vivo environment. The same platform has previously been used to facilitate extended culture periods of up to 84 days [11]. Extending the culture period in future studies with this model would facilitate a conclusive assessment of the feasibility of cartilage repair and allow studies into changes in subchondral bone. To induce a moderate cartilage damage, aggrecanase 1 and aggrecanase 2 were applied to cartilage for 40 min following collagenase treatment. In our preliminary study, a two-hour treatment with 2 µg/mL aggrecanase 1 and 2 µg/mL aggrecanase 2 did not result in sGAG loss observed from SO-FG staining (data not shown). This suggests that aggrecanase alone might not effectively interact with aggrecan when the collagen network remains intact, or that aggrecan fragments diffuse out of the matrix only to a limited extent. Therefore, for inducing a moderate damage, our proposed procedure involves an initial treatment with collagenase for 5 min, followed by a subsequent treatment with aggrecanase 1 and aggrecanase 2 for 40 min followed by culture under static conditions.

sGAG loss was observed in both the Collagenase Treatment (CT) and Collagenase and Aggrecanase Treatment (CAT) groups, while chondrocytes remained viable. Surprisingly, collagen release in the CT and CAT groups was lower than in the untreated group at day 2 (Fig. 1D). At this early time point, the released collagen in the control group is likely derived from the cartilage surface and the freshly cut edges of the explant. We hypothesize that the enzyme treatment and subsequent washing steps removed this initial burst release. However, the enzyme digest solutions and PBS washes were not collected for confirmation. In both the CT and CAT groups, expression of inflammation related markers was observed during the initial two days, as evidenced by elevated levels of IL6 and nitric oxide release. TNF-α release was measured during cartilage damage; however, no significant difference was observed between the damaged and untreated groups (Fig. S1).

Within the CT group, a higher presence of neo-epitopes of aggrecan compared to the untreated group was observed via immunohistochemical staining at day 7. This observation suggests that upon disruption of the collagen network, chondrocytes initiate the activation of endogenous aggrecanases, possibly as part of the cartilage remodeling process. Collagenase subtypes exhibit proteolytic side activities, such as trypsin [24]. A likely mechanism would be pro-ADAMTS4 was activated [25] by the trypsin-like activity of collagenase II. This model provides a mimic of very early mild collagen damage on cartilage.

Notably, we did not observe alterations in mRNA levels of COL2A1, ACAN, SOX9, IL6, ADAMTS4, and ADAMTS5 within the cartilage damage groups. We also assessed the mRNA expression of IL-1β, which was undetectable in both the untreated and CT groups in preliminary study (data not shown). The mRNA expression of IL6, IL-1β, ADAMTS4, and ADAMTS5 did not correspond with IL6, IL-1β release and the immunohistochemical staining of aggrecan neo-epitope. We propose that this discrepancy may stem from the mRNA level reflecting all layers of cartilage, whereas inflammation or catalytic processes might primarily occur on the cartilage surface, where the histological changes are observed. However, it is also known that the correlation of gene expression and protein production is low due to intracellular mechanisms, including further transcriptional and post-translational factors [26].

Joint tissues demonstrate a high sensitivity to mechanical stimuli, and mechanical loading might be a critical external factor in regulating both cartilage development and long-term functional maintenance [27]. In comparison to the CAT group, the mild cartilage damage model induced by collagenase treatment alone appears to be more favorable as it triggers the intrinsic activity of endogenous aggrecanase. Given the important role of mechanical loading, our investigation also delved into the state of mild cartilage damage under varying compression and shear forces. Our findings indicate that chondrocytes in mild damaged cartilage could survive under conditions of 10–20% compression of cartilage thickness and 25 degrees of shear stress applied hourly per day. However, when subjected to a higher magnitude of compression, specifically 20–40% of cartilage thickness, combined with the same shear stress regimen, cell death was detected at the cartilage surface that was TUNEL negative. Further work is required to fully establish the mechanism of cell death, which might include death types such as necrosis, pyroptosis or autophagy [28] rather than apoptotic processes beginning on day 2. The combined treatment of collagenase and mechanical loading at 20–40% compression with superimposed shear stress, may serve as a valuable tool for screening OA drugs. Furthermore, this offers insight into protecting from cell death in osteoarthritis patients subjected to joint overload.

We do not observe cell death in healthy cartilage even under compression levels ranging from 20 to 40%. However, another study reported that healthy bovine cartilage underwent apoptosis upon compression levels of 4.5 MPa, 10 MPa or 20 MPa (30–50% thickness of cartilage) [29]. The reason might be the absence of the superficial layer of cartilage of cartilage in their study. They removed the initial 100–400 μm of cartilage tissue by using a sledge microtome to provide a flat surface, followed by the acquisition of two subsequent 1-mm thick cartilage slices. This methodical difference potentially impacts the comparative analysis, as deep cartilage and superficial cartilage respond differently to loading exercise [30].

Under static conditions, COL2A1 staining increased in the CT group at day 7 but remained unchanged in CAT group. Upon application of 10–20% compression and shear, COL2A1 staining remained unchanged in CT loading group. Notably, reduced COL2A1 staining was consistently observed in the CT group subjected to 20–40% compression and shear. Reduced COL2A1 staining suggests a lower level of COL2A1 protein. Conversely, increased COL2A1 staining could indicate either a higher concentration of COL2A1 protein or enhanced antibody binding due to collagen fragmentation following collagenase treatment. We deduce that compression and shear in the CT group cause collagen fragments to diffuse from the cartilage tissue into the medium, resulting in reduced COL2A1 presence under 20–40% compression and shear conditions. However, detecting collagen content in the medium is necessary to confirm this hypothesis. Unfortunately, we did not collect the medium under loading conditions.

Under static conditions and 10–20% compression, more aggrecan and neo-epitope of aggrecan were observed in CT group comparing to untreated cartilage. The neo-epitope of aggrecan serves as an indicator of aggrecanase activity. We deduce that aggrecanase activity may similarly facilitate interaction between the ACAN antibody and its antigen. The staining pattern of the neo-epitope of aggrecan usually aligns with the intense staining observed in ACAN staining, thus providing partial support for our speculation. Notably, in the CT loading group at 20–40% compression, two out of three bovines showed a gradual decrease of neo-epitope of aggrecan compared to the CT loading group on day 1. However, this trend was not observed in ACAN staining. Aggrecanase cleaves ACAN into peptides [31]. The peptide containing the neo-epitope might be released at 20–40% compression and shear, whereas the ACAN antibody stains the remaining portion of the ACAN molecule that is partially retained within the matrix.

Within this cartilage damage model, the release of inflammation markers reduced with time. However, the inflammation in osteoarthritis is chronic, low-grade, and differs in its clinical progress [32]. Osteoarthritis impacts the entire joint, involving the articular cartilage, subchondral bone, and synovial membrane [33]. Therefore, a further addition of synovium may more accurately represent the clinical situation [34]. Another limitation of this study is its focus on osteochondral tissue responses within an artificial platform and loading setting. Factors such as angiogenesis and cross talk with synovial tissue and fluid could not be adequately considered under the described culture conditions.

There are several studies previously that have examined inflammatory cytokines such as IL-1β + loading in cartilage explants as in vitro models of OA [35]. Although they are also characterised by ECM degradation, the cytokine concentrations used are a 100 to 1000-fold higher than the actual concentrations in OA synovial fluid and may hence more reflect acute inflammation rather than the low-grade inflammation generally found in late OA. Interestingly, the data presented here, including IL-6 (Fig. 3A), IL-1β (Fig. 3B), and NO (Fig. 3C) release, suggest that matrix loss triggers an inflammatory response. Hence, the relationship between IL-1β and matrix degradation appears to be mutually reinforcing, as IL-1β is known to be upstream of matrix-degrading enzymes such as collagenase and aggrecanase, while the degraded matrix fragments in turn activate inflammation [36]. This cycle of inflammation and matrix degradation highlights a bidirectional interaction contributing to joint pathology.

A previous study also demonstrated that the addition of inflammatory cytokines IL-1β and TNF-α for 7 days to human osteochondral plugs could induce osteoarthritic characteristics in ex vivo [37]. Human cartilage is a superior model for investigating the response of cartilage in joint diseases compared to bovine cartilage. However, a potential limitation of their study arises from the acquisition of human osteochondral plugs, which often exhibit varying degrees of degeneration already, thereby complicating comparisons across different groups. Additionally, bovine osteochondral tissue is easier to obtain therefore making the model available to more laboratories. However, bovine tissue tends to be from young animals due to meat consumption rules. These tissues will potentially have different structures and would have a greater capacity for repair. Although adding matrix-degrading enzymes or inflammatory cytokines could not fully represent osteoarthritis in vivo due to the complexity, they remain valuable tools for assessing therapeutic intervention of initial matrix damage and potential later regeneration in ex vivo models before in vivo testing.

Fig. 6

Combined Collagenase II treatment and 20–40% mechanical loading could be a promising tool to assess the response to early matrix damage and responses to potential new therapies

In conclusion, we have successfully established a mild bovine cartilage damage model by subjecting it to collagenase treatment for 5 min (Fig. 6) followed by additional multiaxial load. The applied load exacerbated the damage and could also induce cell death in a magnitude dependent manner. This provides a model for early cartilage damage under mechanical conditions that would be experienced in a patients joint. Moreover, when 20–40% compression is applied, it has the potential to facilitate the screening of therapies for patients with osteoarthritis induced by tissue overload.